Liver

Autoimmune Liver Disease
Autoimmune Liver Disease
Management and Clinical Practice
Edited by
James Neuberger
University Hospital Birmingham
Queen Elizabeth Hospital
Edgbaston
Birmingham, UK
Gideon M. Hirschfield
University of Toronto
Toronto
Ontario, Canada
This edition first published 2020 © 2020 by John Wiley & Sons Ltd
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as
permitted by law. Advice on how to obtain permission to reuse material from this title is available at
http://www.wiley.com/go/permissions.
The right of James Neuberger, Gideon M. Hirschfield to be identified as the author of editorial work has been
asserted in accordance with law.
Registered Office(s)
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products visit us at
www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears
in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty

The contents of this work are intended to further general scientific research, understanding, and discussion only
and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or
treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes
in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and
devices, the reader is urged to review and evaluate the information provided in the package insert or instructions
for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of
usage and for added warnings and precautions. While the publisher and authors have used their best efforts in
preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the
contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of
merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives,
written sales materials or promotional statements for this work. The fact that an organization, website, or product
is referred to in this work as a citation and/or potential source of further information does not mean that the
publisher and authors endorse the information or services the organization, website, or product may provide or
recommendations it may make. This work is sold with the understanding that the publisher is not engaged in
rendering professional services. The advice and strategies contained herein may not be suitable for your situation.
You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in
this work may have changed or disappeared between when this work was written and when it is read. Neither the
publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited
to special, incidental, consequential, or other damages.
Library of Congress Cataloging‐in‐Publication Data

Names: Neuberger, James, editor. | Hirschfield, Gideon M., editor.
Title: Autoimmune liver disease : management and clinical practice / edited
by James Neuberger, University Hospital Birmingham, Queen Elizabeth
Hospital, Edgbaston, Birmingham, UK, Gideon M. Hirschfield, University
of Toronto, Toronto, Ontario, Canada.
Description: Hoboken, NJ : Wiley-Blackwell, 2020. | Includes
bibliographical references and index.
Identifiers: LCCN 2019051484 (print) | LCCN 2019051485 (ebook) | ISBN
9781119532606 (hardback) | ISBN 9781119532613 (adobe pdf ) | ISBN
9781119532620 (epub)
Subjects: LCSH: Liver–Diseases–Immunological aspects. | Autoimmune
diseases.
Classification: LCC RC846 .A87 2020 (print) | LCC RC846 (ebook) | DDC
616.3/62079–dc23
LC record available at https://lccn.loc.gov/2019051484
LC ebook record available at https://lccn.loc.gov/2019051485
Cover Design: Wiley
Cover Images: Immunofluorescence Microscopy © Beano5/Getty Images, Prostate Gland © BeholdingEye/
Getty Images, PET CT scan © springsky/Shutterstock, Human organ liver © Marochkina Anastasiia/Shutterstock
Set in 10/12pt Warnock by SPi Global, Pondicherry, India
10 9 8 7 6 5 4 3 2 1
v
Contents
Preface xv
Contributors xvii
Abbreviations xxi
Introduction: The Paradigm and Paradox of Liver Autoimmunity 1

M. Eric Gershwin
Section I Scientific Basis of Clinical Autoimmune Liver Diseases 3
1 Introduction to the Physiology, Immunology and Pathology

of the Liver and Biliary Tree 5
Marco Carbone and Mario Strazzabosco
Key Points 5
Liver Cell Types and Organization 6
Hepatic Metabolism 7
Bilirubin Metabolism and Transport 7
Carbohydrate Metabolism 7
Lipid Metabolism 8
Protein Metabolism 8
Metabolic Zonation 8
Hepatic Transport Systems 9
Basolateral (Sinusoidal) Transporters 9
Apical (Canalicular) Transporters 11
Drug Metabolism 11
Bile Formation, Secretion and the Enterohepatic Circulation 12
Bile Acid Synthesis and Metabolism 13
Enterohepatic Bile Acid Circulation 13
Death and Regeneration of Hepatocytes 14
Cell Death 14
Liver Regeneration 15
Cholangiocyte Reaction to Biliary Damage 16
Protective Role of Biliary HCO3− Secretion 16
Cholangiocytes and Immunity 16
Biochemical Markers and Patterns of Hepatic Injury 17
Hepatocellular Necrosis 18
Cholestasis 18
References 20
vi Contents
2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 21

Isaiah G. Roepe and John M. Vierling
Key Points 21
Introduction 22
Role of Innate and Adaptive Immunity in Autoimmunity in the Context
of the Liver as an Immune Organ 22
Overview 22
Innate Immunity 23
Liver as an Innate Immune Organ 25
Adaptive Immunity and Adaptive Immune Functions of the Liver 25
Role of the Liver as an Adaptive Immune Organ 28
Generation and Maintenance of Tolerance to Self‐antigens 28
Overview 28
Central Tolerance 29
Central T‐cell Tolerance 29
Central B‐cell Tolerance 29
Peripheral Tolerance 30
T‐ and B‐cell Clonal Anergy 30
T‐cell Mediated Immune Regulation 30
Natural and Inducible T Regulatory Cells 30
Peripheral B‐cell Regulatory Mechanisms 31
Regulatory Dendritic Cells 31
Immunoregulatory Interplay Between Treg and Th17 Cells 31
Risk Factors for Autoimmune Diseases 31
Genetics 31
Complex Genetic and Monogenic Diseases 31
HLA Risk Alleles 32
Non‐HLA Gene Associations 32
Critical Role of Epigenetics 33
Transcription Factor Enhancers and Super Enhancers 33
MicroRNAs 33
Sex and Sex Hormones 33
Vitamin D and Sunlight Exposure 34
Loss of Immune Tolerance to Autoantigens and Perpetuation
of Autoimmune Diseases 34
Overview 34
Role of the Microbiome 35
Mechanisms of Loss of Tolerance to Autoantigens 35
Bacterial and Viral Infections 35
Molecular Mimicry of Autoantigens 36
Neoantigens 36
Failure of Apoptosis to Conceal Autoantigens and Eliminate Autoreactive Cells 36
Immune Deviation of Activated T Cells 37
T‐cell Receptor Revision in the Periphery 37
Perpetuation of Autoimmune Diseases 37
Epigenetics 37
Epitope Spreading 38
Tissue Memory T Cells 38
Contents vii
Cytokines Promoting Chronic Inflammation and Autoimmunity 38

Tertiary Lymphoid Structures and Germinal Centers 39
Epithelial Cell‐induced Transformation of iTreg to Th17 Cells 39
Prevention of Autoimmunity and Therapeutic Control of Autoimmune Diseases 39
Overview 39
Strategies to Prevent Autoimmunity 39
Vitamin D3 39
Gut Microbiota Manipulation in Pregnancy and Infancy 39
Oral Tolerance 39
Strategies to Treat Established Autoimmune Diseases 40
Inducible T Regulatory T Cells 40
Epigenetic Enhancer Regulation 40
References 44
3 Genetics and Risk of Autoimmune Liver Diseases 47

George F. Mells
Key Points 47
Introduction 48
HLA Associations in Autoimmune Liver Disease 50
Non‐HLA Associations in Autoimmune Liver Disease 53
Conclusion 61
References 62
4 Autoantibodies and Understanding of Autoimmune Liver Diseases 65

Benedetta Terziroli Beretta‐Piccoli, Giorgina Mieli‐Vergani , and Diego Vergani
Key Points 65
Introduction 66
Methods of Detection 66
Anti‐nuclear Antibody 68
History 68
Immunofluorescence Reactivities and Antigenic Targets 68
Clinical Significance in Autoimmune Liver Disease 70
Anti‐smooth Muscle and Anti‐actin Antibodies 73
History 73
Immunofluorescence Reactivities, Antigenic Targets, and Clinical Significance 73
Anti‐liver‐kidney Microsomal Antibody 74
History 74
Anti‐liver Cytosol Type 1 Antibody 76
History 76
Anti‐soluble Liver Antigen Antibody 76
History 76
Anti‐neutrophil Cytoplasmic Antibody 77
History 77
Immunofluorescence Reactivities, Antigenic Targets, and their Clinical
Significance 77
viii Contents
Anti‐mitochondrial Antibody 78
History 78
Significance 78
Anti‐asialoglycoprotein Receptor Antibody 79
History 79
Significance 79
Indications for Autoimmune Liver Serology Testing 79
Concluding Remarks 81
References 82
5 Environmental Exposure and Risk in Autoimmune Liver Diseases 83

Ying Qi Li and Andrew L. Mason
Key Points 83
Introduction 84
Autoimmunity 84
Interaction of Genes and Environmental Triggers in Autoimmunity 85
Mechanisms for Triggering Autoimmunity 86
Primary Biliary Cholangitis 89
Geo‐epidemiology, Clusters, and Case–Control Studies of PBC 92
Bacterial Infection and PBC 93
Xenobiotics and PBC 96
Viruses in PBC 97
Autoimmune Hepatitis 99
Prospectus 101
References 102
Section II Autoimmune Liver Diseases and Their Clinical Correlation 103
6 Autoimmune Hepatitis 105

Aliya Gulamhusein and Patrick McKiernan
Key Points 105
Introduction 106
Definition and Pathophysiology 106
Epidemiology 107
Presentation 108
Adult‐onset AIH 108
Pediatric‐onset AIH 109
Diagnosis 109
Autoimmune Sclerosing Cholangitis 112
Treatment 112
Therapeutic Approach in Adults 113
First-line Therapy 113
Withdrawal of Therapy 114
Second‐line Options 114
Therapeutic Approach in Children 115
First-line Therapy 115
Withdrawal of Therapy 115
Contents ix
Second‐line Options 115

Prognosis 116
Adulthood 116
Childhood 116
References 118
7 Primary Biliary Cholangitis 123

Alessio Gerussi and Marco Carbone
Key Points 123
Introduction and Definition 124
Epidemiology 124
Etiopathogenesis 125
Clinical Presentation 126
Asymptomatic Patients 126
Symptomatic Patients 126
Diagnosis 127
Biochemical Tests 127
Autoantibodies 128
Liver Biopsy 128
Imaging 130
Differential Diagnosis 130
Natural History 130
Disease Course in the UDCA Era 131
Risk Stratification 132
Treatment 134
Ursodeoxycholic Acid 134
Obeticholic Acid 135
Fibric Acid Derivatives 136
Budesonide 137
Liver Transplantation 137
Symptom Management 137
Complications of Liver Disease 139
Osteoporosis 139
Advanced Liver Disease 140
Varices 140
HCC 140
Useful Websites 141
References 141
8 Primary Sclerosing Cholangitis 143

Mette Vesterhus, Benedetta Terziroli Beretta‐Piccoli, Kirsten Muri Boberg,
and Giorgina Mieli‐Vergani
Key Points 143
Introduction 144
Definition 145
Adult PSC 145
Epidemiology 145
Diagnosis 146
Diagnosis of Variant Phenotypes 147
Presentation 147
x Contents
Malignancy 149
Cholangiocarcinoma 149
Gallbladder Cancer 150
Colorectal Cancer 150
Prognosis 150
Risk Stratification: Clinical Characteristics 150
Natural History Models 151
ALP and Bilirubin 151
Non‐invasive Evaluation of Fibrosis 151
Treatment 152
Medical Treatment 152
Bile Acid Therapy 153
Microbiota Modulators 153
Anti-inflammatory Treatment 154
Antibiotic Therapy 154
Endoscopic Treatment 154
Surveillance for Malignancy 154
Colorectal Cancer: Colonoscopy 155
Gallbladder Cancer: Ultrasound 155
Hepatocellular Cancer: Ultrasound 155
Cholangiocarcinoma 155
Bone Mineral Density 155
Juvenile Sclerosing Cholangitis 155
Epidemiology in Pediatric Disease 155
Autoimmune Sclerosing Cholangitis 156
Diagnosis in Children 156
Clinical Features in Pediatric Disease 158
Treatment of Pediatric Disease (ASC and Juvenile PSC) 158
Prognosis in Pediatric Disease 159
Conclusion 159
References 160
9 IgG4‐Related Liver and Biliary Disease 163

Eleanor Barnes and Emma L. Culver
Key Points 163
Introduction and Historical Perspective 164
Etiology and Pathogenesis of IgG4‐RD 165
Development and Characteristics of an IgG4 Antibody Response 165
Antigens That May Drive an IgG4‐RD Response 166
IgG4 Antibodies and Pathogenesis 167
Incidence and Prevalence of IgG4‐RD 167
Clinical Characteristics of IgG4‐related Hepatobiliary Disease 168
Hepatobiliary Disease 168
Disease Outside the Hepatobiliary System 168
Diagnostic Criteria, Histologic Features, and Approach to Diagnosis
of IgG4‐related Hepatobiliary Disease 168
Diagnostic Criteria for AIP and IgG4‐RD 168
Histologic Diagnosis of IgG4‐related Hepatobiliary Disease 170
Contents xi
Utility of Blood Tests, Including Serum IgG4 Levels

in IgG4‐RD Diagnostics 171
Common Alternative Diagnoses to Consider 172
Radiologic Characteristics of Hepatobiliary IgG4‐RD 173
IgG4‐RD and Relationship with Malignancy 173
Management and Treatment of Patients with IgG4‐RD 174
Monitoring and Follow‐up 175
Conclusions and Future Directions 175
References 177
Section III Specific Clinical Challenges 181
10 Managing Acute and Chronic Seronegative Liver Disease 183

Marcus C. Robertson and Peter C. Hayes
Key Points 183
Introduction 184
An Approach to Seronegative Acute Liver Failure 184
Potential Causes of Seronegative ALF and Features that Suggest
an Autoimmune Pathogenesis 185
Clinical Features of Seronegative ALF 188
Risk Stratification in Seronegative ALF 188
Management of Seronegative ALF 189
N‐Acetylcysteine 190
Corticosteroids 190
Emergency Liver Transplantation 191
An Approach to Cryptogenic Chronic Liver Disease 192
Potential Causes of Seronegative Chronic Liver Disease and Features that
Suggest an Autoimmune Pathogenesis 192
Non‐alcoholic Fatty Liver Disease 194
Wilson Disease 194
Seronegative Autoimmune Hepatitis 195
Seronegative Primary Biliary Cholangitis 197
Conclusion 197
References 198
11 Managing Pregnant Women with Autoimmune Liver Disease 201

Eleni Theocharidou and Michael A. Heneghan
Key Points 201
Introduction 202
Fertility in AIH 202
Pregnancy Outcomes in AIH 202
Liver‐related Outcomes in Pregnancy 206
Pregnancy in Cirrhosis 209
Safety of Medication in Pregnancy 210
Summary 213
References 214
xii Contents
12 Bone Health in Patients with Autoimmune Liver Diseases 219

Albert Parés and Núria Guañabens
Key Points 219
Introduction 220
Prevalence of Osteoporosis and Fractures 222
Primary Sclerosing Cholangitis 223
Pathogenesis 224
Assessment of Bone Disease 226
Prevention and Treatment of Bone Loss 227
General Measures 227
Pharmacologic Agents 227
Bisphosphonates 227
Other Agents 229
Summary 230
References 231
Section IV Transplantation and Its Role in Autoimmune Liver Disease 233
13 Recurrent Autoimmune Liver Disease and Its Impact on Clinical Practice 235

Carlos Moctezuma‐Velázquez and Aldo J. Montano‐Loza
Key Points 235
Introduction 236
Recurrence of PBC After LT 236
Diagnosis of PBC After LT 236
Risk Factors Associated with PBC Recurrence After LT 237
Treatment of PBC Recurrence After LT 238
Prognostic Impact of PBC Recurrence After LT 239
Primary Sclerosing Cholangitis 239
Recurrence of PSC After LT 240
Diagnosis of Recurrence of PSC After LT 240
Risk Factors Associated with PSC Recurrence After LT 240
Treatment of PSC Recurrence After LT 241
Prognostic Impact of PSC Recurrence After LT 241
Recurrence of AIH After LT 242
Risk Factors Associated with AIH Recurrence After LT 242
Treatment of AIH Recurrence After LT 243
Prognostic Impact of AIH Recurrence After LT 243
Conclusions 243
References 244
14 Recurrent Autoimmune Liver Disease and its Scientific Significance 247

Atsushi Tanaka, Patrick S.C. Leung, and M. Eric Gershwin
Key Points 247
Introduction 247
Recurrence of PBC 248
Incidence and Diagnosis of Recurrent PBC 249
Contents xiii
Risk Factors of Recurrent PBC 250

Impact of Recurrent PBC on Long‐term Outcomes 250
Recurrence of PSC 251
Incidence and Diagnosis of Recurrent PSC 252
Risk Factors of Recurrent PSC 253
Impact of Recurrent PSC on Long‐term Outcomes 254
Recurrence of AIH 254
Incidence and Diagnosis of Recurrent AIH 254
Risk Factors of Recurrent AIH 255
Impact of Recurrent AIH on Long‐term Outcomes 255
Concluding Remarks 256
References 257
Section V Controversies in Autoimmune Liver Diseases 263
15 Making Sense of Overlap and Crossover Syndromes 265

Olivier Chazouillères
Key Points 265
Introduction 266
General Considerations 267
Liver Biopsy 268
AIH Scores 269
Genetics 269
PBC/AIH Overlap Syndrome 269
PSC/AIH Overlap Syndrome 272
Liver Transplantation 273
Conclusions 274
References 275
16 The Role of Extrahepatic Autoimmunity in Autoimmune Liver Disease 277

Ewa Wunsch and Piotr Milkiewicz
Key Points 277
Introduction 278
Epidemiology 278
Autoimmune Thyroid Disease 279
Sjögren Syndrome 279
Systemic Sclerosis 280
Systemic Lupus Erythematosus 281
Rheumatoid Arthritis 282
Celiac Disease 283
Inflammatory Bowel Disease 284
Conclusions 286
Guidelines for Clinicians 287
References 288
17 Symptoms, Chronic Disease, and Patient Management 289

David Jones
Key Points 289
Background 290
Goals of Treatment 290
xiv Contents
Symptoms, Quality of Life, and Health Utility: Key Concepts 291

Symptoms 291
Health‐related Quality of Life 292
Health Utility 292
Symptoms and Their Management in AILD 292
Impact of Disease‐modifying Therapy on Symptoms in AILD 292
Stage‐associated Symptoms 293
Stage‐independent Symptoms 293
Pruritus 293
Fatigue and Cognitive Symptoms 295
Social Isolation Symptoms 298
Effective Care Delivery in AILD 299
References 300
Index 301
xv
Preface
Over the last three decades there have been pathological and clinical aspects of AILD but
many advances in the understanding of auto- also stimulate further research to provide
immune liver disease (AILD). The advances effective therapies. We are indebted to the
in understanding the pathogenesis of these authors who have worked so hard to summa-
diseases is starting to lead to more specific rize complex fields. They represent some of
therapies both for slowing or arresting pro- the world leaders in this area and we are
gression and for treatment of the symptoms. grateful to them for their time. We know
Many people with AILD see delays in diag- there is some overlap and minor conflicts
nosis and are misdiagnosed; treatments tend between chapters but we have retained these
at best to control progression rather than as we would like chapters to be comprehen-
induce resolution. Symptoms are often not sive and self‐contained.
treated because of either poor knowledge or We would also like to thank Deirdre Barry,
inadequate therapies. Liver transplantation, Yogalakshmi Mohanakrishnan and Baskar
a highly effective therapy for those with end‐ Anandraj at Wiley for their support, patience
stage disease, has its inherent risks and com- and editorial skills and, most importantly, our
plications and is really a reflection of a failure families who have been patient during this time.
to treat the original disease. Recurrence of Finally, we thank you for reading this
AILD after transplantation, which manifests volume and hope you find it stimulating and
despite sufficient immunosuppression to provoking. Whilst autoimmune liver dis-
prevent allograft rejection and in a foreign eases are rare, they are impactful, and every
HLA milieu, must provide clues to some of opportunity to raise awareness, standards
the processes that lead to AILD but the and knowledge is an opportunity for better
lessons remain obscure. patient care.
We have been pleased to edit this volume
James Neuberger
which we hope will not only provide a bal-
Gideon M. Hirschfield
anced and comprehensive background to the
xvii
Contributors
Eleanor Barnes PhD, FRCP, F Med Sci M. Eric Gershwin MD, MACR, MACP
Professor of Hepatology and Experimental Distinguished Professor of Medicine, Chief,
Medicine, Translational Gastroenterology Division of Rheumatology, Allergy and
Unit, Nuffield Department of Medicine, Clinical Immunology, The Jack and
University of Oxford; and Oxford NIHR Donald Endowed Professor,
Biomedical Research Centre, Oxford, UK University of California School of Medicine,
Davis, California, USA
Kirsten Muri Boberg MD, PhD
Professor Norwegian PSC Research Alessio Gerussi MD
Center and Section for Gastroenterology, Research Fellow, Division of
Department of Transplantation Medicine, Gastroenterology, University of Milan
Division of Surgery, Inflammatory Medicine Bicocca, Milan, Italy
and Transplantation, Oslo University
Hospital Rikshospitalet; and Institute of Núria Guañabens MD
Clinical Medicine, Faculty of Medicine, Professor of Medicine, University of
University of Oslo, Oslo, Norway Barcelona; and Senior Consultant Physician,
Department of Rheumatology,
Marco Carbone MD, PhD
Metabolic Bone Diseases Unit,
Clinical Lecturer in Gastroenterology,
Hospital Clínic, Institut d’Investigacions
Program for Autoimmune Liver Diseases,
Biomèdiques August Pi Sunyer (IDIBAPS),
Department of Medicine and Surgery,
Centro de Investigación Biomédica en Red
University of Milan Bicocca, Milan, Italy
de Enfermedades Hepáticas y Digestivas
(CIBERehd), Barcelona, Spain
Olivier Chazouillères MD
Director Department of Hepatology,
AP‐HP, Hôpital Saint‐Antoine, Service Aliya Gulamhusein MD, MPH
d’Hépatologie; Reference Center for Toronto Centre for Liver Diseases, Toronto
Inflammatory Biliary Diseases and General Hospital, University Health
Autoimmune Hepatitis; INSERM, UMR‐S Network; and Division of Gastroenterology,
938, CDR Saint‐Antoine; and Sorbonne University of Toronto,
Université, Paris, France Toronto, Ontario, Canada
Emma L. Culver PhD, MRCP Peter C. Hayes MBChB, BMSc, MD, PhD
Consultant Hepatologist and Senior Head of Department, Professor of Medicine
Lecturer, Translational Gastroenterology and Consultant Physician,
Unit, Nuffield Department of Medicine, Department of Hepatology,
University of Oxford; and Oxford NIHR Royal Infirmary of Edinburgh,
Biomedical Research Centre, Oxford, UK Edinburgh, UK
xviii Contributors
Michael A. Heneghan MD, MMedSc, FRCPI Piotr Milkiewicz MD, DSc, MRCP(UK)
Clinical Director for Liver Services and Professor of Medicine/Consultant Physician,
Consultant Hepatologist and Professor of Head, Liver and Internal Medicine Unit,
Hepatology, Institute of Liver Studies, King’s Medical University of Warsaw, Warsaw,
College Hospital, Denmark Hill, London, UK Poland
David Jones MD, PhD, OBE Carlos Moctezuma‐Velázquez MD

Dean for the NIHR Academy, Professor of Mount Sinai Liver Cancer Program, Division
Liver Immunology, Newcastle University; of Liver Diseases, Tisch Cancer Institute,
and Honorary Consultant Hepatologist, Icahn School of Medicine at Mount Sinai,
Newcastle upon Tyne Hospitals, Newcastle New York, USA
upon Tyne, UK Division of Gastroenterology and Liver Unit,
University of Alberta Hospital, Edmonton,
Patrick S.C. Leung PhD Alberta, Canada
Department of Medicine, Teikyo University
School of Medicine, Tokyo, Japan Aldo J. Montano‐Loza MD, PhD
Division of Gastroenterology and Liver Unit,
Ying Qi (Tammy) Li BSc University of Alberta Hospital, Edmonton,
Division of Gastroenterology and Alberta, Canada
Hepatology, University of Alberta,
Edmonton, Alberta, Canada Albert Parés MD
Professor of Medicine, University of
Andrew L. Mason MBBS, FRCPI, FAASLD Barcelona; and Senior Consultant Physician,
Professor of Medicine, and Director of Liver Unit, Hospital Clínic, Institut
the Applied Genomic Core, Division d’Investigacions Biomèdiques August Pi
of Gastroenterology and Hepatology, Sunyer (IDIBAPS), Centro de Investigación
University of Alberta, Edmonton, Alberta, Biomédica en Red de Enfermedades
Canada Hepáticas y Digestivas (CIBERehd),
Barcelona, Spain
Patrick McKiernan MD
Children’s Hospital of Pittsburgh, University Marcus C. Robertson MD, PhD, FRACP
of Pittsburgh Medical Center, Pittsburgh, Gastroenterologist and Hepatologist,
Pennsylvania, USA Monash Health, Melbourne; and Adjunct
Senior Lecturer, Monash University
George F. Mells PhD, MRCP Department of Medicine, Clayton,
Senior Lecturer, Division of Australia
Gastroenterology and Hepatology,
Department of Medicine, and Academic Isaiah G. Roepe BS
Department of Medical Genetics, University Medical Student, Baylor College of
of Cambridge, Cambridge, UK Medicine, Houston, Texas,
USA
Giorgina Mieli‐Vergani MD, PhD, FRCP,
FRCPCH, FAASLD Mario Strazzabosco MD, PhD
Emeritus Professor of Paediatric Hepatology, Professor of Medicine; Clinical Program
Honorary Consultant Paediatric Liver, GI Leader, Liver Cancer Program; and Deputy
& Nutrition Centre, Mowat Labs, King’s Director, Yale Liver Center, Digestive
College London Faculty of Life Sciences Disease Section, Yale Liver Center, Yale
and Medicine at King’s College Hospital, University School of Medicine, New Haven,
Denmark Hill, London, UK Connecticut, USA
Contributors xix
Atsushi Tanaka MD Mette Vesterhus MD, PhD

Associate Professor, Department of Consultant Physician, Norwegian
Medicine, Teikyo University School of PSC Research Center, Department of
Medicine, Tokyo, Japan Transplantation Medicine, Division of
Surgery, Inflammatory Medicine and
Benedetta Terziroli Beretta‐Piccoli MD Transplantation, Oslo University Hospital,
Epatocentro Ticino, Via Soldino 5, 6900 Rikshospitalet, Oslo; and Department of
Lugano, Switzerland Clinical Science, University of Bergen,
Institute of Liver Studies, Mowat Labs, Bergen, Norway
King’s College Hospital, Denmark Hill,
London, UK John M. Vierling MD, FACP, FAASLD, AGAF
Professor of Medicine and Surgery, Chief of
Eleni Theocharidou MSc, PhD Hepatology, Director of Baylor Liver Health,
Senior Clinical Fellow, Institute of Director of Advanced Liver Therapies,
Liver Studies, King’s College Hospital Program Director, Hepatology and Liver
NHS Foundation Trust, Denmark Hill, Transplantation Fellowship
London, UK Division of Abdominal Transplantation,
Michael E. DeBakey Department of
Diego Vergani MD, PhD, FRCP, FRCPath, Surgery, Section of Gastroenterology and
FAASLD Hepatology, Margaret M. and Albert B.
Emeritus Professor of Liver Alkek Department of Medicine, Baylor
Immunopathology, Honorary Consultant, College of Medicine, Houston, Texas, USA
Institute of Liver Studies, Mowat Labs,
King’s College London Faculty of Life Ewa Wunsch MD
Sciences and Medicine at King’s College Translational Medicine Group, Pomeranian
Hospital, Denmark Hill, London, UK Medical University, Szczecin, Poland
Abbreviations
AASLD American Association for the BCOADC branched‐chain 2‐oxo acid

Study of Liver Diseases dehydrogenase complex
ABC ATP‐binding cassette BCR B‐cell receptor
ABD autoimmune biliary disease BEC biliary epithelial cell
ACA anti‐centromere antibody BET bromodomain and
ACR American College of extra‐terminal
Rheumatology BLC B‐lymphocyte
ADCC antibody‐dependent cell‐ chemoattractant
mediated cytotoxicity BMD bone mineral density
AFLP acute fatty liver disease of Breg B regulatory cell
pregnancy BSEP bile salt export pump
AIH autoimmune hepatitis BTSC biliary tree stem/progenitor
AILD autoimmune liver disease cell
AIP autoimmune pancreatitis C′ complement
AIRE autoimmune regulator CA19‐9 carbohydrate antigen 19‐9
AISC autoimmune sclerosing CC cryptogenic cirrhosis
cholangitis CCH cryptogenic chronic hepatitis
ALF acute liver failure CD celiac disease
ALP alkaline phosphatase CDCA chenodeoxycholic acid
ALT alanine aminotransferase CI confidence interval
AMA anti‐mitochondrial antibody CNSDC chronic non‐suppurative
ANA anti‐nuclear antibody destructive cholangitis
ANCA anti‐neutrophil cytoplasmic CREB cAMP response element
antibody binding protein
APC antigen‐presenting cell CS Cesarean section
APECED autoimmune polyendocrinop- CT computed tomography
athy/candidiasis/ectodermal CTD connective tissue disease
dystrophy CTL cytotoxic T lymphocyte
APRI AST‐to‐platelet ratio index CTLA cytotoxic T‐lymphocyte
APS‐1 autoimmune polyendocrine antigen
syndrome type 1 CYP cytochrome P
ASC autoimmune sclerosing D3 1,25‐dihydroxyvitamin D3
cholangitis DAG diacylglycerol
ASGPR asialoglycoprotein receptor DAMP damage‐associated molecular
ASMA anti‐smooth muscle antibody pattern
AST aspartate aminotransferase DC dendritic cell
ATX autotaxin DCD donation after circulatory
BA bile acid death
xxii Abbreviations
DCreg regulatory dendritic cell HLA human leukocyte antigen

DDLT deceased donor liver HMG‐CoA 3‐hydroxy‐3‐methylglutaryl‐
transplant coenzyme A
DILI drug‐induced liver injury HPF high‐power field
DOCK1 dedicator of cytokinesis 1 HpSC hepatic stem/progenitor cell
(protein) HR hazard ratio
DXA dual‐energy X‐ray HRQoL health‐related quality of life
absorptiometry HSC hepatic stellate cell
EASL European Association for the HSEC hepatic sinusoidal endothelial
Study of the Liver cell
EGD esophagogastroduodenoscopy HSV herpes simplex virus
EHAID extrahepatic autoimmune IAC IgG4‐associated cholangitis
disease IAH International Autoimmune
ELF enhanced liver fibrosis (test) Hepatitis (score)
ELISA enzyme‐linked immunosor- IAIHG International IAH Group
bent assay IBD inflammatory bowel disease
ELTR European Liver Transplant ICU intensive care unit
Registry IgG4‐RD IgG4‐related disease
ENA extractable nuclear antigen IIF indirect immunofluorescence
eQTL expression quantitative trait IL interleukin
locus INR international normalized ratio
ERCP endoscopic retrograde IPEX immunodysregulation polyen-
cholangiopancreatography docrinopathy enteropathy
eRNA enhancer RNA X‐linked (syndrome)
EULAR European League Against IRF interferon regulatory factor
Rheumatism ISRE interferon‐stimulated response
EUS endoscopic ultrasound element
EVL endoscopic variceal ligation JAK Janus kinase
FDA Food and Drug Administration LAK lymphokine‐activated killer
FISH fluorescence in situ (cell)
hybridization LC1 liver cytosol type 1 (antibody)
FNA fine needle aspiration LD linkage disequilibrium
FoxP forkhead box P3 LDL low‐density lipoprotein
FXR farnesoid X receptor LDLT living donor liver transplant
GEF guanine nucleotide exchange LE lupus erythematosus
factor LFT liver function test
GGT gamma‐glutamyltranspepti- LKM1 liver kidney microsomal type 1
dase (or gamma‐ (antibody)
glutamyltransferase) lncRNA long non‐coding RNA
GLP glucagon‐like peptide LPA lysophosphatidic acid
GWAS genome‐wide association LPS lipopolysaccharide
studies LSEC liver sinusoidal endothelial cell
HAI hepatic/hepatitis activity LT liver transplantation
index MAGUK membrane‐associated guanyl-
HBRV human betaretrovirus ate kinase
HBV hepatitis B virus MAIT mucosal invariant T (cell)
HCC hepatocellular carcinoma MAPK mitogen‐activated protein
HCV hepatitis C virus kinase
HDL high‐density lipoprotein MDR multidrug resistance
Abbreviations xxiii
MELD Model for End‐stage Liver OR odds ratio

Disease (score) OS overlap syndrome
MHC major histocompatibility PAMP pathogen‐associated molec-
complex ular pattern
MIF macrophage migration pANCA perinuclear anti‐neutrophil
inhibitory factor cytoplasmic antibody
miRNA microRNA pANNA perinuclear anti‐neutrophil
MMF mycophenolate mofetil nuclear antibody
6‐MMP 6‐methylmercaptopurine PBC primary biliary cholangitis
MMTV mouse mammary tumor virus PBG peribiliary gland
6‐MP 6‐mercaptopurine PD‐1 programmed cell death 1
mQTL methylation quantitative trait PDC pyruvate dehydrogenase
locus complex
MRC magnetic resonance PET positron emission tomography
cholangiography PKC protein kinase C
MRCP magnetic resonance P1NP procollagen type 1 N‐terminal
cholangiopancreatography propeptide
MRP multidrug resistance protein PPAR peroxisome proliferator‐
MS multiple sclerosis activated receptor
mTEC medullary thymic PRR pattern recognition receptor
epithelial cell PSC primary sclerosing cholangitis
NAC N‐acetylcysteine PXR pregnane X receptor
NAFLD non‐alcoholic fatty liver QALY quality‐adjusted life‐year
disease RA rheumatoid arthritis
NANA nuclear anti‐neutrophil RAG recombination activating gene
antibody rAIH recurrent autoimmune
NASH non‐alcoholic steatohepatitis hepatitis
NET neutrophil extracellular trap RANKL receptor activator of nuclear
NFκB nuclear factor kappa B factor kappa‐B ligand
NICE National Institute for RDC reactive ductular cell
Health and Care Excellence RF rheumatoid factor
(UK) ROR receptor‐related orphan
NICU neonatal intensive care unit receptor
NK natural killer (cell) rPBC recurrent primary biliary
NKT natural killer T (cell) cholangitis
NLRP3 nucleotide‐binding oligomeri- rPSC recurrent primary sclerosing
zation domain (NOD)‐like cholangitis
receptor P3 RR relative risk
NOD non‐obese diabetic RTK receptor tyrosine kinase
NSAID non‐steroidal anti‐ RXR retinoid X receptor
inflammatory drug SASP senescence‐associated secre-
NTCP sodium taurocholate cotrans- tory phenotype
porting polypeptide SCID severe combined
OATP organic anion‐transporting immunodeficiency
polypeptide SE super enhancer
OCA obeticholic acid SH2B SH2B adapter protein 3
OGDC 2‐oxoglutarate dehydrogenase SLA soluble liver antigen
complex SLE systemic lupus erythematosus
OPG osteoprotegerin SMA smooth muscle antibody
xxiv Abbreviations
SNP single nucleotide polymorphism TORC transducer of regulated CREB

SS Sjögren syndrome activity
SSc systemic sclerosis TPMT thiopurine
STAT signal transducer and activator S‐methyltransferase
TCR T‐cell receptor TRAIL TNF‐apoptosis‐inducing
T1DM type 1 diabetes mellitus ligand
TF transcription factor Treg T regulatory cell
Tfh T follicular helper (cell) TRM tissue‐resident memory T
TGF transforming growth factor (cell)
6‐TGN 6‐thioguanine nucleotide UC ulcerative colitis
TIPS transjugular intrahepatic UDCA ursodeoxycholic acid
portosystemic shunt UGT uridine diphosphate
TLR Toll‐like receptor glucuronosyltransferase
TLS tertiary lymphoid structure ULN upper limit of normal
TNF tumor necrosis factor UTI urinary tract infection
TNFAIP3 TNF‐induced protein 3 VB variceal bleeding
TNIP1 TNFα‐induced protein VDR vitamin D receptor
3‐interacting protein 1 WD Wilson disease
1
Introduction
The Paradigm and Paradox of Liver Autoimmunity
M. Eric Gershwin
Division of Rheumatology, Allergy and Clinical Immunology, University of California School of Medicine, Davis, CA, USA
From an evolutionary perspective, the (TCR). This magnitude of diversity is a result

immune system was one of the first physio of evolutionary pressure to protect our species
logical components of life – even plants from the outside world and therefore pro
have some form of a response against patho tect us from a myriad of potential pathogens.
gens. The evolution of immunity was of Interestingly, however, this diversity is not
course to protect species from the outside limited to humans. People, dogs, cats, fish
world and this is as much true of people as it and even earthworms exhibit evidence of an
is of earthworms. However, on occasion, the immune system, but the pleiotropic nature of
immune system attacks one’s own tissues, a immune responses becomes more sophisti
process known as autoimmunity. Autoim cated in higher animals.
munity affects a variety of species and clearly Ironically, despite the sophistication of
increases with inbreeding and genetic diver the immune system, our recognition of its
sity. Although we now catalog more than 100 importance and the multitude of ways in
different autoimmune diseases, the etiology of which it can harm the body is a relatively
virtually all of autoimmunity remains enig recent development. Although many physi
matic other than the concept, as noted below, cians can be credited with establishing
of genetic predisposition and one or more immunology as a science, it is clear that the
environmental stimuli. Indeed, understanding first giant of immunology was Clemens von
the etiology of autoimmunity, while a major Pirquet, born in 1874 in Vienna. Von Pirquet
problem, has not deterred effective treat devoted his early work to understanding
ments. In the case of rheumatoid arthritis, we the immune response to scarlet fever and, in
have very effective treatments directed at the collaboration with multiple coworkers, was
mediators of inflammation without knowing among the first to identify precipitants in
what initiates the entire process. human sera. He also coined the word “allergy”
To put this in perspective, it should be which essentially means deviation from self.
re‐emphasized that the immune system is However, it was not until the work of Paul
the most pleiotropic and promiscuous physio Ehrlich on hemolytic anemia that scientists
logical component of the human body. Indeed, began to appreciate the possibility that our
the potential individual repertoire may include pleiotropic immune system could attack
up to 1013 arrangements of the T‐cell receptor our own body. Ironically, the first textbook
Autoimmune Liver Disease: Management and Clinical Practice, First Edition. Edited by James Neuberger
and Gideon M. Hirschfield.
© 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.
2 Introduction
devoted to autoimmunity was not published In this volume, Neuberger and Hirschfield
until 1963, or over 60 years after Ehrlich’s have taken up the challenge of providing the
initial observations. This book, authored by scientific basis for liver autoimmunity. Liver
Mackay and Burnett, was published in a autoimmunity has often been a stepchild
series of monographs entitled American compared with other autoimmune diseases.
Lectures in Living Chemistry (Charles The pharmaceutical industry has made enor
Thomas Publishers, Springfield, Illinois). mous efforts in helping patients with
This textbook contains 12 chapters devoted rheumatoid arthritis, with psoriasis, and of
to specific autoimmune diseases. The con course with human viral hepatitis, but the
cept of the forbidden clone was discussed commitment to liver autoimmunity, not only
and, of particular interest to liver specialists, in industry but in academia, has been modest
the terms active chronic and lupoid hepatitis at best. And yet the same generic factors that
were introduced. Absent from this book was lead to both organ‐specific and systemic auto
of course the concept of lymphoid subpopu immune diseases are at play in the liver auto
lations, the genetics of the immune response, immune syndromes, since the development of
major histocompatibility complex (MHC) disease requires a genetic predisposition and
restriction and, especially, discussion of an environmental stimulus. This is a feature
maternal–fetal immune homeostasis. These that I have coined “bad genes and bad luck.”
omissions are easily forgiven when it is real However, this volume goes beyond descriptive
ized that the tools used by immunologists in immunobiology. It provides a rigorous anal
the 1960s had not changed significantly in ysis of multiple critical areas of why patients
over three decades and relied particularly on become susceptible and why their natural his
diffusion and electrophoresis. tory of disease may vary, including for example
This issue of maternal–fetal immune the roles of genetics and epigenetics, environ
homeostasis is critical in the introduction of mental exposures, the specific and usual
liver autoimmunity. Whether it is a human or clinical challenges faced by patients with liver
a mouse, the fetus is a privileged site and disease, and of course helping the particularly
spared from any immune attack from the difficult and challenging patient (i.e. liver dis
mother. The basis for this tolerance resides in ease during pregnancy). This volume also fills
the liver. The liver was known even in ancient a major void in explaining the role of liver
Greek mythology to be an organ capable of transplantation by placing not only its
regeneration. Prometheus stole from the scientific significance in perspective, but also
Gods, giving it away to mortals. As a result, he tackling the difficult subject of disease recur
was chained to a rock by Zeus, where each day rence following liver transplant. There remain
an eagle pecked out his liver. Presumably, many controversies and multiple unanswered
Prometheus survived because of liver regen questions, and the book ends with an attempt
eration. However, with that singular exception, to make sense of overlap and crossover syn
and the use of liver in the everyday kitchen, dromes, and the relationship to extrahepatic
the liver as an organ responsible for immu abnormalities in patients with autoimmune
nity was ignored. Indeed, a major basis liver disease. Finally, it focuses on the
for maternal–fetal tolerance is the portal most important element, which is clinical
circulation. It is therefore ironic that the organ management and how we may help our
which is most important for the critical patients. At the end of the day, whatever
element of immune tolerance can itself the scientific interest may be, the only
become a victim of autoimmunity, manifest in fundamental question of value is whether our
such diseases as autoimmune hepatitis, pri work and our publications will have a positive
mary biliary cholangitis, primary sclerosing impact on the health of people. This volume
cholangitis and IgG4‐associated disease. should fill such a need.
3
Section I
Scientific Basis of Clinical Autoimmune Liver Diseases

5
Introduction to the Physiology, Immunology and Pathology

of the Liver and Biliary Tree
Marco Carbone1 and Mario Strazzabosco2
1
Division of Gastroenterology, University of Milan Bicocca, Milan, Italy
2
Digestive Disease Section, Yale Liver Center, Yale University School of Medicine, New Haven, CT, USA
Abstract
The liver has many functions including metabolic homeostasis, disposal of endotoxins and
xenotoxins, metabolism of bilirubin and urea, and bile formation and secretion. The process of bile
formation depends on the liver synthesis and the canalicular secretion of bile acids. Besides their
roles in dietary lipid absorption and cholesterol homeostasis, bile acids also play a key role as
signaling molecules in the regulation of hepatic metabolism and energy homeostasis. The regener‑
ative ability of the liver lies in the multiple niches of biliary tree stem cells.
Keywords bile formation; carbohydrate metabolism; hepatocyte; cholangiocytes lipid
etabolism; metabolic zonation; protein metabolism; bile acid.
m
Key Points
●● The liver is largely composed of hepato‑ ●● Besides their roles in dietary lipid absorption
cytes and biliary epithelial cells or cholan‑ and cholesterol homeostasis, bile acids
giocytes; both hepatocytes and intrahepatic (BAs) also play a key role as signaling mole‑
cholangiocytes differentiate from bipotent cules in the regulation of hepatic metabo‑
liver progenitors, the hepatoblasts. lism and energy homeostasis.
●● The liver has many functions, among which ●● BAs also have hormonal signaling
metabolic homeostasis, disposal of endo‑ functions and interact with dedicated
toxins and xenotoxins, metabolism of bili‑ receptors such as the nuclear receptor
rubin and urea, and bile formation and farnesoid X receptor and G protein‐coupled
secretion are just examples. The liver also receptors, which regulate BA homeostasis
produces fundamental circulating proteins and BA‐induced injury and/or
and clotting factors and hormones. In inflammation.
addition, the liver receives and processes ●● Multiple niches of biliary tree stem/
the blood coming from the intestine and has progenitor cells reside in different locations
a fundamental role in innate immunity. along the human biliary tree and within the
6 Section I Scientific Basis of Clinical Autoimmune Liver Diseases
liver parenchyma and have a key role in rise to the ductular reaction, a major driver
regeneration of the liver. of the progression of hepatic fibrosis.
●● Cholangiocytes modify the primary bile ●● Cholangiocytes also contribute to the
by secretion of chloride and bicarbonate immune response through antigen pre‑
fluid. This is a major protective mecha‑ sentation to immune cells, being a target
nism for the biliary tree. of immune‐mediated aggression or being
●● Cholangiocytes, a barrier and secretory epi‑ the initiators of an inflammatory reaction
thelium in normal conditions, activate and/ that then progresses to adaptive immune
or proliferate following a liver insult and give activation.
Liver Cell Types and Organization transporters (e.g. ileal BA transporter), and

exchangers (e.g. Cl−/HCO3− or Na+/H+
Liver cells can be classified into the following exchangers). The large cholangiocytes are
groups: located at the level of interlobular and major
bile ducts and they express several different
●● parenchymal cells, which include hepato‑ ion channels and transporters at the basolat‑
cytes and biliary epithelial cells (BECs); eral or apical domain; they are believed to be
●● sinusoidal cells, which include hepatic mostly involved in secretin/cyclic AMP‐regu‑
sinusoidal endothelial cells and Kupffer lated bile secretion. Smaller bile duct branches
cells; and include terminal cholangioles and ductules or
●● perisinusoidal cells, which include the hepatic canals of Hering; the latter is a channel located
stellate cells (HSCs) and the pit cells [1]. at the ductular–hepatocellular junction, lined
partly by hepatocytes and partly by cholangio‑
The hepatocytes, which comprise approxi‑ cytes, and represents the physiologic link bet‑
mately 60% of the liver cell mass, are epithe‑ ween the biliary tree and the hepatocyte
lial cells with two distinct domains on their canalicular system extended within the lobule.
plasma membrane: (i) the sinusoidal (or Their secretory function is believed to be
basolateral) surface, facing the sinusoids, in mostly regulated by intracellular calcium.
contact with plasma through the fenestrated More recently, other important biological
endothelium of the sinusoids, which allows a properties restricted to cholangiocytes lining
bidirectional flow of liquids and solutes the smaller bile ducts have been reported,
though the space of Disse; and (ii) the cana‑ especially with regard to their p lasticity (the
licular (or apical) surface, which encloses the ability to undergo limited phenotypic changes),
bile ductules and represents the beginning of reactivity (the ability to participate in the
the biliary drainage system. inflammatory reaction to liver damage), and
The BECs (or cholangiocytes) are the epithe‑ ability to behave as liver progenitor cells.
lial cells lining the biliary tree. The biliary epi‑ The hepatic sinusoidal endothelial cells
thelium shows a morphologic heterogeneity (HSECs) represent 20% of the total liver cells.
that is associated with a variety of functions Unlike capillary endothelial cells, HSECs do
performed at the different levels of the biliary not form intracellular junctions and simply
tree. Other than funneling bile into the overlap one another. They can secrete pros‑
intestine, BECs are actively involved in taglandins and cytokines [2]. They are also
bile production by performing both absorptive responsible for maintaining a cell niche that
and secretory functions via various m embrane favors the quiescence of HSCs.
transport mechanisms including channels The Kupffer cells are specialized tissue
(e.g. water channels and chloride channels), macrophages and account for up to 90% of
Chapter 1 Physiology, Immunology and Pathology of the Liver and Biliary Tree 7
the total population of fixed macrophages in bone marrow cells. Bilirubin formed in the
the body. These are macrophages attached to monocytic–macrophage cell system of liver,
the endothelial lining of the sinusoid, in spleen and bone marrow and some of the bil‑
greater numbers in the periportal areas. irubin formed in the hepatocytes from
They are responsible for removing old and hepatic heme are released into plasma where
damaged blood cells or cellular debris, bilirubin is bound to albumin at high‐affinity
bacteria, viruses, parasites, and tumor cells. binding sites. In normal conditions, only a
The HSCs, also known as Ito cells or lipo‑ minimal amount of bilirubin is unbound in
cytes, are mesenchymal cells and represent plasma. An increase in free bilirubin would
the major storage site of retinoids in healthy allow the pigment to enter tissues where it
livers. They lie within the subendothelial can have toxic effects; this is what is observed
space, and their long cytoplasmic extensions in neonates with defective conjugation and in
have close contact with parenchymal cells Crigler–Najjar syndrome, when diffusion of
and sinusoids, where they may regulate unbound bilirubin into the brain can cause
blood flow and hence influence portal kernicterus.
hypertension. HSC activation is the central In normal conditions, bilirubin is effi‑
event in hepatic fibrosis. During hepatocyte ciently taken up by the liver whereas the
injury, HSCs proliferate, migrate to zone albumin remains in plasma. In the liver,
three of the acinus, change to a myofibro‑ bilirubin is bound initially to glutathione‐
blast‐like phenotype, and produce collagen S‐transferase, then glucuronidated and
and laminin. excreted into bile. Microsomal bilirubin
Finally, the pit cells represent the natural uridine diphosphate glucuronosyltransfer‑
killer (NK) cells of the liver and are located ase (UGT) is the enzyme that converts
within the sinusoids. Pit cells show unconjugated bilirubin to conjugated bili‑
spontaneous cytotoxicity against tumor‐ and rubin monoglucuronide and diglucuronide.
virus‐infected hepatocytes. Biliary excretion of the glucuronide is
mediated by the adenosine triphosphate
(ATP)‐dependent multidrug resistance
Hepatic Metabolism protein (MRP)‐2 and this is the rate‐lim‑
iting factor in the transport of b ilirubin
The liver is the site of many metabolic path‑ from plasma to bile. Bilirubin diglucuro‑
ways. It stores and makes available many nide is not reabsorbed from the small
nutrients as energy sources. In turns, the intestine; in the colon it may be hydrolyzed
metabolic function of the liver is regulated by by bacterial β‐glucuronidases, producing
hormones secreted by the pancreas, adrenal urobilinogens and urobilin, which are
gland, and thyroid. excreted in the stool or urine.
Bilirubin Metabolism and Transport Carbohydrate Metabolism
Bilirubin is an end‐product of heme catabo‑ The liver plays a key role in carbohydrate
lism. Two enzymes are involved in bilirubin metabolism. It maintains carbohydrate stores
formation: the microsomal heme oxygenase by synthesizing glycogen and generating
converts heme to biliverdin and a cytosolic glucose from precursors such as lactate, pyru‑
reductase subsequently reduces biliverdin to vate, and amino acids. Glycogen stored in the
bilirubin. liver is the main source of rapidly available
The majority (up to 85%) of heme is derived glucose for the whole organism. In the fed
from hemoglobin and only a small fraction state, glycogen synthesis occurs preferentially
from other heme‐containing proteins such as in the perivenous hepatocytes whereas in the
cytochrome P450, myoglobin and immature fasting state, glucose release via glycogenolysis
and gluconeogenesis initially occurs in peri‑ lesterol into the plasma. In patients with
portal hepatocytes. The liver glycogen stores severe chronic cholestasis who are however
contain up to a 2‐day supply of glucose, after malnourished, for example in premature
which glucogenesis occurs mainly from lac‑ ductopenic variant of primary biliary cholan‑
tate. In acute liver failure, the blood glucose gitis (PBC) or in carcinomatous biliary
level may drop whereas this is infrequent in obstruction, the serum cholesterol may be
chronic liver disease, where it is more common lower. Elevated cholesterol levels are clini‑
to observe hyperglycemia and insulin resis‑ cally associated with skin xanthomas and
tance. This may be related to decreased xanthelasma. Hypercholesterolemia is not
glucose uptake by muscle and reduced gly‑ consistently associated with subclinical ath‑
cogen storage in the liver and muscle. erosclerosis in PBC, but should be treated if
other risk factors for cardiovascular disease
are also present.
Lipid Metabolism
In addition, the liver has other roles in lipid
A major role of the liver in lipid metabolism metabolism such as oxidizing triglycerides to
is to synthesize large quantities of cholesterol produce energy; synthesizing lipoproteins;
and phospholipids, many of which are pack‑ and converting excess carbohydrates and
aged with lipoproteins and made available to proteins into fatty acids and triglyceride,
the rest of the body. Free cholesterol also which are then exported and stored in
derives from the uptake of chylomicron rem‑ adipose tissue.
nants and lipoproteins from the circulation.
BAs are synthesized from free cholesterol,
Protein Metabolism
and both BA and cholesterol are secreted
into bile. Bile provides a major route for Amino acids are derived from the diet and
cholesterol excretion. from tissue breakdown and they reach the
The rate‐limiting step of cholesterol liver via the portal vein. Many critical aspects
synthesis is the conversion of 3‐hydroxy‐3‐ of protein metabolism occur in the liver, such
methylglutaryl‐CoA (HMG‐CoA) to meval‑ as the deamination and transamination of
onate by the enzyme HMG‐CoA reductase, amino acids, followed by conversion of the
which is located almost exclusively in peripor‑ non‑nitrogenous parts of these molecules to
tal cells. Synthesis is increased in biliary duct glucose or lipids. Several of the enzymes used
obstruction, terminal ileal resection, biliary or in these pathways, e.g. alanine and aspartate
intestinal lymph fistula, and with medications aminotransferases, are commonly measured
such as colestyramine, corticosteroids and in serum to assess liver damage.
thyroid hormones. Cholesterol synthesis is The liver is also responsible for a number
inhibited by BAs, cholesterol feeding, fasting, of vital metabolic processes, including the
and medications such as fibrates, nicotinic removal of ammonia (an important factor in
acid, and statins. the development of hepatic encephalopathy)
During chronic cholestasis, such as in via the synthesis of urea; the synthesis of
primary biliary cholangitis and primary
non‐essential amino acids; and the synthesis
sclerosing cholangitis (PSC) but also in acute of most plasma proteins such as albumin
cholestasis, cholesterol serum levels are (the major plasma protein), fibrinogen, α1‐
increased. This is mainly secondary to reten‑ antitrypsin, haptoglobin, ceruloplasmin,
tion of cholesterol normally excreted in bile transferrin, and several coagulation factors.
but also to increased hepatic synthesis of
cholesterol, reduced plasma lecithin‐choles‑
Metabolic Zonation
terol acyltransferase activity, and regurgita‑
tion of biliary lecithin, which produces a shift The multiple functions of the liver are
of cholesterol from pre‐existing tissue cho‑ facilitated by a functional specialization
of the liver parenchyma, known as meta‑ Hepatic Transport Systems

bolic zonation, where hepatocytes show
different functional and structural char‑ Hepatic uptake and efflux of electrolytes and
acteristics according to their location in solutes involved in bile formation are main‑
the liver acinus. Within each acinus, the tained by distinct transport systems
functional unit in terms of blood flow, expressed at the two surface domains of
blood rich in nutrients and hormones hepatocytes. After canalicular secretion, bile
enters at the portal triad through the composition undergoes further modification
portal vein, mixes with oxygen‐rich blood in the bile canaliculi, involving reabsorption
from the hepatic artery, flows through the and secretion processes maintained by the
sinusoids and eventually exits the lobule apical and basolateral transport systems in
through the central vein. As the blood cholangiocytes (Figure 1.1).
flows through the sinusoids there is
free exchange of nutrients and metabo‑
Basolateral (Sinusoidal) Transporters
lites between blood and hepatocytes.
Functional variation is observed in hepa‑ The uptake of exogenous and endogenous
tocytes based on their location along the compounds from the portal circulation is
portal–central axis. Hepatocytes exhibit facilitated by a number of basolaterally
a distinct gene expression based on their located, sodium‐dependent and sodium‐
location within the acinus, which mani‑ independent transporters. The sodium‐
fests as diverse availability of substrates dependent transporters include the sodium
and concentration of enzymes in different taurocholate cotransporting polypeptide
parts of the acinus. Based on this organi‑ (NTCP), specific to conjugated bile salts and
zation and heterogeneity of hepatocytes, certain sulfated steroids. NTCP accounts for
the acinus comprises three geographical more than 80% of conjugated but less than
areas or zones: periportal or zone 1, midzonal half of unconjugated bile salt uptake. The
or zone 2, and perivenous or zone 3. Hepatocytes sodium‐independent transporters include
in zone 3 contain the drug‐metabolizing P450 several members of the superfamily of
enzymes, have a reduced oxygen supply, organic anion‐transporting polypeptides
receive a higher concentration of any (OATPs). With the exception of OATP2B1,
toxic product of drug metabolism, and the other OATPs exhibit overlapping trans‑
have a reduced glutathione concentration port activities for conjugated and unconju‑
compared with zone 1. This makes them gated bile salts, neutral steroids, steroid
particularly susceptible to drug‐induced sulfates and glucuronides, selected organic
liver injury. Also, hepatocytes in zone 1 cations, and drugs including the antihista‑
receive blood with a high bile salt mine fexofenadine, opioid peptides, digoxin,
concentration and are therefore particu‑ the HMG‐CoA reductase inhibitor pravas‑
larly important in bile salt‐dependent bile tatin, the angiotensin‐converting enzyme
formation, whereas hepatocytes in zone 3 inhibitor enalapril, and the antimetabolite
are important in non‐bile salt‐dependent methotrexate. In addition, the basolateral
bile formation. Functions such as gluco‑ hepatocyte membrane also localizes several
neogenesis, g lycolysis, and ketogenesis ATP‐dependent efflux pumps belonging to
appear to be dependent on the direction the family of MRPs, multispecific trans‑
of blood flow along the sinusoid. For porters for different organic anions. These
others, such as cytochrome P450 activity, have been implicated in the cellular efflux of
the gene t ranscription rate differs bet‑ drugs conjugated with glutathione, glucu‑
ween perivenular and periportal hepato‑ ronic acid, and sulfate (MRP1); the efflux of
cytes. This functional zonation is often bile salts (MRP3); and the transport of nucle‑
lost in liver diseases. oside analog drugs (MRP4).
Alternative BA export
BA BA BA
MRP4
MRP3
OSTα
OSTβ
CAR
PXR
GR
RAR
MRP2 Bili-glu
CAR PPARα
OATP1B1 HNF4 BSEP BA
BA
FXR MDR2/3 PC
BA Uptake Biliary secretion
PPARα
BA ABCG8
HNF4
NTCP SHP PXR Chol
Na+ GR ABCG5
LXR
Cl–
GR AE2
HCO3–
PPARα CYP7A1
BA synthesis
Hepatocyte
Figure 1.1 Transcriptional regulation of hepatocellular bile formation. Expression of hepatobiliary transporters in
hepatocytes determines hepatic bile acid (BA) flux and hepatocellular concentrations of these potentially toxic
metabolites. To ensure the balance between synthesis, uptake and excretion, expression of hepatobiliary
transporters is tightly regulated by nuclear receptors (NRs). NRs provide a network of negative feedback and
positive feed‐forward mechanisms for the control of intracellular concentration of biliary constituents, which are
often also ligands for these NRs. BA‐activated FXR is a central player in this network, that represses hepatic BA
uptake (NTCP) and (via SHP) synthesis (CYP7A1), promotes bile secretion via induction of canalicular transporters
(BSEP, MRP2, ABCG5/8, MDR3), and induces BA elimination via alternative export systems at the hepatic
basolateral (sinusoidal) membrane (OSTα/β). Several NR pathways converge at the level of CYP7A1 as the rate‐
limiting enzyme in BA synthesis. CAR and PXR facilitate adaptation to increased intracellular BA concentrations by
upregulation of alternative hepatic export routes (MRP3 and MRP4) and induction of detoxification enzymes (not
shown). Together with RAR, these xenobiotic receptors also regulate the canalicular expression of MRP2.
Cholesterol sensor LXR promotes biliary cholesterol excretion via ABCG5/8. Stimulation of AE2 expression by GR
stimulates biliary bicarbonate secretion thus reducing bile toxicity. Green arrows indicate stimulatory and red
lines suppressive effects on target genes. Bili‐glu, bilirubin glucuronide; BSEP, bile salt export pump; CAR,
constitutive androstane receptor; CYP7A1, cholesterol 7α‐hydroxylase; FXR, farnesoid X receptor; GR,
glucocorticoid receptor; HNF4, hepatocyte nuclear factor 4; LXR, liver X receptor; MDR3, multidrug resistance
protein 3, phospholipid flippase; MRP2, multidrug resistance‐associated protein 2; MRP3, multidrug resistance‐
associated protein 3; MRP4, multidrug resistance‐associated protein 4; NTCP, sodium taurocholate cotransporting
polypeptide; OSTα/β, organic solute transporter alpha and beta; PC, phosphatidylcholine; PXR, pregnane X
receptor; PPARα, peroxisome proliferator‐activated receptor alpha; RAR, retinoic acid receptor; SHP, small
heterodimer partner. Source: Halilbasic et al. [3]. Reproduced with permission of Elsevier.
Hepatocellular transporters are subject (FXR)‐mediated signaling, thereby pre‑

to extensive transcriptional and posttran‑ venting the hepatocyte from further accu‑
scriptional regulation, allowing for adapta‑ mulating toxic bile salts. Likewise, the
tional changes in response to the expression of OATP1B1 is downregulated.
intracellular accumulation of bile salts. In contrast, cholestasis leads to FXR‐
During cholestasis, the NTCP is sup‑ mediated activation of hepatic OATP1B3,
pressed through farnesoid X receptor which might constitute an escape mechanism
promoting the hepatocellular clearance of leukotriene C4, and divalent bile salt conju‑
xenobiotics during cholestasis. gates, as well as drug substrates such as
In addition to its role in the uptake of chemotherapeutic agents and antibiotics.
conjugated BAs, NTCP also plays a key role MDR1 contributes to the canalicular excre‑
in hepatitis B and hepatitis D virus entry into tion of drugs and other xenobiotics into bile,
hepatocytes; and, recently, NTCP has also although its exact contribution has yet to be
been shown to modulate hepatitis C virus established. Its broad substrate specificity
infection of hepatocytes by regulating innate and its physiologic expression in various tis‑
antiviral immune responses in the liver. As sues with excretory and protective functions
such NTCP has been established as a novel make MDR1 one of the major determinants
antiviral target. of drug disposition and toxicity. Substrates
are neutral and positively charged organic
compounds and include various chemother‑
Apical (Canalicular) Transporters
apeutic and immunosuppressant agents,
Various ATP‐binding cassette (ABC) trans‑ antiarrhythmic drugs, HIV protease inhibi‑
porters mediate the secretion of bile salts tors, and antifungals.
and xenobiotics across the canalicular Transcriptional regulation of BSEP and
membrane of hepatocytes. These include MDR3 is mediated by FXR and their activation
members of the family of multidrug leads to increased bile salt efflux and the
resistance (MDR) P‐glycoproteins such as formation of mixed micelles in the biliary tree
MDR1 (ABCB1), MDR3 (ABCB4), and the during cholestatic episodes, thereby preventing
bile salt export pump (BSEP or ABCB11). the toxic effects of bile salts on hepatocytes and
Within the family of MDR proteins, BSEP cholangiocytes. In addition, FXR has been
and MDR3 are two highly conserved mem‑ shown to induce MRP2 expression, which
bers that are involved in the secretion of might constitute another compensatory mech‑
cholephilic compounds from the liver cell anism during cholestasis. In contrast, MDR1 is
into the bile canaliculus. upregulated via the pregnane X receptor (PXR),
BSEP constitutes the predominant bile salt which in addition to endogenous ligands is
efflux system of hepatocytes and mediates the activated by different xenobiotics. This pathway
cellular excretion of numerous conjugated bile is part of a general cellular mechanism of detox‑
salts such as taurine‐ or glycine‐conjugated ification, because MDR1 is the key transporter
cholate, chenodeoxycholate, and deoxycho‑ protein involved in the cellular efflux of
late. MDR3 works as an ATP‐dependent numerous drugs and xenobiotics.
phospholipid flippase, translocating phospha‑
tidylcholine from the inner to the outer leaflet
of the canalicular membrane. Canalicular Drug Metabolism
phospholipids are then solubilized by canalic‑
ular bile salts to form mixed micelles, thereby The liver has a major role in drug metabo‑
protecting cholangiocytes from the detergent lism. The main hepatocyte enzymes
properties of bile salts. involved in metabolism belong to the
In addition to these processes, MRP2, the cytochrome P450 group, a large family of
only canalicular member of the MDR‐associ‑ related enzymes housed in the smooth
ated protein family, mediates the canalicular endoplasmic reticulum of the hepatocyte.
transport of glucuronidated and sulfated bile Metabolism is often divided into two phases
salts. MRP2 is the main driving force for bile of biochemical reaction. Phase 1 involves
salt‐independent bile flow through canalic‑ reduction, hydrolysis or oxidation of the
ular excretion of reduced glutathione. drug, the latter being the most common
Furthermore, MRP2 transports a wide process. After phase 1 reactions, the
spectrum of organic anions, including bili‑ resulting drug metabolite is still often chem‑
rubin diglucuronide, glutathione conjugates, ically active. Phase 2 metabolism involves
conjugation with glutathione, methyl or heavy metals, as well as exogenous drugs,

acetyl groups, which usually occurs in the xenobiotics and environmental toxins [4].
cytoplasm of the hepatocyte and makes the Lipophilic constituents are in solution in
metabolite more hydrosoluble. This facili‑ mixed micelle composed of BAs, phospho‑
tates excretion as well as decreasing the lipids, and cholesterol.
pharmacologic activity. Some drugs may Bile is essential for several important
undergo just phase 1 or just phase 2 metab‑ functions:
olism, but more often the drug will undergo
phase 1 and then phase 2 sequentially. ●● the excretion of potentially harmful exoge‑
Many factors can affect liver metabolism of nous lipophilic substances, as well as the
drugs. The numbers of hepatocytes and excretion of endogenous substrates such as
enzyme activity can decline, with a reduction bilirubin and bile salts;
in the metabolic potential of the liver, follow‑ ●● the digestion and absorption of lipid in the
ing aging, acute and chronic liver disease, gut by bile salts;
and conditions that affect hepatic blood flow. ●● cholesterol homeostasis, by facilitating
Metabolism can also be altered due to ge‑ intestinal cholesterol absorption and, on
netic deficiency of a particular enzyme and the other hand, promoting cholesterol
secondary to the use of other drugs as well as elimination;
dietary and environmental factors. ●● the excretion of immunoglobulin A (IgA)
Capillarization of sinusoids during chronic and inflammatory cytokines, thus protect‑
liver disease increases the bioavailability of ing the organism from enteric infections;
drugs at high hepatic extraction, possibly ●● signaling properties of the BAs in the liver
increasing the side effects. Drug‐induced and the intestine, which are mediated by
liver injury is a major clinical problem, is nuclear BA receptors such as FXR, PXR
often favored by exposure to a combination and vitamin D receptor (VDR), as well as
of drugs and, at times, may be mediated by by membrane a5b1 integrin, epidermal
immunologic mechanisms. growth factor receptor, and sphingosine‐1‐
phosphate receptor 2.
Our understanding of cholestatic liver d iseases

ile Formation, Secretion
B has been profoundly advanced by the dis‑
and the Enterohepatic covery of nuclear receptors for BA signaling
Circulation and their role in hepatobiliary excretory
function and the adaptive changes counter‑
Bile is a complex secretion that originates acting the liver injury caused by retained,
from hepatocytes and is modified distally by potentially toxic, and proinflammatory BAs.
absorptive and secretory transport systems Ligand‐activated nuclear receptors such as
in the bile duct epithelium. Bile formation by FXR control a broad range of metabolic
the hepatocytes involves secretion of osmot‑ processes, including hepatic BA transport and
ically active inorganic and organic anions metabolism, lipid and glucose metabolism,
into the canalicular lumen, followed by drug disposition, as well as liver regeneration,
passive water movement. Bile then enters the inflammation, fibrosis, cell differentiation,
gallbladder where it is concentrated or is and tumor formation. Moreover, FXR has
delivered directly to the bowel. anti‐inflammatory and immunomodulatory
Bile comprises about 95% water in which actions and controls intestinal integrity and
are dissolved a number of endogenous con‑ permeability, as well as gut microbiota.
stituents, including bile salts, bilirubin, Conversely, the gut microbiota metabolizes
phospholipid, cholesterol, amino acids, ste‑ BAs, with formation of secondary BAs that, in
roids, enzymes, porphyrins, vitamins, and turn, modulate BA signaling. Based on these
broad physiologic effects in the liver and secondary bile salt species are contained in
intestine, drugs targeting FXR and TGR5 there‑ human bile, although primary bile salts are
fore open important perspectives for pharma‑ usually predominant.
cotherapy of cholestatic and metabolic liver
disorders, including the complications of liver
Enterohepatic Bile Acid Circulation
cirrhosis such as portal hypertension and
hepatocellular carcinoma (HCC). BAs undergo an enterohepatic circulation that
In addition, BAs stimulate glucagon‐like depends on active transport systems in the
peptide (GLP)‐1 production via TGR5 liver and the intestine (Figure 1.2). More spe‑
activation. GLP‐1 is known to promote cifically, BAs are excreted from hepatocytes
insulin secretion and thus regulate glucose into bile through BSEP/ABCB11 at the bile
homeostasis. Because GLP‐1 mimetics and canaliculus, reabsorbed in the ileum by the
receptor agonists are currently under clinical apical sodium‐dependent bile salt transporter
development and have shown promise in (ASBT/SLC10A2), and return through the
improving glucose homeostasis in diabetes, portal blood to the liver, where they are taken
BA‐based TGR5 agonists may be a potential up by hepatocytes via the basolateral transport
therapeutic to stimulate GLP‐1 secretion in systems NTCP/SLC10A1 for conjugated BAs
diabetic patients. and OATPs/SLCO/SLC21 family for unconju‑
gated BAs, thus limiting the amount of BA
spillover into the systemic circulation. BAs
Bile Acid Synthesis and Metabolism
complete the enterohepatic circulation six to
BAs are synthesized from cholesterol. In eight times a day and are highly efficiently con‑
humans, the “primary” BAs are cholic acid served. BAs that escape ileal reabsorption
(CA) and chenodeoxycholic acid (CDCA). reach the colon, where they are deconjugated
Before secretion into the bile, both CA and and metabolized (e.g. dehydroxylated) by gut
CDCA are conjugated to the amino group of microbiota to secondary BAs, which can still
taurine or glycine. Conjugation enhances the be passively absorbed as unconjugated BAs in
hydrophilicity of the BA, the major function the colon. Unconjugated BAs are partially
of this being to decrease the passive diffusion reconjugated (and rehydroxylated) during
of BAs across the cell membranes during their passage through the liver before being
their transit through the biliary tree and excreted into bile again, which completes their
intestine. Therefore, conjugated BAs are enterohepatic cycle. In addition, BAs are

absorbed only if a specific membrane carrier filtered by the glomeruli and then reabsorbed
is present. The process of bile formation in renal tubules, again limiting their renal loss.
depends on the liver synthesis and the cana‑ BAs may also cycle between cholangio‑
licular secretion of BAs. The active transport cytes and hepatocytes through a cholehepatic
of BAs across the canalicular membranes of shunt pathway. Unconjugated BAs induce a
hepatocytes is a primary driving force for bile greater degree of bile flow per BA molecule
flow. The majority of the BAs in the intestine excreted in bile. To account for this hyper‑
are absorbed intact. Approximately 15% are choleretic effect, it was proposed that uncon‑
deconjugated by the bacterial flora in the jugated BAs may be passively absorbed by
distal small intestine, with the production of bile ducts, enter the peribiliary plexus
“secondary” BAs by the conversion of CA to adjacent to intrahepatic bile ducts, and then
deoxycholic acid and of CDCA to lithocholic forwarded to the hepatic sinusoids to be

acid. Most of the conjugated and deconju‑ returned to cholangiocytes by hepatocyte
gated BAs are reabsorbed in the distal secretion. Cholehepatic shunting initiated by
intestine and undergo enterohepatic passive absorption of non‐ionized bile salt
circulation that maintains the BA pool. Thus, results in the generation of HCO3–‐rich
at least 12 major conjugated primary and hypercholeresis [4,5].
Liver Cholesterol
CYP7A1
NTCP
Bile acids
Systemic
circulation BSEP
Active
transport Kidney
Enterohepatic
circulation Bile
Intestine
ASBT
Urine
Feces
Figure 1.2 Enterohepatic circulation of bile acids (BAs). After hepatic synthesis and biliary secretion, BAs
undergo enterohepatic circulation. The BSEP (ABCB11) is the main canalicular transporter for BAs. After
reabsorbtion by ASBT/SLC10A1 in the ileum, BAs (and exit through OSTα/β from enterocytes; not shown) are
transported back to the liver through the portal blood. Reuptake of conjugated BA from portal blood through
NTCP/SLC10A1 (and OATPs for unconjugated BAs; not shown) into the hepatocytes completes the
enterohepatic circulation. ASBT, apical sodium‐dependent bile salt transporter; BSEP, bile salt export pump;
NTCP, sodium taurocholate cotransporting polypeptide; OATPs, organic anion transporters; OST, organic solute
transporter α/β. Source: Trauner et al. [5]. Reproduced with permission of John Wiley and Sons.
Death and Regeneration
recognition. Apoptosis is essential to avoid the
of Hepatocytes outflow of intracellular contents and to limit
the immunologic response against intracel‑
Cell Death
lular autoantigens. Nevertheless, apoptotic
Hepatocytes can die because of either necrosis bodies and fragments can under some cir‑
or apoptosis. Necrosis is the loss of plasma cumstances constitute a major source of
membrane integrity with release of the cel‑ immunogens in autoimmune diseases that
lular contents locally, which triggers an involve the targeting of ubiquitous autoanti‑
inflammatory response. Apoptosis is a highly gens. This has been described in PBC. In the
regulated process in which cells that are dam‑ BECs of patients with PBC there is increased
aged, senescent or deregulated self‐destruct DNA fragmentation, implying increased
with a lower release of inflammatory prod‑ apoptosis, when compared with normal con‑
ucts. Dying cells undergo morphologic modi‑ trols. While mitochondrial proteins are pre‑
fications including chromatin condensation, sent in all nucleated cells, in PBC there is a
nuclear fragmentation, and generation of apo‑ highly specific multilineage immune response
ptotic bodies. Furthermore, they express sig‑ directed to the nominal mitochondrial auto‑
nals on the cell surface that allow macrophage antigenic epitope, the inner lipoyl domain of
the E2 subunit of the pyruvate dehydrogenase which far exceeds the capacity of remaining
complex (PDC‐E2) of the BECs. Apoptosis of healthy hepatocytes to replicate and restore
BECs has been proposed as a potential source liver function, resident liver progenitor cells
of “neoantigens” that could be responsible for (i.e. oval cells) are activated to support or take
activation of autoreactive T lymphocytes or a over the role of regeneration. In adults, there
target for effector cells or antibodies. PDC‐E2 are multiple niches of biliary tree stem/pro‑
is not only immunologically intact during genitor cells (BTSCs) residing in different loca‑
apoptosis in BECs, but it localizes in the apo‑ tions along the human biliary tree and niches
ptotic bodies of BECs where it is accessible to found within the liver parenchyma, with a key
recognition by anti‐mitochondrial antibodies. role in regeneration of the liver. Figure 1.3
shows the location of stem/progenitor cell
niches in the human biliary tree. Canals of
Liver Regeneration
Hering harbor hepatic stem/progenitor cells
Liver possesses a unique capacity to replace its (HpSCs), while peribiliary glands (PBGs) con‑
mass after tissue injury or loss. The regenera‑ stitute the niche for BTSCs.
tive process involves a cascade of events that Those within the biliary tree are found in
moves cells from their resting G0 phase PBGs and contain especially primitive stem
through G1, S phase (DNA synthesis), and G2 cell populations, expressing endodermal tran‑
to M phase (mitotic cell division). A typical scription factors relevant to both liver and
example is hepatic growth after partial hepa‑ pancreas, pluripotency genes, and even
tectomy. The majority of research on liver markers indicating a genetic signature over‑
regeneration has focused on cytokine‐ and lapping with that of intestinal stem cells [7].
growth factor‐ mediated pathways involved in The distribution of PBGs is not uniform,
initiation and progression through the cell varying along the biliary tree: PBGs are mostly
cycle. During more extensive acute liver injury, found in the hepatopancreatic ampulla and
Canals of Hering Peribiliary glands
Hepatic stem/progenitor cells Biliary tree stem/progenitor cells

• Origin: Ductal plate • Origin: Hepatic diverticulum – caudal part
• Location: Canals of Hering • Location: Peribiliary glands – large intrahepatic

and extrahepatic bile ducts
• Potency: Hepatocytes and cholangiocytes • Potency: Hepatocytes, cholangiocytes and
pancreatic β-cells
• Disease: NAFLD, ASH, viral cirrhosis,
acute hepatitis, cholangiopathies • Disease: PSC, PSC-CCA, NAS, BA
Figure 1.3 Stem/progenitor cell niches in the human biliary tree. Canals of Hering harbor hepatic stem/
progenitor cells (HpSCs), while peribiliary glands (PBGs) constitute the niche for biliary tree stem/progenitor
cells (BTSCs). ASH, alcoholic steatohepatitis; BA, bile acid; CCA, cholangiocarcinoma; NAFLD, non‐alcoholic fatty
liver disease; NAS, non‐anastomotic strictures; PSC, primary sclerosing cholangitis. Source: Overi et al. [6].
are less numerous in the common bile duct; follows paracrine cross‐talk mediated by the
they are not present in the gallbladder, but a ability of RDCs to secrete profibrotic and
stem cell‐like compartment is located in the proinflammatory growth factors, and cross‐
epithelial crypts. Biliary progenitors support talk with cells of mesenchymal origin, in
the renewal of large intrahepatic and extrahe‑ particular Kupffer cells and portal fibro‑
patic bile ducts. Stem cells are present in the blasts, which are the main effectors of
canals of Hering, and participate in the renewal fibrosis, as stimulators of the deposition of
of the small intrahepatic bile ducts and in the extracellular matrix by activated myofibro‑
regeneration of liver parenchyma. Small hepa‑ blasts. In addition, RDCs also establish para‑
tocytes located in pericentral positions are also crine communications with endothelial cells
believed to act as progenitor cells on certain that provide the vascular support necessary
occasions. Distinct subpopulations of mature for the growth and arborization of the ductal
hepatocytes and stem/progenitor cell com‑ structures themselves [8].
partments are differentially activated depend‑
ing on the nature and duration of the liver
damage versus different human pathologies. rotective Role of Biliary
P
HCO3− Secretion
Cholangiocyte Reaction The cholangiocyte is exposed to millimolar
to Biliary Damage concentrations of hydrophobic bile salts, which
are toxic to other cells such as hepatocytes
BECs are usually quiescent, but following a at micromolar levels. Resistance against these
liver insult they activate and/or proliferate. A noxious compounds and their cytolytic poten‑
typical element of the repair response to liver tial is therefore essential. One of the strategies
damage is the ductular reaction (DR), a ste‑ that cholangiocytes have developed to survive is
reotyped histopathologic lesion of the biliary the biliary HCO3− umbrella.
epithelium that plays a fundamental role in Biliary HCO3− secretion sustains bile flow
the progression of hepatic fibrosis. The DR and confers its appropriate viscosity, gener‑
is characterized by a marked proliferation ates part of the alkaline tide necessary for
of cholangiocytes, leading to formation of optimal digestion of various nutrients within
reactive ductular cells (RDCs), with poor
the intestine, and protects the apical surface
cytoplasm and arranged in cell cords without of cholangiocytes against protonated apolar
a lumen or in richly anastomosed small‐ hydrophobic BA monomers by maintaining
diameter ducts (<10 μm) with almost unrec‑ an alkaline pH above the apical membrane.
ognizable lumens. RDCs are activated Isoforms of the Cl−/HCO3− exchanger, AE2,
epithelial cells that secrete a vast array of are responsible for the vast majority of biliary
factors, including cytokines, chemokines, HCO3− secretion. Dysfunction of any of the
growth factors, and angiogenic factors. They elements involved in HCO3− formation
may derive from hepatocytes undergoing a might weaken the biliary HCO3− umbrella
process of ductular metaplasia, or from and contribute to the development of chronic
activation of the hepatic progenitor cell cholestatic liver disease such as sclerosing
compartment and/or from proliferation and cholangitis.
dedifferentiation of preexisting cholangio‑
cytes. The increase in RDCs is generally
associated with a significant increase in Cholangiocytes and Immunity
inflammatory infiltrate and portal fibrosis.
RDCs are considered the major driver of BECs are a first line of defense in liver innate
portal fibrosis during parenchymal and/or immunity: they can present antigen to
biliary injury. The deposition of fibrosis immune cells, be a target of immune‐mediated
aggression, or be the initiators of an involved in the immune response of BECs.

inflammatory reaction that then progresses This is a receptor superfamily that includes
to adaptive immune activation [9]. The con‑ the glucocorticoid receptor, the retinoic acid
tribution of BECs to liver immune responses receptor, the VDR, the liver X receptors, and
was initially believed to be limited to the the peroxisome proliferator‐activated recep‑
secretion of immunoglobulin A (IgA) into tors (PPARs). Nuclear receptors control
the bile, but it is now clear that their role in several cell functions including cell prolifera‑
the immune response is far more complex. tion and apoptosis, cell metabolism, cell–cell
The biliary epithelium stands as a first line of interaction, detoxification from BAs, and bile
defense against bacteria, fungi and other secretion.
pathogens by secreting antimicrobial pep‑ In addition, continuous exposure to DAMPs
tides such as defensin and cathelicidin. A and PAMPs could promote cellular senescence.
major role in epithelial innate immunity in Cell senescence is a mechanism of irreversible
BECs is played by Toll‐like receptors (TLRs) cell arrest in G1 stage induced by different
and by nuclear receptors. TLRs can recog‑ stimuli. The main causes responsible for the
nize pathogen‐associated molecular pat‑ onset of senescence are DNA damage (particu‑
terns (PAMPs), i.e. bacterial elements such larly but not exclusively) to the telomeres, the
as lipopolysaccharide (LPS), DNA and RNA activation of mitogenic signals induced by
fragments, but also respond to endogenous oncogene activation, epigenetic modifications,
components or damage‐associated molec‑ and expression of tumor suppressor genes.
ular patterns (DAMPs), such as hyaluronan All these signals lead to different physiologic
and high mobility group box 1 (HMGB1) responses generally leading to tumor suppres‑
that are released from damaged cells. TLR4‐ sion; however, in some cases it could promote
mediated signaling is the better studied in cancer development or induce a fibrosing
cholangiocytes. Once activated by LPS or response and mediate age‐related degenerative
other ligands, TLR4 activates two different diseases. Once senescent, cells not only cease
pathways, one mediated by NF‐κB, which proliferation but assume a senescence‐associ‑
stimulate the expression of a number of pro‑ ated secretory phenotype (SASP) characterized
inflammatory cytokines and chemokines, by the secretion of a plethora of peptides
and one mediated by mitogen‐activated with profibrogenic, proinflammatory, and
protein kinase (MAPK)/activator protein 1 tumorigenic properties. This indicates that
(AP‐1), which requires the nuclearization of senescence could not only act as a barrier to
the AP‐1 complex. In normal cholangiocytes, tumor growth, but also paracrinally stimulate
TLR4 signaling is repressed by protective the activation of aberrant reparative/
mechanisms aimed at maintaining LPS toler‑ regenerative responses. In chronic biliary
ance. Since the biliary epithelium is continu‑ diseases, cholangiocyte senescence is likely the
ously in contact with bacterial products of result of ongoing inflammation, a sort of
intestinal origin, changes in one or more “exhaustion” of the activated cholangiocytes.
regulatory checkpoints may trigger an exag‑ This is particularly important in PSC, given the
gerated inflammatory response in the liver. association with cholangiocarcinoma.
For example in cystic fibrosis‐related liver
disease, a decrease in LPS tolerance plays a
major role in the development of the disease. iochemical Markers and Patterns
B
A continuous stress in the absence of a of Hepatic Injury
correct modulation of the TLR‐mediated
responses could be the trigger for a chronic Contrary to the kidney, no single test can be
inflammatory or autoimmune response. used to assess liver function. The liver bio
Nuclear receptors, particularly the retinoid chemical tests, or liver function tests (LFTs),
X receptor (RXR), have recently been provide indirect evidence of hepatobiliary
isease. LFTs that more accurately reflect liver

d bloodstream. However, their increase is not
synthesis include serum albumin, serum bili‑ always pathologic: they can be raised by vig‑
rubin, and prothrombin time, which is stan‑ orous physical activity, and rarely an isolated
dardized to the international normalized ratio AST can be the result of the binding of the
(INR). The enzymes aspartate aminotrans‑ enzyme with an immunoglobulin forming a
ferase (AST) and alanine aminotransferase macro‐enzyme (macro‐AST) complex. The
(ALT) are markers of hepatocellular disease, diagnostic specificity of mild‐to‐moderate
whereas alkaline phosphatase (ALP) and increases in aminotransferases is poor, with
gamma‐glutamyltranspeptidase (GGT) are many differential diagnoses being possible,
markers of cholestasis. The clinical evaluation whereas the spectrum of liver conditions
of patients with abnormal LFTs involves an indicated by markedly elevated aminotrans‑
accurate interpretation of the pattern of liver ferase levels (>2000 IU/l) narrows to viral
damage in the context of an accurately col‑ (mostly hepatitis A and hepatitis B virus),
lected medical history and physical examina‑ ischemic (shock liver), and drugs. Autoimmune
tion. The severity and pattern of LFT hepatitis can sometimes have an acute outset
abnormality, assessed by serial measurements, with striking elevation of aminotransferases.
can be distinctive or aspecific. Different pat‑ Rarely, bile duct stones can manifest as marked
terns of damage can be observed: hepatocel‑ rise in aminotransferase, although this is fol‑
lular damage, with predominant elevations in lowed by a rapid fall within 48 hours.
AST and ALT; cholestatic damage, with The AST/ALT ratio can often provide a clue
predominant increases in ALP, GGT, and bili‑ to the diagnosis. In the majority of cases of
rubin; and infiltrative damage with ALP and hepatitis, the AST/ALT ratio is less or equal to
GGT increased disproportionately to bili‑ 1. The AST/ALT ratio is typically greater than
rubin. These patterns are valuable in directing 2 during alcoholic hepatitis. This occurs
specific serologic tests, imaging, and liver because damage is primarily mitochondrial
biopsy. However, they are not diagnostic for (thus more AST is released systemically) and
a specific cause, nor are they able to distin‑ ALT synthesis is more sensitive than AST to
guish whether cholestasis is intrahepatic or pyridoxal 5‐phosphate deficiency, a common
extrahepatic. finding in alcoholics, leading to lower serum
ALT levels. Supplementation with pyridoxine
in patients with alcoholic hepatitis results in a
Hepatocellular Necrosis
rise in the level of ALT. An AST/ALT ratio
Elevation of ALT and AST indicates hepato‑ greater than 1 in patients without a history of
cellular necrosis. The interpretation of these alcoholism is suggestive of advanced fibrosis
increases should consider the rate of rise, the or cirrhosis. An AST/ALT ratio greater than 4
severity (peak level), the AST/ALT ratio, and is observed in patients with fulminant Wilson
coexisting abnormalities in other LFTs and disease.
other investigations. ALT and AST are
enzymes that catalyze the transfer of amino
Cholestasis
groups from alanine or aspartic acid to
ketoglutaric acid to form pyruvic acid and Cholestasis is an overarching term applied to
oxaloacetic acid, respectively, during gluco‑ conditions in which there is impairment of bile
neogenesis. ALT is localized primarily in the formation and/or bile flow. It occurs where
liver and confined to the cytoplasm, while there is a failure at any point along the biliary
AST can be released by the liver, myocar‑ tree, between the basolateral (sinusoidal)
dium, skeletal muscle, kidney, pancreas, membrane of the hepatocyte and the ampulla
and blood cells, and can be found in the of Vater, as a result of congenital or acquired
cytoplasm and mitochondria. During hepa‑ injuries, that leads to impaired secretion of bile
tocellular injury they are released into the such that biliary constituents spill into blood. It
may result from (i) hepatocellular and/or chol‑ failure can result in elevation of the intestinal
angiocellular secretory defects or (ii) obstruc‑ ALP isoenzyme. In patients with raised ALP,
tion of bile ducts by bile duct lesions, stones or hyperthyroidism should be ruled out. Rarely,
tumors, but may also be related to mixed ALP can be identified in patients with under‑
mechanisms in conditions such as PBC or PSC. lying malignancy not involving either liver or
ALP and GGT are markers of cholestasis. bones. This is the Regan isoenzyme, bio‑
ALP is a ubiquitous membrane‐bound glyco‑ chemically different from the liver isoen‑
protein that catalyzes the hydrolysis of phos‑ zyme, that has been described in lung cancer,
phate monoesters at basic pH values. Liver Hodgkin disease, and renal cell carcinoma.
and bone are the major source of serum ALP. Finally, ALP should be tested after fasting
The liver isoenzyme is located on the cana‑ since its level can rise after a fatty meal.
licular side of the hepatocyte plasma mem‑ GGT is an enzyme that can be induced by
brane and the luminal surface of bile duct several stimuli such as drugs and alcohol. It
epithelium. Serum ALP elevation more than is mainly localized in hepatocytes and bil‑
three times normal strongly suggests chole‑ iary epithelia, and is also present in
stasis if bone disease is absent and GGT is extrahepatic tissues such as kidney, spleen,
elevated. In patients with cholestasis, the pancreas, heart, lung, and brain, but not
ALP elevation is triggered by increased syn‑ bone. The lack of GGT in bone can be used
thesis and release of the enzyme into serum to distinguish a liver source from a bone
rather than impaired biliary secretion. BAs source of a raised ALP. GGT is more liver
build up in hepatocytes and solubilize the specific than ALP, although during chole‑
plasma membrane, thereby resulting in stasis is less specific since it can be influ‑
release of ALP. The half‐life of serum ALP is enced by other factors such as alcohol, fat,
5–7 days, and therefore serum ALP remains and drugs. A GGT/ALP ratio over 2.5 may
elevated for several days after resolution of point to alcohol abuse, although up to one‐
the biliary obstruction. ALP is not used as a third of those who abuse alcohol (>80 g/day)
marker of cholestasis in adolescent and preg‑ have a normal GGT. A normal GGT in
nant women since ALP in these conditions patients with elevated liver ALP isoenzyme
can be raised as a consequence of rapid bone should raise the suspicion of benign recur‑
growth and placental growth. Chronic renal rent intrahepatic cholestasis.
References
1 Ober, E.A. and Lemaigre, F.P. (2018). 6 Overi, D., Carpino, G., Cardinale, V. et al.
Development of the liver: insights into organ (2018). Contribution of resident stem cells
and tissue morphogenesis. J. Hepatol. 68: to liver and biliary tree regeneration in
1049–1062. human diseases. Int. J. Mol. Sci. 19 (10):
2 Roskams, T.A., Theise, N.D., Balabaud, C. 2917. https://doi.org/10.3390/ijms
et al. (2004). Nomenclature of the finer 19102917.
branches of the biliary tree: canals, ductules, Cardinale, V., Wang, Y., Carpino, G. et al.
7
and ductular reactions in human livers. (2011). Multipotent stem/progenitor cells in
Hepatology 39: 1739–1745. human biliary tree give rise to hepatocytes,
3 Halilbasic, E., Claudel, T., and Trauner, M. cholangiocytes, and pancreatic islets.
(2013). Bile acid transporters and regulatory Hepatology 54: 2159–2172.
nuclear receptors in the liver and beyond. J. 8 Fabris, L., Spirli, C., Cadamuro, M. et al.
Hepatol. 58 (1): 155–168. (2017). Emerging concepts in biliary repair
4 Boyer, J.L. (2013). Bile formation and and fibrosis. Am. J. Physiol. Gastrointest.
secretion. Compr. Physiol. 3: 1035–1078. Liver Physiol. 313: G102–G116.
5 Trauner, M., Fuchs, C.D., Halilbasic, E., and 9 Strazzabosco, M., Fiorotto, R., Cadamuro,
Paumgartner, G. (2017). New therapeutic M. et al. (2018). Pathophysiologic
concepts in bile acid transport and signaling implications of innate immunity and
for management of cholestasis. Hepatology autoinflammation in the biliary epithelium.
65 (4): 1393–1404. Biochim. Biophys. Acta 1864: 1374–1379.
21
Concepts of Autoimmunity Relevant to Autoimmune

Liver Diseases
Isaiah G. Roepe1 and John M. Vierling2
1
Baylor College of Medicine, Houston, TX, USA
2
Departments of Medicine and Surgery, Section of Gastroenterology and Hepatology, Division of Abdominal Transplantation,
Baylor College of Medicine, Houston, TX, USA
Abstract
Knowledge of the concepts of autoimmunity can aid gastroenterologists and hepatologists in

management and counseling of patients with autoimmune liver diseases and in understanding the
rationales and sites of action of therapies. This chapter addresses major themes: innate and adaptive
immunity in the context of the liver as an immune organ; generation and maintenance of tolerance
to autoantigens; and risk factors for autoimmunity. The themes discussed also include: loss of
immune tolerance to autoantigens and perpetuation of autoimmune diseases; and prospects for
prevention of autoimmunity and therapeutic control of autoimmune diseases. Bacterial, fungal or
viral infections can instigate innate and adaptive immune responses that result in autoimmunity.
Vitamin D deficiency is epidemiologically associated with risk of autoimmunity. The increasing
incidence of autoimmunity and inflammatory diseases observed worldwide is correlated with
changes in environmental factors, including a more modern lifestyle, improved hygiene, a Western
diet, use of antibiotics, and elimination of childhood parasitic infections.
Keywords adaptive immunity; autoantigens; autoimmune liver diseases; autoimmunity prevention;

immune homeostasis maintenance; immune tolerance; innate immune organ; therapeutic control
Key Points
●● Autoimmunity results from the complex autoreactive T‐cell receptors, and auto
interplay of genetic, epigenetic, immuno reactive B cells to develop a specific
logic, and environmental factors. autoimmune disease.
●● Environmental triggers initiate loss of tol ●● Subsequently, additional immunologic
erance to autoantigens in genetically sus mechanisms perpetuate chronic progres
ceptible individuals. sive inflammatory disease.
●● Susceptible individuals must have ●● Novel therapies based on immunopatho
inflammatory innate immune responses, genic mechanisms in autoimmunity are in
human leukocyte antigen (HLA) alleles development.
capable of autoantigen presentation to
Introduction cholangitis (PSC) are classified as autoimmune

liver diseases (AILDs). However, comparison
Until the identification of human autoim of each of these AILDs with classical autoim
mune diseases at the dawn of the twentieth mune diseases reveals shared and atypical fea
century proved Paul Ehrlich’s concept of hor- tures (Table 2.1). Knowledge of the concepts of
ror autotoxicus to be correct, clinicians autoimmunity can aid gastroenterologists and
believed that the immune system was inca hepatologists in management and counseling
pable of reactions against self‐tissues and of patients with AILDs and in understanding
organs. To date, over 80 human autoimmune the rationales and sites of action of future
diseases have been identified, ranging in therapies. Thus, this chapter addresses five

nature from systemic to localized. The genesis major themes: (i) innate and adaptive immu
of autoimmunity requires loss of immune tol nity in the context of the liver as an immune
erance to self‐antigens and results from the organ; (ii) generation and maintenance of tol
complex interplay of genetic, epigenetic, erance to autoantigens; (iii) risk factors for
immunologic, and environmental factors autoimmunity; (iv) loss of immune tolerance
(Figure 2.1) [1]. Genetics confers susceptibility to autoantigens and perpetuation of
for autoimmunity, not for a specific autoim autoimmune diseases; and (v) prospects for
mune disease. Thus, susceptible patients often prevention of autoimmunity and therapeutic
have more than one autoimmune disease [1]. control of autoimmune diseases.
While complex genetic susceptibility and
environmental triggers suggest that autoim
munity results from “bad genes, bad luck,” this ole of Innate and Adaptive
R
notion neglects the evolutionary advantage of Immunity in Autoimmunity
robust immune responses to a broad array of in the Context of the Liver
antigens, including some autoantigens. An as an Immune Organ
evolutionary survival advantage likely explains
the unexpectedly high frequency of autoim Overview
mune reactions in healthy humans that fail to
progress to autoimmune diseases [2]. The evolutionarily ancient innate immune
Autoimmune hepatitis (AIH), primary bil system provides immediate responses to
iary cholangitis (PBC), and primary sclerosing pathogen‐associated molecular patterns
Autoimmune diseases
Cumulative
adaptive
immune
repertoire
Genetics Environment
Thymic B/T cell selection Microbiome
PAMPs, DAMPs
Sex Gut
Pathogens
HLA alleles Skin
Xenobiotics/drugs
Non-HLA genes Nasopharynx
Hepatic micro-
Epigenetic enhancers Genitourinary
environment
Immune
regulatory
capacity
Figure 2.1 Interaction of factors required for generation of autoimmune diseases. Autoimmune diseases,
including AILDs, result from the complex interaction among factors involving genetic susceptibility, the status
of the cumulative immune response repertoire, an individual’s immune regulatory capacity, environmental
triggers, and the influence of the microbiome.
Chapter 2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 23
Table 2.1 Comparison of classic autoimmunity with AIH, PBC or PSC.
Classic autoimmunity AIH PBC PSC
Disease‐specific autoantigens Yesa Yes Yesb

Disease‐specific autoantibodies Yes, types 1 and 2 Definite Yes
Epitope determinant spreading Unclear No Unclear
Female predilection Yes Yes No
Occurrence in children and adults Yes No, adults only Yes
Polygenetic predisposition Yes Yes Yes
HLA immunogenetic associations Yes Yes Yes
Non‐HLA genetic associations Yes Yes Yes
Environmental triggers Yes Yes Yes
Tissue/organ‐specific pathology Yes Yes Yes
Associated extrahepatic autoimmune diseases Yes Yes Yes
Associated IBD Yes (weak) No Yes (strong)
Response to immunosuppression Yes Yes No
Autoantigen‐specific animal models Yes for type 2 Yes No
No for type 1
a
T and B cell epitopes defined in type 2, not type 1 AIH.
b
β‐Tubulin isoform 5, not rigorously tested to determine prevalence, sensitivity or specificity.
AIH, autoimmune hepatitis; IBD, inflammatory bowel disease; PBC, primary biliary cholangitis; PSC, primary
sclerosing cholangitis.
(PAMPs) and damage‐associated molecular instead are active participants in the immu
patterns (DAMPs) associated with microbes nopathogenesis of their own injury and
and injured or dying cells, respectively [3]. In demise.
contrast, the adaptive immune system
responds to specific peptide antigens through
Innate Immunity
antigen‐specific T‐cell responses and B‐cell
production of antigen‐specific antibodies [4]. The innate immune system responds imme
It is now clear that cytokines produced by an diately to PAMPs, comprising microbial con
innate immune response dictate both the stituents, and to DAMPs, comprising cellular
type and magnitude of adaptive responses constituents from stressed, neoplastic or
and that both innate and adaptive immunity dying cells (Table 2.2). Both PAMPs and
play important roles in the immunopatho DAMPs bind to pattern recognition recep
genesis of all autoimmune diseases (Table 2.2 tors (PRRs) and other receptors expressed on
and Figure 2.2). neutrophils, dendritic cells (DCs), activated
Innate immune inflammasome responses, macrophages (including Kupffer cells), and
especially those mediated by nucleotide‐ natural killer (NK) and natural killer T (NKT)
binding oligomerization domain (NOD)‐like cells [3]. NK cells express killer receptors for
receptor P3 (NLRP3), occur in both innate stressed, dying or neoplastic cells and Fc
immune cells and non‐immune cells, receptors that mediate antibody‐dependent
including hepatocytes, cholangiocytes, and cell‐mediated cytotoxicity (ADCC) of target
hepatic stellate cells (HSCs) [5]. Thus, clini cells coated with antibodies [3]. Antigen‐pre
cians must embrace a new paradigm: cells or senting cells (APCs) activate NKT and γδT
tissues are not passive “targets” of autoim cells by invariant HLA class I‐like molecule
mune diseases (including AILDs), but (CD‐1) presentation of lipid or glycolipid
Table 2.2 Comparison of selected features of innate and adaptive immunity.
Innate immunity Adaptive immunity
Characteristics
Onset Rapid responses mediated by Delayed due to requirement for antigen
preformed PRRs for microbial activation, clonal proliferation and maturation
PAMPs and DAMPs of effector cell functions
Specificity PAMPs, DAMPs, ROS, Epitopes of peptide antigens recognized by
activated C′ proteins, TCRs or B‐cell immunoglobulins
apoptotic bodies
Genetics Restricted, germline‐encoded Complex with TCRs and antigen‐binding
domains of immunoglobulins produced by
somatic recombination of gene segments
Diversity Limited Virtually infinite due to TCR rearrangement
Evolutionarily conserved
Memory No Yes. Memory T‐ and B‐cell responses capable
of anamnestic reactivation
Self‐tolerance Discriminates pathogens and Positive and negative selection of TCR and
endogenous DAMPs. Not B‐cell immunoglobulin reduce autoreactive
responsive to autoantigens capacity
Components
Physical barriers Skin, mucosal epithelia of Intraepithelial lymphocytes within gut, Peyer’s
oropharynx, gut and patches
reproductive tract,
antimicrobial proteins
Cells Dendritic cells, monocytes, Professional APCs, αβT cells, γδT cells,
macrophages (Kupffer cells), natural and inducible Treg cells, B cells,
neutrophils, NK cells, NKT plasma cells
cells
Proteins C′ proteins, IFN‐α, ‐β, ‐γ, T‐cell cytokines
cytokines, chemokines IgM, IgG, IgA, IgE antibodies
Role of liver as Yes Yes
an immunologic DCs, neutrophils, Kupffer Antigen presentation by Kupffer cells, LSECs,
organ cells, excessive concentrations HSCs, cytokines stimulate hepatocytes and
of NK cells, NKT cells cholangiocytes, PD‐L1/2 induce senescent
IFN‐α, ‐β, ‐γ, balance between T‐cell apoptosis and inhibition of activated
proinflammatory and CD8 T cells
immunosuppressive
chemokines
C′, complement; DAMPs, damage‐associated molecular patterns; HSC, hepatic stellate cell; IFN, interferon; LSEC,
liver sinusoidal endothelial cell; NK, natural killer; PAMPs, pathogen‐associated molecular patterns; PD‐L1/2,
programmed death ligands 1 and 2; PRRs, pattern recognition receptors; ROS, reactive oxygen species; TCR, T cell
receptor.
Examples of common PAMPs: (i) lipopolysaccharide (LPS), a cell wall constituent of all Gram‐negative bacteria;
(ii) lipotechoic acid, a cell wall constituent of all Gram‐positive bacteria; (iii) peptidoglycans, essential cell wall
components of all bacteria; (iv) viral proteins; and (v) unmethylated bacterial CpG dinucleotide motifs.
Examples of common DAMPs: (i) nuclear proteins; (ii) cytosolic proteins and lysosomal enzymes; and
(iii) non‐proteins, such as ATP, DNA, RNA, mitochondrial DNA, U1 ribonucleoprotein (U1RNP) and uric acid.
antigens [3]. γδT cells protect from autoim secrete collagen, resulting in fibrosis. PAMPs
mune diseases and microbial infections and and DAMPs also induce NLRP3 inflamma
participate in mucosal immunity and tumor some responses [5] in hepatocytes, cholan
surveillance. giocytes and HSCs, resulting in their
Microbial PAMPs bind predominantly to secretion of proinflammatory cytokines and
PRRs called Toll‐like receptors (TLRs). chemokines.
DAMPs bind to PRRs and receptors for The reticuloendothelial function of the
chemical ligands or nuclear constituents. liver normally prevents PAMPs, DAMPs and
Still other innate receptors bind apoptotic microbial antigens in portal venous blood
bodies and activated complement (C′) mole from entering the systemic circulation. This
cules. PAMPs and/or DAMPs activate DCs barrier function of the liver also plays a role
and macrophages (including Kupffer cells), in oral tolerance. Hepatic lymph formed in
increasing phagocytic activity and inducing the space of Disse flows retrograde into
secretion of chemokines and proinflamma portal tract lymphatics. Thus, hepatic lymph
tory cytokines interleukin (IL)‐1β, IL‐6, exposes portal bile ducts, arterioles and
IL‐12, and IL‐18 and tumor necrosis factor portal vein branches to variable amounts of
(TNF)‐α, as well as the immunosuppressant PAMPs, DAMPs, cytokines, chemokines,
cytokine IL‐10 [3]. The relative amounts of proteins, lipoproteins, and lipids.
specific PAMPs, DAMPs, cytokines and che
mokines creates a dynamic balance between
Adaptive Immunity and Adaptive
proinflammatory and immunosuppressive
Immune Functions of the Liver
cytokines that determines whether the
adaptive immune response promotes immu As shown in Figure 2.2, the adaptive immune
nopathology (Figure 2.2). system in autoimmunity generates multiple,
delayed, autoantigen‐specific effector immune
responses [4]. Most adaptive immune
Liver as an Innate Immune Organ
responses occur in lymph nodes and spleen;
The liver is an essential innate immune organ thus, effector cells must circulate to tissues
that must either tolerate or respond to expressing autoantigens to mediate immuno
microbial PAMPs, toxins, xenobiotics, food pathology. Class II HLA‐DR, ‐DQ or ‐DP
antigens, microbial pathogens, and neo molecules present peptide autoantigens to T‐
plastic cells [6]. The liver contains special cell receptors (TCRs) of autoreactive CD4 T
ized liver sinusoidal endothelial cells (LSECs) cells. In contrast, class I HLA‐A, ‐B or ‐C mol
and large quantities of Kupffer cells, DCs and ecules present peptide autoantigens to TCRs
NK, NKT and γδT cells. Most intrahepatic B of autoreactive cytotoxic CD8 T cells, also
cells are B1 cells, an innate B‐cell subset that known as cytotoxic T lymphocytes (CTLs).
secretes low‐affinity IgM antibodies against Polymorphisms in HLA class I and II alleles
glycoproteins, called “natural antibodies.” determine which autoantigens can be
The liver also synthesizes C′ proteins, presented to T cells, which explains the strong
acute‐phase reactant proteins, and multiple association of autoimmunity with specific
cytokines. PAMP and DAMP activation of HLA alleles [2]. Polymorphic class III HLA
TLRs on Kupffer cells, LSECs, hepatocytes, alleles encode immune response proteins,
DCs and HSCs primarily generates IL‐10, including TNF‐α/β, complement (C′) factors
providing an immunosuppressive environ and heat‐shock proteins. The balance bet
ment. However, high quantities of PAMPs or ween proinflammatory and anti‐inflammatory
DAMPs induce the proinflammatory cyto cytokines produced by local innate immune
kines IL‐1β, IL‐6, IL‐12, IL‐18 and TNF‐α. responses determines the polarization of
HSCs normally produce matrix proteins, but cytotoxic mechanisms of the adaptive immune
once transformed into myofibroblasts they response (Figure 2.2).
IFNy
CD8 TNFα
Thymus CTL
IL-2
nTreg
IL-12 CD4
IFNγ Mechanisms of
Th1
IFNγ Cytotoxicity
Macrophage TNFα
nTreg nTreg HLA IL-4 IL-9 Cell-Mediated
CD4 IL-1β
APC Class II IL-10
TGFβ Th9
IL-21 IL-23 • CD8 cytotoxic T cells
Costimulation:
• NK or NKT cels
CD80/86 CD40 Fc • Macrophages
Ag NK
HLA Receptor
TCR Class I CD4
cell
CD28 CD154 IL-4 IL-4 Antibody-Mediated
Th2
Naive
CD8 IL-10 B Plasma Target
CD4 • IgG, IgM, or IgA
CTL IL-13 cell cell cell
Th0 • Complement
CD4 IL-21
IL-6 NKT • Immune complexes
Tfh
• Antibody-Dependent
cell cellular cytotoxicity
Autoantibodies Complement (ADCC) mediated by NK
IL-23 activation Fc receptors
IL-4 IL-12 CD4
IL-6
IL-10 IL-6
Anti-inflammatory
IL-6 Th17 IL-17, IL-22, TNFα Cytokine-Mediated

Proinflammatory
(TGFβ) IL-1β TGFβ

Transformation of • TNFα
IFNγ Inhibit all
cytokines
cytokines
CD4 iTregs to
CD4 Th
CD4 Th17 cells
subsets
CD4
TGFβ iTreg
IL-23
Local innate CD4
IL-6
CD4
immune responses Tr1 Th3 IL-1β
TGFβ
Figure 2.2 Adaptive immune responses generate multiple mechanisms of cytotoxicity in autoimmunity. Failure of thymic nTregs allows professional APC
presentation of peptide autoantigens to autoreactive naive CD4 Th0 and CD8 T cells. APC class II HLA molecules, encoded by HLA‐DR, −DQ, and ‐DP alleles, present
peptide autoantigens to TCRs of CD4 Th0 cells. APC class I HLA molecules, encoded by HLA‐A, ‐B and ‐C alleles, present different peptide autoantigens to TCRs of
naive CD8 CTLs. APC processing of autoantigen–immunoglobulin immune complexes dramatically increases autoantigen‐specific activation of CD8 CTLs.
Polymorphisms in HLA class I and II alleles determine which autoantigens are presented to TCRs. Differentiation of functional subsets of autoantigen‐specific CD4 Th
cells and cytotoxic CD8 T cells depends on two factors: costimulation of HLA–antigen activated T cells and the composition of the cytokine milieu produced by local
innate immune responses. Potent costimulation is mediated by binding of CD80/CD86 expressed on APCs to CD28 expressed on T cells or binding of CD40 expressed
on APCs to CD154 (aka CD40 ligand, CD40L) expressed on T cells. If costimulation is absent or inadequate, antigen‐activated CD4 Th0 and CD8 T‐cell clones become
anergic to subsequent reexposures to the activating peptide antigens. When local innate immune responses produce abundant proinflammatory cytokines IL‐12,
IL‐6, IL‐1β and IFN‐γ, activated CD4 Th0 cells predominantly differentiate into Th1 cells, rather than Th2 cells. Secretion of other cytokines, acting alone or in
combination, promotes CD4 Th0 differentiation into other CD4 T‐cell subsets. CD4 Th1 cytokines and growth factors drive the proliferation and differentiation of
cytotoxic antigen‐specific CD8 CTLs. Th1 cytokines also activate tissue macrophages and induce B‐cell secretion of C′‐fixing IgG2a. IL‐4 produced by local innate
immune responses favors CD4 Th0 differentiation into Th2 cells that secrete IL‐4, IL‐5, IL‐10, and IL‐13 and induce B‐cell secretion of IgG, IgM, IgA, and IgE. CD4 Th1
and Th2 cells achieve a dynamic balance by secreting signature cytokine profiles that inhibit the proliferation and cytokine secretion of the other subset. TGF‐β and
IL‐6 play important roles, both alone and in combination, in the differentiation of CD4 Tfh, Th17, and iTreg cells. IL‐21, IL‐23 and RORγ, RORα promote differentiation
of Th17 cells, which intensify and perpetuate tissue inflammation. Tfh and Th9 cells secrete the pluripotent cytokine IL‐21. IL‐21 is a pivotal cytokine in autoimmune
disease pathogenesis mediated by autoantibodies and cytotoxic cells and cytokines. Specifically, IL‐21 induces B‐cell differentiation into plasma cells, increases
proliferation of CD4 subsets, especially Th17 cells, increases resistance to iTregs and proliferation and cytokine production of NKT cells, increases NK cell cytotoxicity
and ADCC of immunoglobulin‐coated target cells, and increases IFN‐γ secretion and proliferation and cytotoxicity of CD8 CTLs. IL‐9 stimulates cell proliferation and
inhibits apoptosis, while IL‐13 is an anti‐inflammatory cytokine that induces IgE and enhances APC survival. Tfh cells localize within B‐cell follicles in lymph nodes and
Peyer’s patches where they promote selection and survival of B‐cell clones by secreting IL‐4 and IL‐21. Tfh cells play a central role in the formation of tertiary germinal
centers characteristic of autoimmune diseases. CD4 iTreg cells suppress antigen‐specific T‐cell responses of all CD4 T‐cell subsets by secreting immunosuppressive
IL‐10 and TGF‐β. Local secretion of IL‐23, Il‐6, IL‐1β, and TGF‐β by cytokine‐stimulated epithelial cells (e.g. cholangiocytes) transform iTregs into proinflammatory Th17
cells, defeating the immunoregulatory function of iTregs. Innate NK, NKT cells and activated macrophages and adaptive immune effector B and T cells act together to
mediate target cell cytotoxicity. NK cells express killer inhibitory receptors (KIRs) that inhibit killing of normal cells but target abnormal cells requiring elimination.
Activated NK cells secrete abundant IFN‐γ, a synergizing cytokine in the proinflammatory cytokine milieu. NKT cells react to lipid or glycolipid antigens presented by
CD1d molecules and secrete IFN‐γ, IL‐2, IL‐4, TNF‐α, and granulocyte–macrophage colony‐stimulating factor (GM‐CSF). NK group 2 member D (NKG2D) receptors on
activated macrophages, NK, NKT and γδT cells induce apoptosis by engaging MICA and MICB ligands produced by HLA class I‐like genes in abnormal cells. Activated
macrophages and NK and NKT cells also express Fc receptors that mediate antibody‐dependent cellular cytotoxicity.
After exiting lymph nodes, circulating Furthermore, binding of Kupffer cell PD‐L1
activated T cells and B cells undergo transen to PD‐1 expressed by activated CD4 T cells
dothelial migration to enter inflamed tissues induces immunosuppressive IL‐10 and
expressing chemokines [1]. Activated T cells, reduces apoptosis of CD4 T regulatory cells
B cells, and other leukocytes expressing che (Tregs).
mokine receptors migrate through the tight Mucosal invariant T (MAIT) cells have
junctions of endothelial cells, and their che characteristics of both innate and adaptive
mokine receptors mediate migration toward immune cells that are relevant in AILDs [7].
the cellular sources of chemokine secretion. MAIT cells have invariant TCR α‐chains that
This results in mixed inflammatory infil respond to bacterially processed vitamin B
trates that mediate cytotoxicity (Figure 2.2). antigens presented by unique major
histocompatibility complex (MHC) class I‐
related molecules (MR‐1) on APCs. Normally
Role of the Liver as an Adaptive
found only in blood, gut mucosa and mesen
Immune Organ
teric lymph nodes, MAIT cells congregate in
In addition to the innate immune functions, the portal tracts of patients with chronic liver
the liver has important roles in adaptive diseases, including AIH, PBC, PSC, alcoholic
immunity [6]. The normal liver is a distinct and non‐alcoholic fatty liver diseases, and
“lymphoid” compartment containing CD8 αβ chronic hepatitis C virus (HCV) infection.
T cells, activated CD4 and CD8 T cells, γδT Proinflammatory cytokines IL‐12 and IL‐18
cells, memory T cells and B cells. Indeed, the activate MAIT cells to display dual charac
proportions of highly activated T cells in the teristics of CD4 Th1 and Th17 cells by
liver exceed those in blood. In contrast, secreting interferon (IFN)‐γ, TNF‐α and
normal liver contains scant numbers of naive IL‐17. Up to 25% express cytotoxic granzyme
T cells and B cells. B, a feature of cytotoxic NK, NKT, and CD8
Most often, antigen‐activated DCs migrate T cells. Importantly, MAIT cell cytokines
to local lymph nodes to activate non‐hepatic stimulate cholangiocyte secretion of cyto
T cells, which subsequently circulate and kines capable of transforming CD4 iTregs
enter tissues through transendothelial migra into proinflammatory CD4 Th17 cells
tion. However, hepatic DCs, LSECs, and (Figure 2.2).
HSCs also function as APCs for intrahepatic
T‐cell activation of adaptive immunity.
Hepatocytes and cholangiocytes may also Generation and
serve as APCs after cytokine stimulation.
Maintenance of Tolerance
CD4 T cells activated by hepatic APCs pri
marily differentiate into Th2 cells, secreting to Self‐antigens
immunosuppressive IL‐10 and IL‐4.
Overview
Frequently, CD8 T cells activated in the liver
have functional deficiencies leading to pre Immune tolerance is the ability of the
mature apoptosis. However, LSECs can immune system to identify and respond to
directly activate functional antigen‐specific foreign or non‐self‐antigens, while aborting
CD8 CTLs in the liver by cross‐presentation or controlling potentially deleterious
of soluble exogenous antigens in class I HLA responses to autoantigens [1]. Despite ge
molecules. netic variability within the population and
Hepatic Kupffer cells regulate T‐cell substantial differences in environmental
homeostasis by inducing apoptosis of senes exposures that shape an individual’s immune
cent T cells expressing CD95 (Fas), TNF‐ repertoire, the vast majority of people main
apoptosis‐inducing ligand (TRAIL), and tain immune tolerance to self‐antigens. As a
programmed cell death 1 (PD‐1) receptors. result, autoimmune diseases are rare, even
among persons with genetic susceptibility tissue‐specific self‐antigen expression

[2]. In contrast, rare monogenetic defects required for deletion of T cells with TCRs
predictably result in the development of that react strongly with autoantigens. In
autoimmune diseases in early life [1]. Broadly, addition, thymic B class switching also pro
immune tolerance is divided into two cate motes central tolerance through negative
gories: natural and inducible tolerance. In selection of CD4 T cells. The importance of
addition, natural tolerance has been subclas AIRE in central tolerance is underscored by
sified as either central or peripheral, based the consequences of genetic AIRE defi
the anatomic site of development. ciency in humans, which results in an
autoimmune disease characterized by auto
immune polyendocrine syndrome type 1
Central Tolerance
(APS‐1) [1]. AIRE is also involved in
Central tolerance involves clonal deletion, peripheral T‐cell tolerance and is expressed
clonal anergy or receptor editing to eliminate in peripheral lymph nodes and the spleen
autoreactive αβT cells in the thymus [8] and where it mediates elimination of autoreac
autoreactive B cells in the bone marrow [9]. tive extrathymic T cells.
The tolerogenic mechanisms for different Thymic editing of TCRs plays a seminal
immune cell lineages show similarities and role in the genesis of the TCR repertoire and
differences. is mediated by recombination activating gene
(RAG)‐1 and RAG‐2 proteins. After RAG
Central T‐cell Tolerance proteins are induced during T‐cell
T cells with high affinity TCRs for self development, they cause TCR β‐chain rear
HLA–self peptide complexes are eliminated rangement and expression of prefabricated
in the thymus by clonal deletion [8]. This TCR α‐chains to form pre‐TCRs. A
process of negative selection favors the subsequent round of RAG protein induction
survival of T cells with TCRs that exhibit transcribes rearranged TCR α‐chains that
either no or very low affinity for self‐HLA– displace prefabricated α‐chains. Yet another
autoantigen peptide complexes. Among the round of induction of RAG proteins finalizes
natural (n) αβT cells selected in the thymus rearrangements of α‐chains, forming
are subsets of CD4 T cells, including nTregs, functional α/β TCRs to undergo thymic selec
NKT cells, and IL‐17‐secreting nTh17 cells. tion. This thymic process of TCR editing has
Given the dichotomous functions of nTregs a counterpart in the periphery, referred to as
and nTh17 cells to suppress versus augment TCR revision, which can generate an
inflammatory pathology, thymic regulation unlimited array of antigen‐specific TCRs,

of the balance of these subsets is an impor overcoming the effects of thymic TCR selec
tant determinant of the risk for autoimmu tion. These newly generated TCRs likely con
nity. Overall, more than 95% of T cells tribute to peripheral creation and evolution
generated in the thymus undergo deletion of autoreactive T cells (discussed later).
because of excessive self‐reactivity, which
underscores the high frequency of autoreac Central B‐cell Tolerance
tive TCRs among T cells. B cells develop tolerance to self‐antigens in
Negative selection of autoreactive T cells the bone marrow by both B‐cell receptor
by is mediated by the coordinated display of (BCR) editing and clonal deletion of autore
self‐HLA molecules containing self‐pep active B cells [9]. BCR editing results from
tides generated by expression of genes con the induction of VDJ recombinase and rear
trolled by the autoimmune regulator (AIRE) rangements which lead to production of new
transcription factor. In addition to DCs, immunoglobulin‐edited light chains that
medullary thymic epithelial cells (mTECs) modify the specificity of BCRs to prevent
selectively express AIRE, which drives recognition of autoantigens.
Peripheral Tolerance FoxP3 Tregs. FoxP3 Tregs secrete immuno

regulatory and immunosuppressive cyto
Central immunologic tolerance for both T
kines, including, IL‐9, IL‐35, IL‐10 and
and B cells occurs in the early developmental
transforming growth factor (TGF)‐β. These
stages of both lineages. Additional mecha
cytokines immunosuppress activated CD4
nisms are required, however, to achieve toler
and CD8 T cells and promote T‐cell anergy
ance to self‐antigens that are not expressed
by downregulating APC expression of
during fetal or neonatal life or are expressed
costimulatory CD80 and CD86. In addition,
exclusively in peripheral non‐lymphoid
Treg upregulation of intracellular cyclic
organs [10].
AMP directly inhibits cell proliferation and
T‐ and B‐cell Clonal Anergy reduces production of IL‐2 required for
Clonal anergy is a major mechanism in the T‐cell proliferation.
generation and maintenance of peripheral
tolerance to self‐antigens (Figure 2.2) [10]. Natural and Inducible T Regulatory Cells
Autoreactive T‐cell clones not deleted in the Tregs are classified as natural Tregs (nTregs)
thymus or the peripheral lymphoid tissues or inducible Tregs (iTregs) [11]. nTregs are
may become anergic to autoantigens if TCR produced in the thymus and represent
activation occurs in the absence of obliga 5–10% of the total CD4 T‐cell population.
tory costimulation. Lack of costimulation Despite having a relatively high affinity for
results in inadequate secretion of IL‐2 and autoantigens, nTregs escape clonal deletion
growth factors required for T‐cell clonal in the thymus. In the periphery, nTregs act
proliferation and maturation of effector as autoantigen‐specific sentinels within the
functions. Without costimulation, autoanti lymph nodes and spleen to maintain
gen‐specific T cells cease to differentiate and peripheral tolerance. They do so by inhibit
fail to respond if reexposed to autoantigen. ing autoantigen activation of T cells by
Anergic B cells have reduced surface IgM, APCs, causing direct cytotoxicity of auto
impaired signal transduction and short lifes antigen‐activated T cells and secreting
pans, and exhibit anergy to autoantigen anti‐inflammatory cytokines IL‐10 and
reexposure. TGF‐β.
iTregs are a subset of CD4 T cells activated
T‐cell Mediated Immune Regulation in the periphery (Figure 2.2). When naive
A subclass of CD4 T cells referred to as CD4 T cells (CD4 Th0 cells) are activated in
Tregs are essential for peripheral tolerance the presence of IL‐10, IL‐4 and TGF‐β, they
and maintenance of immune homeostasis differentiate into antigen‐specific CD4,
(Figure 2.2) [11]. Tregs mediate immune CD25, FoxP3 iTregs. Both foreign antigens
tolerance by functionally suppressing all
and autoantigens can generate antigen‐
activated CD4 T‐cell subsets and CD8 CTLs. specific iTregs, conferring immunoregula
Thus, reductions in the numbers or tory importance in both normal immunity
functional capacity of Tregs can facilitate and autoimmunity. iTregs suppress effector
autoimmunity and autoimmune diseases. CD4 T‐cell subsets and cytotoxic CD8 T cells
Phenotypically, Tregs are CD4, CD25, fork by secreting immunosuppressive IL‐10 and
head box P3 (FoxP3) T cells. FoxP3 is the key TGF‐β, inducing cell cycle arrest and effector
regulator of Treg function, as shown by T‐cell apoptosis. In addition, iTregs block
monogenic deficiency in FOXP3 that results the costimulation and maturation of DCs.
in immunodysregulation, polyendocrinopa Among iTregs (Figure 2.2), the T regulatory 1
thy and enteropathy X‐linked (IPEX) syn (Tr1) subclass exclusively secretes IL‐10 but
drome [1]. FoxP3‐deficient mice also develop does not express FoxP3. In contrast, the T
autoimmune diseases, which can be pre helper 3 (Th3) subclass exclusively secretes
vented by adoptive transfer of CD4, CD25, TGF‐β.
Peripheral B‐cell Regulatory regulation [14]. Low concentrations of IL‐2

Mechanisms promote iTreg proliferation and survival but
Activated B regulatory cells (Bregs) also are suboptimal for Th17 proliferation and
secrete immunosuppressive cytokines IL‐10, differentiation. However, the adverse effects
IL‐35 and TGF‐β. IL‐10 and IL‐35 render of low concentrations of IL‐2 on Th17 cells
CD4 Th1 and Th17 cells incapable of medi can be overcome by proinflammatory IL‐1β.
ating immunopathology or producing proin Conversely, high concentrations of IL‐2 drive
flammatory cytokines [12]. In contrast, Breg proliferation of CD4 and CD8 effector cells,
secretion of TGF‐β promotes antigen‐ including Th17 cells. iTregs also exhibit plas
activated CD4 T‐cell differentiation into ticity based on the cytokine milieu. TGF‐β
iTregs (Figure 2.2), secreting IL‐10 that promotes differentiation of either iTregs or
inhibits TNF‐α production. Additional Th17 cells by inducing expression of FOXP3
mechanisms also contribute to B‐cell immu and receptor‐related orphan receptor
noregulation. Secreted IgM induces anti‐ (ROR)γ, respectively. The cytokine environ
inflammatory apoptotic bodies to reduce ment dictates differentiation toward either
proinflammatory cytokines. Contact bet regulatory iTreg or the proinflammatory
ween B and CD4 T cells reduces CD4 T‐cell Th17 phenotypes. In the absence of IL‐6 and
proliferation and secretion of Th1. Binding IL‐21 (Figure 2.2), FoxP3 binds to RORγ sup
of B cell CD95L (Fas ligand) to CD95 (Fas) on pressing its transcriptional activity and pre
T cells induces T‐cell apoptosis. Secretion of venting Th17 differentiation. In the presence
autoantigen‐specific IgG4 inhibits ADCC, C′ of IL‐6 and IL‐21, FoxP3 dissociates from
activation, and formation of immune RORγ, allowing Th17 differentiation.
complexes because of the low affinity of IgG4 Retinoic acid from gut DCs also suppresses
for Fc receptors, low capacity to activate C′, Th17 cells, while expanding iTregs. The
and its competition with other antibodies. resulting balance between iTregs and Th17
dictates tissue immunopathology in autoim
Regulatory Dendritic Cells mune diseases.
DCs are potent professional APCs with the In addition, iTregs and Th17 cells can
dual roles of initiating adaptive immune undergo interconversion [14]. Thus, Th17
responses against deleterious antigens and cells can convert to immunosuppressive
maintaining immunologic homeostasis by IL‐10‐secreting cells, and FoxP3‐positive
inducing antigen‐specific tolerance in the iTregs can convert to Th17 cells. Conversion
periphery [13]. Only immature DCs exhibit of iTregs to Th17 cells requires a milieu con
regulatory functions, and immature DCs are taining IL‐1β, IL‐6, IL‐23, and TGF‐β.
concentrated in the liver. Immunosuppressive Activated epithelial target cells, including
cytokines or drugs, microbial products, cholangiocytes, secrete cytokines favoring
interactions with Tregs or NKT cells, and conversion of iTregs to Th17. Thus, target tis
phagocytosis of apoptotic blebs activate sues can produce an imbalance of Th17 and
regulatory DCs (DCregs). DCreg presenta iTregs, intensifying chronic inflammation.
tion of peptide antigens to CD4 and CD8 T
cells contributes to antigen‐specific toler
ance by inducing T‐cell anergy, inhibiting
isk Factors for
R
T‐cell function, promoting generation of
iTregs, and enhancing T‐cell apoptosis. Autoimmune Diseases
Genetics
Immunoregulatory Interplay Between
Treg and Th17 Cells Complex Genetic and Monogenic Diseases
The interplay and balance between iTregs and Complex genetic diseases require an inter
pathogenic Th17 cells is pivotal for immuno play of genes and environmental factors that
may manifest as more than one clinical dis HLA‐DRB1*03:01 haplotypes associated
ease (Figure 2.1) [1]. Overall, environmental with autoimmune diseases. It is notable that
exposures are more important than genetics even healthy persons with the 8.1 haplotype
in shaping the immune repertoire. Thus, a exhibit an enhanced immunologic pheno
person’s “exposome” – the lifelong sum of all type with autoantibodies, circulating
endogenous and exogenous environmental immune complexes, activated T cells, and
exposures interacting with genetic and increased apoptosis of lymphocytes. The
epigenetic factors – can favor autoimmunity HLA‐DRB1 gene is of interest because it is
or protect against it. the most polymorphic gene within the HLA
Genome‐wide association studies (GWAS) class II region. Thus, it is not surprising that
in autoimmune diseases, including AILDs, PBC is associated with the DR allele HLA‐
have demonstrated shared associations with DR8. Future studies are needed to define how
HLA alleles and single nucleotide polymor different HLA alleles contribute to a break in
phisms (SNPs) of genes related to innate tolerance and development of specific auto
immunity and adaptive immunity [15]. immune diseases.
These include genes encoding interleukins,
interleukin receptors, proinflammatory Non‐HLA Gene Associations
cytokines and their receptors, cytotoxic T Interpreting the risks for autoimmune dis
lymphocyte antigen (CTLA)‐4 (a mediator eases conferred by non‐HLA genes is prob
of senescence and apoptosis of activated T lematic because the odds ratios for risk of
cells), signal transducer and activator autoimmune diseases are far lower than
(STAT)‐4, chemokine receptors and CD69 those for HLA alleles [15]. Polymorphisms
(involved in migration of lymphocytes from in genes encoding CTLA‐4, TNF‐α, Fas
lymph nodes and development of immuno (CD95), TNF‐induced protein 3 (TNFAIP3),
logic “memory”). However, odds ratios for macrophage migration inhibitory factor
each gene’s contribution to risk for autoim (MIF), and SH2B adapter protein 3 (SH2B3)
munity are quite small, suggesting that mul have been implicated in susceptibility to
tiple SNPs are required for development of autoimmune diseases, including AILDs.
one or more of the 80 known autoimmune CTLA‐4 is critical for downregulating
diseases. effector immune responses and producing
T‐cell senescence. TNF‐α is a potent
HLA Risk Alleles proinflammatory effector cytokine and

The strongest HLA‐associated risks for auto therapeutic target in specific autoimmune
immune diseases, including AIH and PSC, diseases. Dysfunction of genetic variants or
but not PBC, lie within the evolutionarily deficient levels of TNFα‐induced protein
conserved 8.1 ancestral haplotype: HLA‐A1, 3‐interacting protein 1 (TNIP1) predispose
Cw7, B8, TNFAB*a2b3, TNFN*S, C2*C, Bf*s, normal innate cells to produce abnormal
C4A*Q0, C4B*1, DRB1*03:01, DRB3*01:01, inflammatory responses to innocuous TLR
DQA1*05:01, DQB1*02:01 [2]. The extended ligands. TNFAIP3 is a modifying enzyme for
8.1 haplotype is the result of linkage disequi ubiquitin, which may influence antigenicity
librium, indicating an evolutionary advantage of self‐proteins. CD95 (Fas) and CD95L
to sequestration of these alleles within class I, (FasL) variants promote survival of autore
II and III HLA loci. These alleles are impor active T and B cells and modify apoptosis of
tant for robust CD4 and CD8 T‐cell activation target cells. MIF is a proinflammatory cyto
and generation of TNF‐α/β and C′ factors. kine in both innate and adaptive immunity.
The 8.1 haplotype is associated with AIH and SH2B3 suppresses cytokine‐induced inflam
PSC. Most Caucasians with HLA‐B8, mation by downregulating Janus kinases
DR3 have the 8.1 haplotype; however, (JAKs) and receptor tyrosine kinases (RTKs).
recombinations have produced additional
Several autoimmune diseases also are linked
to SNPs in cytokine and cytokine receptor immune cells. lncRNAs promote suscepti
genes. A pertinent example is the IL23R bility to autoreactivity and perpetuation in
gene, which encodes a protein that forms a autoimmune diseases, including systemic
dimer with IL‐12Rβ1 to create the IL‐23 lupus erythematosus (SLE), rheumatoid
receptor complex required for IL‐23 arthritis (RA), multiple sclerosis (MS), and
induction of proinflammatory Th17 cells. psoriasis.
MicroRNAs
Critical Role of Epigenetics
MicroRNAs (miRNAs) function epigeneti
Transcription Factor Enhancers cally as posttranscriptional and posttransla
and Super Enhancers tional regulators of gene expression [17].
Gene transcription is an essential process in miRNAs and TFs regulate each other’s
autoimmunity and autoimmune disease expression in feedback loops: TFs are nega
specificity. Emerging evidence indicates that tively regulated by miRNAs, while miRNAs
the generation and perpetuation of autoim are positively regulated by TFs. Compelling
munity is primarily driven by epigenetics evidence indicates that miRNAs control the
[16]. Over 90% of the thousands of disease‐ activation state of innate macrophages and
associated SNPs detected by GWAS are DCs, activation of T cells, apoptotic elimina
localized within non‐coding regions of DNA, tion of activated T and B cells, and the critical
that produce regulatory enhancers, some balance between iTregs and Th17 cells in
times occurring in clusters spanning 50 kb of sites of inflammation. To date multiple
DNA, called super enhancers (SEs). Multiple miRNAs have been implicated in the

enhancers and SEs control the expression of immunopathogenesis of autoimmune dis

protein‐coding genes. These enhancers, eases, including AILDs.
especially SEs, are especially important
drivers of cell‐ and tissue‐specific genes.
Sex and Sex Hormones
Enhancer functions are mediated through
transcription‐factor (TF) binding motifs that Most autoimmune diseases have a female
require TF recruitment and nucleation of predilection [1,18]. However, PSC is a
transcriptional machinery. Thus, enhancer notable exception among AILDs (Table 2.1).
activity depends on the local structure of Sex is genetically binary and imprinting of
chromatin responsible for the packaging of XX female and XY male genomes occurs in
DNA in association with histone proteins. all cells from embryogenesis throughout
Since tightly packed DNA is inaccessible to life [18]. Both innate and adaptive immune
DNA‐binding TFs, additional epigenetic cells have sex hormone receptors, and sex
mechanisms are required to render DNA hormones contribute to the development
accessible to enhancer TFs. These processes and activity of the immune repertoire
include methylation of DNA and posttran (Figure 2.1). Yet it is unclear whether female
scriptional modification of histones through predominance in autoimmune diseases
methylation, acetylation, or ubiquitination. reflects regulatory mechanisms related to
Epigenetic proteins also interact directly sex‐linked genes, sex‐specific hormones or
with transcriptional mechanisms and a combination of both. Females normally
enhancer DNA can also generate enhancer produce higher titers of antibodies, more
RNA (eRNA) to increase the amount of autoantibodies and greater cell‐mediated
enhancer proteins. Epigenetic long non‐cod immunity to infections and immunization
ing RNAs (lncRNAs) are expressed within than do normal males. In addition, there is
innate immune cells and CD4 Th1, Th2, a gender bias in AIRE expression in the
Th17, Tregs, CD8 and B cells. lncRNAs thymus, which may impact deletion of
directly regulate activation and functions of autoreactive clones and formation of nTregs
and Bregs. Estrogen levels are also immu c orrelation between the incidence of autoim
noregulatory: high levels inhibit CD4 Th1‐ mune disease and increasing latitude.
mediated cellular responses and promote SNPs in two VDR genes, BsmI and TaqI,
CD4 Th2 antibody‐mediated responses. are associated with AILDs. SNPs in these
Prolactin, progesterone and testosterone genes and FokI are associated specifically
also regulate immune responses by modu with AIH. Dietary vitamin D and vitamin D
lating expression of estrogen receptors and synthesized in response to ultraviolet‐B
altering cytokine secretion. However, the radiation in sunlight determine serum
effects of estrogen and other sex hormones vitamin D levels. Sequential hydroxylation
in adult women do not explain the female in the liver and kidney produce D3, the most
predilection for autoimmunity in children potent ligand for VDRs. Innate and adaptive
or elderly adults. Postulated explanations immune cells express VDRs that induce
include skewed X chromosome activation, multiple immunologic effects. For example,
X monosomy, and microchimerism. D3 regulates production of antimicrobial
Beginning during embryogenesis and peptides and reactive oxygen species in
continuing through life, either the maternal innate immune cells and promotes NKT cell
or paternal X chromosome within each cell immunomodulatory functions. D3 also
is inactivated, forming a Barr body adjacent downregulates immunopathogenic CD4
to the nucleolus. Thus, females are a mosaic Th1 and Th17 cells, while upregulating
of two cell types, based on the parental immunosuppressive CD4 Th2 cells and
source of the active X chromosome. iTregs. By reducing Th1 secretion of IL‐2,
Skewing of X inactivation varies widely and D3 not only inhibits proliferation of CD4
increases with age. Thus, some X‐linked and CD8 effector T cells but also generates
autoantigens might theoretically escape low IL‐2 levels, favoring proliferation of
central or peripheral tolerance. The functional iTregs.
proximity of Barr bodies to the nucleolus
suggested that they may interact to expose
cryptic autoantigens. X monosomy might L oss of Immune Tolerance to
promote autoimmunity by preventing tol Autoantigens and Perpetuation
erance to autoantigens encoded by both X
chromosomes. Entry of fetal hematopoietic
of Autoimmune Diseases
stem cells into the maternal circulation
Overview
during pregnancy can cause fetal–maternal
microchimerism that could promote loss of Autoimmune diseases are uncommon,
self‐tolerance. despite the high frequency of genetic suscep
tibility and environmental exposure to
microbial pathogens and xenobiotics,
Vitamin D and Sunlight Exposure
including drugs. Autoimmunity requires
In addition to its roles in bone mineralization four essentials. First, environmental triggers
and calcium homeostasis, vitamin D modu are required to initiate loss of tolerance to
lates immune reactions and risk for autoim autoantigens in susceptible individuals
munity [19]. Both vitamin D deficiency and (Figure 2.1). Second, the individual must
genetic polymorphisms in the vitamin D have HLA alleles that facilitate autoantigen
receptor (VDR) increase the incidence and activation of T cells. Third, T cells must
severity of multiple autoimmune diseases. express autoreactive TCRs. Fourth, the
Conversely, high vitamin D levels reduce the individual must fail to immunoregulate the
risk of MS. Reduction in sunlight exposure response to autoantigens. Evidence that low
and synthesis of 1,25‐dihydroxyvitamin‐D3 levels of autoimmune reactions are common
(D3) might explain, in part, the strong in healthy people highlights the fact that
these initial steps are common but only dele As discussed earlier, bacterial metabolites of
terious when not immunoregulated. GWAS vitamin B are the activating antigens for
indicates that failure to immunoregulate an MAIT cells. Thus, the gut microbiome and
autoimmune reaction is dictated by epige intestinal immune responses appear to
netic SNPs for enhancers and SEs of gene influence systemic immunity, and risk for
expression more than SNPs for HLA and autoimmunity, through the innate and
non‐HLA proteins. Moreover, these SNPs adaptive immune responses of the liver.
are critical drivers of the cell and tissue spec Altered composition of the gut microbiota,
ificity of autoimmune diseases. Multiple referred to as “dysbiosis,” has been reported
factors and mechanisms to break tolerance in autoimmune diseases, including AILDs.
to self‐antigens have been identified. Whether dysbiosis is the cause or the effect
The increasing incidence of autoimmunity of autoimmunity remains a key unresolved
and inflammatory diseases observed world issue. Female sex hormones can also alter the
wide is correlated with changes in environ gut microbiome, which may contribute to
mental factors, including a more modern the female predilection in autoimmunity. If
lifestyle, improved hygiene, a Western diet, an altered gut microbiota or increased gut
use of antibiotics, and elimination of permeability were causally related to autoim
childhood parasitic infections. Each of these mune diseases, then restoration of a normal
affect the composition and function of the microbiota and mucosal integrity could be
gut microbiota. Although the proposed therapeutic.
causal link between the gut microbiome and
autoimmunity has not been proved, available
Mechanisms of Loss of Tolerance
data indicate that interplay between the gut
to Autoantigens
microbiota and the innate and adaptive
immune systems of the intestine and liver Bacterial and Viral Infections
plays key roles in both normal and dysfunc Bacterial, fungal or viral infections can insti
tional systemic immunity. gate innate and adaptive immune responses
that result in autoimmunity [21]. Migration
Role of the Microbiome of neutrophils to sites of infection or sterile
The gut microbiome has been implicated in inflammation mediated by IL‐8, IFN‐γ, C3a,
autoimmunity (Figure 2.1) [20]. The gut C5a, and leukotriene B4 plays a key role.
microbiota change dynamically under the Neutrophil phagocytosis of microbial organ
influence of aging, sex hormones, diet, water isms and particles leads to release of soluble
purity, hygiene, sanitation, pollution, antimicrobials and formation of neutrophil
exposure to pathogens, use of antibiotics, extracellular traps (NETs) comprising
and the integrity of the gut–blood barrier. chromatin and proteases that enmesh and
Gut microbiota are composed of commensal kill microbes. The proteolytic function of
bacteria, fungi and viruses, as well as poten NETs also can generate neoantigens or
tially pathogenic bacteria and fungi. In microbial peptide molecular mimics of auto
addition, the microbiota process food, antigens (see next section).
xenobiotic pollutants and drugs to generate Superantigens are proteins produced by
micronutrients and metabolites. Intestinal pathogenic bacteria and viruses that trigger
immune responses and the leakiness of the non‐antigen‐specific polyclonal T‐cell
mucosa dictate the types and concentrations activation of up to 20% of the host’s T cells.
of PAMPs, DAMPs, cytokines, microbes, These activated T cells create a cytokine
and food antigens entering the portal vein for storm of secreted cytokines, especially IFN‐γ,
processing in the liver. PAMPs and DAMPs which activates macrophages to secrete
stimulate enterocyte inflammasomes leading proinflammatory IL‐1β, IL‐6 and TNF‐α.
to secretion of proinflammatory cytokines. This proinflammatory environment could
romote adaptive immune responses to

p and a minority of patients with chronic HCV
autoantigens in dead or dying cells. infection, whereas haptens bound to UDP‐
glucuronosyltransferase (UGT) in those
Molecular Mimicry of Autoantigens infected with HCV or hepatitis D virus, and
Molecular mimicry is a key mechanism with Addison disease, APS‐1 syndrome and
leading to loss of tolerance to autoantigens some drug‐induced liver injuries result in
[21]. It is defined as the cross‐reactivity of anti‐UGT reactions. A subset of NK cells
immune responses to antigenic epitopes of develop antigen‐specific memory to haptens,
microbial peptides with autoantigenic epi which may be important for immune
topes of host peptides. Molecular mimicry responses after hapten reexposure. Oral sup
has been observed between human autoanti plements of lipoic acid act as a haptens in the
gens and peptide antigens of several viral, immunogenicity of the lipoic acid‐binding
bacterial and fungal pathogens. Xenobiotics site of PDC‐E2, the principal autoantigen in
may also function as molecular mimics for PBC. Biochemical modifications of self‐
autoantigens. For example, some haloge antigens can also increase the immunoge
nated xenobiotics exhibit immunogenic sim nicity of autoantigens. The best example is
ilarity with the lipoic acid‐binding domain of citrullination, produced by posttranslational
pyruvate dehydrogenase complex (PDC)‐E2, conversion of arginine to citrulline. Since
the primary autoantigen in PBC. Yet molec DNA does not encode the amino acid citrul
ular mimicry is an insufficient explanation line, citrullinated self‐proteins become auto
for a sustained loss of tolerance, which also antigenic, as in RA.
requires failure to control the autoimmune
response. Thus, molecular mimicry repre Failure of Apoptosis to Conceal
sents an environmental trigger capable of Autoantigens and Eliminate
progressing to autoimmunity. Molecular Autoreactive Cells
mimicry to adjuvants (substances that Apoptosis is normally an immunologically
enhance immune responses to an antigen) silent form of cell death [23]. Phagocytic
involves nucleosomes or ribonucleoproteins removal of the corpse (efferocytosis) of apo
released after cell death. These molecules ptotic cells and blebs by APCs normally
mimic viruses by stimulating an antiviral‐ induces release of anti‐inflammatory mole
like, innate immune response with produc cules, inhibits NF‐κB stimulation of proin
tion of type 1 interferon. flammatory cytokines, and promotes
secretion of anti‐inflammatory IL‐10 and
Neoantigens TGF‐β. Failure of these normal APC mecha
Neoantigens, also called cryptic antigens, can nisms is an attractive hypothesis for the
elicit autoimmune responses against autoan genesis of autoimmunity, because apoptotic
tigenic epitopes that are not immunogenic blebs contain disproportionately high
until modified by either somatic hypermuta concentrations of known autoantigens.

tions or binding of haptens [22]. Haptens are However, non‐apoptotic forms of cell death,
small molecules, most often metabolites of such as necrosis or necroptosis, result in
drugs or environmental xenobiotics, that are APC phagocytosis, processing and presenta
incapable of eliciting an immune response tion of autoantigenic peptides to autoreactive
unless bound to host carrier proteins. T and B cells in immunogenetically suscep
Hapten–carrier protein complexes elicit tible persons. This is the proposed explana
immune responses against a single hapten tion for AIH after documented infections
epitope and multiple autoepitopes of the car with hepatotrophic or other viruses causing
rier protein. For example, haptens generate hepatocyte necrosis. Defective apoptosis might
autoimmunity against cytochrome P450 also contribute to the persistence of autoreactive
(CYP)2D6 in type 2 AIH and autoantibodies T and B cells in autoimmune diseases. Genetic
defects in Fas (CD95), Fas‐L (CD95L) and RAS T‐cell Receptor Revision
pathways cause lymphoproliferative disorders in the Periphery
and autoimmunity due to inability to eliminate The fact that the TCR repertoire selected in the
autoreactive effector cells. thymus can be altered in the periphery helps
explain the presence of rogue autoreactive T
Immune Deviation of Activated T Cells cells [24]. Surface expression of the costimula
Immune deviation refers to the evolution of tory molecule CD40, thought to be restricted
dominant populations of effector T cells, to APCs, has been observed recently on neural
which alter local immune responses and com cells, endothelial cells, adipocytes, and subsets
promise tolerance [1]. Activated CD4 Th0 of CD4 and CD8 T cells. Activation of CD40
differentiate into multiple functional Th1 expressed by the mature CD4 Th subset,
subsets differing in cytokine production designated Th40 cells, induces RAG‐1/RAG‐2‐
(Figure 2.2). CD4 Th1 cells secrete IL‐2, the mediated rearrangement of both TCR α‐ and
mitogen for all activated CD4 and CD8 T β‐chains in the periphery. This refutes early
cells, as well as IFN‐γ and TNF‐α. Th1 cells dogma of the immutability of the TCR reper
activate macrophages and stimulate B cells to toire and its antigen specificity after T cells
secrete C′‐fixing IgG2. In contrast, Th2 cells emigrate from the thymus. The fact that Th40
secrete IL‐4, IL‐5, IL‐10, and IL‐13 that stim cells can undergo multiple cycles of TCR revi
ulate B cells to secrete IgG, IgM and IgA anti sion multiplies the risk of developing autoreac
bodies, while immunosuppressing the effects tive TCRs, including ones recognizing unique
of Th1 cytokines. Th1 predominance pro self‐neoantigens. The finding that adoptively
duces greater immunopathology associated transferred Th40 cells induced antigen‐specific
with autoimmune diseases that cause tissue type 1 diabetes mellitus (T1DM) in non‐obese
damage. The signature cytokines of Th1 and diabetic severe combined immunodeficiency
Th2 cells inhibit the proliferation and secre mice confirms their immunogenic function.
tion of the cytokines of each other, resulting Thus, TCR revision to recognize autoantigens
in a dynamic balance. Skewing of this balance represents a new paradigm, explaining autore
contributes to either the maintenance or the active peripheral T cells.
loss of tolerance. The T follicular helper (Tfh)
cell secretes IL‐21, which is best known for
Perpetuation of Autoimmune
inducing differentiation of activated B cells
Diseases
into memory B cells and plasma cells.
Additional pluripotent effects of IL‐21 impact The primary factor in perpetuation of autoim
both innate and adaptive immune functions mune diseases is failure to immunoregulate
to increase immunopathology. In innate and terminate the initial activation of B and
immunity, IL‐21 increases antigen processing T cells to autoantigens or their mimics [1].
and presentation by APCs, a ctivates macro Multiple factors contribute to perpetuation.
phages to chemoattract neutrophils, increases
NK cell cytotoxicity, including ADCC and Epigenetics
secretion of IFN‐γ, and induces NKT cell pro Enhancers and Super Enhancers
liferation and secretion of IFN‐γ, IL‐2, IL‐4, As discussed earlier, epigenetics plays a crucial
IL‐13, and IL‐17A. In adaptive immunity, role in regulating TFs in innate and adaptive
IL‐21 induces proliferation and differentiation immune responses involved in autoimmune
of Th17 cells and increases both the cytotox diseases and determining specificity for differ
icity of CD8 CTLs and their secretion of ent cell types and tissues [16]. While the mech
IFN‐γ and TNF‐α. Thus, polarization of anisms already discussed focus on factors
immune responses toward Th1 and Tfh cells leading to autoantigen recognition, it is now
greatly increases the consequences of autore clear that epigenetics dictates the subsequent
active T and B cell activation. obligatory failure to regulate and terminate
autoimmune reactions. GWAS showed that mechanisms of apoptosis or other forms of

SNPs associated with immune‐mediated dis efferocytosis to prevent presentation of auto
eases encode enhancers and SEs that cluster antigens. The hypothesis that apoptosis is
within the immunoregulatory regions of involved in epitope spreading is attractive
immune cells. Indeed, analysis of the causal because apoptotic blebs do not contain
variants in 21 autoimmune diseases showed random samples of intracellular constitu
that 60% of SNPs mapped to immune cell ents, but instead contain high concentrations
enhancers and eRNA induced by immune cell of known autoantigens.
activation. Disease‐specific profiles are
beginning to be recognized. For example, Tissue Memory T Cells
T1DM‐associated SNPs are present in Tissue‐resident memory T (TRM) cells are a
enhancers active in the thymus, T cells, B cells, newly identified subset of non‐circulating
and hematopoietic stem cells. The finding that antigen‐specific memory T cells that persist
SNPs for SEs in CD4 Thf cells are present in long term in epithelial barriers (e.g. skin, lung,
multiple autoimmune diseases, including gut and reproductive tract) and in brain,
T1DM, MS, RA, celiac sprue, and ulcerative kidney, joints, and pancreas [25]. TRM cells
colitis, strongly indicates their importance in are distinct from circulating memory T cells
pathogenesis. In addition, epigenetics controls and provide immediate protective responses
suppression of autophagy, which dysregulates upon reexposure to antigens. However, auto
immune modulation in effector cells and pro reactive TRM cells have been implicated in
motes persistence of cytokine‐activated target chronic autoimmune diseases, including
cells/tissues in autoimmune diseases. AIH, vitiligo, psoriasis, and RA. The longevity
and privileged location of autoreactive TRM
MicroRNA cells within tissues favors perpetuation of
Numerous studies indicate that miRNAs are autoinflammation and tissue destruction.
important epigenetic regulators of proteins
and networks involved in the generation and Cytokines Promoting Chronic
perpetuation of autoimmunity [17]. The Inflammation and Autoimmunity
mechanisms contributing to progressive IL‐12 Family
autoimmunity include hyperactivation of The family of IL‐12 cytokines – IL‐12, IL‐23,
innate macrophages and DCs, hyperstimula IL‐27 and IL‐35 – are secreted by activated
tion of autoreactive T cells, inhibition of apo innate DCs and macrophages [1]. Each
ptotic elimination of autoreactive T and B cytokine can perpetuate autoimmune

cells, and alteration of regulatory balance. inflammatory diseases by regulating CD4 T‐
The impact of miRNAs alters the balance cell subsets in the adaptive immune response
between pathogenic Th17 and iTregs by pro (Figure 2.2). IL‐12 and IL‐23 are among the
moting dominance of Th17 cells. most potent proinflammatory cytokines, while
IL‐27 and IL‐35 are anti‐inflammatory and
Epitope Spreading immunosuppressive. IL‐12 drives initial and
Autoimmunity begins with loss of tolerance subsequent differentiation of naive CD4 Th0
to a specific autoantigen, but later expands to cells into Th17 cells and γδ T cells to produce
include reactions to additional autoantigens, inflammatory IL‐17. Recent studies indicate
a phenomenon called epitope spreading [1]. that IL‐21, secreted by Thf and Th9 cells, is the
By extending the autoreactive T‐cell and B‐ most dominant factor in the g eneration and
cell repertoire mediating autoimmune dis maintenance of immunopathology (Figure 2.2).
ease, epitope spreading greatly reduces the
prospect that host immunoregulatory mech IL‐20 Receptor Cytokines
anisms can resume control. The causes of IL‐19, IL‐20 and IL‐24 signal through the
epitope spreading are poorly defined but may IL‐20RA/RB receptor complex [26]. Theses
involve failure of normal anti‐inflammatory cytokines are expressed by both immune and
epithelial cells and result in bilateral stimula ation of autoimmunity provide conceptual as
tion. These cytokines have a dual potential well as realistic targets for interventions to
to immunoregulate innate and adaptive prevent and treat autoimmune diseases [1].
immune responses or to promote immuno Clearly, genetic and epigenetic SNPs and envi
pathogenesis. These cytokines are most often ronmental exposures cannot be eliminated as
pathogenic in autoimmune diseases, perpet risk factors. However, it may become possible
uating chronic tissue inflammation. to identify children at risk of autoimmunity
and develop strategies to reduce their risk of
Tertiary Lymphoid Structures autoimmune diseases. The development of
and Germinal Centers new immunosuppressive medications, inhibi
Tertiary lymphoid structures (TLSs) frequently tion of cytokine production and function, and
develop in autoimmune diseases that contain epigenetic inhibitors increase the probability
T‐cell and B‐cell zones capable of producing of controlling a variety of autoimmune dis
adaptive cellular and humoral immune eases in the near future.
responses [27]. TLSs portend poor prognosis
due to increased autoactivation of T‐ and B‐cell Strategies to Prevent Autoimmunity
clones. Excessive production of autoantibodies
against initiating autoantigens and those Vitamin D3
involved with epitope spreading augment anti Vitamin D deficiency is epidemiologically
body‐mediated cytotoxicity during the evolu associated with risk of autoimmunity [19].
tion of autoimmune diseases (Figure 2.2). In Achieving and maintaining high normal
addition, APC phagocytosis and processing of serum levels of vitamin D3 is a realistic and
immune complexes composed of autoantigens achievable goal, which could reduce the
and autoantibodies dramatically increase the incidence and severity of autoimmune
quantity of autoantigen‐specific CD8 CTLs. diseases.
Tfh cells and their signature cytokine IL‐21
drive B‐cell production of high‐affinity autoan Gut Microbiota Manipulation
tibodies, differentiation of memory B cells, and in Pregnancy and Infancy
APC functions of activated B cells within the It remains unclear if changes in the fecal
expanded number of germinal centers in TLSs. microbiota associated with specific autoim
mune diseases represent causes or effects
Epithelial Cell‐induced Transformation [20]. Were causal relationships identified for
of iTreg to Th17 Cells either initiation or perpetuation of autoim
As discussed earlier, cytokines produced by munity, several strategies theoretically could
MAIT cells can stimulate epithelial cells, decrease the risk of autoimmunity, espe
including cholangiocytes, to secrete IL‐6, IL‐1β, cially if used during pregnancy and infancy.
IL‐23 and TGF‐β [7]. This combination of cyto These include probiotics to shape the evolu
kines can transform iTregs into activated Th17 tion of the gut microbiota and sustain the
cells that promote a proinflammatory Th17‐ mucosal barrier, fecal microbiota transplan
mediated inflammation (Figure 2.2). tation to create a preventive gut microbi
ome, and either deliberate infection with
non‐pathogenic helminths or ingestion of
revention of Autoimmunity
P specific parasite peptides to promote a
systemic immunosuppressive Th2 environ
and Therapeutic Control
ment [28].
of Autoimmune Diseases
Oral Tolerance
Overview
Ingestion of antigens leads to absorption and
Our current understanding of the mecha interaction with the reticuloendothelial
nisms involved in the generation and perpetu system of the liver. Repeated ingestion can
result in oral tolerance, defined as the T1DM, HCV‐associated vasculitis, and SLE.
inability to respond to the ingested antigen The second strategy involves generating
when given parenterally [29]. Low doses of antigen‐specific iTregs from peripheral
antigen induce iTregs, while higher doses blood mononuclear cells ex vivo for infusion.
create anergic tolerance. Studies involving iTregs specific for CYP2D6 autoantigens in
the induction of oral tolerance to glutamic type 2 AIH have been produced, but not yet
acid decarboxylase, the instigating autoanti infused.
gen in T1DM, are ongoing. The applicability
of oral tolerance depends on identification of Epigenetic Enhancer Regulation
disease‐specific initiating autoantigens; thus, SNPs within enhancers, especially SEs, play
it would be currently relevant in AILDs only dominant roles in autoimmune pathogenesis
for PBC and type 2 AIH. (see section Epigenetics) [16]. The theoret
ical fear that targeting epigenetic regulatory
proteins might cause severe toxicity has
Strategies to Treat Established
Autoimmune Diseases
not been observed with first‐generation
inhibitors. Among SE proteins regulating

At the time of diagnosis, the initiating events genes associated with activation, stress and
of most autoimmune diseases likely occurred differentiation, the bromodomain and extra‐
months to years earlier and mechanisms of terminal (BET) family has emerged as a
perpetuation are well established. This poses prime target for inhibition. BET inhibition
challenges for clinical management and decreased macrophage expression of
focuses attention on therapeutic strategies to inflammatory genes induced by lipopolysac
control inflammation, modify symptoms and charide by preferentially inhibiting de novo
signs, and retard progression. Progress in expression of SE genes. BET inhibitors also
understanding the pathogenesis of autoim blocked de novo SE gene expression in endo
mune diseases provides rationales for addi thelial cells, markedly reduced CD4 Th cells
tional conventional and novel therapies differentiation to polarized subsets, and pre
(Table 2.3). Two therapeutic approaches vented effector cytokine production by
warrant additional comment. already polarized Th cells. BET inhibitors
also impacted B cells by inhibiting prolifera
Inducible T Regulatory T Cells tion and ability to switch the isotype of their
Studies of two strategies to produce iTregs antibodies. More importantly, BET inhibi
specific for autoantigens are in progress tors have reduced inflammation and pro
(Table 2.3) [1]. The first strategy is to increase tected from disease in animal models of
the proliferation of existing iTreg popula T1DM, MS, RA, and psoriasis. These data
tions by infusion of low doses of IL‐2, which indicate that BET‐specific, and more gener
is being studied in hematopoietic stem cell ally SE‐related, mechanisms are promising
transplantation, graft‐versus‐host disease, therapeutic targets.
Table 2.3 Current and future therapeutic approaches to control autoimmune diseases.
Approaches Goal Status
Increase regulatory Infuse low‐dose IL‐2 Expand existing autoantigen‐specific iTregs in vivo POC established. Clinical trials ongoing
control of Infuse iTregs generated ex Ex vivo generation of disease and autoantigen‐ POC of iTreg generation established. No
autoimmune vivo to control Th cell and specific iTregs infusion trials to date. Viability, function
responses cytokine responses and distribution after infusion unknown
Bromodomain and Inhibit disease‐specific epigenetic enhancers, POC established. Clinical trials ongoing
extra‐terminal (BET) super enhancers and eRNA production
family of proteins
Mesenchymal stem cells Inhibit innate immune cells, effector T cells. POC established. Clinical trials ongoing
Induction of iTregs. Reduction of
proinflammatory TNF‐α
Decrease numbers Immunosuppressive Reduce functional mass of activated T cells using SOC in multiple autoimmune diseases.
and/or functions of drugs: CNI, mTOR, immunosuppressive drugs Active research to refine target specificity
autoimmune antiproliferative agents and reduce toxicities. Combinations of
effector cells and drugs working at different sites of action
pathogenic using subtoxic doses preferred
autoantibodies Anti‐CD20 B‐cell depletion SOC and off‐label uses
Anti‐BAFF Initial B‐cell depletion followed by mobilization of SOC in SLE. Broad future potential,
memory B cells. Concurrent inhibition of BAFF especially in combination regimens
signaling in T cells
Anti‐BAFF, followed by Depletion Sequential use to eliminate mobilized
anti‐CD20 of memory B cells mobilized by anti‐BAFF memory B cells to increase efficacy
Anti‐CD40 Block CD40–CD40L (CD154) costimulation of T POC. Clinical trial initiated in
cells and B cells transplantation
Block IgG recycling and First in class antibody fragment to block FcRn POC to reduce pathogenic autoantibodies
increase IgG clearance (efgartigimod) and formation of immunoglobulin–
autoantigen complexes
Prevent egress of activated Block sphingosine‐1‐phosphate receptors SOC in MS, new agents in development
T cells from lymph nodes
MDSCs Inhibit activation and proliferation of autoreactive POC in preclinical models. Plans for
T cells clinical trials in RA
(Continued )
Table 2.3 (Continued)
Decrease Anti‐TNF‐α or TNF‐α Decrease TNF‐α mediated tissue injury and SOC in multiple autoimmune diseases
proinflammatory receptor proinflammatory signaling
cytokines Anti‐IL‐6 or anti‐IL‐6R Decrease pathological consequences of SOC in RA, clinical trials ongoing
proinflammatory IL‐6 signaling in innate and
adaptive immune response
Anti‐IL‐12 Decrease pathological consequences of Monoclonal antibody against p40 subunit.
proinflammatory IL‐12 signaling in innate and SOC in psoriasis and Crohn’s disease.
adaptive immune response Also blocks IL‐23 signaling
Anti‐IL‐23 Decrease pathological consequences of Monoclonal antibody against p40 subunit.
proinflammatory IL‐23 SOC in psoriasis and Crohn’s disease.
Blocks IL‐23 stimulation of Th17 cells
Anti‐IL‐17 Decrease pathological consequences of Th17 SOC for psoriasis and psoriatic arthritis.
production of IL‐17 Clinical trials designed
Anti‐IL‐21 Decrease multiple pathological consequences of Clinical trials in RA, T1DM, Crohn’s
IL‐21 in innate and adaptive immune responses disease
Inhibition of IL‐2 Decrease proliferation of activated CD4 and CD8 SOC CNI and mTOR inhibitors. POC
proinflammatory T cells JAK3 inhibition
cytokine signaling IL‐6 Decrease proinflammatory IL‐6R‐mediated POC and SOC indications for JAK1/2
signaling inhibition
IL‐12/IL‐23 Decrease proinflammatory IL‐12 and IL‐23 POC and SOC indications for JAK2
signaling that polarize to a Th1 response, increase inhibition
secretion of IFN‐γ and TNF‐α, increase
cytotoxicity of NK cells and CD8 CTLs and drive
differentiation of pathogenic Th17 cells
IFN‐α/IFN‐β Decrease gene expression induced by type 1 IFNs POC and SOC indications for JAK1
inhibition
IFN‐γ Decrease proinflammatory actions of IFN‐γ POC and SOC indications for JAK1/JAK2
produced by NK, NKT, CD4 and CD8 T cells inhibition
Immunosuppressant rHuIL‐10 Antagonize Th1‐mediated pathology SOC to prevent pancreatitis post‐ERCP.

cytokines Trial in ulcerative colitis terminated for
concern over Guillain–Barré syndrome
Inhibition of Anti‐chemokine receptors Prevent injury by blocking egress of effector cells Block of α4β7 integrin ineffective in PSC.
transendothelial or integrins from blood into target tissues Potential for studies of other FDA‐
migration of effector approved chemokine/integrin inhibitors
cells
Physiologic PIF Administration to recreate immunosuppressive Phase 1b trial of synthetic PIF in AIH
immunoregulation and immunomodulatory environment of the fetus completed
and mother during pregnancy
AIH, autoimmune hepatitis; BAFF, B‐cell‐activating factor; CNI, calcineurin inhibitor; ERCP, endoscopic retrograde cholangiopancreatography; IL, interleukin; iTregs,
inducible T regulatory cells; JAK, Janus kinase; MDSCs, myeloid‐derived suppressor cells; MS, multiple sclerosis; mTOR, mechanistic target of rapamycin; PIF, pre‐
implantation factor; POC, proof of concept; PSC, primary sclerosing cholangitis; RA, rheumatoid arthritis; rHuIL‐10, recombinant human IL‐10; SLE, systemic lupus
erythematosus; SOC, standard of care; T1DM, type 1 diabetes mellitus; Th, T helper cell; TNF, tumor necrosis factor.
References
1 Devarapu, S.K., Lorenz, G., Kulkarni, O.P. 10 Theofilopoulos, A.N., Kono, D.H., and
et al. (2017). Cellular and molecular Baccala, R. (2017). The multiple pathways to
mechanisms of autoimmunity and lupus autoimmunity. Nat. Immunol 18 (7): 716–724.
nephritis. Int. Rev. Cell Mol. Biol 332: 11 Dominguez‐Villar, M. and Hafler, D.A.
43–154. (2018). Regulatory T cells in autoimmune
2 Gambino, C.M., Aiello, A., Accardi, G. et al. disease. Nat. Immunol 19 (7): 665–673.
(2018). Autoimmune diseases and 8.1 12 Taher, T.E., Bystrom, J., Ong, V.H. et al.
ancestral haplotype: an update. HLA 92 (3): (2017). Intracellular B lymphocyte
137–143. signalling and the regulation of humoral
3 Toubi, E. and Vadasz, Z. (2019). Innate immunity and autoimmunity. Clin. Rev.
immune‐responses and their role in driving Allergy Immunol 53 (2): 237–264.
autoimmunity. Autoimmun. Rev 18 (3): 13 Phillips, B.E., Garciafigueroa, Y., Engman,
306–311. C. et al. (2019). Tolerogenic dendritic cells
4 Caso, F., Costa, L., Nucera, V. et al. (2018). and T‐regulatory cells at the clinical trials
From autoinflammation to autoimmunity: crossroad for the treatment of autoimmune
old and recent findings. Clin. Rheumatol 37 disease: emphasis on type 1 diabetes
(9): 2305–2321. therapy. Front. Immunol. 10: 148.
5 Shen, H.H., Yang, Y.X., Meng, X. et al. 14 Lee, G.R. (2018). The balance of Th17
(2018). NLRP3: a promising therapeutic versus Treg cells in autoimmunity. Int. J.
target for autoimmune diseases. Mol. Sci 19 (3): 730.
Autoimmun. Rev 17 (7): 694–702. 15 Inshaw, J.R.J., Cutler, A.J., Burren, O.S.
6 Grant, C.R. and Liberal, R. (2017). Liver et al. (2018). Approaches and advances in
immunology: how to reconcile tolerance the genetic causes of autoimmune disease
with autoimmunity. Clin. Res. Hepatol. and their implications. Nat. Immunol 19
Gastroenterol 41 (1): 6–16. (7): 674–684.
7 Chiba, A., Murayama, G., and Miyake, S. 16 Tough, D.F. and Prinjha, R.K. (2017).
(2018). Mucosal‐associated invariant T cells Immune disease‐associated variants in
in autoimmune diseases. Front. Immunol. 9: gene enhancers point to BET epigenetic
1333. mechanisms for therapeutic intervention.
8 Cheng, M. and Anderson, M.S. (2018). Epigenomics 9 (4): 573–584.
Thymic tolerance as a key brake on 17 Di Marco, M., Ramassone, A., Pagotto, S.
autoimmunity. Nat. Immunol 19 (7): et al. (2018). MicroRNAs in autoimmunity
659–664. and hematological malignancies. Int. J. Mol.
9 Ray, A. and Dittel, B.N. (2017). Mechanisms Sci 19 (10): 3139.
of regulatory B cell function in autoimmune 18 Moulton, V.R. (2018). Sex hormones in
and inflammatory diseases beyond IL‐10. acquired immunity and autoimmune
J. Clin. Med 6 (1): 12. disease. Front. Immunol. 9: 2279.
19 Dankers, W., Colin, E.M., van Hamburg, how unique costimulatory molecules
J.P., and Lubberts, E. (2016). Vitamin D in contribute to diabetogenesis. Front.
autoimmunity: molecular mechanisms Endocrinol 8: 208.
and therapeutic potential. Front. Immunol. 25 Molodtsov, A. and Turk, M.J. (2018). Tissue
7: 697. resident CD8 memory T cell responses in
20 De Luca, F. and Shoenfeld, Y. (2019). The cancer and autoimmunity. Front. Immunol.
microbiome in autoimmune diseases. Clin. 9: 2810.
Exp. Immunol 195 (1): 74–85. 26 Chen, J., Caspi, R.R., and Chong, W.P.
21 Rojas, M., Restrepo‐Jimenez, P., Monsalve, (2018). IL‐20 receptor cytokines in
D.M. et al. (2018). Molecular mimicry and autoimmune diseases. J. Leukoc. Biol 104
autoimmunity. J. Autoimmun 95: 100–123. (5): 953–959.
22 White, K.D., Chung, W.H., Hung, S.I. et al. 27 Pipi, E., Nayar, S., Gardner, D.H. et al.
(2015). Evolving models of the (2018). Tertiary lymphoid structures:
immunopathogenesis of T cell‐mediated autoimmunity goes local. Front. Immunol.
drug allergy: the role of host, pathogens, 9: 1952.
and drug response. J. Allergy Clin. Immunol 28 Wang, M., Wu, L., Weng, R. et al. (2017).
136 (2): 219–234. quiz 235. Therapeutic potential of helminths in
23 Abdolmaleki, F., Farahani, N., Gheibi autoimmune diseases: helminth‐derived
Hayat, S.M. et al. (2018). The role of immune‐regulators and immune balance.
efferocytosis in autoimmune diseases. Parasitol. Res 116 (8): 2065–2074.
Front. Immunol. 9: 1645. 29 Commins, S.P. (2015). Mechanisms of oral
24 Wagner, D.H. Jr. (2017). Overlooked tolerance. Pediatr. Clin. North Am 62 (6):
mechanisms in type 1 diabetes etiology: 1523–1529.
47
Genetics and Risk of Autoimmune Liver Diseases

George F. Mells1,2
1
Division of Gastroenterology and Hepatology, Department of Medicine
2
Academic Department of Medical Genetics, University of Cambridge, Cambridge, UK
Abstract
There have been major advances in understanding the human leukocyte antigen (HLA) associations
with autoimmune liver disease using serotyping, phenotyping and sequencing based HLA association
studies, and more from imputation of classical HLA alleles from genomewide association studies
(GWAS) or iChip genotype data. This chapter summarizes genetic associations that have been identi-
fied largely as a result of GWAS and related study designs. GWAS and related study designs have sub-
stantially improved knowledge of the genetic architecture of primary biliary cholangitis (PBC) and
primary sclerosing cholangitis (PSC), especially outside the HLA region. Numerous genomewide
significant non-HLA risk loci for PBC or PSC have been identified. Although genetic studies of PBC or
PSC have successfully identified numerous risk loci, it is estimated that less than 20% of PBC heritability
and less than 10% of variance of PSC liability has been explained, the so-called ‘missing heritability’.
Keywords Autoimmune liver disease; genomewide association studies; human leukocyte

antigen (HLA) a ssociations; non-HLA associations; primary biliary cholangitis; primary sclerosing
cholangitis; risk loci
Key Points
●● There have been major advances in under- ●● Many of these HLA associations are
standing the human leukocyte antigen shared with other autoimmune diseases.
(HLA) associations with autoimmune liver ●● In addition to HLA‐associated genes,
disease using serotyping, phenotyping and non‐HLA gene loci have been identi-
sequencing‐based HLA association studies, fied in association with presence of
and more recently from imputation of disease.
classical HLA alleles from genome‐wide ●● Only a relatively small proportion of
association studies or iChip genotype data. overall genetic risk of developing autoim-
●● For primary biliary cholangitis (PBC), pri- mune liver diseases is explained by the
mary sclerosing cholangitis (PSC) and new loci determined so far.
autoimmune hepatitis (AIH), both risk ●● Such loci, particularly in the case of PBC,
and protective HLA associations have do however support an autoimmune eti-
been identified. ology for disease.
Introduction that they result from the complex interaction

of genetic and environmental factors.
Primary biliary cholangitis (PBC), primary Since 2006, genome‐wide association
sclerosing cholangitis (PSC), and autoim- studies (GWAS) and related study designs
mune hepatitis (AIH) are heritable traits. For have revolutionized the study of complex
example, the proportion of PBC patients disease genetics. A GWAS is a case–control
with a family history of the disease is esti- association study in which association is
mated to be 1.33–9.00%, in contrast to an sought between a given phenotype and ge-
estimated population prevalence of 0.002– netic loci across the genome. To achieve this,
0.04% [1]. Having a first‐degree relative with high‐density microarrays are used to geno-
PBC is a strong independent risk factor for type multitudinous single nucleotide poly-
developing the disease, with an odds ratio morphisms (SNPs), carefully selected (mainly
(OR) of 6.8–10.7 [2, 3], and the relative based on linkage disequilibrium [LD] with
sibling risk (λS) in PBC is estimated to be neighboring SNPs) to capture genetic varia-
10.5, comparable to that of other complex tion across the entire genome. This leads to
disorders (Table 3.1) [4]. Similarly, the risk of the problem of multiple testing, i.e. a high
PSC is substantially increased in first‐degree number of false‐positive associations. To
relatives of PSC patients, with hazard ratios overcome this, a stringent significance level
(HRs) estimated as 11.5 for offspring, 11.1 for is adopted, so‐called “genome‐wide level of
siblings, and 2.3 for parents of probands [5]. significance,” which is typically P <5 × 10−8. In
Family studies of AIH have not been reported addition, it is generally recommended that
but there are numerous case reports of AIH genome‐wide significant associations identi-
occurring in two or more relatives [6–8]. fied in a “discovery panel” be confirmed by
These observations suggest that PBC, PSC replicating the association in an independent
and AIH are complex disorders, meaning “validation panel” (albeit at a less stringent
level of significance). A disadvantage of this
approach, however, it that very large sample
Table 3.1 Heritability of selected autoimmune or sizes – thousands of cases and controls – are
immune‐mediated disorders, measured by required to identify variants at genome‐wide
the relative sibling risk.
level of significance. The iChip association
study is a related study design. The iChip is
Relative sibling an Illumina Infinium array containing
Disease risk
196 524 rare to common variants across 186
Primary sclerosing 9–39 known autoimmunity risk loci. Such dense
cholangitis coverage enables fine mapping of risk loci to
Crohn disease 5–35 refine the association signal and pinpoint
candidate genes.
Type 1 diabetes mellitus 15
In the last decade, there have been major
Primary biliary cirrhosis 10
efforts across the globe to establish large,
Psoriasis 4–12 well‐characterized patient cohorts for GWAS
Rheumatoid arthritis 5–10 and related study designs of PBC, PSC and
Ulcerative colitis 6–9 AIH. High‐throughput genetic studies of
Ulcerative colitis in PSC 8 these conditions undertaken to date are
siblings listed in Tables 3.2–3.4 [9–14, 26]. These
studies have undoubtedly been successful,
The relative sibling risk (or λS) refers to the risk of
identifying numerous risk loci for these dis-
disease in siblings of disease probands compared to the
risk of disease in the general population. Note that λS orders that were hitherto unknown,
for PBC and PSC is comparable to that of other providing fresh insight into the genetic basis
autoimmune complex disorders. of autoimmune liver disease. This chapter
Chapter 3 Genetics and Risk of Autoimmune Liver Diseases 49
Table 3.2 Genome‐wide association studies (GWAS) or related study designs of primary biliary
cholangitis (PBC).
Discovery panel Validation panel
Study Design Ancestry Cases Controls Cases Controls
Hirschfield et al. (2009) [9] GWAS European 505 1507 526 1206
a
Liu X et al. (2010) [10] GWMA European 958 2452 NA NA
Mells et al. (2011) [11] GWAS European 1840 5163 620 2514
Nakamura et al. (2012) [12] GWAS Japanese 487 476 787 615
Liu JZ et al. (2012) [13] iChip European 2861 8514 NA NA
Juran et al. (2012) [14] iChip European 2426 5731 NA NA
Cordell et al. (2015)b [15] GWMA European 2764 10 475 3716 4261
Paziewska et al. (2017) [16] GWAS Polish 443 934 NA NA
Kawashima et al. (2017) [17] GWAS Japanese 1381 1505 512 6512
Qiu et al. (2017) [18] GWAS Chinese 1122 4036 907 2127
a
GWMA of discovery panel from the GWAS by Hirschfield et al. [9] and a GWAS dataset consisting of 453 Italian
cases and 945 Italian controls.
b
GWMA of discovery panels from the GWAS by Hirschfield et al. [9], Liu et al. [10] and Mells et al. [11].
GWMA, genome‐wide meta‐analysis; iChip, Illumina Immunoarray association study.
Table 3.3 Genome‐wide association studies (GWAS) or related study designs of primary sclerosing
cholangitis (PSC).
Discovery Validation
Karlsen et al. (2010) [19] GWAS Norwegian 285 298 766 2935
Melum et al. (2011) [20] GWAS European 715 2962 1025 2174
Ellinghaus et al. (2013)a [21] GWAS European 392 2977 1012 11 659
Liu et al. (2013) [22] iChip European 3789 25 079 NA NA
Ji et al. (2016) [23] GWAS European 2871 12 019 1925 7936
Ellinghaus et al. (2016)b [24] iChip European 3408 34 213 NA NA
Paziewska et al. (2017) [16] GWAS Polish 120 934 NA NA
a
In this study, Ellinghaus et al. [50] undertook combined analyses of GWAS for PSC and ulcerative colitis (UC)
comprising 392 PSC cases, 987 UC cases and 2977 controls and followed up top association signals in 1012 PSC
cases, 4444 UC cases and 11 659 controls.
b
In this study, Ellinghaus et al. [42] undertook cross‐phenotype subset‐based meta‐analysis of Illumina
Immunoarray (iChip) genotype data from 52 262 cases with ankylosing spondylitis (8726), Crohn’s disease (19 085),
psoriasis (6530), PSC (3408) or UC (14 513), and 34 213 healthy controls.
iChip, Illumina Immunoarray association study.
Table 3.4 Genome‐wide association studies (GWAS) or related study designs of type 1 autoimmune hepatitis.
Discovery Validation
de Boer et al. GWAS European 649 13 436 451 4103

(2014) [25]
summarizes genetic associations that have from imputation of classical HLA alleles
been identified largely as a result of GWAS from GWAS or iChip genotype data [29].
and related study designs in the last decade. Several HLA associations with PBC, PSC or
AIH have been identified (Table 3.5).
Consistent with other autoimmune condi-
tions, HLA effects in PBC, PSC or AIH are
HLA Associations large, with ORs up to approximately 3.
in Autoimmune Liver In PBC, the most robust associations in
Disease populations of European origin are with the
risk haplotypes DRB1*08:01‐DQA1*04:01‐
The human leukocyte antigen (HLA) com- DQB1*04:02 and DRB1*04:04‐DQB1*03:02,
plex is located on the short arm of and the protective haplotypes DRB1*11:01‐
chromosome 6 and harbors genes encoding DQA1*05:01‐DQB1*03:01 and DRB1*15:01‐
the HLA molecules that are essential for DQA1*01:02‐DQB1*0602. In the study by
antigen presentation. It is characterized by Invernizzi et al. [30], conditional analysis
clusters of highly polymorphic genes that suggested that these haplotype associations
are related by sequence homology and/or were most likely driven by their respective
function: class I genes (HLA‐A, ‐B and ‐C) DRB1 alleles, i.e. the risk alleles DRB1*08 and
and class II genes (HLA‐DP, ‐DQ and ‐D2) DRB1*14 and the protective allele DRB1*11.
encode proteins involved in antigen presen- In PSC, key risk haplotypes include Ancestral
tation, while class III genes encode a large Haplotype 8.1 (AH8.1), which carries the B*08
number of other immune proteins, such as and DRB1*03:01 alleles, and the DRB3*01:01‐
tumor necrosis factor (TNF)‐α [27]. The DRB1*13:01‐DQA1*01:03‐DQB1*06:03 haplo-
HLA complex is characterized by strong type [31–33]. In admixed African‐American
LD extending across large fractions of the populations, LD between PSC associated
region, giving rise to extended HLA haplo- HLA‐B and DRB1 alleles is low and association
types [28]. This means that when HLA is reported for HLA‐B*08 but not DRB1*03,
associations are detected, it may be difficult suggesting that HLA‐B*08 accounts for the
to determine which allele within an AH8.1 association [34]. More generally, it high-
extended haplotype accounts for the lights that class I effects are important in PSC.
association. Patterns of LD across the HLA However, class II effects are also important,
region are different in different ethnicities, shown for example by association with the
so a potential solution to this problem is to protective haplotypes DRB1*04‐DQB1*03:02
study HLA associations across different and DRB1*07:01‐DQB1*03:03 [33, 35]. Strong
ethnic groups. LD across the class II region makes it difficult
Knowledge of HLA associations with auto- to identify which allele drives association with
immune liver disease comes from serotyp- these haplotypes, but circumstantial evidence
ing, phenotyping and sequencing‐based suggests that DRB1 alleles may account for
HLA association studies, and more recently protective effects [36].
Table 3.5 HLA locus associations in primary sclerosing cholangitis (PSC), primary biliary cirrhosis (PBC)
and autoimmune hepatitis (AIH).
Disease Effect HLA allele/haplotype
PSC R B*08
B*07
C*06
A*01‐C*07‐B*08‐DRB3*01:01‐DRB1*03:01‐DQA1*05:01‐DQB1*02:01
DRB1*13:01‐DQA1*01:03‐DQB1*06:03
DRB1*15:01‐DQB1*06:02
rs116212904/A at NOTCH4
P DRB1*04‐DQB1*03:01
DRB1*04‐DQB1*03:02
DRB3*03:01‐DRB1*13:02‐DQB1*06:04
DRB1*07:01‐DQB1*03:03
DRB1*13:XX‐DQB1*06
PBC (European) R B*39:06‐DRB1*08:01‐DQA1*04:01‐DQB1*04:02

DRB1*04:04‐DQA1*03:01‐DQB1*03:02
DRB1*14‐DQB1*05:03 (Italian population only)
DPB1*03:01
C8*04:01
DPB1*10:01
DPB1*17:01
DPA*02:01
P DRB1*11:04‐DRB1*11:01‐DQA1*05:01‐DQB1*03:01
B*07:02‐DRB1*15:01‐DQA1*01:02‐DQB1*06:02
DPB1*04:01
PBC (Japanese) R DRB1*08:03‐DQB1*06:01

DRB1*04:05‐DQB1*04:01
P DRB1*13:02‐DQB1*06:04
DRB1*11:01‐DQB1*03:01
DRB1*15:01
Type 1 AIH R A1‐B8‐DRB3*01:01‐DRB1*03:01‐DQAl*05:01‐DQBl*02:01

DR4 (DRB1*04:01, Europe; DRB1*04:04 & DRB1*04:05, China, Japan, Mexico)
DRB1*13:01‐DQB1*06 (Latin America)
DRB1*14 (India)
P DRB5*01:01‐DRB1*15:01 (UK)
DQB1*04 DQB1*03:01 (Latin America)
DQB1*04
Type 2 AIH R DRB1*03:01‐DQB1*02

DRB1*07‐DQB1*02
DQB1*06
In northern European populations, type 1 The mechanism by which HLA alleles or

AIH is associated with the risk alleles haplotypes confer risk of autoimmune liver
DRB1*03:01 and DRB1*04:01, and with the disease has not been fully elucidated, but it is
protective allele DRB1*15:01 [37]. In Chinese, generally supposed that risk alleles make the
Japanese and Mexican populations, type 1 AIH major histocompatibility complex (MHC)
is associated with the risk alleles DRB1*04:04 molecule more prone to bind specific self‐
and DRB1*04:05 [38]. In Latin American pop- antigens. Consistent with this hypothesis,
ulations, type 1 AIH is associated with the risk Hov et al. [43] have shown that residues 37
allele DRB1*13:01 [39, 40]. In European popu- and 86 in HLA‐DRB1 are independently
lations, type 2 AIH is associated with the risk associated with PSC, and that both influence
alleles DRB1*03, DRB1*07, and DQB1*02:01 the electrostatic properties of pocket P9
[41, 42]. Thus an apparent overlap between of the peptide‐binding groove. Thus,
HLA associations in AIH and PBC is that of DRB1*13:01, which confers risk of PSC, has a
the protective DRB1*15:01 association, which positive P9 pocket, while DRB1*13:02, which
is found in both conditions. Of note, the protects from PSC, has a negative P9 pocket.
DRB1*15:01 allele maps to the AH7.1 haplo- Similarly, Cordell et al. [44] have shown that
type, which has been shown to confer risk of HLA associations with PBC can be explained
PSC in AH8.1 and DRB1*13:01 negative indi- by variation at five separate amino acid posi-
viduals [31]. An apparent overlap between AIH tions, and have demonstrated correlations
and PSC is that of the AH8.1 (represented by between the electrostatic potential of
DRB1*03:01) as well as DRB1*13:01. pocket P6 in HLA‐DP molecules and the
Table 3.6 Risk loci and candidate genes for primary biliary cholangitis.
Locus Study Lead SNP P OR Candidate gene(s)
1p36.32 Hirschfield et al. (2010) [26] rs3748816 3 × 10−8 1.33 MMEL1

−11
1p31.3 Hirschfield et al. (2009) [9] rs3790567 3 × 10 1.51 IL12RB2
1p13.1 Qiu et al. (2017) [18] rs2300747 2 × 10−12 1.29 CD58
−14
1q31.3 Mells et al. (2011) [11] rs12134279 2 × 10 1.34 DENND1B
2q12.1 Cordell et al. (2015) [15] rs12712133 5 × 10−9 1.14 IL1RL2, IL1RL1
−19
2q32.2 Mells et al. (2011) [11] rs10931468 2 × 10 1.50 STAT4
2q33.2 Qiu et al. (2017) [18] rs4675369 1 × 10−13 1.31 CD28, CTLA4
−10
2q36.3 Cordell et al. (2015) [15] rs4973341 2 × 10 1.22 CCL20
3p24.3 Mells et al. (2011) [11] rs1372072 2 × 10−8 1.20 PLCL2
−11
3q13.33 Mells et al. (2011) [11] rs2293370 3 × 10 1.35 CD80
3q25.33 Hirschfield et al. (2009) [9] rs6441286 2 × 10−14 1.54 IL12A
−12
4p16.3 Cordell et al. (2015) [15] rs11724804 9 × 10 1.22 DGKQ
4q24 Mells et al. (2011) [11] rs7665090 4 × 10−12 1.26 NFKB1
−11
4q27 Qiu et al. (2017) [18] rs17005934 1 × 10 1.29 IL21
5p13.2 Mells et al. (2011) [11] rs860413 1 × 10−11 1.30 IL7R
−8
5q21.1 Cordell et al. (2015) [15] rs526231 1 × 10 1.15 PAM
5q33.3 Cordell et al. (2015) [15] rs2546890 1 × 10−10 1.15 IL12B
−10
6q23.3 Cordell et al. (2015) [15] rs6933404 1 × 10 1.18 TNFAIP3
7p14.1 Mells et al. (2011) [11] rs6974491 4 × 10−8 1.25 ELMO1
7q32.1 Hirschfield et al. (2010) [26] rs10488631 9 × 10−13 1.57 IRF5

9q32 Nakamura et al. (2012) [12] rs4979462 3 × 10−14 1.56 TNFSF15
−10
11q13.1 Mells et al. (2011) [11] rs538147 2 × 10 1.23 CCDC88B
11q23.1 Nakamura et al. (2012) [12] rs4938534 2 × 10−8 1.39 POU2AF1
−12
11q23.3 Mells et al. (2011) [11] rs6421571 3 × 10 1.37 CXCR5
12p13.31 (1) Mells et al. (2011) [11] rs1800693 2 × 10−9 1.22 TNFRSF1A
−9
12p13.31 (2) Liu et al. (2012) [13] rs11064157 2 × 10 1.23 LTBR
12q24.12 Liu et al. (2012) [13] rs11065979 3 × 10−9 1.20 ATXN2, SH2B3
−8
13q14 Juran et al. (2012) [14] rs3862738 2 × 10 1.33 TNSF11
13q14.2 Cordell et al. (2015) [15] rs9591325 1 × 10−10 1.63 DLEU1
−11
14q24.1 Mells et al. (2011) [11] rs911263 2 × 10 1.29 RAD51B
14q32.32 Mells et al. (2011) [11] rs8017161 3 × 10−13 1.22 TNFAIP2
−9
15q25.1 Qiu et al. (2017) [18] rs11556218 9 × 10 1.29 IL16
16p13.13 Mells et al. (2011) [11] rs12924729 3 × 10−12 1.29 SOCS1, CLEC16A
−9
16p12.2 Kawashima et al. (2017) [17] rs7404928 4 × 10 1.26 PRKCB
16p12.1 Qiu et al. (2017) [18] rs2189521 4 × 10−16 1.41 IL4R, IL21R
−8
16q21 Qiu et al. (2017) [18] rs2550374 2 × 10 1.23 CSNK2A2, CCDC113
16q24.1 Mells et al. (2011) [11] rs11117432 5 × 10−11 1.31 IRF8
−9
17q21.1 Liu et al. (2010) [10] rs9303277 2 × 10 1.38 IKZF3
17q21.31 Liu et al. (2012) [13] rs17564829 2 × 10−9 1.25 CRHR1, MAPT
−12
19p13.3 Qiu et al. (2017) [18] rs10415976 4 × 10 1.30 ARID3A
19p13.2 Liu et al. (2012) [13] rs34536443 1 × 10−12 1.91 TYK2
−11
19q13.33 Liu et al. (2010) [13] rs3745516 8 × 10 1.46 SPIB
22q13.1 Mells et al. (2011) [11] rs968451 1 × 10−9 1.27 SYNGR1
HLA‐DPB1 amino acid substitutions that small (OR ~1.5). Non‐HLA risk loci for AIH
confer PBC susceptibility or protection. have not been identified at genome‐wide
level of significance, but GWAS of AIH have
so far been comparatively underpowered.
Non‐HLA Associations Although genetic studies of PBC or PSC
in Autoimmune Liver Disease have successfully identified numerous risk
loci, it is estimated that less than 20% of
GWAS and related study designs have sub- PBC heritability and less than 10% of vari-
stantially improved knowledge of the ge- ance of PSC liability has been explained, the
netic architecture of PBC and PSC, so‐called “missing heritability” [13, 45]. It is
especially outside the HLA region. plausible that better‐powered studies of
Numerous genome‐wide significant non‐ PBC or PSC would identify additional risk
HLA risk loci for PBC or PSC have been loci for these conditions, but it may be
identified (Tables 3.6 and 3.7). Consistent impossible to achieve the sample sizes
with other complex disorders, the effects of required to identify additional risk loci with
non‐HLA risk loci in these conditions are ever‐smaller effects [46–49].
Table 3.7 Risk loci and candidate genes for primary sclerosing cholangitis.
1p36.32 Liu et al. (2013) [22] rs3748816 7 × 10−12 1.21 MMEL1, TNFRSF14
2q13 Melum et al. (2011) [20] rs6720394 4 × 10−8 1.29 BCL2L11
−20
2q33.2 Liu et al. (2013) [22] rs7426056 2 × 10 1.30 CD28, CTLA4
2q36.3 Ellinghaus et al. (2016) [24] rs7556897 5 × 10−9 1.18 CCL20
−9
2q37.3 Ellinghaus et al. (2013) [21] rs4676410 2 × 10 1.38 GPR35
3p21.31 Melum et al. (2011) [20] rs3197999 1 × 10−16 1.39 MST1
−15
3p13 Ji et al. (2017) [23] rs80060485 3 × 10 1.44 FOXP1
4q24 Ellinghaus et al. (2016) [24] rs17032705 4 × 10−10 1.18 NFKB1
−13
4q27 Liu et al. (2013) [22] rs13140464 9 × 10 1.30 IL2, IL21
6q15 Liu et al. (2013) [22] rs56258221 8 × 10−12 1.23 BACH2
−17
10p15.1 Liu et al. (2013) [22] rs4147359 8 × 10 1.24 IL2RA
11q13.1 Ji et al. (2017) [23] rs663743 2 × 10−13 1.20 CCDC88B
−9
11q23.1 Liu et al. (2013) [22] rs7937682 3 × 10 1.17 SIK2
12q13 Liu et al. (2013) [22] rs11168249 5 × 10−9 1.15 HDAC7
−9
12q23.3 Ellinghaus et al. (2016) [24] rs12369214 1 × 10 1.18 RFX4, RIC8B
12q24.12 Liu et al. (2013) [22] rs3184504 6 × 10−11 1.18 SH2B3, ATXN2
−11
16p13.13 Ellinghaus et al. (2016) [24] rs11649613 1 × 10 1.19 CLEC16A, SOCS1
18q21.2 Ellinghaus et al. (2013) [21] rs1452787 3 × 10−8 1.33 TCF4
−8
18q22.2 Liu et al. (2013) [22] rs1788097 3 × 10 1.15 CD226
19q13.32 Liu et al. (2013) [22] rs60652743 6 × 10−10 1.25 PRKD2, STRN4
−17
21q22.2 Liu et al. (2013) [22] rs2836883 3 × 10 1.28 PSMG1
21q22.3 Ji et al. (2017) [23] rs1893592 2 × 10−12 1.22 UBASH3A
−8
22q11.23 Ji et al. (2017) [23] rs145832854 3 × 10 1.87 SGSM1
Candidate genes for PBC and PSC are listed AIH are briefly described in Boxes 3.1–3.3,
in Tables 3.6 and 3.7; however, it should be respectively. The reader is reminded that the
noted that GWAS and related studies identify potential role of these genes in the pathogen-
disease‐associated SNPs, but do not identify esis of autoimmune liver disease is obvious
causal variants or causal genes (Figure 3.1). but also speculative.
Causal variants and genes may be prioritized The mechanism by which causal variants
through functional annotation of risk loci but lead to disease is unknown and may be differ-
laboratory experiments linking genotype and ent for each risk locus. It has been observed,
molecular phenotype are required to confirm however, that many candidate causal variants
(and explain) the contribution of the variant for complex disorders are regulatory vari-
and gene to disease pathogenesis [50]. ants, meaning that they overlap regulatory
Nevertheless, based on functional annotation, elements (e.g. promoter regions) active in
risk loci for PBC, PSC and AIH appear to be relevant tissues. In the genome‐wide meta‐
enriched for gene products involved in innate analysis of PBC by Cordell et al. [15], for
or acquired immune responses, consistent example, up to one‐third of candidate causal
with an autoimmune component to patho- variants were regulatory, overlapping with
genesis. Candidate genes for PBC, PSC and regulatory elements active in cells of the
Plotted SNPs
25 r2 100
chr16:11188930
0.8
Recombination rate (cM/Mb)

20 0.6 80
0.4
0.2
–log10(p-value)
15 60
10 40
5 20
0 0
NUBP1 CIITA CLEC16A SOCS1 RMI2

TVP23A DEXI TNP2 LOC101927131
PRM3
PRM2
PRM1
10.8 11 11.2 11.4

Position on chr16 (Mb)
Figure 3.1 LocusZoom plot of the PBC risk locus 16p13.13, illustrating that, while the index variant may
strongly suggest a causal gene (in this case, CLEC16A), this cannot be assumed: at any risk locus, the index
variant is the variant most strongly associated with disease, while the causal variant is an unknown variant
with functional effects that influence disease expression. Association is detected with the index variant owing
to correlation with the causal variant. The index variant is correlated to numerous other variants, however,
making it difficult to know which one is causal. Risk loci encompass the index variant and neighboring variants
correlated to it. They may harbor several genes (in this case, DEXI, CLEC16A, SOCS1, TNP2, PRM3, and so forth),
any of which might be the gene involved in causing disease. Bioinformatics approaches are helpful for
prioritizing candidate genes but functional studies are ultimately required to confirm causality and elucidate
mechanism.
immune system. This suggests that many clinically associated autoimmune disorders –
variants contribute to disease by influencing ulcerative colitis (UC), Crohn disease (CD),
the regulation of gene expression. Consistent type 1 diabetes mellitus, celiac disease, psori-
with this, many candidate causal variants are asis (PS), rheumatoid arthritis and sarcoid-
correlated with gene expression (expression osis – suggesting close similarity in the
quantitative trait loci or eQTLs) or methyla- genetic architecture of PSC and each of these
tion (methylation quantitative trait loci or conditions [45]. This was further explored in
mQTLs). the study by Ellinghaus et al. [24], who under-
A striking observation from GWAS is the took a cross‐phenotype, subset‐based meta‐
extent to which risk loci are shared by mul- analysis of iChip genotype data from 52 262
tiple autoimmune conditions. Pervasive cases with ankylosing spondylitis (8726), CD
sharing of genetic effects has been formally (19 085), PS (6530), PSC (3408) or UC
demonstrated in PSC. In the iChip study by (14 413), and 34 213 healthy controls. They
Liu et al. [22], pleiotropy analysis revealed identified 244 independent multidisease sig-
strong positive correlation between associa- nals and showed that comorbidities among
tions with PSC and associations with seven the five disorders were best explained by
Box 3.1 Candidate Genes for Primary Biliary Cholangitis

Membrane metalloendopeptidase like 1 teraction of CD28 on T cells with CD80 or CD86
(MMEL1, 1p36.32) is a member of the mem- on APCs is required for T‐cell activation, prolif-
brane metalloendopeptidase family. It eration, cytokine production, and survival.
degrades a broad variety of small peptides, Without this CD28–B7 interaction, the T cell
especially those containing neutral bulky ali- becomes anergic.
phatic or aromatic amino acid residues.
Cytotoxic T‐lymphocyte associated protein
Interleukin‐12 receptor subunit beta 2 4 (CTLA4, 2q33.2) is a member of the immuno-
(IL12RB2, 1p31.3) combines with IL12RB1 to globulin superfamily. It is expressed by
form the receptor for IL‐12. IL12RB2 is the sig- activated T cells and functions as an immune
naling component of the receptor, coupling to checkpoint, whereby binding of CTLA4 by its
the JAK2/STAT4 pathway. IL12RB2 is upregu- ligands, CD80 and CD86, expressed by APCs,
lated by interferon (IFN)‐γ. Signaling via generates an inhibitory signal that downregu-
IL12RB2 promotes the proliferation of T cells lates T‐cell activation and diminishes the
and natural killer (NK) cells, and promotes immune response.
differentiation of T cells to Th1 cells by strongly
enhancing IFN‐γ production. C–C motif chemokine 20 (CCL20, 2q36.3) is
the ligand for C–C chemokine receptor 6
Lymphocyte function‐associated antigen 3 (CCR6). CCL20–CCR6 is involved in the chemo-
(LFA3, CD58, 1p13.1) is a ligand of the CD2 gly- taxis of dendritic cells (DCs) and lymphocytes.
coprotein on T lymphocytes. CD58–CD2 It plays an important role in skin and mucosal
functions in adhesion and activation of T lym- immunity, contributing to the formation and
phocytes, including interaction of thymocytes function of mucosal lymphoid tissues by
with thymic epithelial cells, and antigen‐ attracting lymphocytes and DCs toward epi-
independent and ‐dependent interactions of T thelial cells. CCL20–CCR6 is required for the
cells with target cells and antigen‐presenting normal migration of Th17 cells in Peyer’s
cells (APCs). patches; regulates effector T‐cell balance and
DENN domain‐containing protein 1B distribution in inflamed intestine; is essential
(DENND1B, 1q31.3) is a guanine nucleotide for B‐cell localization in the subepithelial dome
exchange factor (GEF) for RAB35. It is thus of Peyer’s patches, where it is also required for
involved in clathrin‐mediated endocytosis. efficient B‐cell isotype switching to IgA; and is
DENND1B controls the rate of T‐cell receptor essential for appropriate anatomic distribution
(TCR) internalization (via endocytosis) in Th2 of memory B cells in the spleen and for the
cells, thereby influencing Th2 cell cytokine secondary recall response of memory B cells.
production.
Phospholipase C‐like protein 2 (PLCL2,
Signal transducer and activator of tran- 3p24.3) may play a role in the regulation
scription 4 (STAT4, 2q32.2) is a member of the of
inositol 1,4,5‐trisphosphate around the
STAT family of transcription factors, which have endoplasmic reticulum.
the dual function of signal transduction and
CD80 (CD80, 3q13.33) is a membrane receptor
activation of transcription. STAT4 mediates sig-
expressed by APCs that binds CD28 or CTLA4,
naling via IL12RB2 (see above).
expressed by activated T cells. CD80 is involved
Cluster of differentiation 28 (CD28, 2q33.2) in the costimulatory signal essential for T‐cell
is expressed by T cells and provides the activation. Thus, binding of CD80 with CD28
costimulatory signal necessary for T‐cell induces T‐cell proliferation and cytokine pro-
activation. Thus, when the TCR interacts with duction, while binding of CD80 with CTLA4
the MHC–antigen complex on APCs, the coin- inhibits T‐cell activation.
Interleukin‐12 subunit alpha (IL12A, 3q25.33) ment and cell motility protein family. It inter-
is the 35‐kDa subunit of IL‐12, a cytokine that acts with dedicator of cytokinesis 1 (DOCK1)
functions as a growth factor for activated T and protein to promote phagocytosis of apoptotic
NK cells, enhances the lytic activity of NK cells cells and cell migration.
and lymphokine‐activated killer (LAK) cells,
Interferon regulatory factor 5 (IRF5, 7q32.1)
and stimulates production of IFN‐γ. IL‐12 sig-
is a transcription factor involved in the
naling is mediated by STAT4.
induction of interferons IFNA and INFB and
Nuclear factor kappa B p105 subunit inflammatory cytokines following viral infec-
(NFKB1, 4q24) is the 105‐kDa subunit of NFκB, tion. It is activated by Toll‐like receptor (TLR)7
a pleiotropic transcription factor expressed in or TLR8 signaling.
most cell types. NFκB functions in the tran-
Tumor necrosis factor ligand superfamily
scriptional regulation of diverse biological
member 15 (TNFSF15, 9q32) is the ligand for
processes, including inflammation, immunity,
TNFRSF25 and TNFRSF6B. It mediates
differentiation, cell growth, tumorigenesis,
activation of NFκB and mitogen‐activated pro-
and apoptosis.
tein kinases (MAPK), promotes activation of
Interleukin‐21 (IL21, 4q27) is a cytokine that caspases and apoptosis, and may inhibit endo-
functions in both innate and adaptive immune thelial cell proliferation.
responses by inducing the differentiation, pro-
Coiled‐coil domain containing 88B
liferation and activity of multiple target cells
(CCDC88B, 11q13.1) is a member of the hook‐
including macrophages, NK cells, B cells, and
related protein family. It is required for several
cytotoxic T cells.
CD4+ and CD8+ T‐cell functions, including
Interleukin‐7 receptor subunit alpha (IL7R, expression of cell surface markers of activation,
5p13.2) is a subunit of the receptor for IL‐7, a proliferation and cytokine production in
hematopoietic growth factor that is important response to specific or non‐specific stimula-
for B‐ and T‐cell development. IL7–IL7R is tion. It enhances NK cell cytotoxicity and sup-
necessary for V(D)J rearrangement of the TCRβ presses endoplasmic reticulum stress‐induced
(TCRB) during early T‐cell development. It may apoptosis.
also be involved in regulation of intestinal
POU domain class 2‐associating factor 1
mucosal lymphocytes.
(POU2AF1, 11q23.1) is a transcriptional coacti-
Interleukin‐12 subunit beta (IL12B) is the 40‐ vator that specifically associates with either
kDa subunit of the cytokines IL‐12 (see above) OCT1 or OCT2, which is essential for the
and IL‐23, a heterodimeric cytokine which response of B cells to antigens and required for
binds to a receptor complex composed of the formation of germinal centers.
IL12RB1 and IL23R, activating a Jak–Stat sig-
C‐X‐C chemokine receptor type 5 (CXCR5,
naling cascade. IL‐23 signaling stimulates
11q23.3) is a cytokine receptor that binds to
memory T cells and promotes production of
B‐lymphocyte chemoattractant (BLC). It is
proinflammatory cytokines.
involved in B‐cell migration into B‐cell follicles
Tumor necrosis factor alpha‐induced proof spleen and Peyer’s patches.
tein 3 (TNFAIP3, 6q23.3) is a ubiquitin‐editing
Tumor necrosis factor receptor superfamily
enzyme involved in cytokine‐mediated
member 1A (TNFRSF1A, 12p13.31) is the
immune and inflammatory responses, which
receptor for TNF‐α and lymphotoxin‐α. Binding
inhibits NFκB activation and TNF‐mediated
of membrane‐bound TNF‐α to membrane‐
apoptosis.
bound TNFRSF1A induces receptor trimeriza-
Engulfment and cell motility protein 1 tion and activation, which plays a role in cell
(ELMO1, 7p14.1) is a member of the engulf- survival, apoptosis, and inflammation.
Tumor necrosis factor receptor superfamily protein kinase that is involved in various cel-
member 3 (LTBR, 12p13.31) is the receptor for lular processes, including regulation of the B‐
the heterotrimeric lymphotoxin containing cell receptor (BCR) signalosome and oxidative
LTA and LTB, and for TNFS14/LIGHT. It pro- stress‐induced apoptosis.
motes apoptosis and may play a role in the
Interleukin‐21 receptor (IL21R, 16p12.1) is
development of lymphoid organs.
the receptor for IL‐21. Ligand binding of IL21R
SH2B adapter protein 3 (SH2B3, 12q24.12) is leads to activation of multiple downstream
a member of the SH2B adaptor family of pro- signaling molecules, including JAK1, JAK3,
teins, which are involved in a range of signaling STAT1, and STAT3. Signaling via IL21R is impor-
activities by growth factor and cytokine recep- tant for the proliferation and differentiation of
tors. SH2B3 links TCR activation signal to phos- T cells, B cells, and NK cells.
pholipase Cγ1, GRB2 and phosphatidylinositol
Coiled‐coil domain‐containing protein 113
3‐kinase. It is a key negative regulator of cyto-
(CCDC113, 16q21) is a component of centrio-
kine signaling and plays a critical role in
lar satellites contributing to primary cilium
hematopoiesis.
formation.
Leukemia‐associated protein 1 (DLEU1,
Interferon regulatory factor 8 (IRF8, 16q24.1)
13q14.2) is an RNA gene that may act as a
is a transcription factor of the interferon
tumor suppressor.
regulatory factor (IRF) family that binds to the
DNA repair protein RAD51 homolog 2 IFN‐stimulated response element (ISRE) and
(RAD51B, 14q24.1) is involved in the homolo- regulates expression of genes stimulated by
gous recombination repair pathway of double‐ type I IFNs (i.e. IFN‐α and IFN‐β). IRF8 has a
stranded DNA breaks arising during DNA negative regulatory role in cells of the immune
replication or induced by DNA‐damaging system, and is involved in CD8+ DC
agents. differentiation.
Tumor necrosis factor alpha‐induced pro- Zinc finger protein Aiolos (IKZF3, 17q21.1) is
tein 2 (TNFAIP2, 14q32.32) can be induced by a transcription factor involved in lymphocyte
TNF‐α and may be a mediator of inflammation differentiation. It plays an essential role in reg-
and angiogenesis. ulation of B‐cell differentiation, proliferation
and maturation to an effector state. It is
Interleukin‐16 (IL16, 15q25.1) is a proinflam-
involved in regulating BCL2 expression and
matory cytokine that signals via CD4. It is che-
controlling apoptosis in T cells in an IL2‐
motactic for CD4+ T lymphocytes, monocytes,
dependent manner.
and eosinophils; primes CD4+ T cells for IL‐2
and IL‐15 responsiveness; upregulates IL‐2 Microtubule‐associated protein tau (MAPT,
receptor and HLA‐DR4 expression; and inhibits 17q21.31) promotes microtubule assembly
TCR/CD3‐dependent activation. and stability, and might be involved in the
establishment and maintenance of neuronal
C‐type lectin domain family 16 member A
polarity.
(CLEC16A, 16p13.13) is involved in mitophagy
through the upstream regulation of the AT‐rich interactive domain‐containing pro-
RNF41/NRDP1‐PRKN pathway, mitophagy tein 3A (ARID3A, 19p13.3) is a transcription factor
being a selective form of autophagy necessary which may be involved in B‐cell differentiation.
for mitochondrial quality control.
Non‐receptor tyrosine‐protein kinase (TYK2,
Protein kinase C beta type (PRKCB, 16p12.2) 19p13.2) is a member of the Janus kinase (JAK)
is a calcium‐activated, phospholipid‐ and diac- family of protein. TYK2 associates with the
ylglycerol (DAG)‐dependent serine/threonine cytoplasmic domain of type I and type II cytokine
receptors and propagates cytokine signals by duce large amounts of interferon and block
phosphorylating receptor subunits. It is a com- viral replication. SPIB may be required for BCR
ponent of both the type I and type III inter- signaling, which is necessary for normal B‐cell
feron signaling pathways. development and antigenic stimulation.
Transcription factor Spi‐B (SPIB, 19q13.33) is Synaptogyrin‐1 (SYNGR1, 22q13.1) is an inte-
a transcriptional activator which binds to the gral membrane protein associated with pre-
PU‐box, a purine‐rich DNA sequence (5′‐ synaptic vesicles in neuronal cells, which may
GAGGAA‐3′) that acts as a lymphoid‐specific play a role in synaptic‐like microvesicle
enhancer. SPIB promotes development of plas- formation and/or maturation, and may play a
macytoid DCs, which have the capacity to pro- role in regulated exocytosis.
Box 3.2 Candidate Genes for Primary Sclerosing Cholangitis

Membrane metalloendopeptidase like 1 inhibition of caspase‐dependent apoptosis;
(MMEL1, 1p36.32). See Box 3.1. and is involved in regulation of monocyte
differentiation and macrophage functions.
BCL2 like 11 (BCL2L11, 2q13) is a member of
the BCL‐2 protein family. It induces apoptosis Nuclear factor kappa B subunit 1 (NFKB1,
and anoikis, and may function as an essential 4q24). See Box 3.1.
initiator of apoptosis in thymocyte‐negative
Interleukin‐2 (IL2, 4q27) is a cytokine secreted
selection.
by T cells in response to antigenic or mitogenic
Cytotoxic T‐lymphocyte associated protein stimulation, and is required for T‐cell prolifera-
4 (CTLA4, 2q33.2). See Box 3.1. tion. IL‐2 promotes differentiation of immature
T cells into Tregs, which is critical for tolerance.
C–C motif chemokine 20 (CCL20, 2q36.3). See
However, it also promotes the differentiation
Box 3.1.
of activated T cells into effector T cells and
G protein‐coupled receptor 35 (GPR35, memory T cells, which is critical for immunity.
2q37.3) is a receptor for kynurenic acid, an IL‐2 also stimulates B cells, monocytes, LAK
intermediate in the tryptophan metabolic cells, and NK cells.
pathway.
BTB domain and CNC homolog 2 (BACH2,
Serine/threonine kinase 4 (STK4, MST1, 6q15) is a transcription factor that binds to
3p21.31) is a stress‐activated, pro‐apoptotic MAF recognition elements (MARE). One of its
kinase which, following caspase cleavage, functions is to induce apoptosis in response to
enters the nucleus and induces chromatin oxidative stress through repression of the anti‐
condensation followed by DNA fragmentation. apoptotic factor HMOX1.
Notably, STK3/MST2 and STK4/MST1 are
Interleukin‐2 receptor subunit alpha (IL2RA,
required to repress proliferation of mature
CD25, 10p15.1), together with IL2RB and IL2RG,
hepatocytes and prevent activation of faculta-
forms the receptor for IL‐2, which is critical for
tive adult liver stem cells (oval cells).
tolerance as well as immunity (see above).
Forkhead box P1 (FOXP1, 3p13) belongs to
Coiled‐coil domain containing 88B (CCDC88B,
subfamily P of the forkhead box (FOX) tran-
11q13.1). See Box 3.1.
scription factor family. Among other functions,
FOXP1 is an essential transcriptional regulator Salt inducible kinase 2 (SIK2, 11q23.1) phosphor-
of B‐cell development; negatively regulates ylates Ser‐794 of IRS1 in insulin‐stimulated adipo-
the differentiation of Tfh cells; cooperates with cytes. It inhibits cAMP response element binding
NFκB signaling to promote B‐cell expansion by protein (CREB) activity by phosphorylating and
repressing transducers of regulated CREB activity Protein kinase D2 (PRKD2, 19q13.32) is a

(TORCs), the CREB‐specific coactivators. serine/threonine protein kinase that converts
transient DAG signals into prolonged
Histone deacetylase 7 (HDAC7, 12q13) is
physiological effects downstream of protein
responsible for the deacetylation of lysine resi-
kinase C (PKC). It is involved in oxidative stress‐
dues on the N‐terminal part of the core his-
induced NFκB activation; production of IL‐2 and
tones (H2A, H2B, H3 and H4). Histone
IFN‐γ following TCR engagement in T cells; reg-
deacetylation gives a tag for epigenetic repres-
ulation of cell proliferation via MAPK1/3
sion and plays an important role in transcrip-
(ERK1/2) signaling; inhibition of HDAC7 tran-
tional regulation, cell cycle progression, and
scriptional repression; Golgi membrane traffick-
developmental events.
ing; angiogenesis; secretory granule release;
RIC8 guanine nucleotide exchange factor B and cell adhesion.
(RIC8B, 12q23.3) can activate some, but not all, Gα
Proteasome assembly chaperone 1 (PSMG1,
proteins by exchanging bound GDP for free GTP.
21q22.2) is a chaperone protein which promotes
SH2B adaptor protein 3 (SH2B3, 12q24.12). assembly of the 20S proteasome.
See Box 3.1.
Ubiquitin associated and SH3 domain con-
C‐type lectin domain containing 16A taining A (UBASH3A, 21q22.2) is a member of
(CLEC16A, 16p13.13). See Box 3.1. the T‐cell ubiquitin ligand (TULA) family. It
functions to negatively regulate T‐cell signaling
CD226 antigen (CD226, 18q22.2) is a glyco- and facilitate growth factor withdrawal‐induced
protein expressed at the surface of NK cells, apoptosis in T cells, which may occur via its inter-
platelets, monocytes, and T cells, and is action with AIF, an apoptosis‐inducing factor.
involved in cellular adhesion, lymphocyte sig-
naling, cytotoxicity, and lymphokine secretion. Small G protein signaling modulator 1 (SGSM1,
In T cells, ligand binding stimulates prolifera- 22q11.23) interacts with numerous Rab family of
tion and cytokine production, including IL‐2, small GTPases, functioning as a Rab effector for
IL‐5, IL‐10, IL‐13, and IFN‐γ. some, and as GTPase activator for others.
Box 3.3 Candidate Genes for Autoimmune leiotropy analysis has not been undertaken
p
Hepatitis in PBC or AIH, but it is anticipated that
findings would be similar. Pervasive sharing
SH2B adaptor protein 3 (SH2B3, 12q24.12). of risk variants of genes across a variety of
See Box 3.1. autoimmune diseases is relevant because
cross‐phenotype analyses might therefore
Caspase recruitment domain family
be helpful in identifying molecular targets
member 10 (CARD10, 22q13.1) belongs to
for therapies with widespread applications
the membrane‐associated guanylate kinase
(although this cannot be assumed).
(MAGUK) family and activates NFκB via
Conversely, pervasive sharing of genetic
BCL10 and IKK.
effects across multiple autoimmune condi-
tions emphasizes that GWAS of PBC, PSC
biological pleiotropy rather than heteroge- and AIH have not identified candidates that
neity, suggesting, for example, that the would explain the tissue specificity of these
strong comorbidity between PSC and diseases. Paradoxically, genetic studies of
inflammatory bowel disease (IBD) may rep- autoimmune liver disease highlight the likely
resent a unique disease, genetically distinct importance of environmental factors in
from classical IBD phenotypes. Formal shaping disease.
Consistent with this, some of the strongest inflammatory risk loci). The pleiotropy of
associations in GWAS of immune‐mediated the GWAS outcomes is poorly understood
conditions identify risk loci directly involved and varies between different geographic and
in gene–environment interactions (e.g. HLA‐ ethnic groups. Importantly, only a fraction
DQB1 in celiac disease and NOD2 in Crohn of the heritability of liability is explained by
disease) [51, 52]. It is conceivable that common loci determined to date, partly because
causal variants for complex disorders have study panels have been relatively modest
escaped neutral processes (e.g. genetic drift) given the low prevalence of autoimmune
because of selective pressure exerted by envi- liver diseases. However, the latter may also
ronmental influences. GWAS findings may suggest that interacting genetic or non‐
therefore represent “hotspots” where the genetic factors (environmental factors in
gene–environmental homeostasis of an particular) are involved in causing each
individual is disturbed. Interpretation of ge- specific clinical condition on the background
netic findings may therefore require better partly determined by GWAS. Following
knowledge of the environmental factors GWAS, there is therefore a strong need to
involved in disease, justifying ongoing efforts search for these factors, along with further
to identify these factors. genetic characterization of risk loci
determined (e.g. by means of sequencing to
determine rare pathogenic variants).
Conclusion Ultimately, the accumulating knowledge
from these studies needs to be explored in
The genetic architecture of PBC, PSC and the context of existing and novel in vitro and
AIH is similar to most autoimmune diseases in vivo model systems with the aim of gener-
(i.e. a strong HLA association accompa- ating a basis for novel therapeutics for
nied by associated groups of general autoimmune liver diseases.
References
1 Boonstra, K., Beuers, U., and Ponsioen, C.Y. 10 Liu, X., Invernizzi, P., Lu, Y. et al. (2010).
(2012). Epidemiology of primary sclerosing Genome‐wide meta‐analyses identify three
cholangitis and primary biliary cirrhosis: a loci associated with primary biliary
systematic review. J. Hepatol. 56: cirrhosis. Nat. Genet. 42: 658–660.
1181–1188. 11 Mells, G.F., Floyd, J.A., Morley, K.I. et al.
2 Gershwin, M.E., Selmi, C., Worman, H.J. (2011). Genome‐wide association study
et al. (2005). Risk factors and comorbidities identifies 12 new susceptibility loci for
in primary biliary cirrhosis: a controlled primary biliary cirrhosis. Nat. Genet. 43:
interview‐based study of 1032 patients. 329–332.
Hepatology 42: 1194–1202. 12 Nakamura, M., Nishida, N., Kawashima, M.
3 Corpechot, C., Chretien, Y., Chazouilleres, et al. (2012). Genome‐wide association
O., and Poupon, R. (2010). Demographic, study identifies TNFSF15 and POU2AF1 as
lifestyle, medical and familial factors susceptibility loci for primary biliary
associated with primary biliary cirrhosis. cirrhosis in the Japanese population. Am. J.
J. Hepatol. 53: 162–169. Hum. Genet. 91: 721–728.
4 Jones, D.E., Watt, F.E., Metcalf, J.V. et al. 13 Liu, J.Z., Almarri, M.A., Gaffney, D.J. et al.
(1999). Familial primary biliary cirrhosis (2012). Dense fine‐mapping study identifies
reassessed: a geographically‐based new susceptibility loci for primary biliary
population study. J. Hepatol. 30: 402–407. cirrhosis. Nat. Genet. 44: 1137–1141.
5 Bergquist, A., Lindberg, G., Saarinen, S., and 14 Juran, B.D., Hirschfield, G.M., Invernizzi, P.
Broome, U. (2005). Increased prevalence of et al. (2012). Immunochip analyses identify
primary sclerosing cholangitis among first‐ a novel risk locus for primary biliary
degree relatives. J. Hepatol. 42: 252–256. cirrhosis at 13q14, multiple independent
6 Yoshida, O., Abe, M., Furukawa, S. et al. associations at four established risk loci
(2009). A familial case of autoimmune and epistasis between 1p31 and 7q32 risk
hepatitis. Intern. Med. 48: 315–319. variants. Hum. Mol. Genet. 21: 5209–5221.
7 Nozic, D., Dimitrijevic, J., and Knezevic‐ 15 Cordell, H.J., Han, Y., Mells, G.F. et al.
Usaj, S. (2005). Occurrence of autoimmune (2015). International genome‐wide meta‐
chronic hepatitis in siblings. Vojnosanitetski analysis identifies new primary biliary
pregled Military‐Medical and cirrhosis risk loci and targetable pathogenic
Pharmaceutical Review 62: 591–594. pathways. Nat. Commun 6: 8019. https://
8 Ohira, H., Shinzawa, J., Suzuki, T. et al. doi.org/10.1038/ncomms9019.
(1998). Two sister cases of autoimmune 16 Paziewska, A., Habior, A., Rogowska, A.
hepatitis. Fukushima J. Med. Sci. 44: 113–120. et al. (2017). A novel approach to genome‐
9 Hirschfield, G.M., Liu, X., Xu, C. et al. wide association analysis identifies genetic
(2009). Primary biliary cirrhosis associated associations with primary biliary
with HLA, IL12A, and IL12RB2 variants. cholangitis and primary sclerosing
N. Engl. J. Med. 360: 2544–2555. cholangitis in Polish patients. BMC Med.
Genomics 10 (1): 2. https://doi.org/10.1186/ 26 Hirschfield, G.M., Liu, X., Han, Y. et al.

s12920‐016‐0239‐9. (2010). Variants at IRF5‐TNPO3, 17q12‐21
17 Kawashima, M., Hitomi, Y., Aiba, Y. et al. and MMEL1 are associated with primary
(2017). Genome‐wide association studies biliary cirrhosis. Nat. Genet. 42: 655–657.
identify PRKCB as a novel genetic 27 Kelley, J. and Trowsdale, J. (2005). Features
susceptibility locus for primary biliary of MHC and NK gene clusters. Transplant
cholangitis in the Japanese population. Immunol. 14: 129–134.
Hum. Mol. Genet. 26 (3): 650–659. 28 Traherne, J.A. (2008). Human MHC
18 Qiu, F., Tang, R., Zuo, X. et al. (2017). A architecture and evolution: implications for
genome‐wide association study identifies disease association studies. Int. J.
six novel risk loci for primary biliary Immunogenet. 35: 179–192.
cholangitis. Nat. Commun. 8: 14828. 29 Dilthey, A.T., Moutsianas, L., Leslie, S., and
19 Karlsen, T.H., Franke, A., Melum, E. et al. McVean, G. (2011). HLA*IMP: an
(2010). Genome‐wide association analysis integrated framework for imputing
in primary sclerosing cholangitis. classical HLA alleles from SNP genotypes.
Gastroenterology 138: 1102–1111. Bioinformatics 27: 968–972.
20 Melum, E., Franke, A., Schramm, C. et al. 30 Invernizzi, P., Ransom, M., Raychaudhuri,
(2011). Genome‐wide association analysis S. et al. (2012). Classical HLA‐DRB1 and
in primary sclerosing cholangitis identifies DPB1 alleles account for HLA associations
two non‐HLA susceptibility loci. Nat. with primary biliary cirrhosis. Genes
Genet. 43 (1): 17–19. Immunity 13: 461–468.
21 Ellinghaus, D., Folseraas, T., Holm, K. et al. 31 Donaldson, P.T., Farrant, J.M., Wilkinson,
(2013). Genome‐wide association analysis M.L. et al. (1991). Dual association of HLA
in primary sclerosing cholangitis and DR2 and DR3 with primary sclerosing
ulcerative colitis identifies risk loci at cholangitis. Hepatology 13: 129–133.
GPR35 and TCF4. Hepatology 58: 32 Spurkland, A., Saarinen, S., Boberg, K.M.
1074–1083. et al. (1999). HLA class II haplotypes in
22 Liu, J.Z., Hov, J.R., Folseraas, T. et al. primary sclerosing cholangitis patients
(2013). Dense genotyping of immune‐ from five European populations. Tissue
related disease regions identifies nine new Antigens 53: 459–469.
risk loci for primary sclerosing cholangitis. 33 Donaldson, P.T. and Norris, S. (2002).
Nat. Genet. 45 (6): 670–675. Evaluation of the role of MHC class II
23 Ji, S.G., Juran, B.D., Mucha, S. et al. (2017). alleles, haplotypes and selected amino acid
Genome‐wide association study of primary sequences in primary sclerosing
sclerosing cholangitis identifies new risk cholangitis. Autoimmunity 35: 555–564.
loci and quantifies the genetic relationship 34 McElroy, J.P., Cree, B.A., Caillier, S.J. et al.
with inflammatory bowel disease. Nat. (2010). Refining the association of MHC
Genet. 49 (2): 269–273. with multiple sclerosis in African
24 Ellinghaus, D., Jostins, L., Spain, S.L. et al. Americans. Hum. Mol. Genet. 19:
(2016). Analysis of five chronic 3080–3088.
inflammatory diseases identifies 27 new 35 Donaldson, P.T. (2004). Genetics of liver
associations and highlights disease‐specific disease: immunogenetics and disease
patterns at shared loci. Nat. Genet. 48 (5): pathogenesis. Gut 53: 599–608.
510–518. 36 Folseraas, T., Melum, E., Franke, A., and
25 de Boer, Y.S., van Gerven, N.M., Zwiers, A. Karlsen, T.H. (2011). Genetics in primary
et al. (2014). Genome‐wide association sclerosing cholangitis. Best Pract. Res. Clin.
study identifies variants associated with Gastroenterol. 25: 713–726.
autoimmune hepatitis type 1. 37 Strettell, M.D., Donaldson, P.T., Thomson,
Gastroenterology 147 (2): 443–452.e5. L.J. et al. (1997). Allelic basis for HLA‐
encoded susceptibility to type 1 45 Liu, J.Z., Folseraas, T., Ellinghaus, E. et al.

autoimmune hepatitis. Gastroenterology (2012). Identification of multiple novel
112: 2028–2035. susceptibility regions for primary
38 Yoshizawa, K., Ota, M., Katsuyama, Y. et al. sclerosing cholangitis through dense
(2005). Genetic analysis of the HLA region genotyping of autoimmunity risk loci.
of Japanese patients with type 1 Hepatology 56: S86.
autoimmune hepatitis. J. Hepatol. 42: 46 Franke, A., McGovern, D.P., Barrett, J.C.
578–584. et al. (2010). Genome‐wide meta‐analysis
39 Bittencourt, P.L., Goldberg, A.C., Cancado, increases to 71 the number of confirmed
E.L. et al. (1999). Genetic heterogeneity in Crohn’s disease susceptibility loci. Nat.
susceptibility to autoimmune hepatitis Genet. 42: 1118–1125.
types 1 and 2. Am. J. Gastroenterol. 94: 47 Anderson, C.A., Boucher, G., Lees, C.W.
1906–1913. et al. (2011). Meta‐analysis identifies 29
40 Goldberg, A.C., Bittencourt, P.L., Mougin, additional ulcerative colitis risk loci,
B. et al. (2001). Analysis of HLA haplotypes increasing the number of confirmed
in autoimmune hepatitis type 1: identifying associations to 47. Nat. Genet. 43: 246–252.
the major susceptibility locus. Hum. 48 Bradfield, J.P., Qu, H.Q., Wang, K. et al.
Immunol. 62: 165–169. (2011). A genome‐wide meta‐analysis of six
41 Czaja, A.J., Kruger, M., Santrach, P.J. et al. type 1 diabetes cohorts identifies multiple
(1997). Genetic distinctions between types associated loci. PLoS Genet. 7: e1002293.
1 and 2 autoimmune hepatitis. Am. J. 49 Jostins, L., Ripke, S., Weersma, R.K. et al.
Gastroenterol. 92: 2197–2200. (2012). Host–microbe interactions have
42 Jurado, A., Cardaba, B., Jara, P. et al. (1997). shaped the genetic architecture of
Autoimmune hepatitis type 2 and hepatitis inflammatory bowel disease. Nature 491:
C virus infection: study of HLA antigens. 119–124.
J. Hepatol. 26: 983–991. 50 Mells, G. and Hirschfield, G.M. (2013).
43 Hov, J.R., Kosmoliaptsis, V., Traherne, Genetics of primary biliary cirrhosis. eLS:
J.A. et al. (2011). Electrostatic Wiley online library. doi:https://doi.
modifications of the human leukocyte org/10.1002/9780470015902.a0024406.
antigen‐DR P9 peptide‐binding pocket 51 Lundin, K.E., Scott, H., Fausa, O. et al.
and susceptibility to primary sclerosing (1994). T cells from the small intestinal
cholangitis. Hepatology 53 (6): mucosa of a DR4, DQ7/DR4, DQ8 celiac
1967–1976. disease patient preferentially recognize
44 Darlay, R., Ayers, K.L., Mells, G.F. et al. gliadin when presented by DQ8. Hum.
(2018). Amino acid residues in five separate Immunol. 41: 285–291.
HLA genes can explain most of the known 52 Hugot, J.P., Chamaillard, M., Zouali, H.
associations between the MHC and et al. (2001). Association of NOD2 leucine‐
primary biliary cholangitis. PLoS Genet. rich repeat variants with susceptibility to
14 (12): e1007833. Crohn’s disease. Nature 411: 599–603.
65
Autoantibodies and Understanding of Autoimmune

Liver Diseases
Benedetta Terziroli Beretta‐Piccoli1,3, Giorgina Mieli‐Vergani 2, and Diego Vergani 3
1
Epatocentro Ticino, Lugano, Switzerland
2
Faculty of Life Sciences and Medicine, Paediatric Liver, GI and Nutrition Centre, Mowat Labs, King’s College Hospital, London, UK
3
Institute of Liver Studies, Mowat Labs, King’s College Hospital, London, UK
Abstract
Autoimmune liver serology is an essential tool in the diagnosis and management of autoimmune
liver disease. However, autoantibodies are not diagnostic on their own, and the serologic laboratory
reports must always be interpreted in the clinical context of the single patient. In autoimmune
hepatitis (AIH), autoantibodies are a key disease feature: according to the simplified diagnostic cri-
teria published in 2008 by the International AIH Group, a definite diagnosis cannot be made in the
absence of positive autoimmune serology. Antinuclear antibody (ANA), originally named antinu-
clear factor, was the first antibody associated with AIH. Extrahepatic and systemic conditions may
also be associated with autoantibodies positivities, especially ANA, antismooth muscle antibody,
and antineutrophil cytoplasmic antibody (ANCA), thus multidisciplinary patient management is
key to avoid misdiagnosis. ANCA is tested on ethanol-fixed and formaldehyde-fixed human
neutrophils, in contrast to the remaining liver-relevant autoantibodies, which are detected on tri-
ple rodent tissue.
Keywords anti-mitochondrial antibody; anti-neutrophil cytoplasmic antibody; anti-nuclear antibody;

anti-smooth muscle antibody; autoimmune hepatitis; autoimmune liver serology; triple rodent tissue
Key Points
●● The recommended first‐line tests for are characterized by anti‐nuclear and/or
autoimmune liver serology are indirect anti‐smooth muscle antibody.
immunofluorescence (IIF) on triple rodent ●● AIH type 2 is characterized by anti‐liver‐
tissue coupled with a solid‐phase assay kidney microsomal type 1 and/or anti‐liver
for anti‐soluble liver antigen antibody. cytosol type 1 antibody.
●● IIF using human neutrophils as a substrate ●● Anti‐mitochondrial antibody is the serologic
should be used to detect anti‐neutrophil hallmark of primary biliary cholangitis (PBC).
cytoplasmic antibody (ANCA). ●● Anti‐nuclear antibody with a rim‐like/
●● Autoimmune hepatitis (AIH) type 1 and membranous or multiple nuclear dots
autoimmune sclerosing cholangitis (ASC) pattern on HEp2 cells is specific for PBC.
●● Atypical p‐ANCA is frequently found in ●● Autoimmune liver serology is not diag-

primary sclerosing cholangitis, ASC and nostic by itself, but must be interpreted in
inflammatory bowel disease; it may rarely the clinical context of the individual patient.
be the only serologic marker in AIH type 1. Viral hepatitis must always be ruled out.
Introduction immunosuppressive treatment is usually

required in type 2 AIH.
Autoantibodies are an essential tool in the Primary biliary cholangitis (PBC) is diag-
diagnosis and management of autoimmune nosed in the presence of anti‐mitochon
liver disease. They can be classified into dis- drial antibody (AMA) and of a cholestatic
ease‐specific, i.e. occurring almost only in biochemical profile; these diagnostic criteria
patients affected by a specific disease, and underscore the specificity of AMA. A liver
non‐specific, i.e. occurring in patients with a biopsy is needed only if one of these two
variety of hepatic and extrahepatic diseases criteria is not satisfied.
or even in healthy individuals. Some of them In primary sclerosing cholangitis (PSC), the
also have prognostic value. A basic knowledge most frequently detected autoantibody is anti‐
of the methods of detection and conse- neutrophil cytoplasmic antibody (ANCA),
quently an interaction with the laboratory which is found in up to 94% of the patients [3].
are crucial for the clinician in order to exploit ANCA positivity is found also in some 75% of
the value of autoimmune liver serology. patients with the pediatric condition referred
In autoimmune hepatitis (AIH), autoanti- to as autoimmune sclerosing cholangitis
bodies are a key disease feature: according to (ASC), distinguished from AIH by abnormal
the simplified diagnostic criteria published in cholangiogram at diagnosis. ANA and/or
2008 by the International AIH Group (IAIHG), SMA positivity in ASC is universal.
a definite diagnosis cannot be made in the De novo AIH is a condition indistinguish-
absence of positive autoimmune serology [1]. able from AIH that occurs in liver transplant
recipients who have undergone transplanta-
●● Type 1 AIH is characterized by the presence tion for conditions different from AIH.
of anti‐nuclear antibody (ANA) and/or De novo AIH is characterized serologically by
anti‐smooth muscle antibody (SMA); these either ANA and/or SMA or typical/atypical
specificities are present in over 95% of anti‐LKM1 antibody. Alloantigens (i.e.
patients, provided that they are tested antigens different in individuals of the same
according to established guidelines [2]. species) are only a minor component of
●● Type 2 AIH is diagnosed in the presence of the antigenic composition of the human body.
anti‐liver kidney microsomal type 1 (LKM1) The transplanted liver is soon colonized
antibody and/or anti‐liver cytosol type 1 by the recipient’s antigen‐presenting cells
(LC1) antibody. capable of presenting autoantigens as well
This subclassification of AIH into type 1 and as alloantigens within HLA restriction con-
type 2 is not merely a laboratory feature, but straints. In both recurrent and de novo AIH,
has clinical relevance. Type 2 AIH affects the peptide antigen presented within the cleft
mainly children (including infants) and ado- of HLA class 2 is likely to be an autoantigen.
lescents and is much rarer than type 1 AIH
and more aggressive, though it responds very
well to immunosuppressive treatment, which Methods of Detection
needs to be started as soon as the diagnosis is
made in order to avoid progression to liver The gold standard for detecting autoanti-
cirrhosis or fulminant liver failure. Therefore, bodies in the context of liver disease is indirect
the diagnosis should not be missed. Lifelong immunofluorescence (IIF) on fresh triple
Chapter 4 Autoantibodies and Understanding of Autoimmune Liver Diseases 67
rodent tissue (i.e. liver, kidney and stomach), region is added, and the preparation is placed
though some observer‐independent immuno- under an ultraviolet microscope: fluoro-
chemical techniques are either already avail- chrome‐labeled antibodies emit light allowing
able or under development [2]. IIF, a technique the detection of any bound antibody
introduced in 1954 by Weller and Coons, (Figure 4.1). Positive sera should be titrated to
allows the simultaneous detection of ANA, extinction; the process should be repeated,
SMA, AMA, anti‐LKM1, and anti‐LC1, the doubling the dilution until the reactivity disap-
latter specificity being usually masked by pears. Any nuclear reactivity (i.e. sera which
the concomitant presence of anti‐LKM1. stain the cell nuclei on triple rodent tissue)
However, IIF is observer‐dependent and should be further characterized on HEp2 cells,
poorly standardized. derived from a human laryngeal carcinoma,
In an attempt to maximize consistency which have prominent nuclei that allow the
among laboratories, a consensus statement detection of IIF nuclear patterns. These pat-
was published in 2004 by the Committee for terns have high clinical s ignificance and should
Autoimmune Serology of the IAIHG [2] always be reported by the laboratory to the cli-
providing detailed guidelines on substrate nician. An additional very important advantage
preparation, application of test sera, dilution of of IIF is the ability to detect autoantibodies
sera and fluorochrome‐labeled reagents, use of directed against still molecularly undefined
control sera, and interpretation of immunoflu- antigenic targets (see later).
orescence reactivities. The quality of the tissue ANCA are directed against antigens local-
substrate is important: besides appropriate ized in neutrophils, hence human neutrophils
cutting of the kidney tissue to ensure presence are used as an IIF substrate to look for this
of both cortex and medulla on the preparation, specificity.
rodent tissues should be freshly frozen and In adults, the positivity cutoff at IIF on
unfixed. Commercially available slides with triple rodent tissue is 1 : 40, while in chil-
triple rodent tissue are usually treated with fix- dren and adolescents the cutoff is lower,
atives to extend shelf‐life, hampering the inter- being 1 : 20 for ANA and SMA, and 1 : 10
pretation of fluorescence patterns due to for anti‐LKM1 and anti‐LC1. The target
background staining. The patient’s serum, at a antigens of anti‐LKM1, AMA, anti‐LC1,
starting dilution of 1 : 10, is incubated with the and partially ANA, SMA, and ANCA have
tissue sections, which are subsequently washed been identified, leading to the establish-
to remove unbound antibodies. A second ment of solid‐phase immunoassays. Such
fluorochrome‐labeled antibody specifically
assays are based on purified or recombinant
targeting the human immunoglobulin constant antigens coupled to a solid phase; patient
Antibody-containing
patient serum
Fluorochrome-
labeled anti-human
antibody
Figure 4.1 Diluted patient serum is incubated with triple rodent tissue sections, which are subsequently
washed to remove unbound antibodies. A second fluorochrome‐labeled antibody specifically targeting the
human immunoglobulin constant region is added, and the preparation is placed under an ultraviolet
microscope. Fluorochrome‐labeled antibodies emit light allowing the detection of any bound antibodies that
give specific patterns.
serum is added, and any circulating anti- large variety of reported IIF staining patterns
body targeting the specific antigen binds to and of molecularly identified ANA antigens.
it. A second antibody is added in order to ANA, originally named anti‐nuclear factor,
detect any bound autoantibody; this second was the first antibody associated with AIH. It
antibody may be enzyme‐labeled, radiola- was detected for the first time in 1949 in the
beled, or labeled with a chemiluminescence blood of patients with lupus erythematosus
or fluorescent agent (Figure 4.2). The test (LE), being responsible for the presence of
result may be quantitative or semiquantita- the LE cells, which are neutrophils engulfed
tive, depending on the technique employed. with nuclear debris of damaged cells. The
Enzyme‐linked immunosorbent assay phagocytosis is mediated by ANA. LE cells
(ELISA) is widely used because it can be had been reported by Ian Mackay in 1956 in
fully automated and allows the simulta- the ascites of patients with “chronic hyper-
neous analysis of large numbers of samples, gammaglobulinemic hepatitis,” which led to
in contrast to immunoblot (also called the original naming of AIH as lupoid
Western blot) which is demanding for the hepatitis. Subsequently, it became clear that
laboratory but has the main advantage of LE and lupoid hepatitis (or “chronic hyper-
being highly sensitive and specific. It is gammaglobulinemic hepatitis”) are different
important to note that the AIH‐specific diseases, both being characterized by ANA
anti‐soluble liver antigen antibody (anti‐ positivity. The cumbersome LE test was
SLA) is not detectable by IIF, and therefore replaced by IIF in the early 1960s.
recent international guidelines recommend
combining a solid‐phase assay for anti‐SLA Immunofluorescence Reactivities
with IIF on triple rodent issue in the initial and Antigenic Targets
diagnostic work‐up of patients with sus-
The recommended technique for investi-
pected autoimmune liver disease [4].
gating ANA in the context of liver disease is
IIF on triple rodent tissue (Figure 4.3) [2].
The use of HEp2 cells at a screening level is
Anti‐nuclear Antibody not advisable owing to a reported frequency
of ANA positivity (≥1 : 40) as high as 30%
History
in healthy adults using this very sensitive
ANA, as the name implies, is directed against technique. In IIF on triple rodent substrate,
antigenic targets located in the cell nucleus, ANA stains nuclei in each of the three tissues,
which is a complex structure, explaining the and the staining pattern should be further
Antigen-coated microplate well

Antibody-containing patient serum
Enzyme-labeled
Anti-human
immunoglobulin
Figure 4.2 ELISA (enzyme‐linked immunosorbent assay) as an example of a solid‐phase assay. Purified or
recombinant antigen is coupled to a microplate well. Patient serum is added and any circulating antibody
targeting the specific antigen binds to it. A second, enzyme‐labeled antibody is added in order to detect any
bound autoantibody. In a third step, a chromogenic enzymatic substrate solution is added, which promotes a
color reaction. The intensity of the color produced is proportional to the concentration of antibodies in the
serum sample.
ANA SMA Anti-LKM1 Anti-LC1
Figure 4.3 Immune serology in autoimmune hepatitis: ANA, anti‐nuclear antibody (rat liver), and/or SMA,
anti‐smooth muscle antibody (rat kidney), define autoimmune hepatitis type 1; anti‐LKM1, anti‐liver
microsomal antibody type 1 (rat liver, bottom; rat kidney, top), and/or anti‐LC1, anti‐liver cytosol type 1 (rat
liver), define autoimmune hepatitis type 2.
characterized on HEp2 cells. A variety of ANA The most frequently detected IIF pattern
patterns can be recognized in IIF: homoge- in AIH type 1 is the homogeneous one,
neous, speckled, nucleolar, membranous/rim‐ being present in some three‐quarters of the
like, and (multiple) nuclear dots. An updated patients, the remainder displaying a speck
tentative international consensus on ANA led or nucleolar pattern. Identified ANA
patterns is available at https://anapatterns. antigenic targets in AIH include double‐ and
org. The extractable nuclear antigens (ENAs) single‐stranded DNA, histones, centromere,
are a heterogeneous group of nuclear antigens chromatin, ribonucleoproteins, and cyclin A.
mostly corresponding to a speckled IIF pattern Commercial kits for a variety of identified
on HEp2 cells. They can be detected by ELISA nuclear antigens are available; however, they
or immunodiffusion. Anti‐ENA antibodies should not be used as a screening tool in AIH
are frequently named after the patients in because about one‐third of ANA‐positive
whose serum they were first identified: anti‐ AIH patients do not react with any of the
Sm for Smith (specific for LE), anti‐Ro for known nuclear molecular targets. Therefore,
Robert (also referred to as anti‐Sjögren syn- ANA in this substantial proportion of
drome antigen A), and anti‐La for Lane (also patients is detected only by IIF.
referred to as anti‐Sjögren syndrome antigen Two IIF ANA patterns are of particular
B). Anti‐ribonucleotide (typically associated importance in the context of liver disease,
with mixed connective tissue disease), anti‐ being pathognomonic for PBC:
Jo‐1 (typically associated with polymyositis),
●● the multiple nuclear dots (MND); and
and anti‐Scl‐70 and anti‐centromere anti-
●● the rim‐like/membranous pattern (Figure 4.4).
bodies (ACA) (both typically associated with
systemic sclerosis) are also part of the anti‐ This highlights the crucial relevance of the
ENA antibodies. The term “anti‐synthetase laboratory reporting the IIF pattern on HEp2
antibodies” is used to identify a subset of ANA cells of any ANA‐positive serum sample.
directed against aminoacyl‐tRNA synthetases Anti‐MND at IIF on HEp2 cells appears as
(including anti‐Jo‐1), which is highly specific 3–20 nuclear dots sized 0.2–1 μm, corres
for a form of immune‐mediated inflammatory ponding to the nuclear bodies. The main
myopathy associated with multiorgan involve- antigenic targets of anti‐MND are the nuclear
ment, primarily interstitial lung disease. proteins referred to as sp‐100 and promyelocytic
AMA RLM - ANA MND - ANA
Figure 4.4 Immune serology in primary biliary cholangitis: AMA, anti‐mitochondrial antibody (rat kidney, bottom;
rat stomach, top); RLM, rim‐like membranous antibody (HEp2 cells); MND, multiple nuclear dots (HEp2 cells).
leukemia protein. Recently, sp140 and small mitotic plate. This is in contrast to anti‐
ubiquitin‐related modifiers have also been MND, which do not stain metaphase cells,
reported to be targeted by anti‐MND. Anti‐ allowing differentiation between the two.
rim‐like/membranous ANA stains the The ACA target is the kinetochore, a com-
nuclear envelope, a complex structure con- plex protein localized to the chromosome
sisting of a double bilayer membrane, the centromere, the immunodominant autoanti-
nuclear lamina layer, and the nuclear pore gen being CENP‐B.
complex. It is therefore not surprising that a ANA are frequently detected also in PSC
variety of molecular targets have been identi- patients, with a reported prevalence rang-
fied, including gp210, lamin‐B receptor, and ing from 7.4 to 77%. However, their diag-
nucleoporin p62. ELISA and immunoblotting nostic and prognostic role, as well as
assays based on recombinant gp210 and specific antigenic targets, have been poorly
sp100 have been developed; however, they investigated. Reported antigenic targets in
should not replace IIF as they do not include PSC are double‐stranded DNA, SSA/B,
all antigens targeted by anti‐rim‐like/ ENA, RNP, Scl‐70, single‐stranded DNA
membranous and anti‐MND ANA, respec- and Sm (Table 4.1).
tively. They have a role in confirming IIF
results, as well as in detecting anti‐gp210 and
Clinical Significance in Autoimmune
anti‐sp100, being more sensitive than IIF,
Liver Disease
partially because at IIF concomitant AMA
may mask ANA. Non‐specific ANA IIF ANA characterizes type 1 AIH, being found
patterns detected in PBC patients include in two‐thirds of the patients, where it is asso-
ACA, speckled, homogeneous, and nucleolar ciated with anti‐SMA in half of the cases and
patterns. ACA is best identified on the
is found in isolation in the remainder
Hep‐20‐10 form of HEp2 cells which con- (Tables 4.1 and 4.2). ANA titers do not corre-
tains a high number of mitotic cells and late with disease activity in neither children
appears on interphase cells as multiple dis- nor adults [5,6]. ANA are very rarely found in
crete dots scattered all over the nucleus, association with anti‐LKM1 in type 2 AIH
corresponding to the chromosome centro- [6]. There is no association of specific ANA
meres. ACA stains mitotic cells as well, with IIF patterns or molecular targets in AIH with
a speckled/dot pattern localized to the clinical parameters, with the exception of a
Table 4.1 Clinical significance of anti‐nuclear antibody in autoimmune liver disease.
AIH PBC PSC
Frequency 75% in AIH type 1 10–40% 8–77%

IIF pattern on Homogeneous in ~75% Nuclear‐rim/membranous pattern in 10–40% Homogeneous
HEp2 cells Speckled or nucleolar in ~25% Multiple nuclear dots in 20–40% Speckled
Centromere in 9–30%
Target antigens Unknown in 30% Nuclear‐rim/membranous pattern: gp210, nucleoporin p62, Double‐stranded
Chromatin lamin B receptor DNA
Histones Multiple nuclear dots pattern: Sp100, promyelocytic leukemia SSA and SSB
Centromeres protein, Sp140, small ubiquitin‐related modifiers RNP
Cyclin A Ribonucleoproteins Centromeric pattern: CENP‐A, CENP‐B, CENP‐C, CENP‐D, Scl70
Double‐stranded DNA CENP‐E, CENP‐F Sm
Single‐stranded DNA Single‐stranded DNA
Comments Coexisting with anti‐SMA in Nuclear‐rim/membranous and multiple nuclear dots patterns Overlap with AIH
50% of type 1 AIH are virtually diagnostic of PBC has to be excluded
Positive in 75% of ASC patients Anti‐centromere pattern may be the only serologic marker in
AMA‐negative PBC
AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ASC, autoimmune sclerosing cholangitis; PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis;
RNP, ribonucleoprotein; Sm, Smith; SMA, smooth muscle antibody; SS, Sjögren’s antigen.
Table 4.2 Main features of the liver‐relevant autoimmune serologic specificities.
Other
Methods of associated
Specificity detection Antigenic targets Disease specificity Frequency liver diseases Comments
SMA IIF Filamentous actin VG and VGT IIF patterns 85% in AIH‐1 DILI V pattern is seen in 20% of type 1
Vimentin specific for AIH‐1 75% in ASC NAFLD AIH patients
Desmin Up to 83% in PSC Viral hepatitis Titers correlate with disease activity
Unknown in PBC Wilson disease
Anti‐actin Molecular‐ Actin Specific for AIH‐1 at high titers 60% in AIH‐1 DILI Sensitivity and specificity depend
antibody based assays NAFLD on the chosen cutoff point. Less
Viral hepatitis specific for AIH than the VG/VGT
Wilson disease IIF pattern
Anti‐LKM1 IIF Epitopes of CYP2D6 Specific for AIH‐2 in absence Up to 90% in Hepatitis C Titers correlate with disease activity
Molecular‐ of hepatitis C AIH‐2
based assays
Anti‐LC1 IIF Formiminotransferase Specific for AIH‐2 in absence 60% of AIH‐2 Hepatitis C Only serologic marker in 10–30%
Molecular‐ cyclodeaminase of hepatitis C of AIH‐2 cases
based assays Titers correlate with disease activity
Anti‐SLA Molecular‐ O‐Phosphoseryl‐tRNA (Sec) Highly specific for AIH. Low Up to 58% in Very rare in Associated with more severe
based assays selenium transferase disease sensitivity when tested AIH‐1, AIH‐2 hepatitis C disease course
with commercial solid phase and ASC
assays
pANNA IIF β‐Tubulin isotype 5; HMG1; Specific for AIH‐1, PSC and 40–96% in AIH‐1 May be the only serologic marker in
HMG2; other yet unknown IBD 26–94% in PSC AIH‐1
autoantigens 75% in ASC IBD to be excluded
Anti‐ IIF ASGPR Liver‐specific but not disease‐ 24–82% in AIH‐1 DILI Titers correlate with disease activity
ASGPR specific 16–57% in AIH‐2 Viral hepatitis
6–100% in PBC NAFLD
33% in PSC Alcoholic liver
disease
AMA IIF E2 subunits lipoyl domains Specific of PBC 95% in PBC Acute liver Predicts PBC if detected in absence
Molecular‐ of PDC, OGDC and failure of liver disease
based assays BCOADC, E3‐binding PBC/AIH
protein of PDC overlap
AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ASGPR, asialoglycoprotein receptor; BCOADC, branched‐chain 2‐oxo acid dehydrogenase complex; DILI, drug‐
induced liver injury; HMG1, high mobility group non‐histone chromosomal protein; IBD, inflammatory bowel disease; IIF, indirect immunofluorescence; LC1, liver cytosol type 1;
LKM, liver kidney microsomal; NAFLD, non‐alcoholic fatty liver disease; OGDC, 2‐oxoglutarate dehydrogenase complex; pANNA, perinuclear anti‐neutrophil nuclear antibody;
PBC, primary biliary cholangitis; PDC, pyruvate dehydrogenase complex; PSC, primary sclerosing cholangitis; SLA, soluble liver antigen; SMA, smooth muscle antibody; staining
patterns: V, vessel; G, glomerulus; T, tubulus.
reported association between anti‐double‐ remains to be established. Interestingly, some

stranded DNA positivity, characteristic of data suggest that in AMA‐negative patients,
systemic LE, and AIH with PBC overlapping ACA and anti‐Ro/SSA are specific serologic
features. PBC markers.
It is important to bear in mind that ANA
lacks disease specificity, being detected in a
variety of systemic autoimmune, infectious, Anti‐smooth Muscle
neoplastic, toxic, and metabolic conditions. In
the context of liver disease, ANA may be
and Anti‐actin Antibodies
detected in a wide range of acute and chronic
History
conditions, such as viral hepatitis B, C, D,
and E, acute liver failure, non‐alcoholic fatty The detection of SMA in the blood of patients
liver disease (NAFLD), alcohol‐induced with “chronic active hepatitis,” reported in
liver disease, and hepatocellular carcinoma. 1965, contributed significantly to the
Further more, it is found in a substantial differentiation of AIH from LE, since in the
proportion of liver‐graft recipients who were latter condition SMA was absent. Ten years
seronegative before transplantation. To add later, Gianfranco Bottazzo recognized that
more complexity, ANA are often positive in SMA shows three different staining patterns
patients with drug‐induced liver injury (DILI) on kidney substrate: serum from some
which, in the presence of autoantibodies, is patients stained only arterial vessels (V),
referred to as AIH‐like drug‐induced hepatitis, whereas serum from other patients stained
highlighting the close similarity with genuine both arterial vessels and glomerular mesan-
AIH. Indeed, the two conditions may be distin- gium (VG), and serum from still other
guished only during follow‐up, being identical patients stained peritubular fibers as well
at initial presentation; in AIH‐like drug‐ (VGT). While the V pattern was detected in
induced hepatitis, the immunosuppressive the sera of patients with a variety of hepatic
treatment can be discontinued without recur- and extrahepatic diseases, mainly at low
rence. Therefore, in order to avoid long‐term titers, the VG and VGT patterns were
unnecessary treatment, a careful medical his- restricted to patients with severe chronic
tory about the use of medicinal and herbal hepatitis responsive to steroids, often at high
products is an essential part of the diagnostic titers. This early work showed the high diag-
work‐up of patients presenting with suspi- nostic value of the VG and VGT patterns, in
cion of AIH. Lastly, it should be noted that contrast to the V pattern.
some nuclear reactivities reported in AIH are
characteristic of other extrahepatic autoim-
Immunofluorescence Reactivities,
mune diseases, such as ACA for systemic
Antigenic Targets, and Clinical
sclerosis, or anti‐double‐stranded DNA for
Significance
systemic LE. Therefore, any ANA positivity
should be interpreted in the specific clinical At IIF on triple rodent tissue, SMA stains the
context. smooth muscle of the arterial walls on all
As already mentioned, detection of the three substrates (Figure 4.3). In addition, on
rim‐like/membranous or MND IIF patterns stomach sections, it stains the muscularis
is virtually diagnostic of PBC (Table 4.1). mucosae and the lamina propria of the gastric
Moreover, positivity of these PBC‐specific wall and, on kidney tissue, it shows three
ANA subtypes is associated with a more different staining patterns (as described in
severe disease course. In clinical practice, they the previous section), namely the V, VG, and
may be found in patients without overt liver VGT patterns. If cultured fibroblasts or vascular
disease; whether this predicts the development smooth muscle 47 cells from rat embryonic
of PBC akin to isolated AMA (see below) thoracic aorta are used as a substrate, the
VGT pattern corresponds to the F‐actin or from AIH. For this reason, the laboratory
microfilament pattern. These reactivities are should report the SMA IIF pattern on kidney
highly characteristic, but not entirely specific tissue of every positive sample. Nevertheless,
for type 1 AIH. Furthermore, some 20% of the clinician needs to be aware that some 20%
type 1 AIH patients lack the VGT/microfila- of the type 1 AIH SMA‐positive patients have
ment pattern, and this finding should not only the V pattern. Recent data suggest that
preclude a diagnosis of AIH. The observed SMA‐positive patients with VG/VGT pat-
F‐actin IIF reactivity suggested the establish- terns and any increase of transaminase levels
ment of molecular‐based assays using puri- are at risk of developing type 1 AIH on long‐
fied filamentous actin as an antigen source. term follow‐up. Molecular assays based on
These assays are less specific than the VG/ purified actin are available and should be
VGT IIF patterns, their specificity being used as complementary to IIF. Lastly, it should
higher if the anti‐actin antibody is detected be recalled that ASC is characterized by the
at high concentration. Molecularly tested low‐ same serologic profile as type 1 AIH, namely
concentration anti‐actin antibody is found in by SMA and ANA positivity.
a wide variety of conditions, including non‐
alcoholic steatohepatitis, type 2 AIH, PBC,
PSC and, importantly, viral hepatitis. Attempts Anti‐liver‐kidney
have been made to establish higher cutoffs of
the assays, but the improved specificity is
Microsomal Antibody
achieved at the expense of sensitivity. The cur-
History
rently used cutoffs are low, leaving it to the cli-
nician to correctly interpret the laboratory Anti‐LKM was first reported by Mario
results in the clinical context. The molecular Rizzetto in 1973 while working in Deborah
target of VGT SMA is probably part of fila- Doniach’s laboratory in London. He detected
mentous actin, as suggested by the IIF pattern in the serum of 16 patients, the majority hav-
seen on fibroblasts. However, this assumption ing liver disease, an antibody staining liver
is challenged by the fact that about 20% of and kidney tissue of both human and animal
AIH type 1 patients with VG or VGT SMA origin. This antibody was named anti‐LKM
fail to react with purified actin when tested by since the reactivity was abolished by
molecular‐based assays. Possible causes for incubation of the serum with a “microsomal
this incomplete overlap are presence of other, fraction” obtained by ultracentrifugation of a
yet unknown, SMA antigenic targets, or loss liver homogenate. This fraction contains the
of epitopes during antigen purification. endoplasmic reticulum, where the antigenic
SMA titers correlate with disease activity protein is located, as subsequently demon-
both in adults and in children with AIH type strated. The liver disease associated with the
1 and should be used, together with IgG anti‐LKM antibody was characterized by
levels, to monitor treatment response. Daniel Alagille’s group in Paris in 1987, who
In conclusion, SMA characterizes type 1 described this condition affecting mainly
AIH, being detected in some 85% of patients, children. In Alagille’s series, the outcome was
where it is associated with ANA in half of the poor despite treatment with prednisone and
cases. However, it lacks disease specificity, azathioprine. This aggressive juvenile liver
being detected in viral hepatitis, malig- disease was called type 2 AIH.
nancies, rheumatic diseases, DILI, NAFLD, Anti‐LKM was renamed anti‐LKM1 after
and alcohol‐induced liver disease. The IIF the description of antibodies with slightly dif-
patterns on kidney tissue are very helpful to ferent IIF patterns on triple rodent tissue,
the clinician, since the VG and VGT patterns which were referred to as anti‐LKM2 and anti‐
are specific for AIH, while the V pattern is LKM3, respectively. Anti‐LKM2, targeting
seen also in a variety of conditions different CYP2C9, is of historical interest, since it is
associated with hepatitis induced by ticrynafen levels and higher histologic activity has been
(also known as tielinic acid), an uricosuric anti- reported. Interestingly, two case reports have
hypertensive drug withdrawn from the market been published showing AIH type 2
in the 1980s for its hepatotoxicity. Anti‐LKM3, development following spontaneous HCV
also originally reported by Mario Rizzetto’s clearance. Both patients were females; the
group, is detected in about 13% of patients time elapsed from viral clearance to type 2
with chronic hepatitis delta infection and in a AIH diagnosis was 17 months and 8 years,
minority of type 2 AIH patients. respectively. These cases suggest that, though
rarely, HCV may have a role in triggering
AIH in predisposed individuals, a concept
Immunofluorescence Reactivities, supported by the identification of HCV
sequences that are shared in common with
Significance
CYP2D6, possibly leading to cross‐reactivity
Anti‐LKM1 is the serologic hallmark of type of anti‐HCV antibodies with self‐antigens.
2 AIH, provided that hepatitis C virus (HCV) Therefore, LKM1‐positive HCV‐infected
infection is ruled out. At IIF on triple rodent patients may be predisposed to autoimmu-
tissue, it stains brightly hepatocyte cyto- nity. Because of interferon’s potential to
plasm, while on kidney tissue it stains the induce/enhance autoimmunity, anti‐LKM1‐
larger, proximal tubules (Figure 4.3). It does positive HCV patients were a challenging
not stain stomach tissue, at variance with clinical issue until a few years ago. In the
AMA, thus allowing differentiation of the current era of direct‐acting antiviral agents
two reactivities. This difference is the main this is no longer the case, and patients can be
reason why it is not advisable to use slides safely treated for HCV infection, avoiding
with only kidney tissue as an IIF substrate. interferon and its side effects. However, long‐
Indeed, there are fine differences in the IIF term follow‐up data of anti‐LKM1‐positive
pattern on kidney tissue between anti‐LKM HCV patients after successful treatment with
and AMA, but they are appreciated only be direct‐acting antiviral agents are lacking.
experienced observers. Furthermore, the use In addition to type 2 AIH and HCV infec-
of triple tissue substrate allows the detection tion, anti‐LKM1 is also detected in the rare
of anti‐LC1, which stains only liver tissue. patients with AIH type 2‐like hepatitis in the
The antigenic target of anti‐LKM1 is the context of autoimmune polyendocrinopathy/
hepatic cytochrome P4502D6 (CYP2D6). candidiasis/ectodermal dystrophy (APECED),
Reported immunodominant linear epitopes an autosomal recessive disease due to muta-
within this large protein are CYP2D6193–212, tions in the autoimmune regulator (AIRE)
CYP2D6254–271 and CYP2D6321–351. Molecular‐ gene. Despite the fact that on IIF anti‐LKM1
based assays are commercially available and in this condition is indistinguishable from
have been shown to have high sensitivity and anti‐LKM1 seen in AIH type 2, it targets a
specificity. However, they should not be used different cytochrome, namely CYP2A6, in

at a screening level for the previously men- addition to CYP2D6.
tioned advantages of IIF on triple rodent tissue. The target antigen of anti‐LKM3 is the
In type 2 AIH, anti‐LKM1 levels correlate with uridine 5′‐diphosphate glucuronosyltrans-
disease activity and should be tested every 6 ferase 1 (UGT‐1) family. In rare cases of type
months to monitor the treatment response. 2 AIH this is the only serologic positivity.
Anti‐LKM1 is usually not detectable after liver However, it is not routinely tested by IIF
transplantation for type 2 AIH, and its reap- because it stains only human tissue, and
pearance predicts post‐transplant relapse. molecular‐based assays are lacking.
Anti‐LKM1 is detected in up to 13% of Lastly, it is worth mentioning that anti‐
patients with chronic HCV infection, where LKM1 giving an atypical IIF staining pattern
an association with higher transaminase on triple rodent tissue has been reported in
de novo AIH, a disease entity indistinguish- There are no clinical features associated
able from AIH arising in patients who have with either reactivity. Although more rarely
undergone liver transplantation for liver dis- than anti‐LKM1, anti‐LC1 has been reported
eases different from AIH. Atypical anti‐LKM1 in patients with chronic HCV infection,
stains only kidney tissue. mainly in association with anti‐LKM1,
thus again underlying the crucial impor-
tance of excluding viral hepatitis before
nti‐liver Cytosol Type 1
A diagnosing AIH.
Antibody
History
nti‐soluble Liver Antigen
A
Anti‐LC1 was identified in Paris in 1988 in Antibody
21 patients with chronic hepatitis (age
range 3–21 years), the majority being History
positive also for anti‐LKM1. Anti‐LC1 was
In 1981, Peter Berg in Germany reported an
not found in healthy controls and in a large
antibody reacting with liver and pancreas
cohort of patients with systemic and
rodent homogenates detected in the blood
hepatic well‐characterized conditions,
of 20 patients with an aggressive form of
thus already suggesting its high disease
chronic hepatitis, responding satisfactorily
specificity.
to immunosuppression with steroids and
azathioprine. He suggested naming this anti-
Immunofluorescence Reactivities, body anti‐liver pancreas (LP). The history of
what later became known as anti‐SLA con-
Significance
tinues in Germany, where in 1986 Michael
The IIF pattern of anti‐LC1 on triple rodent Manns reported an antibody reacting with
tissue is one of bright staining confined to rat liver homogenates which he called anti‐
the liver tissue, with a weakening of the SLA, also detected in the blood of patients
staining around the central vein (Figure 4.3). with chronic hepatitis, mostly young women,
As mentioned before, anti‐LKM1 and anti‐ with hypergammaglobulinemia and good
LC1 often coexist in the same patient; in response to immunosuppressive therapy.
these cases, in IIF anti‐LC1 is masked by Neither anti‐LP nor anti‐SLA was detectable
anti‐LKM1, making the availability of molec- by IIF. Soon it became clear that anti‐LP and
ular‐based assays particularly useful. Indeed, anti‐SLA were one and the same. Both
the identification of the liver enzyme original reports proposed that patients
formiminotransferase cyclodeaminase as positive for anti‐SLA/LP represent a third
the anti‐LC1 antigenic target allowed the type of AIH, since this reactivity was the
establishment of such assays, which should only one detected in a substantial proportion
be used to investigate anti‐LC1 in anti‐ of patients with AIH. This proposal was later
LKM1‐positive patients. declined by the IAIHG, since in the original
About two‐thirds of type 2 AIH patients reports the ANA positivity cutoff was greater
are positive for both anti‐LKM1 and anti‐ than 1 : 80, thus significantly higher than the
LC1. Few data are available about the ≥1 : 40 recommended by the IAIHG diag-
proportion of type 2 AIH patients with iso- nostic scoring systems. Therefore, some
lated anti‐LC1 reactivity, but eported anti‐SLA/LP patients in the original reports
frequencies range from 10 to 30%. All type 2 were likely to be positive for ANA as well,
AIH patients are positive for anti‐LKM1 and thus would be classified as type 1 AIH.
and/or anti‐LC1, apart from the very rare The current name has been simplified to
cases positive for anti‐LKM3 in isolation. anti‐SLA.
Immunofluorescence Reactivities, soon after ANCA was reported also in the

Antigenic Targets, and Clinical context of AIH.
Significance
Anti‐SLA is not detectable by IIF; accordingly, Immunofluorescence Reactivities,

current guidelines [4,7] recommend including Antigenic Targets, and their Clinical
a solid‐phase assay in the initial work‐up of Significance
suspected AIH, in addition to IIF on triple
rodent tissue. Anti‐SLA specificity for AIH is ANCA is tested on ethanol‐fixed and formal-
described to be as high as 98.9% [8] and this dehyde‐fixed human neutrophils, in contrast
high disease specificity was already evident in to the remaining liver‐relevant autoanti-
the early reports previously mentioned, where bodies, which are detected on triple rodent
anti‐SLA was extremely rare outside AIH. It tissue, as already discussed. According to the
can be present in both type 1 and type 2 AIH, IIF patterns on fixed neutrophils, ANCA are
and is associated with a more severe disease subdivided into cANCA, which stains the
course and a higher risk of relapse after therapy neutrophil cytoplasm of both ethanol‐fixed
withdrawal. The target antigen is O‐phospho- and formaldehyde‐fixed neutrophils, and
seryl‐tRNA (Sec) selenium transferase, a sele- pANCA, which give a perinuclear staining
nocysteine synthase. Available solid‐phase pattern only on ethanol‐fixed neutrophils.
assays for anti‐SLA have variable sensitivity, The antigenic target of pANCA is mainly the
explaining the high variability of the reported cytoplasmic enzymatic protein myeloperoxi-
frequency of anti‐SLA. The highly sensitive dase and the perinuclear staining pattern on
radioligand assays, which preserve conforma- ethanol‐fixed neutrophils is due to an artifact
tional epitopes, remain confined to research leading to migration of the positively charged
laboratories. If highly sensitive assays are used, myeloperoxidase toward the negatively charged
the frequency is as high as 58% in type 1 and nuclear membrane. The main antigenic target
type 2 AIH, and 41% in ASC. of cANCA is leukocyte proteinase 3. cANCA is
Interestingly, anti‐SLA is often associated mainly associated with granulomatosis with
with anti‐Ro52, which is a subtype of the ANA polyangiitis, while pANCA is mainly seen in
anti‐SSA. This association probably explains patients with microscopic polyangiitis and
the reported higher risk of fetal loss in anti‐ eosinophilic granulomatosis with polyangi-
SLA‐positive patients, since anti‐SSA are able to itis. In AIH, ASC, PSC, and IBD the most
cross the placental barrier causing fetal disease, commonly detected IIF pattern on a neutro-
including potentially fatal high‐grade heart phil substrate is a different one, namely a
block. This condition goes under the descrip- pANCA pattern that remains perinuclear
tion of neonatal systemic LE syndrome, which is whether neutrophils are fixed in ethanol or
somehow a confusing nomenclature, being dif- formaldehyde; this reactivity is referred to as
ferent from pediatric and adult systemic LE. atypical pANCA, perinuclear anti‐neutro-
phil nuclear antibody (pANNA), or nuclear
anti‐neutrophil antibody (NANA). The
Anti‐neutrophil target antigen is located to the nuclear
envelope, and has been identified as β‐
Cytoplasmic Antibody tubulin isotype 5, but probably there are
more yet unknown nuclear antigenic targets.
History
Indeed, in AIH, ulcerative colitis and PSC, a
ANCA was first reported in the 1980s in con- variety of other nuclear antigenic targets of
ditions not related to the liver, namely small pANNA have been described, including the
and medium‐vessel vasculitis. The association high‐mobility group non‐histone chromosomal
with PSC and inflammatory bowel diseases proteins HMG1 and HMG2, and histone.
(IBDs) was recognized in the early 1990s, and Cytoplasmic ANCA antigenic targets reported
in PSC patients include H1, bactericidal/ published a case report demonstrating high
permeability increasing protein, lactoferrin, titers of autoantibodies to tissue homoge-
elastase, cathepsin G, catalase, and human nates in a PBC patient. Seven years later, a
lysosomal‐associated membrane protein 2. landmark paper by Sheila Sherlock’s group
The clinician should bear in mind that described a characteristic IIF reactivity in
ANA stains the neutrophil nucleus and the serum of 32 PBC patients, which was
therefore may mask ANCA if both antibodies not seen in patients with cholestatic dis-
coexist in the same serum. Substrate con- eases of other origins. That this autoanti-
taining both HEp2 cells and neutrophils may body was directed against a mitochondrial
be helpful in detecting both specificities. antigen was first suggested in 1967 in a
Some laboratories do not routinely test ANCA paper reporting its reactivity against mito-
by IIF on fixed neutrophils, but in stead use chondrial fraction of liver homogenates
molecular‐based assays containing recombinant from differential centrifugation. Different
or purified antigenic targets, mainly proteinase 3 mitochondrial reactivities were subse-
and myeloperoxidase. pANNA, however, is not quently reported, leading to a nomenclature
detected by such assays. ranging from M1 to M9, proposed by Peter
The reported proportion of type 1 AIH Berg, which is no longer used.
patients with ANCA positivity ranges bet- The identification of the AMA molecular
ween 40 and 96%. Importantly, ANCA may be target as the E2 subunit of the pyruvate
the only serologic positivity in rare type 1 AIH dehydrogenase complex (PDC) protein by
patients. Current guidelines recommend Eric Gershwin in 1987 was a further major
looking for ANCA in patients with suspected step forward [10], allowing the establish-
AIH who test negative in IIF on triple rodent ment of assays based on the recombinant
tissue and with solid‐phase assays for anti‐ antigen.
SLA. Of note, ANCA is absent in type 2 AIH.
ANCA is frequently detected in children
with ASC, being present in some three‐quar- Immunofluorescence Reactivities,
Antigenic Targets, and their Clinical
ters of the cases [9]. There are no associa-
Significance
tions between ANCA and specific clinical
features in type 1 AIH and ASC. In PSC, the Since AMA targets mitochondria, it stains all
reported frequency of ANCA is up to 94%. mitochondria‐rich tissues. On triple rodent
Unfortunately, patients with small duct PSC tissue, it stains the hepatocyte cytoplasm
were excluded from studies, so no data are faintly and the gastric parietal cells brightly,
available on ANCA frequency in this sub- whereas on kidney tissue sections it stains
group. Some studies in adults with PSC have preferentially the distal, mitochondria‐rich
suggested that ANCA‐positive patients may tubules (Figure 4.4). This is in contrast to
have a more benign disease course. anti‐LKM1, which stains the proximal, larger
Because of the strong association of renal tubules. The cytoplasm of HEp2 cells is
pANNA with IBD, its detection should also stained by AMA, and laboratory reports
prompt endoscopic investigations even in often suggest AMA positivity of sera giving
asymptomatic patients. a diffuse, reticular cytoplasmic IIF pattern
on HEp2 cells. However, it is not advisable to
screen for AMA on HEp2 cells, since the IIF
pattern on HEp2 cells of AMA does not
Anti‐mitochondrial Antibody always mirror the IIF results on triple rodent
tissue or solid‐phase assays.
History
PDC is part of the 2‐oxo‐acid dehydrogenase
The first report of autoantibody positivity multienzyme complex, together with 2‐oxo-
in PBC dates back to 1958, when Ian Mackay glutarate dehydrogenase complex (OGDC)
and branched‐chain 2‐oxo acid dehydrogenase Immunofluorescence Reactivities,

complex (BCOADC). These are key enzymes Antigenic Targets, and their Clinical
in cellular energy metabolism and share the Significance
E2 subunit, which contains the immunodomi- ASGPR is a lectin constitutively expressed on
nant epitope recognized by AMA, namely the the hepatocyte membrane, and is the only
lipoyl domain. The epitope most commonly liver‐specific autoantigen reported so far.
targeted by AMA is PDC‐E2, this reactivity Data suggest that anti‐ASGPR is pathogenic.
being found in 80–95% of PBC patients, often These two features made anti‐ASGPR very
in association with reactivity against OGDC attractive for both clinical and research pur-
and/or BCOADC. Isolated reactivity against poses. The detection of anti‐ASGPR requires
one of the two latter antigens is rare. A purified or recombinant ASGPR. Reliable
recombinant protein termed MIT3 was devel- molecular‐based assays are missing, owing to
oped in 1993 by Gershwin’s group, having the loss of conformational epitopes during
main advantage of coexpressing the three antigen purification and lack of success in
immunodominant epitopes of the E2 subunits obtaining the recombinant protein. In
of PDC, OGDC, and BCOADC. MIT3‐based addition, anti‐ASGPR is organ‐ but not dis-
assays are more sensitive than IIF or assays ease‐specific, being detected in a variety of
based on conventional antigens, and a sub- liver diseases other than AIH, including PBC,
stantial proportion of patients previously clas- PSC, viral hepatitis, non‐alcoholic steato-
sified as AMA‐negative PBC were found to be hepatitis, alcoholic liver disease, and DILI.
AMA‐positive by this methodology. However, For these reasons, it is not used in clinical
MIT3‐based assays are less specific, leading to practice. Anti‐ASGPR is detected in both
low‐level AMA detection in patients with type 1 and type 2 AIH, its frequency varying
conditions different from PBC. according to the assays employed, ranging
AMA are increasingly seen also in subjects from 24 to 82% and from 16 to 57% in type 1
without overt liver disease. There is evidence and type 2 AIH, respectively. In PSC, one
that such individuals are at risk of developing study reported a prevalence of anti‐ASGPR
PBC on long‐term follow‐up, thus probably of 33.3%, but other studies have reported
deserving monitoring. lower prevalences. In PBC, reported preva-
AMA is considered the most specific auto- lence of anti‐ASGPR ranges from 6.2 to
antibody in human immunopathology. 100%, depending on the employed assays.
However, it is important for the clinician to
be aware that, beside low‐titer positivity when
tested with highly sensitive assays in patients
with conditions other than PBC, AMA may I ndications for Autoimmune
be transiently positive in acute liver failure. Liver Serology Testing
The indication for testing autoimmune liver
serology in patients with liver disease
Anti‐asialoglycoprotein
depends on the clinical context, but testing is
Receptor Antibody probably never wrong, provided that results
are interpreted in the clinical context of the
History
individual patient. A diagnostic algorithm
Anti‐asialoglycoprotein receptor (ASGPR) anti- based on IIF on triple rodent tissue is pro-
body targets a component of the “liver‐specific posed in Figure 4.5. The importance of an
protein” (LSP). Anti‐LSP antibody was shown accurate investigation of viral hepatitis in
in the 1980s to correlate with inflammatory every patient with liver disease cannot be
disease activity and with relapse after
stressed enough. The significance of autoim-
treatment withdrawal in patients with AIH. mune liver serology is different in patients
Indirect immunofluorescence on triple rodent tissue for

ANA, SMA, AMA, LKM1, LC1
negative
SMA positive:
ANA positive: characterize AMA positive:
characterize LKM1 and/or
pattern on HEp2 cells LC1 positive:
pattern on kidney PBC highly
tissue probable AIH still possible: check
diagnostic for
AIH-2 for anti-SLA (ELISA/line
Multiple nuclear blot) and ANCA (IIF)
dots or
Membranous/rim-
like: PBC highly VG or VGT pattern:
probable highly suggestive of
AIH SLA positive: pANNA positive highly negative
diagnostic of AIH suggestive of AIH-1 or
ASC
Homogeneous :
suggestive of AIH MIT3-based AMA
V pattern: AIH still assay,
possible, particularly if recombinant-
high titer based gp210/
Any positivity: highly sp100 assays
Other patterns: PBC, AIH, suggestive for PBC
PSC still possible
Figure 4.5 Diagnostic algorithm for autoimmune liver serology based on indirect immunofluorescence on
triple rodent tissue. AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ANA, anti‐nuclear
antibody; ANCA, anti‐neutrophil cytoplasmic antibody; ASC, autoimmune sclerosing cholangitis; LC1, anti‐liver
cytosol type 1; LKM1, anti‐liver kidney microsomal type 1; pANNA, perinuclear anti‐neutrophil nuclear
antibody; PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis; SLA, soluble liver antigen; SMA,
smooth muscle antibody; staining patterns: V, vessel; G, glomerular; T, tubular.
with or without acute or chronic viral It is crucial to accurately assess drug/herbal

hepatitis A, B, C, D, and E. In addition to viral remedy intake, and it is advisable in non‐
hepatitis, two further liver diseases deserve severe cases to observe the clinical course
special consideration in this context, namely after drug/herbal remedy withdrawal for a
NAFLD and DILI. The first condition affects few weeks before introducing immunosup-
as much as one quarter of the general pressive treatment. Biologics, which are
population in Western countries, increasing increasingly used, can also induce DILI with
with age. NAFLD is associated with ANA, autoimmune features, a knowledge essential
SMA and anti‐actin positivity, and, to a lesser in clinical practice.
extent, also AMA positivity. Since the liver Distinction between specific and non‐
histology differs in AIH and NAFLD, a biopsy specific liver autoantibodies is important; for
may be helpful in doubtful cases, bearing in example, a liver biopsy should be considered
mind that, though rarely, patients may be in SLA‐positive patients even with mild
affected by both conditions. DILI may repre- hepatitis, provided that viral hepatitis has
sent a clinical challenge as well: medications been ruled out or treated if present.
and herbal remedies are widely used by Extrahepatic and systemic conditions may
patients, sometimes from undocumented also be associated with autoantibodies posi-
manufacture, and a clinical picture of AIH in tivities, especially ANA, SMA, AMA, and
a patient taking drugs/herbal remedies is fre- ANCA, thus multidisciplinary patient
quent in clinical practice. Widely used drugs management is key to avoid misdiagnosis.
such as statins and various antibiotics have Autoimmune liver serology may be nega-
been recognized as potential triggers of AIH. tive at initial presentation, so that in patients
with unclear liver disease, particularly those tibodies are not diagnostic on their own, and
presenting with acute/fulminant hepatitis, it the serologic laboratory reports must always
is advisable to repeat testing during follow up. be interpreted in the clinical context of the
single patient. Exclusion of viral hepatitis is of
particular importance, since it is a frequent
Concluding Remarks condition and may be associated with positive
autoimmune serology. The clinician should
Autoimmune liver serology is an essential be familiar with the basic principles of the
tool in the diagnosis and management of laboratory techniques, both to order the
autoimmune liver disease. However, autoan- appropriate tests and to interpret the results.
References
1 Hennes, E.M., Zeniya, M., Czaja, A.J. et al. 6 Couto, C.A., Bittencourt, P.L., Porta, G. et al.
(2008). Simplified criteria for the diagnosis of (2014). Antismooth muscle and antiactin
autoimmune hepatitis. Hepatology 48: antibodies are indirect markers of
169–176. https://doi.org/10.1002/hep.22322. histological and biochemical activity of
2 Vergani, D., Alvarez, F., Bianchi, F.B. et al. autoimmune hepatitis. Hepatology 59:
(2004). Liver autoimmune serology: a 592–600. https://doi.org/10.1002/hep.26666.
consensus statement from the committee 7 Manns, M.P., Czaja, A.J., Gorham, J.D. et al.
for autoimmune serology of the (2010). Diagnosis and management of
international autoimmune hepatitis group. autoimmune hepatitis. Hepatology 51:
J. Hepatol. 41: 677–683. https://doi. 2193–2213. https://doi.org/10.1002/
org/10.1016/j.jhep.2004.08.002. hep.23584.
3 Hov, J.R., Boberg, K.M., and Karlsen, T.H. 8 Mieli‐Vergani, G., Vergani, D., Czaja, A.J.
(2008). Autoantibodies in primary et al. (2018). Autoimmune hepatitis. Nat.
sclerosing cholangitis. World J. Rev. Dis. Primers 4: 18017. https://doi.
Gastroenterol. 14: 3781–3791. https://doi. org/10.1038/nrdp.2018.17.
org/10.3748/wjg.14.3781. 9 Gregorio, G.V., Portmann, B., Karani, J.
4 European Association for the Study of the et al. (2001). Autoimmune hepatitis/
Liver (2015). EASL clinical practice sclerosing cholangitis overlap syndrome in
guidelines: autoimmune hepatitis. J. childhood: a 16‐year prospective study.
Hepatol. 63: 971–1004. https://doi. Hepatology 33: 544–553. https://doi.
org/10.1016/j.jhep.2015.06.030. org/10.1053/jhep.2001.22131.
5 Gregorio, G.V., McFarlane, B., Bracken, P. 10 Gershwin, M.E., Mackay, I.R., Sturgess, A.,
et al. (2002). Organ and non‐organ specific and Coppel, R.L. (1987). Identification and
autoantibody titres and IgG levels as specificity of a cDNA encoding the 70 kd
markers of disease activity: a longitudinal mitochondrial antigen recognized in
study in childhood autoimmune liver primary biliary cirrhosis. J. Immunol. 138:
disease. Autoimmunity 35: 515–519. 3525–3531.
83
Environmental Exposure and Risk in Autoimmune

Liver Diseases
Ying Qi Li and Andrew L. Mason
Division of Gastroenterology and Hepatology, University of Alberta, Edmonton, Alberta, Canada
Abstract
It is generally accepted that autoimmune liver diseases arise as a result of a complex interaction of
environmental factors on the background of genetic predisposition to disease. This chapter reviews
the research into the environmental triggers linked with primary biliary cholangitis (PBC) and
factors linked with autoimmune hepatitis. Human leukocyte antigen polymorphisms have been
closely linked with the development of autoimmune disease, drug hypersensitivity, and persistence
of viral infection, whereas other polymorphisms provide protection against developing autoim
mune disease, vasculitis, and clearance of viral infection. There are several models and clinical
observations illustrating how an environmental agent can trigger an autoimmune response, some
of which are not mutually exclusive. The involvement of xenobiotics in the pathogenesis of PBC is
supported by epidemiologic clustering studies around coal mines and toxic waste sites as well as by
animal models modified by specific chemical compounds to elicit the PBC specific antimitochon
drial antibody response.
Keywords animal models; antimitochondrial antibody; autoimmune hepatitis; autoimmune liver

diseases; environmental triggers; human leukocyte antigen polymorphisms; primary biliary
cholangitis; viral infection; xenobiotics
Key Points
●● Autoimmune liver diseases likely arise ●● The possibility of an environmental factor(s)
from a complex interaction between in triggering autoimmune liver diseases is
genetic and environmental factors. suggested by several epidemiologic studies
●● It is possible that the genetic predisposi and laboratory models.
tion provides a relative state of increased ●● Mechanisms by which agents could
risk to specific infectious or otherwise induce loss of tolerance to self include
noxious chemical agents that mediate loss molecular mimicry, presentation of
of tolerance to self and immune mediated cryptic antigens, epitope spreading,
injury. and microbial superantigens triggering
autoimmunity and immunodeficiency.
Introduction antibodies are detected in the majority of

PSC patients, they are not disease specific
It is generally accepted that autoimmune liver and, curiously, their presence is associated
diseases arise as a result of a complex interac with a significantly diminished risk of chol
tion of environmental factors on the angiocarcinoma. AIH is much more difficult
background of genetic predisposition to to phenotype diagnostically and immuno
disease. However, opinions differ of how
logically. Furthermore, there exist AIH
the genes and environment interact to firstly overlap syndromes with similar immune and
trigger autoimmunity and then cause disease clinical features of PBC as well as PSC. It is
[1]. There is a good argument for maintaining possible that these syndromes occur as a
a distinction between the disease process result of shared genetic and environmental
and loss of tolerance to self. For example, risk factors with the cholestatic disorders.
autoimmunity is often physiologic and not It appears as though the AIH overlap
pathogenic. It is well documented that syndromes have a worse prognosis.
IgM complexes help to remove senescent
and harmful substances from the body. On the
other hand, autoimmune responses can Autoimmunity
clearly instigate pathology. Then there are also
examples where autoimmune responses are It is often assumed that the autoimmune
pathologically irrelevant but still diagnosti response actually causes the disease
cally useful. Accordingly, the recognition of an pathology, but this is not always the case. In
autoimmune response does not necessarily most diseases, sorting out the exact role of
indicate immune pathology [2]. As such, auto autoimmunity in generating pathology has
immunity is a phenotypic manifestation of an proven to be challenging. Several criteria
autoimmune disease but not necessarily the have been proposed for this purpose, such
cause. Finally, there are clear examples where as Witebsky’s postulates, which are loosely
an autoimmune response triggers disease; in based on Koch’s postulates. These have
Graves disease, for example, where the auto been used to establish whether a character
antibodies bind the thyroid stimulating hor ized autoantigen and the subsequent auto
mone receptor. In this chapter we overview immune response has a role in generating
the research into the environmental triggers the disease process [2]. These postulates
linked with primary biliary cholangitis (PBC) stipulate that the autoantigen has first to be
and briefly review factors linked with autoim characterized and then should be shown to
mune hepatitis (AIH). generate a similar autoimmune process that
The cardinal features of most autoimmune results in the disease phenotype. This can
diseases include the demonstration of loss of be demonstrated in experimental models by
tolerance to self as well as an increased risk passaging the disease‐specific autoanti
of developing other autoimmune diseases. In bodies, autoreactive T lymphocytes, or the
other words, each disease has an immune and autoantigen itself.
a genetic component. For example, patients These criteria are more easily satisfied
with PBC produce the diagnostic and highly for disorders where autoantibodies clearly
disease‐specific anti‐mitochondrial anti trigger a pathologic state, such as by receptor
bodies (AMAs) and are also predisposed to engagement. Patients with Graves disease,
developing autoimmune thyroid disease, for example, generate autoantibodies to the
Raynaud syndrome, Sjögren syndrome, thyroid‐stimulating hormone receptor and
scleroderma, and other autoimmune disor antibody‐mediated receptor engagement
ders. In contrast, patients with primary then directly promotes the state of hyperthy
sclerosing cholangitis (PSC) have a markedly roidism. In other diseases, however, the role
increased risk of inflammatory bowel disease that autoimmunity plays in generating dis
(IBD); while anti‐neutrophil cytoplasmic ease is less clear. For example, the role that
Chapter 5 Environmental Exposure and Risk in Autoimmune Liver Diseases 85
AMAs play in causing PBC is questionable Accordingly, the immune system can be
because patients with AMA‐negative PBC manipulated into creating organ‐specific dis
have similar disease outcomes as AMA‐ ease. However, this procedure has obvious
positive patients and some AMA‐positive limitations with non‐organ‐specific autoanti
individuals have no disease at all. As a result, gens. For example, the administration of
AMAs appear to fall into the category of PDC‐E2 with adjuvant failed to produce a
pathologically irrelevant but diagnostically PBC‐like disease in rodent models. This
useful antibodies. Furthermore, Witebsky’s then raises the question: how does an autoim
postulates have not been satisfied in PBC for mune response to the PDC‐E2 antigen, which
the characterized mitochondrial antigens, is found in every nucleated cell, cause an
pyruvate dehydrogenase complex (PDC)‐E2, autoimmune disease limited to bile ducts,
and related inner mitochondrial proteins tear ducts and salivary glands [5]?
[3]. Experiments using the passage of AMA,
autoreactive T lymphocytes, or transgenic
Interaction of Genes and Environmental
expression of PDC‐E2 in mice have all failed
Triggers in Autoimmunity
to generate autoimmune biliary disease
(ABD) [2]. Currently, we lack models to demonstrate
It is well known that experimental models reproducibly how any specific environmental
may be manipulated into inducing autoim agent triggers a complex disease on a
mune responses by inoculation of an auto particular genetic background. Koch’s postu
antigen with an adjuvant, such as Freund lates address acute disease and Hill’s criteria
adjuvant containing inactivated mycobacteria chronic disease but both criteria lack tools
emulsified in mineral oil. Stimulants such as for evaluation of the genetic influence [2].
these were referred to as the “immunologist’s Recent genome‐wide association studies
dirty little secret” by the revered immunolo (GWAS) have identified a number of genes
gist Charles Janeway. He noted that “foreign involved in various immune pathways that
antigen alone was insufficient to elicit the contribute to the pathogenesis of autoim
adaptive immune response and that scientists mune liver disease [1]. One common feature
instead had to routinely pepper their experi is the linkage with polymorphisms within the
ments with crude extracts like mineral oil, major histocompatibility complex (MHC) on
mycobacteria, and aluminum hydroxide in the sixth chromosome that contains the
order to get T and B cells to do their bidding” human leukocyte antigen (HLA) class I and
[4]. He questioned why non‐self antigens II regions. HLA polymorphisms have been
alone failed to be recognized and stimulate an closely linked with the development of auto
immune response. His subsequent studies led immune disease, drug hypersensitivity, and
to a long career investigating the need for persistence of viral infection, whereas other
costimulatory molecules and the role of the polymorphisms provide protection against
innate immune system in regulating adaptive developing autoimmune disease, vasculitis,
immune responses. The “secret” was that and clearance of viral infection. However, the
adjuvant was an absolute requirement for role that HLA plays in the development of
breaking tolerance to self to create autoim PBC and PSC remains obscure and fine map
mune responses. Many autoimmune models ping studies are underway to try to identify
have been established using this maneuver. the relevant genetic risk.
For example, autoimmune arthritis models The PBC genomic studies of patients of
have been created by injecting mice with col European descent revealed a strong linkage
lagen and adjuvant. Similarly, experimental with several polymorphisms along the inter
autoimmune allergic encephalomyelitis leukin (IL)‐12 cytokine signaling pathway.
models have been created by inoculation of However, a clinical trial using the IL‐12 inhib
mice with myelin basic p rotein and adjuvant itor ustekinumab failed to show any appre
to produce a multiple sclerosis‐like disorder. ciable biochemical improvement of disease.
Accordingly, how the genetic risk contrib Notably, microbes share many highly
utes to disease remains the subject of conserved proteins with humans and so
intense investigation and debate, with some microbial molecular mimicry is very common.
studies showing that the risk alleles provide In the case of PBC, many bacteria have the
diminished IL‐12 transcription. Looking at highly conserved proteins that share similar
other idiopathic inflammatory disorders, ities with the PDC‐E2 autoantigens linked
the genetic risk of patients with extreme phe with PBC (Table 5.1). In these studies, addi
notypes appears to be linked with immuno tional proof of infection is often lacking or
deficiency and increased risk of infectious unconfirmed. Furthermore, the antigenic
disease. Notably, children with very early association may be obvious, but proof of an
onset IBD are often found to have monogenic autoimmune disease pathogenesis is not.
disease linked with immunodeficiency When the possibility of molecular mimicry is
syndromes (e.g. IL‐10 deficiency, Wiskott– proposed as a potential mechanism for
Aldrich syndrome, immunodysregulation induction of autoimmunity, discussion
polyendocrinopathy enteropathy X‐linked concerning the role of infection itself is
syndrome, and atypical severe combined seldom, if ever, discussed. Another
immunodeficiency to name a few). Indeed, consideration is that viruses use mimicry to
several of the candidate genes associated with hijack some cellular machinery, but then the
PSC and PBC, such as IL2, IL2RA, IL12B, host may then use the very same mechanism
TNFRSF1A, and TYK2, are all linked with to restrict viral infection. In summary, we lack
primary immunodeficiency syndromes [1]. definitive processes to demonstrate that
Similarly, most of the mouse models that shared host and microbial antigenic determi
develop spontaneous autoimmune liver nants trigger an autoimmune response that
disease are severely immune deficient [1,2]. then causes chronic disease, when microbial
Therefore, we have put forward the model pathology and antimicrobial immune
that the genetic predisposition provides a responses are also present.
relative state of increased risk to specific The presentation of cryptic antigens that
infectious agents, permitting establishment are not usually exposed to the immune
of chronic infection that in turn triggers the system has also been linked with auto
autoimmunity, rather than an over‐exuberant immune responses (Figure 5.1b). A good
response to self [5]. example of this process is the demonstration
of increased and aberrant expression of
PDC‐E2 and other oxo‐acid dehydrogenase
Mechanisms for Triggering Autoimmunity
proteins on the cell surface of cholangiocytes,
There are several different models and and salivary glands, and within perihepatic
clinical observations illustrating how an lymph nodes of PBC patients [3]. Variations
environmental agent can trigger an autoim of this process include the modification of
mune response (Figure 5.1), some of which autoantigens such as neoantigens that may
are not mutually exclusive. One such mecha arise in tumors, for example. A second pro
nism, first conceived over 35 years ago, is posed mechanism is the modification of
molecular mimicry, which is thought to occur enzymes and other autoantigens by xenobi
when viral or bacterial epitopes resemble host otics or drugs that have been shown to trigger
proteins and infection triggers an autoim autoimmune responses.
mune response with loss of tolerance to self Epitope spreading may occur as a result of
(Figure 5.1a) [6]. On first sight, the hypothesis bystander activation. This process of “guilt
appears attractive. Indeed, an exhaustive list by association” occurs when a microbial
of autoantibodies that recognize microbial antigen is copresented with a self‐antigen,
antigens and host autoantigens has accumu with subsequent induction of an antimicro
lated throughout the years. However, this bial and an autoimmune response to self
hypothesis has proven very difficult to confirm. (Figure 5.1c) [6]. Bystander activation has
(a) (b)
(c) (d)
Figure 5.1 Models of microbial induction of autoimmune responses to pyruvate dehydrogenase complex
(PDC)‐E2. (a) The antigen‐presenting cell (APC) is processing bacterial PDC‐E2 (red) to CD4+ lymphocytes that
provides the necessary secondary signaling to create CD8+ and B‐cell immune responses to the microbial
PDC-2 antigen. By the mechanism of molecular mimicry, the T cells and anti‐mitochondrial antibody (AMA)
recognize and break tolerance to self PDC‐E2. (b) A leaky mitochondrial syndrome occurs as a result of
infection or cell damage and the mitochondrial PDC‐E2 antigens (red) usually sequestered in the inner
mitochondrial membrane become exposed or possibly modified as a cryptic antigen to trigger an autoimmune
response. (c) As retroviruses incorporate cellular proteins (red) when budding from the cell surface, APC
present both self (red) and viral (gray) proteins to the adaptive immune system. Bystander activation occurs
when CD4+ lymphocytes reacting to viral proteins provide secondary signals to B cells to produce AMA and to
CD8+ lymphocytes to enable epitope spreading to PDC‐E2. (d) The betaretrovirus‐infected B cell expresses a
superantigen (gray) that binds the class II MHC to the T‐cell receptor outside of the antigenic grove and
activates T lymphocytes of different antigenic specificity. This potent immune‐stimulating mechanism is
hypothesized to trigger an autoimmune response [6].
Table 5.1 Bacteria linked with primary biliary cholangitis (PBC) and the development of anti‐mitochondrial antibody (AMA) through the mechanism of
molecular mimicry.
Bacterial protein ID
(percent of PBC patients AMA reactivity (antigenic
with serologic reactivity to similarity to amino acids Pathogen detection (PBC
Bacteria linked with PBC bacterial protein) 150–173 of human PDC‐E2) vs. other liver disease) Clinical relevance
Chlamydophila 100% vs. 8.5% in liver Data not reproduced

pneumoniae 91% by serology
Escherichia coli ClpX (33%) Yes and cellular immunity 100% vs. 25% in feces Associated with UTI
PDC‐E2 (100%) (73–93%)
Haemophilus influenzae TRMA (NA) NA (53%)
Helicobacter pylori, DNA PCR detected in Data not reproduced
Helicobacter hepaticus liver
Lactobacillus delbrueckii BGAL (~50%) Yes (67%) Non‐infectious Food product
Mycobacterium gordonae Hsp65 (100%) PDC‐E2 Yes (47%) Hsp65 serology 100% PBC Hsp65 serology in 47% Spanish
(avium‐intracellulare, (AMA detected in patients DNA 82% PBC liver vs. 4% British PBC patients
kansasii, leprae, with Mycobacterium DNA not reproducibly detected
tuberculosis, xenopi) infection without PBC) in liver but found in controls
Novosphingobium PDC‐E2 (100%) Yes (40–92%) Serum reactivity to PDC‐
aromaticivorans E2 25% vs. 25% in feces
Pseudomonas aeruginosa DCDA (4%) NA (60%) (CD8 epitope
cross‐reactive with PDC‐E)
Salmonella typhimurium ClpX.S (47%)a NA (93%)
Yersinia enterocolitica ClpX.E (47%)a NA (87%)
a
PBC patients with serologic reactivity to one of the Clp peptides.
been clearly documented in both infectious lymphopenia‐induced autoimmunity is found

disease and in experimental models where in patients with HIV infection. When the
mice inoculated with an autoantigen and CD4+ lymphocyte counts drop below 100–
adjuvant develop immune responses to 200/μl, a spectrum of autoantibodies may
unrelated epitopes. When no environmental emerge, such as anti‐small nuclear ribonucleo
agent can be identified, some investigators proteins in approximately 65% of patients and
have evoked a “hit and run” hypothesis. This anti‐phospholipid antibodies in over 50% of
process is thought to occur when a path patients. Furthermore, as HIV incorporates
ogen has been cleared but has generated a cell surface proteins (e.g. HLA, TCR and
chronic autoimmune response. While epi immunoglobulins) when budding from cells, a
tope spreading is commonly observed, the process similar to bystander activation may
hit‐and‐run process has not. No examples occur as patients start to generate antibodies
have been reported of microbial infection to components incorporated into the virion
leading to epitope spreading and the estab (Figure 5.1c). In support of the hypothesis that
lishment of autoimmune disease after the autoimmune diseases are linked with states of
pathogen has departed. immunodeficiency, patients with HIV are
Microbial superantigens have been pro reported to develop a range of immune‐related
posed to trigger autoimmunity by the acti disorders, including immune complex disease,
vation of CD4+ lymphocytes. In classical vasculitides, sarcoidosis, autoimmune cytope
T‐cell responses, approximately one in a nias as well as organ‐specific autoimmune syn
million lymphocytes may recognize a dromes and rheumatic disorders such as
specific MHC–peptide complex, whereas systemic lupus erythematosus (SLE), Sjögren
superantigens may stimulate up to 30% of syndrome and rheumatoid arthritis. Also, the
the T‐lymphocyte population. Following mechanisms of lymphopenia‐induced autoim
infection of B lymphocytes, for example, the munity have been partially characterized.
mouse mammary tumor virus (MMTV) They involve lymphopenia‐induced prolifera
expresses a viral superantigen that binds the tion, in part stimulated by HLA–TCR engage
MHC class II molecule to the lateral side of ment with IL‐7 production, in the absence of
the v‐β domain of the T‐cell receptor (TCR), the regulatory cytokine transforming growth
outside of the antigenic grove (Figure 5.1d). factor (TGF)‐β.
The process stimulates CD4+ lymphocytes
to proliferate and produce cytokines that
induce B cells to replicate as well, thus per Primary Biliary Cholangitis
mitting viral replication within lympho
cytes. Accordingly, superantigen activity PBC is an idiopathic biliary disease character
induces a massive activation of T cells of ized by non‐suppurative cholangitis with
unknown antigen specificity and therefore granulomatous destruction of 30–80 μm
may stimulate autoimmune responses. interlobular bile ducts. The diagnosis is made
However, this hypothesis has not been vali by documentation of cholestatic liver tests,
dated experimentally to cause autoimmunity. serum AMA and by liver biopsy in AMA‐
In contrast, lymphopenia is sufficient to negative patients. The worldwide prevalence
induce autoimmunity. Lymphopenic states are of PBC ranges from 1 in 2000 to 1 in 50 000.
strongly associated with the development of There is a marked female predominance, with
organ‐specific and systemic autoimmune dis an estimated ratio of 9 : 1 with men. Up to 1 in
eases. This process may be linked with supe 700 middle‐aged women develop PBC,
rantigen stimulation because the lymphocytes accounting for 5% of patients requiring liver
subsequently undergo apoptosis following transplantation in North America. Notably,
superantigen activation without secondary presentation at a younger age is associated
survival signals. An excellent example of with a worse prognosis as well an increased
risk of developing recurrent disease following PBC is an autoimmune disease because

liver transplantation. Recurrent PBC is diag patients generate humoral and cellular auto
nosed histologically following liver transplan immune responses to the mitochondrial
tation and recent reports suggest that this protein complexes located on the inner mito
may be observed in up to 50% of patients. chondrial membrane. A range of 80–95% of
Notably, patients developing cholestatic liver PBC patients have been reported to generate
tests soon after liver transplantation have a AMAs. These autoantibodies are highly
higher risk of recurrent disease, often diag specific for PBC, but AMAs may also be
nosed several years later. These biochemistry observed in other autoimmune diseases such
dynamics are consistent with an infectious as AIH and SLE, as well as a proportion of
disease, as an autoimmune process would be patients with hematologic malignancy. Both
less likely in the setting of higher levels of the humoral and cellular immune responses
immunosuppression used in the first year fol are directed toward PDC‐E2, PDC‐E3
lowing transplantation [7]. Indeed, the very binding protein, and other lipoylated oxo‐
fact that immunosuppression is of little utility acid proteins. In healthy individuals, PDC‐E2
in the management of PBC prior to transplan is sequestered away from the immune system
tation questions whether PBC has an autoim in mitochondria but in PBC patients it is
mune etiology/pathogenesis in the first place. aberrantly expressed on the cell surface, as
PBC is considered an archetypal autoim observed in cultured cholangiocytes in vitro
mune disease because of the marked female as well as biliary epithelium and lymph node
predominance, high prevalence of autoanti monocytes in vivo [3]. In fact, the PDC‐E2‐
bodies, and frequent concurrence with other like protein is also found in patients with
autoimmune disorders. Patients often have a disease recurrence following liver transplan
history of other autoimmune disorders, such tation, on the salivary gland epithelium as
as thyroid disease, Raynaud syndrome, SLE, well as biliary epithelium of PBC patients
and Sjögren syndrome and an increased without AMA. This finding is unique and
prevalence of PBC and related autoimmune considered a phenotypic manifestation of
disorders in first‐degree relatives [8]. Current PBC. Indeed, it is likely that the development
and prior use of hormone replacement of this mitochondrial phenotype is central to
therapy is more prevalent in PBC patients the disease process and it is commonly
and younger age of first pregnancy, impli thought that the appearance of PDC‐E2 on
cating a role for female hormones in the etio the cell surface breaks immune tolerance to
pathogenesis of disease. Cigarette smoking is trigger AMA production (Figure 5.1b).
also related to the development of PBC. Based on this aberrant expression of
Notably, patients have increased suscepti mitochondrial antigens, a good argument

bility to specific infections and cancers, sug can be made that whatever triggers cell
gesting the disorder is linked with a degree surface PDC‐E2 expression likely triggers
of immunodeficiency. Epidemiologic studies the inflammatory biliary disease as well. To
report that patients have a higher preva date bacteria, viruses and xenobiotics have
lence of tonsillectomy, vaginal infection, and been proposed as the main environmental
urinary tract infection (UTI) [5]. Furthermore, agents linked with PBC (Table 5.2).
a meta‐analysis of development of cancer in Throughout the years, there has been debate
PBC patients revealed an increased preva concerning the nature of the AMA‐reactive
lence of melanoma and non‐melanoma skin antigens on the cell surface. Some of the early
cancers, lymphomas, and thyroid, cervical, hypotheses suggested the presence of a
esophageal, colorectal, kidney and breast microbial mimic or IgA complexes contain
cancers [5]. These data suggest a level of ing bacterial antigens; no such pathogen has
immunocompromise in PBC patients’ ability been identified, however. Others have sug
to limit infectious disease and cancer. gested that modified oxo‐acid dehydrogenase
Table 5.2 Comparison of environmental factors linked with PBC.
Disease Sustained clinical

Ability to corrected by Detected with Triggers with improvement
Serologic Serologic Cellular trigger antimicrobial Detected in Detected in mitochondrial mitochondrial with antimicrobial
reactivity reactivity immune disease in therapy in PBC biliary PBC lymph phenotype phenotype in therapy in PBC
agent to AMA response animal model animal model epithelium node in vivo vitro patients
Bacteria ✓ ✓ ✓ ✓ ✓ ✗ ✗ ✗ ✗ ✗
Xenobiotics NA Bound to ✗ ✓ NA ✗ ✗ ✗ ✗ ✗
hapten
b a
Betaretrovirus 15% to ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
HBRV Env
a
Electron microscopy, proviral integrations, RNA hybridization, in situ hybridization.
b
Lytvyak et al. [9].
proteins are relocated to the cell surface as a the new risk factors associated with the region
result of apoptosis or xenobiotic modifica they move to. For example, British migrants
tion. However, biliary apoptosis appears to to Australia experience a three to five times
be limited to immune attack, whereas AMA‐ decrease in the risk of developing PBC com
positive biliary epithelium are observed in pared to people in their home country. The
otherwise normally appearing cholangio opposite trend is observed in migrants
cytes without any evidence of apoptosis [3]. from India moving to the UK, where PBC is
Furthermore, xenobiotic modification of far more common. Given these findings, it
PDC‐E2 has been linked with the production reasonable to suspect that exposure to a
of AMA in animal models but not been common environmental factor may be cause
shown to promote cholangiocyte PDC‐E2 of disease.
expression. The most parsimonious explana A substantial contribution to the hypo
tion for cell surface AMA reactivity is that thesis of an environmental trigger for PBC
unmodified oxo‐acid proteins are actually came from a study by David Triger in 1980
aberrantly expressed from leaky mitochon suggesting a waterborne risk factor [8]. The
dria [5]. report noted that Sheffield in northern
England had a particularly high PBC preva
lence of 54 per million, and that PBC clus
Geo‐epidemiology, Clusters,
tered within specific districts throughout the
and Case–Control Studies of PBC
city. While originally thought to be associ
In support of the influence of environmental ated with the coal mining and steel industry,
factors in the pathogenesis of PBC, the dis the source of water supply was noticed to be
ease has been reported to occur in non‐ a common factor linked to the development
related family members and caregivers. of PBC after ruling other potential environ
Furthermore, the genetic input to disease is mental factors. Indeed, individuals receiving
estimated to be less than 25%, implicating a water from the Rivelin reservoir experienced
role for external agents [1]. In agreement a 10‐fold increased risk of PBC. Water quality
with this idea, epidemiologic reports, cluster tests revealed no significant findings other
analyses and case–control studies demon than having soft water and lower concentra
strate geographic differences in disease prev tions of fluoride. Despite the study failing
alence rates, the formation of PBC clusters to narrow down a conclusive environmental
within small regional areas, and potential trigger, the association of PBC clustering
lifestyle variables contributing to PBC [8]. with a water supply system resulted in suc
Foremost, prevalence rates of PBC vary bet cessive studies investigating PBC clusters
ween national cohort studies. Of the coun and potential water‐related risk factors [8].
tries where studies have been conducted, the Clustering of PBC has also been discovered
incidence of PBC is higher in the USA with around areas of former coal mining sites in
400 per million in 2000, and in China with the UK and toxic waste sites in the USA. A
492 cases per million in 2010. In comparison, study analyzing the spatial variation of PBC
27–54 cases per million have been reported around the old coal mining city of Newcastle
in Japan, with the lowest rate of 19 per mil found three definite clusters, where the
lion reported in Australia. In Europe, there is spatial variation could not be accounted for
a greater incidence of PBC in northern by referral bias, familial clustering, or water
regions, with a prevalence of 200 cases per supply. A higher prevalence of men with PBC
million in southern Wales and 240–251 per was seen within each cluster, where coal
million in northern England [8]. mining was a predominant occupation.

Also of interest, observations from inde Other studies also observed geographic clus
pendent migration studies reveal a correla tering of PBC in northeast of England with
tion between the risk of PBC in migrants and evident space–time grouping indicative of
the potential for a transient environmental into a bacterial pathogenesis of PBC has
agent. Similar studies within the USA have yet to emerge.
also reported increased prevalence of PBC in The concept of molecular mimicry has
close proximity to Superfund toxic waste been the most popular mechanism linking
sites [8]. Two of these groups were found to the introduction of foreign microbial proteins
be proximal to sites contaminated with to a loss of tolerance to self‐proteins in PBC
high concentrations of halogenated aromatic (Table 5.1). This hypothesis suggests that
hydrocarbons. microbes presenting epitopes homologous to
While chemical involvement in the patho mitochondrial proteins can break down
genesis of PBC may be inferred from these immune tolerance to PDC‐E2 in the biliary
studies, it is important to note that hidden epithelium through induction of cross‐reac
confounding factors such as socioeconomic tive autoantibodies and effector T cells
status may also be an explanation. Different (Figure 5.1a). Damage of infected bile ducts
risk factors have been uncovered in epidemi by initial T‐cell responses are then believed to
ologic studies from other countries. For release intact PDC‐E2 and mitochondrial
example, a higher risk of PBC was reported unrelated autoantigens, stimulating the initial
among atomic bomb survivors in Japan. This autoimmune attack on bile duct tissue
observation suggests that either radiation (Figure 5.2a). Activation of antigen‐present
may be a trigger for PBC or that patients with ing cells and other facets of the innate immune
depressed immunity had a lower threshold system are thought to propagate the autoim
for infection. The recent Canada‐wide study mune process, where processing of microbial
documenting a higher prevalence of PBC in mimics, PDC‐E2 and other autoantigens may
the eastern provinces is also in keeping with increase cytokine production and expansion
a relationship with PBC and lower socioeco of autoreactive cells [11].
nomic status. Escherichia coli has been studied as a can
didate agent for PBC because of frequent
reports of UTI in PBC patients. There have
Bacterial Infection and PBC
been multiple reports of increased incidence
Supported mainly by frequent concurrence of recurrent UTI in PBC patients, and E. coli
of UTI in PBC patients and experimental has been isolated in many cases. In one study,
AMA and T‐lymphocyte immune responses E. coli was detected in the stool of all PBC
to PDC‐E2‐related microbial proteins, the patients, though also found present in healthy
role of bacterial infection as an environ controls and 25% of patients with other
mental trigger of PBC has been explored for hepatic diseases [10]. Most studies proposed
several decades [10]. For reasons unknown, the mechanism of molecular mimicry to
lipopolysaccharide (LPS) found on the outer be responsible for production of AMA and
membrane of Gram‐negative bacteria has lymphocyte autoreactivity to PDC‐E2, fol
been reported to induce portal lymphocyte lowing several reports of PBC sera cross‐
infiltration and cholangiocyte degeneration reacting to human and E. coli PDC‐E2 [10].
in mice models [11]. Similarly, lipoteichoic Though the affinity of human AMAs to E.
acid from the outer membrane of Gram‐ coli PDC‐E2 is shown to be 100 times higher
positive bacteria has also been found around in PBC patients, a low titer of AMA has also
damaged bile ducts of PBC patients with been observed in women with recurrent
higher IgA titers in sera [10]. Bacterial DNA UTI, independent of a liver disease diag
with unmethylated CpG motifs has also nosis. Notably, very few E. coli proteins unre
been reported to trigger PDC‐E2‐specific lated to human PDC‐E2 are recognized
T‐helper 1 immune response in PDC‐E2 by PBC sera. The cross‐reactive PDC‐E2
immunized mice [11]. Despite the tanta epitope is also recognized by the cellular

lizing clues, further mechanistic insights immune system, suggesting that E. coli and
(a)
(b)
Figure 5.2 Model for loss of tolerance to PDC‐E2 and PBC pathogenesis. (a) In a two‐hit process: (i) biliary
epithelial cells undergo apoptosis or other modifications, with xenobiotics for example, and present intact
PDC‐E2 on the biliary epithelium; and (ii) bacterial infection results in PDC‐E2 presentation on antigen‐
presenting cells (APC), which breaks tolerance to human PDC‐E2 (arrows represent T‐cell help). The subsequent
autoimmune response homes in on the aberrant expression of PDC‐E2 in biliary epithelium (? represents factors
that may precipitate cell surface PDC‐E2 in (a)). (b) In a single‐hit process, the mouse mammary tumor virus
(MMTV)‐like human betaretrovirus infects biliary epithelium leading to PDC‐E2 expression on the cell surface.
Then the virus either incorporates PDC‐E2 while budding from the cell surface or exits with PDC‐E2 in exosomes
(not shown). APCs then present PDC‐E2 and viral proteins, resulting in a bystander immune response to PDC‐E2
and an immune response to viral and self‐proteins expressed on biliary epithelium. PBC, primary biliary
cirrhosis; PDC, pyruvate dehydrogenase complex. Source: adapted from Wasilenko et al. [12] with permission.
human PDC‐E2 cross‐react at the CD4 and and 2‐oxoglutarate dehydrogenase complex
CD8 T‐cell level. As an example, clones of (OGDC)‐E2 peptides, while T‐cell clones
human PDC‐E2‐reactive T cells were also specific to E. coli OGDC‐E2 responded to
found to be responsive to E. coli PDC‐E2 human PDC‐E2 autoantigens as well [11].
Similar interest has been provoked by initiate bile duct injury if not specifically
mycobacterial species. The Mycobacterium observed in PBC patients nor found in the
gordonae heat shock protein 65 (Hsp65) was liver. Thus, the autoimmune damage may
noted to share similarity with human PDC‐ presumably be represented by a two‐hit
E2, and PBC patient AMAs reacted to the model (Figure 5.2a) incorporating a “hit‐and‐
protein. In small series, 100% of PBC run” process as the bacteria cannot be
patients were found to have serologic reac detected in the majority of PBC patients.
tivity to the M. gordonae Hsp65, while in Notably in the NOD.1011 mouse model, N.
other reports both DNA and proteins from aromaticivorans inoculation induced AMA
M. gordonae were detected in the livers of a production and IgG and IgA antibodies
proportion of PBC patients (Table 5.1). against bacterial and mammalian PDC‐E2
However, studies in the UK only observed [11]. Following autoimmune T‐cell responses
antibody reactivity to Hsp65 in 4% of PBC against small bile ducts, the mice appear to
patients and other follow‐up studies failed develop bile duct lesions resembling PBC.
to confirm the clinical association of PBC The destruction of small bile ducts is seen
with M. gordonae. Of note, similar to obser even in transfer experiments of monocytes
vations with E. coli, approximately 40% of from infected to control mice, and antibiotic
patients with Mycobacterium tuberculosis treatments prevented progression of the dis
were found to have low‐level AMAs without ease. While these studies firmly link bacterial
any additional evidence of having PBC. infection with ABD, it should be noted that
To date, Novosphingobium aromaticivorans the non‐obese diabetic (NOD)‐derived
has been heavily promoted as the most likely mouse models harbor MMTV. This is impor
bacterial candidate [11]. This aerobic Gram‐ tant because viral expression is linked with
negative bacterium has the ability to metabo the mitochondrial phenotype of PBC, the
lize halogenated compounds and estrogens. spontaneous production of AMA as well as
Therefore, it has been hypothesized that this the development of cholangitis. As this
bacterial infection can cause aberrant expres betaretrovirus is linked to the development
sion of PDC‐E2, possibly aided by xenobiotic of PBC, it is difficult to make firm conclu
modification. The metabolism of estrogens sions about the role of other candidate agents
provides a linked with female preponderance using the NOD‐derived mouse models.
in PBC. As a molecular mimic, two proteins The long and extensive search for a bacte
have been discovered to have a high degree of rial mimic as a trigger for PBC has been
homology with the human PDC‐E2 epitope, questioned [12]. While molecular mimicry is
reacting with the sera of all AMA‐positive an attractive hypothesis linking involvement
and some AMA‐negative PBC patients. It has of bacteria to PBC, the mechanism remains a
also been reported that first‐degree relatives theoretical model. Indeed, there has yet to be
of PBC patients in Iceland harbor antibodies concrete evidence showing that human auto
against N. aromaticivorans which reacted immune disorders in general can be triggered
with mitochondrial autoantigens in PBC by molecular mimicry [6]. In fact, only her
patients as well [11]. In comparison with pesvirus‐induced stromal keratitis in a
other candidates, serum reactivity to N. aro- mouse model has been able to demonstrate a
maticivorans was found to be 100–1000 times definitive role for the induction of disease
higher compared with E. coli which was though molecular mimicry. Furthermore, in
already shown to have high affinity to AMAs. that specific model, viral replication was still
However, N. aromaticivorans was not really needed to perpetuate the disease [12].
detected in PBC patients and molecular sur One key criticism brought up previously
veys only observed bacterial DNA in 25% of regarding the use of a complex molecular
both PBC and control fecal samples. mimicry model in PBC is its requirement for
Accordingly, it is unclear how the agent could the combination of two unrelated hits to
cause PBC. First the bacterium must provoke PBC. As Koch’s postulates relate to linking
the mitochondrial phenotype of PBC with infection with an acute disease process
expression of PDC‐E2 on surface of bile without genetic predisposition, other criteria,
ducts and then it needs to trigger loss of such as Bradford Hill, can be employed to pro
immune tolerance to PDC‐E2 autoantigens vide a causal relationship for infection and dis
(Figure 5.2a) [12]. Notably, no bacterial infec ease. These criteria address the strength,
tion has been found to induce the mitochon consistency and specificity of the relationship,
drial phenotype, central to the development which most bacterial infections fail to meet
of PBC, or has been convincingly detected in for PBC [2]. The plausibility of an agent trig
PBC biliary epithelium. While the model gering disease and the demonstration of a
covers how immune tolerance may be lost, it biological gradient are also factored in, which
does not provide a definitive cause for the again are lacking for the bacterial hypothesis
development of increased extracellular of PBC [2]. Experimental data are also encom
expression of PDC‐E2 in bile ducts. This passed by the Bradford Hill criteria as well as
model also separates the site of infection the coherence of the bacterial hypothesis,
from the onset of autoimmunity in the liver. which is also incoherent at best (Figure 5.2a)
The second issue is that the data linking [12]. Analogy with similar factors can also be
bacterial pathogens with PBC are circum considered for these criteria, but no bacteria
stantial and heavily reliant on positive have been shown to trigger organ‐specific dis
serology to hypothesized bacterial mimics. ease with autoimmune responses to common
Numerous strains of bacteria have been proteins found within all nucleated cells. As a
investigated as potential infectious agents in result, there are very sparse data to support a
PBC (Table 5.1) with little convincing evi bacterial etiology of PBC (Table 5.2).
dence to link infection with disease. It is true
that many of these bacterial candidates dem
Xenobiotics and PBC
onstrate molecular mimicry to the mitochon
drial autoantigen, and cross‐reactivity with The involvement of xenobiotics in the patho
PDC‐E2 antibodies and/or initiation of T‐cell genesis of PBC is supported by epidemiologic
response. Even parasites such as trypano clustering studies around coal mines and
somes and Ascaridia galli have been linked to toxic waste sites as well as by animal models
PBC by molecular mimicry and antibody modified by specific chemical compounds to
cross‐reactivity to human PDC‐E2. Indeed, elicit the PBC‐specific AMA response [10,11].
the detection of low‐level AMAs in patients Xenobiotics are small foreign compounds
with chronic M. tuberculosis and E. coli infec that can bind and modify self‐proteins and
tion without any histologic or biochemical change their molecular structure sufficiently
evidence of PBC argues against the likelihood to induce an autoimmune response. Notably,
of bacteria (or AMA) inducing PBC. xenobiotics can be found in common
These data lead one to question the reli household detergents, preservatives and
ability of determining the causation of PBC pharmaceuticals and they are mainly metab
based on this serologic activity alone, as olized in the liver. One hypothesis suggests
there are evidently many infectious pathogens that xenobiotics induce autoimmunity in
that fit this criterion. Modifications of Koch’s PBC through a direct toxic effect causing cell
postulates in vitro or in animal models are death through apoptosis or necrosis.
required to demonstrate that a specific Autoantigens will then be found concen
microbe can trigger the mitochondrial pheno trated around the surface of apoptotic cells,
type of PBC, for example. Then the infectious which have been found to generate
agent should be directly linked with the autoantibody responses in systemic autoim
development of cholangitis and shown to munity (Figure 5.2a). Experimental models
reproduce the mitochondrial phenotype of have focused on xenobiotic modification of
the PDC‐E2 complex to alter the molecular protein to trigger loss of tolerance. Notably,
structure and trigger loss of self‐tolerance; T xenobiotics have not been found in PBC
cells and B cells primed against these neoan cholangiocytes nor shown to reproduce the
tigens are then hypothesized to cross‐react mitochondrial phenotype specific for PBC.
with mitochondrial autoepitopes [11]. Current support for the proposed effect of
In support of this hypothesis, a halogenated xenobiotics in PBC has mainly been indirect
organic compound attached to the mitochon evidence based on AMA binding to hapten‐
drial epitope was shown to react with PBC bound xenobiotic or chemically modified
sera at a higher affinity than the native form PDC‐E2. Mouse models treated with xenobi
of the antigen in vitro [11]. Additionally, otics demonstrate a degree of liver disease,
patients repeatedly given halothane‐derived but it is unclear whether this is a result of
anesthesia have been shown to develop anti autoimmune damage or merely chemically
bodies cross‐reacting with lipoylated PDC‐ mediated toxic damage to the liver. The latter
E2. Lipoic acid, a critical component for the is an important consideration as some
reactivity of PDC‐E2 complex, is found on the models also develop peritonitis. Future work
exterior of the PDC‐E2 complex, increasing is still required to determine whether the
the susceptibility for chemical modification. identified halogenated compounds that elicit
Accordingly, it has been hypothesized that AMA and liver lesions in mice are truly
organic compounds with similar molecular linked with the development of PBC in any
structure to lipoic acid could independently shape or form (Table 5.2).
trigger an AMA response. For example, 6‐
bromohexonate, a halogenated xenobiotic
Viruses in PBC
complex, was found to induce AMA produc
tion without the adjunction of the peptide It is not intuitive that a retrovirus should
backbone of PDC‐E2 in a rabbit model [11]. trigger a gramulomatous hepatobiliary dis
However, the model failed to produce liver ease but, as discussed, retroviruses including
lesions and the phenotype dissipated once the HIV have been closely linked with autoim
stimulus was removed. However, a more munity and autoimmune diseases in gen
sustained study using guinea pigs reported eral. Viral discovery studies started with
PBC‐like liver lesions. electron microscopy detection of virus‐like
Animal models of PBC have been designed particles in isolated cholangiocytes from
to evaluate the potential role of xenobiotics. PBC patients as well as serologic studies
One such study reported the development of showing antibody reactivity to retroviruses.
a PBC‐like disease in the NOD.1101 and Then, a human betaretrovirus (HBRV)
C57BL/6 models immunized with 2‐octynoic closely related to MMTV was cloned from a
acid that lost tolerance to PDC‐E2 but did PBC biliary epithelium library using primers
not develop progressive disease. Another capable of amplifying any retroviral pol gene
group also tested the effect of 2‐nonynoic sequence. The full‐length proviral genome
acid, and this compound not only shows a was then cloned from a PBC patient’s peri
high affinity to AMAs in PBC sera but also hepatic lymph node [3].
induces cholangiopathy in treated mice. This The female biology of betaretrovirus infec
artificial compound is frequently used in sev tion is of interest. Betaretroviral replication is
eral cosmetic products and nail polish, which turned on by female hormones during preg
suggests this lifestyle factor may be a reason nancy and the virus is produced in large
for observed female predominance. amounts in breast milk. The reliance for
Similar to the bacterial model for PBC, the female hormones on viral production may be
xenobiotic hypothesis relies on external com intimately linked with the female predisposi
pounds molecularly mimicking or altering tion to PBC. Neonatal mice become tolerant
the highly conserved PDC‐E2 mitochondrial to infection when infected at birth from
infected breast milk and make poor antibody infection (i.e. demonstration of viral integra
responses to MMTV. Initially, B cells become tions into host DNA) showed that the
infected within the gut‐associated lymphoid majority of patients had infection in cholan
tissue and the infected cells produce a supe giocytes. In situ hybridization and RNA
rantigen that stimulates lymphocyte repli measurement assays also clearly showed
cation (Figure 5.2d), which enables MMTV active viral infection in biliary epithelium as
infection as betaretroviruses can only repli well, providing the necessary proof of infec
cate in dividing cells. In mice and humans, tion at the site of disease. In follow‐up
betaretroviruses are cell associated, mainly studies, HBRV was isolated by coculture of
detected in lymphoid tissue, expressed at a PBC lymph nodes with HS578T cells, a
low level, and difficult to detect. human breast cancer cell line.
Because of difficulty with viral detection, One of the most interesting facets linking
the early studies in PBC patients were con betaretrovirus with PBC is that viral infection
sidered somewhat controversial. While is intimately linked with the mitochondrial
HBRV was detected in most perihepatic phenotype [3]. In the first instance, HBRV
lymph nodes, only one‐third of PBC patients and MMTV infection have been shown to
were found to have evidence of viral infec trigger the mitochondrial phenotype and
tion within the liver. Follow‐up studies using have addressed Koch’s postulates in an in
the gold standard for detecting retroviral vitro model (Figure 5.3). In these studies,
PBC
lymph node
γ irradiation
MMTV
Control
lymph node
Control
viruses
Figure 5.3 Primary biliary epithelial cells (BEC) develop the mitochondrial phenotype when incubated with
lymph node homogenates from PBC patients, conditioned supernatants and MMTV. Isolated perihepatic
lymph nodes from PBC and control patients were homogenized and cocultured with normal BEC isolated
from liver transplant recipients. Subsequently, the PBC lymph node‐conditioned BEC demonstrated
anti‐mitochondrial antibody (AMA) reactivity as observed by immunoelectron microscopy and
immunofluorescence, with evidence of human betaretrovirus (HBRV) p27 capsid proteins by
immunofluorescence. Conditioned supernatants had the same effect, whereby only the PBC‐conditioned
supernatants triggered AMA reactivity. This process was abrogated by gamma‐irradiation. Normal BEC
incubated with supernatant from mouse mammary tumor virus (MMTV)‐producing MM5MT cells also
showed AMA reactivity and anti‐p27 capsid immunofluorescence, whereas control viruses had no such
effect. Source: Xu et al. [3]. Copyright 2003 National Academy of Sciences.
lymph node homogenates from PBC patients Western blot and immunoprecipitation
were shown to trigger the appearance of cell studies suggest that betaretroviruses take
surface PDC‐E2 in cultured biliary epithelial up the mitochondrial autoantigens while
cells. The conditioned supernatants could budding from the cell surface, which can
also induce the mitochondrial phenotype and trigger an autoimmune response by epitope
this property could be abrogated by gamma‐ spreading and bystander activation. In
irradiation. Subsequently, the HBRV was addition, betaretroviruses express a
characterized within the conditioned media superantigen that non‐specifically activates
and, as the final experiment, pure isolates of lymphocytes in an unrestricted pattern.
MMTV were shown to trigger the mitochon Notably, a search of the betaretrovirus
drial phenotype in cultured biliary epithelial genome reveals no antigens resembling the
cells (Figure 5.3). Although reported in one mitochondrial oxo‐acid dehydrogenase or
publication, these findings have been repro other known autoantigens associated with
duced in three separate laboratories. PBC. Rather, the autoantigens may be intri
Betaretroviruses have also been linked cately associated with the viral biology, such
with the mitochondrial phenotype in vivo the nuclear pore protein gp210 that may
[2,3]. For example, perihepatic lymph nodes guide a pre‐integration complex into the
from PBC patients have been shown to har nucleus. Accordingly, a different model for
bor monocytes with evidence of HBRV loss of tolerance to PDC‐E2 has been pro
infection and PDC‐E2 expression posed for betaretrovirus infection. HBRV
(Figure 5.4a,b). Similarly, MMTV infection triggers cell surface PDC‐E2 exposure on the
has been linked with the mitochondrial surface of infected cholangiocytes, which is
phenotype in all the spontaneous mouse incorporated into the virion particle that
models of ABD tested (Figure 5.4c–j) [2]. may then trigger AMA production and loss
One of the mouse models, NOD.c3c4, was of tolerance to PDC‐E2 (Figure 5.2b). In this
instrumental for demonstrating a central example, the pathogenesis of PBC no longer
role of MMTV in the development of ABD revolves around autoimmunity, but rather
because the mice had clear evidence of centers on the role of betaretroviral infec
MMTV cholangitis in bile ducts demon tion that masks itself by manipulating
strating the mitochondrial phenotype and the host’s immune system. Indeed, a recent
the disease could be abrogated using report of cellular immune responses
combination antiretroviral therapy. Indeed, involving intrahepatic lymphocytes revealed
a recent clinical trial using the combination that PBC patients harbor a 100 fold excess of
of emtricitabine, tenofovir, and lopinavir lymphocytes reacting with HBRV peptides
boosted with ritonavir demonstrated the versus the characterized immunodominant
utility of this approach for producing sus PDC-E2 epitope.
tained biochemical and histologic responses
for patients who could tolerate therapy.
However, the side‐effect profile of the protease Autoimmune Hepatitis
inhibitors was far in excess of that reported
for the treatment of HIV and enrollment AIH is an idiopathic liver disease character
was terminated prematurely. ized by specific autoantibodies, elevated
Betaretroviruses appear to be capable of IgG levels, histologic evidence of hepatitis,
inducing autoimmunity by several mecha and subsequent response to corticosteroid
nisms (Figure 5.1b–d). First, betaretrovirus treatment following diagnosis. Usually AIH
infection induces changes in metabolism can be diagnosed once other causes of liver
and mitochondrial biogenesis leading disease have been excluded, such as viral
to unveiling of cryptic epitopes usually hepatitis. However, case reports have linked
sequestered within mitochondria. Secondly, AIH with a spectrum of hepatotropic viruses
(a) (b)
(c) (d)
(e) (f) (g)
(h) (i) (j)
Figure 5.4 Immunochemistry studies show betaretrovirus proteins in the same tissue distribution as anti‐
mitochondrial antibody (AMA) reactivity. (a) In the PBC lymph node, green anti‐p27CA signal is observed in the
cytoplasm with aberrant expression of PDC‐E2 plasma membrane localization of mitochondrial antigen with
AMA (red) but only mitochondrial staining can be seen in the (b) PSC lymph node (magnification ×400). (c)
Anti‐mouse mammary tumor virus (MMTV) p27CA reactivity was observed in mononuclear cells in the spleen
of the AMA‐producing interleukin‐2 receptor alpha deficient mouse with (d) AMA reactivity observed in a
similar distribution (magnification ×200, with ×600 in insets). (e) Anti‐MMTV gp52SU, (f ) anti‐MMTV p27CA,
and (g) AMA are all observed in an apical distribution in NOD.c3c4 bile ducts (arrowed) but not in veins (*).
In the healthy C57Bl negative control, only background reactivity is seen with each antibody (h, i, and j)
(magnification ×600). PBC, primary biliary cholangitis; PSC, primary sclerosing cholangitis. Source: (a,b)
Xu et al. [3]. Copyright 2003 National Academy of Sciences.
including Epstein–Barr virus, cytomegalo PBC‐specific phenotype. It is also notable

virus, hepatitis A virus, hepatitis E virus, and that a proportion of patients with AIH
even HIV infection. On occasion, AIH has (without the PBC overlap syndrome) have
been linked with chronic viral infection evidence of HBRV integration in their biliary
secondary to relapsing hepatitis A or hepatitis epithelium. While these data are provocative,
E virus infection. Indeed, some serologic further studies will be required to link HBRV
surveys have identified an increased preva with the development of AIH.
lence of antibodies to hepatitis E virus.
Patients with giant cell hepatitis may present
with features of AIH and respond to cortico Prospectus
steroid therapy. Notably, this condition has
been linked with human herpesvirus 6 infec HBRV is associated with many of the patho
tion. In fact, before serologic diagnostics logic processes involved with the development
were available for hepatitis C virus (HCV) of PBC (Table 5.2). However, we still lack
infection, it was apparent that patients usable diagnostics to perform epidemiologic
with chronic HCV infection were regularly studies and better understand the prevalence
misdiagnosed with AIH. For example, of infection in patients with and without liver
approximately 5% of patients with chronic disease. Accordingly, cellular immunoassays
HCV infection make anti‐nuclear and other are being assessed for clinical use as most
autoantibodies and in the early 1990s, reports PBC patients lack serologic reactivity to stan
emerged of AIH patients undergoing liver dard ELISA assays (Table 5.2), whereas all
transplantation for hepatocellular carcinoma PBC patients tested to date have intrahepatic
who were found to “develop” a high preva lymphocyte reactivity to HBRV peptides.
lence of HCV infection once serologic Other incremental steps have been made to
diagnostics became routinely available.
strengthen the relationship of HBRV with
Accordingly, some of the viral associations PBC. For example, a recent long‐term obser
with AIH may be attributed to misdiagnosis. vation study of combination antiretroviral
One of the main issues with diagnosing therapy revealed that patients derived
AIH is that there are no highly specific dis sustained and clinically meaningful improve
ease features, such as the mitochondrial phe ment. However, better regimens are being
notype observed in patients with PBC. assessed as two‐thirds of PBC patients were
Previously, a positive AMA and histologic unable to tolerate HIV protease inhibitors.
evidence of cholangitis were once thought to After confirming that betaretroviruses are
exclude the diagnosis of AIH. However, there susceptible to HIV integrases in vitro, a sec
are clearly patients who present with features ond randomized controlled trial been initi
of both PBC and AIH, and these patients are ated to find better therapies for symptomatic
now considered to have an overlap syn patients and better link betaretrovirus infec
drome. Indeed, patients with the AIH/PBC tion with PBC. Ultimately, studies such as
overlap syndrome who respond to cortico these are required to help unravel the com
steroid therapy have evidence of AMA reac plex interaction of genetic and environmental
tivity on their bile ducts, indicative of the factors that contribute to trigger disease.
References
1 Hirschfield, G.M., Chapman, R.W., Karlsen, after liver transplantation and effects on
T.H. et al. (2013). The genetics of complex graft and patient survival. Gastroenterology
cholestatic disorders. Gastroenterology 144: 156: 96–107.
1357–1374. 8 Griffiths, L., Dyson, J.K., and Jones, D.E.
2 Sharon, D. and Mason, A.L. (2015). Role of (2014). The new epidemiology of primary
novel retroviruses in chronic liver disease: biliary cirrhosis. Semin. Liver Dis. 34:
assessing the link of human betaretrovirus 318–328.
with primary biliary cirrhosis. Curr. Infect. 9 Lytvyak, E., Montano‐Loza, A.J., Saxinger,
Dis. Rep. 17: 460. L., and Mason, A. (2015). Combination
3 Xu, L., Shen, Z., Guo, L. et al. (2003). Does a anti‐retroviral therapy provides reduction
betaretrovirus infection trigger primary in human betaretrovirus load and durable
biliary cirrhosis? Proc. Natl. Acad. Sci. USA biochemical responses in patients with
100: 8454–8459. primary biliary cirrhosis. Hepatology 62
4 Gayed, P.M. (2011). Toward a modern (Suppl): 528A.
synthesis of immunity: Charles A. Janeway 10 Koutsoumpas, A.L., Kriese, S., and
Jr. and the immunologist’s dirty little secret. Rigopoulou, E.I. (2012). Popular and
Yale J. Biol. Med. 84: 131–138. unpopular infectious agents linked to
5 Mason, A. (2018). Is PBC a viral infectious primary biliary cirrhosis. Autoimmun.
disease? Best Pract. Res. Clin. Gastroenterol. Highlights 3: 95–104.
34–35: 27–39. https://doi.org/10.1016/j. 11 Selmi, C., De Santis, M., Cavaciocchi, F.
bpg.2018.06.001. et al. (2010). Infectious agents and
6 Sfriso, P., Ghirardello, A., Botsios, C. et al. xenobiotics in the etiology of primary
(2010). Infections and autoimmunity: the biliary cirrhosis. Dis. Markers 29: 287–299.
multifaceted relationship. J. Leukoc. Biol. 87: 12 Wasilenko, S.T., Mason, G.E., and Mason,
385–395. A.L. (2009). Primary biliary cirrhosis,
7 Montano‐Loza, A., Hansen, B., Roccarina, bacteria and molecular mimicry: what’s the
D. et al. (2019). Factors associated with molecule and where’s the mimic? Liver Int.
recurrence of primary biliary cholangitis 29: 779–782.
103
Section II
Autoimmune Liver Diseases and Their Clinical Correlation

105
Autoimmune Hepatitis
Aliya Gulamhusein1,2 and Patrick McKiernan3
1
Toronto Centre for Liver Diseases, Toronto General Hospital, University Health Network, Toronto, Ontario, Canada
2
Division of Gastroenterology, University of Toronto, Toronto, Ontario, Canada
3
Children’s Hospital of Pittsburgh, University of Pittsburgh Medical Center, Pittsburgh, PA, USA
Abstract
Autoimmune hepatitis (AIH) is a progressive inflammatory hepatopathy characterized by hyper

gammaglobulinemia, specific autoantibodies, interface hepatitis, and normal cholangiography in
the absence of viral infection. AIH is rare in both adults and children though epidemiologic data are
limited and may be biased by underrecognition, particularly in underdeveloped countries.
Histologic evaluation is required for the diagnosis of AIH. Unique to childhood is the particularly
challenging issue of distinguishing between AIH and autoimmune sclerosing cholangitis. Longterm
immunosuppression is the cornerstone of therapy in pediatric and adult AIH and requires cortico
steroid based induction followed by, ideally, steroid sparing maintenance of remission with azathio
prine. Firstline therapy relies on corticosteroid based induction usually at a daily dose of 0.5 mg/kg,
though effective use of lower doses has been suggested. Secondline therapy may need to be consid
ered in 10–15% of patients as a result of either intolerance or suboptimal response to standard
regimens.
Keywords autoimmune hepatitis; autoimmune sclerosing cholangitis; epidemiologic data; firstline
therapy; histologic evaluation; longterm immunosuppression; secondline therapy
Key Points
●● Autoimmune hepatitis (AIH) is rare and ●● Diagnosis is clinical and takes into account
characterized by liver‐directed immune serologic profile including anti‐nuclear
injury. and anti‐smooth muscle antibodies in the
●● Clinical presentation can vary from acute case of type 1 AIH, and anti‐liver kidney
severe hepatitis with fulminant liver microsomal and anti‐liver cytosol type 1
failure to an asymptomatic elevation in antibodies in the case of type 2 AIH.
liver tests. In either case a high index of Whether type 3 AIH exists is controver
suspicion must be maintained to allow for sial but the disease‐specific anti‐soluble
confident, expedient diagnosis and timely liver antigen antibody may characterize
delivery of effective therapy. this subtype.
106 Section II Autoimmune Liver Diseases and Their Clinical Correlation
●● Histology is mandatory for diagnosis and ●● Prednisone with or without azathioprine

can be useful for directing therapy, partic is the standard induction regimen of
ularly if there is a suboptimal response to choice and these are the only agents that
standard therapy. have been robustly demonstrated to
●● Variant syndromes are uncommon: improve clinical outcomes.
specific to adulthood is AIH with features ●● Suboptimal response is infrequent and
of primary biliary cholangitis and unique most commonly driven by non‐adher
to pediatrics is autoimmune sclerosing ence. In the case of true refractory disease,
cholangitis. second‐line therapy may be considered.
Introduction by hypergammaglobulinemia, specific

autoantibodies, interface hepatitis, and
Autoimmune hepatitis (AIH) is a chronic normal cholangiography in the absence of
inflammatory liver disease that presents with viral infection. Classically, two types of AIH
characteristic circulating autoantibodies, are most widely described and distinguished
hypergammaglobulinemia, and histologic according to their autoantibody profile.
evidence of immune‐mediated liver injury. It
●● Type 1 AIH is characteristically associated
can affect people across the spectrum of age,
with anti‐nuclear antibody (ANA) and/or
gender, and ethnicity but is most typically
smooth muscle antibody (SMA) and is
seen in women with a bimodal age distribu
more typical in adults.
tion and peak incidence in childhood and
●● Type 2 AIH is more common in children
teenage years as well as between the fourth
and associated with the presence of anti‐
and sixth decades of life [1,2]. Clinical mani
liver kidney microsomal (LKM) or anti‐
festations vary widely, ranging from incidental
liver cytosol type 1 (LC1) antibodies [5,6].
elevation of liver tests in the absence of
symptoms to fulminant hepatic failure in the Whether a third subtype, characterized by
minority [3]. Confidence in diagnosis is anti‐soluble liver antigen (SLA) antibodies
imperative to ensure timely initiation of truly exists remains a matter of debate as
effective therapies. To date, standard therapy these patients are similar to those with type 1
is limited to immunosuppression with AIH though their phenotype may be more
prednis(olo)ne with or without azathioprine severe. Despite these distinctions, their
which, while imprecise in action, are effec clinical, genetic, and histologic characteris
tive in improving outcome in the majority of tics are similar [5].
patients [1,4]. That said, therapeutic options Genetic predisposition, molecular mim
have remained stagnant for decades and icry, and immune dysregulation character
novel options are needed for the important ized by an imbalance between effector and
subset of patients intolerant or insufficiently regulatory immune cells are important path
responsive to current therapy. Herein, we ogenic mechanisms of disease in AIH [7,8].
review the entity of AIH and its clinical Candidate gene and genome‐wide association
course in the adult and pediatric setting. studies (GWAS) have demonstrated the
strongest associations at the human leuko
cyte antigen (HLA) locus, with class II DRB1
Definition and alleles demonstrating the most robust associ
Pathophysiology ations [9,10]. In North American and
European cohorts, HLA‐DRB1*0301 and
Autoimmune hepatitis is a progressive HLA‐DRB1*0401 show particularly strong
inflammatory hepatopathy characterized associations in AIH, with the former also
Chapter 6 Autoimmune Hepatitis 107
being associated with younger age of disease (Treg) cells are of particular interest in their
onset, higher immunoglobulin G (IgG) levels, counteracting functions, with Th17 cells
treatment failure, relapse, and progression to involved in host defense and Treg suppres
liver transplantation, and the latter being sion, and Tregs responsible in large part for
more typical of an older, female cohort with a maintaining immune homeostasis and self‐
more benign course [11]. Indeed, associations tolerance [26,27]. Indeed, increased hepatic
with HLA‐DR3 and ‐DR4 are sufficiently Th17 cells are seen in patients with AIH as
strong to be included in some diagnostic cri compared to viral hepatitis B, with degree of
teria including the revised score developed by infiltration correlating with degree of inflam
the International Autoimmune Hepatitis mation and fibrosis [28]. While Tregs are
Group (IAIHG) [12]. Non‐HLA risk loci have likely to be highly relevant, their precise role
also been identified through GWAS efforts, in disease remains undefined. Early studies
with Src homology 2 adaptor 3 (SH2B3) and demonstrated reduced numbers and dys
the caspase recruitment domain family functional Tregs at diagnosis; however,
member 10 (CARD10) identified as genes subsequent efforts suggested normal levels
potentially implicated in disease [13–15]. of circulating Tregs and intrahepatic enrich
Interestingly, while pan‐ethnic associations ment, though this discrepancy may be attrib
have been seen with the HLA‐DRB1 locus, utable to measurement of different Treg
other HLA subtypes differ between ethnic subsets or to aberrant homing in an inflamed
groups and in a cohort of Japanese patients, liver [22,29,30]. It has been suggested that a
the non‐HLA CARD10 gene variant was not low frequency of CD39+ Tregs with impaired
associated with disease, highlighting the fact immunosuppressive ability and heighted ten
that genetic susceptibility may be vary bet dency to produce interferon (IFN)‐γ and
ween and across ethnicities [16]. interleukin (IL)‐17 on proinflammatory
It is likely that in a genetically susceptible challenge leads to increased conversion of
host, exposure to an as yet undefined Tregs into effector cells, driving the patho
antigenic trigger results in a dysregulated genesis of disease [24,31]. Such lines of evi
liver‐targeted immune response. While the dence emphasize the importance of immune
triggers in AIH are largely unclear, some tar dysregulation in disease and further under
gets are well recognized with cytochrome standing of these biologic mechanisms offers
P450 2D6 as the principal autoantigen recog potential for novel targeted therapeutic
nized by the anti‐LKM1 antibody, sep (O‐ strategies.
phosphoserine) tRNA: Sec (selenocysteine)
tRNA synthetase (SepSecS) as the main
target for anti‐SLA antibody, and formimi Epidemiology
notransferase cyclodeaminase as the
principal target for anti‐LC1 antibodies AIH is rare in both adults and children
[17–19]. Once initiated, both the innate and though epidemiologic data are limited and
adaptive immune responses interact leading may be biased by under‐recognition, particu
to hepatitic liver injury. Several lines of evi larly in underdeveloped countries. That said,
dence highlight the importance of T‐cell‐ in adults the prevalence ranges between 15
mediated immunity in disease [20–24]. and 25 cases per 100 000, though this seems
Presentation of an antigenic peptide bound to be rising [5]. A recent large nationwide
to HLA class II molecules to naive CD4+ T‐ population‐based Danish study noted a dou
helper cells leads to widespread differentiation bling in incidence of AIH between 1994 and
to Th1, Th2, and Th17 subsets which subse 2012 which could not be attributed to diag
quently produce an array of distinct cyto nostic ascertainment [32]. In children, the
kines that contribute to downstream immune reported minimum incidence is 0.2–0.4 per
dysregulation [25]. Th17 and T regulatory 100 000 children and prevalence rates vary
between 3 and 10 per 100 000 based on two fueled the hypothesis of ethnic variability in
recent North American studies [33]. outcome in AIH. That said, such differences
AIH is a female predominant disease, may also reflect distinct environmental expo
occurring in a 3 :
1 ratio [5]. A bimodal sures or intricate gene–environment
distribution of presentation is seen in adults interactions, or be driven by complex social
with peaks in the teenage years and another determinants of health.
in middle age between the fourth and sixth
decade, though recent data suggest an
increasing number of elderly patients diag
nosed beyond 65 years [34,35]. Type 1 AIH is
Presentation
by far most frequent in adults, with type 2
Adult‐onset AIH
AIH accounting for at most 10% of adult AIH
cases [5]. In children, type 1 AIH is at least The spectrum of clinical manifestations in
twice as common as type 2 and females are AIH are widely variable, ranging from
more commonly affected in both types. In asymptomatic to extreme forms of severe
children, the mean age of onset for type 1 acute hepatitis and fulminant hepatic failure.
AIH is between 10 and 11 years of age and for In the latter case, it is important to note
type 2 is younger at 6–7 years of age [36]. that autoantibodies and hypergammaglobu
Concurrent autoimmune diseases in patients linemia may be absent at first screen and thus
and their first‐degree relatives are common, a heightened index of suspicion for the possi
in both childhood and adulthood reported in bility of AIH must be maintained [46].
20–40%. While autoimmune thyroid disease Clinical presentation can be broadly grouped
is most common, a spectrum of manifesta into those with an acute severe presentation,
tions ranging from celiac disease to type 1 those with non‐specific symptoms, and
diabetes, inflammatory bowel disease, and asymptomatic patients with often subclinical
rheumatologic, dermatologic, and hemato disease. Up to 25% of patients present with
logic autoimmune manifestations has been an acute onset and can be further subdivided
described [37,38]. into those with an acute‐on‐chronic presen
While typical AIH is seen in all ethnic tation in contrast to true acute AIH in the
backgrounds, there are insufficient data to absence of clinical or histologic evidence of
determine whether significant inter‐ethnic chronicity [47,48]. In one‐third of patients,
differences exist in disease presentation and/ chronic non‐specific symptoms ranging
or outcome, though some observations have from, but not limited to, fatigue and gener
been reported. For instance, Alaskan natives alized malaise, right upper quadrant pain,
have a high frequency of acute icteric disease anorexia, weight loss, amenorrhea, and poly
at presentation, and a more frequent and arthralgias may be the primary complaints
severe phenotype has been reported in North [49]. Finally, one‐third of patients with AIH
American Aboriginal populations [39,40]. are asymptomatic at presentation with an
African‐American patients fare worse than insidious course typically characterized by
their white American counterparts, with a chronically unexplained elevations in liver
higher frequency of cirrhosis, treatment tests [5,50]. Importantly, up to half of patients
failure, and progression to liver transplanta at diagnosis have histologic evidence of
tion [41,42]. Patients of Hispanic origin may cirrhosis at presentation, attesting to the
also have an aggressive presentation with cir likelihood of long‐standing subclinical dis
rhosis and cholestatic features being reported ease [47,51]. This highlights the importance
more frequently, and Asian patients also tend of timely and confident diagnosis to avoid
toward poorer outcomes [43–45]. These delays in initiation of immunosuppressive
observations are limited to retrospective therapy that may delay or prevent poor
series of small patient numbers but have outcomes (Table 6.1). Physical findings are
Table 6.1 Clinical Presentation of autoimmune hepatitis in adults and children
Adults Pediatrics
Acute severe presentation (25‐30%) Acute hepatitis (50%)

●● sub‐fulminant presentation without underlying ●● minority progress to fulminant failure with
chronic liver disease encephalopathy within 2 months
●● acute on chronic AIH with histologic evidence of ●● more common in Type 2 AIH
chronicity
Chronic non‐specific symptoms (35%) Chronic non‐specific symptoms (30‐40%)
●● malaise ●● malaise
●● right upper quadran pain ●● weight loss
●● anorexia ●● poor growth
●● weight loss ●● intermittent jaundice
●● amenorrhea
●● polyarthralgias
Asymptomatic (35%) Complication of chronic liver disease (10‐20%)

●●chronically unexplained elevated liver tests ●● hematemesis
●● bruising
●● incidentally indentified organomegaly
●● incidental abnormal liver tests (rare)
often absent except in the case of established Occasionally, children come to attention
cirrhosis where typical findings may be seen, because of an incidental finding of abnormal
including palmar erythema, spider angioma, liver function tests. Disappointingly, some
and sarcopenia. In decompensated disease children will give a history of a previous
with portal hypertension, obvious manifesta unexplained seronegative hepatitis, which
tions of splenomegaly, ascites, pedal edema, emphasizes the importance of identifying an
or encephalopathy may be seen. etiology for childhood hepatitis. Frequently,
there is clinical evidence of chronic liver dis
ease including spider nevi, finger clubbing,
Pediatric‐onset AIH
and palmar erythema. Hepatomegaly is usual
Four basic phenotypes can be distinguished and splenomegaly is common.
in pediatric AIH. The commonest presenta
tion, in about half of cases, is with an acute
hepatitis. Symptoms are similar to acute viral Diagnosis
hepatitis, which is itself now rare in most
developed countries. A small propor tion The diagnosis of AIH is a summation of
may progress to develop encephalopathy and clinical, biochemical, immunologic, imaging,
acute liver failure within the first 2 months of and histologic features once diseases that
presentation. This is commoner is type 2 mimic AIH, including viral hepatitis and
AIH and is often seen in younger children. drug‐induced liver injury, have been
For another 30–40%, there is a longer and excluded. Typical biochemical abnormalities
often non‐specific history including malaise, include at times severe elevation in transam
weight loss, poor growth, and intermittent inase levels and hyperbilirubinemia, often in
jaundice. The remainder of patients pre the absence of significant elevation of chole
sent initially with a complication of chronic static enzymes, though these too may be
liver disease including hematemesis, bruis moderately deranged. Elevated gamma‐glob
ing, or incidentally identified organomegaly. ulin (IgG) levels are found in 80–85% of
patients even in the absence of cirrhosis and possibility of a variant syndrome with PBC
usually with normal corresponding IgA and but this should only be considered in the
IgM levels [52,53]. Concomitant elevation in clinical context of associated cholestasis and
IgA and/or IgM may be suggestive of super the predominant pathology based on clinical
imposed additional disease processes such as phenotype should be treated [65]. Periodic
alcoholic liver disease or primary biliary reassessment of serologic profile may be of
cholangitis (PBC). Isolated elevated IgG value particularly in patients initially diag
levels are distinctive to AIH and are a useful nosed with seronegative disease as antibody
biomarker of treatment response [54,55]. titers may vary over time. While in adults
The absence of viral hepatitis is included as a autoantibody titers correlate weakly with dis
diagnostic criterion but poses challenges in ease activity and clinical course, in children
endemic settings where the two entities may these are useful biomarkers of treatment
coexist. Particularly in the era of direct‐act response and should be followed.
ing antiviral therapy for hepatitis C virus Histologic evaluation is required for the
(HCV) infection, treatment of underlying diagnosis of AIH. As much as possible, it
viral hepatitis with subsequent careful reas should be performed before treatment initia
sessment of biochemical activity, autoanti tion and is useful in guiding treatment
body profile, and histologic findings may aid decisions and staging fibrosis. Hepatitis at the
in diagnosis and treatment decisions. portal–lobular boundary, so‐called “inter
Autoantibodies are an essential compo face” hepatitis, exhibits typical manifestations
nent of diagnostic evaluation. Indirect immu including plasma‐cell rich infiltrates, hepato
nofluorescence is the preferred technique cyte rosettes, emperipolesis, and hepatocyte
[56,57]. ANA and SMA are not disease necrosis; while these can be etiologically non‐
specific and show wide heterogeneity but are specific, in the right clinical context they are
hallmarks of type 1 AIH and are present in 43 highly supportive of the diagnosis (Figure 6.1)
and 40–60% of patients, respectively [66]. Plasma cells are usually abundant,
[56,58,59]. Anti‐LKM1 and/or anti‐LC1 are though their absence does not preclude diag
hallmarks of type 2 AIH, often coexist, and nosis [67]. In the acute setting, severe pan‐
have been reported in 66 and 53% of patients, lobular hepatitis with bridging and/or massive
respectively [58]. Interestingly, these anti necrosis with histologic collapse may be seen
bodies have been described in 5–10% of that may closely resemble drug‐induced toxic
patients with HCV infection, likely due to the injury and thus needs to be considered in the
presence of homologous sequences between correct clinical context [68].
CYP2D6 and HCV proteins [60,61]. Anti‐ A comprehensive scoring system was
SLA is the only disease‐specific autoantibody originally developed by the IAIHG in 1999 to
though is detected in only 30% of patients, define homogeneous cohorts of patients for
and often coexists with other autoantibodies, clinical trials rather than for diagnostic
particularly anti‐Ro52 [62]. As mentioned purposes and utilizes clinical, laboratory,
previously, the presence of anti‐SLA has serologic, and histologic factors along with
been associated with severe disease and response to therapy to classify patients’ proba
poorer prognosis in AIH [63]. In adults, bility of having disease (Table 6.2) [12]. Though
positive titers are considered at levels above robust, this tool is rather cumbersome to use
1 :
40 dilution by immunofluorescence, in clinical practice given its complexity,
whereas in children weaker titers of 1 : 20 for inability to distinguish cholestatic syndromes,
ANA and SMA or 1 : 10 for anti‐LKM are and use of response to therapy as a diagnostic
considered sufficiently positive in the correct criterion. Subsequently, in 2008 a simplified
clinical context [64]. Anti‐mitochondrial scoring system was proposed for everyday
antibodies (AMAs) are seen in 8–12% of clinical practice by the IAIHG and is based
patients with adult AIH and may raise the on four easy‐to‐use parameters including
Panel A Panel B
Figure 6.1 A and B. Panel A ‐ Histologic depiction of marked chronic portal inflammation with extension into
interface and lobules. Panel B – magnified view of severe portal and interface lymphoplasmacytic infiltration Images
provided by: Dr. Oyedele Adeyi, MD. Department of Pathology, Toronto General Hospital, University Health Network
Table 6.2 Revised Original Scoring System of the International Autoimmune Hepatitis Group
Criteria Score
Sex Female +2
AP:AST (or ALT) ratio >3 –2
< 1.5 +2
γ globulin or IgG level > ULN >2 +3
1.5‐2.0 +2
1.0‐1.5 +1
<1 0
ANA, SMA, or anti‐LKM‐1 titres >1:80 +3
1:80 +2
1:40 +1
<1:40 0
AMA Positive –4
Viral markers Positive –3
Negative +3
Drugs Yes –4
No +1
Alcohol <25 g/day +2
>60 g/day –2
HLA DR3 or DR4 +1
Immune Disease Thyroiditis, colitis, etc +2
Other markers Anti‐SLA, anti‐actin, anti‐LC1, pANCA +2
Histologic Features Interface hepatitis +3
Plasmacytic +1
Rosettes +1
None of the above –5
Biliary changes –3
Other features –3
Treatment Response Complete +2
Relapse +3
Pre‐treatment score: Definite diagnosis >15, Probable diagnosis 10‐14

Post‐treatment score: Definite diagnosis > 17, probable 12‐17
IgG concentration, autoantibody titers, his profile. Biliary histologic features are com
tologic findings, and absence of viral hepatitis moner in ASC but may be very subtle and
[54]. The diagnostic performance of the sim can also occur in AIH. There are epidemio
plified score is very good, with a sensitivity of logic differences in that ASC shows an equal
95%, specificity of 90%, and diagnostic accu sex distribution and is more frequently
racy of 92% [69,70]. Its ease of use, utility in associated with inflammatory bowel disease.
patients with cholestatic syndromes, and However, in individual cases distinguishing
lack of requirement for treatment response between these entities requires cholangiog
make it a useful tool in the correct clinical raphy. Abdominal ultrasound frequently
context to serve as a guide for the initiation demonstrates a large gallbladder in ASC but
of therapy. The revised IAIHG or simplified intrahepatic bile duct abnormalities are
score is not ideal for use in childhood due to rarely detected. Magnetic resonance cholan
age‐related differences in alkaline phospha giography is the investigation of choice with
tase levels and autoantibody cutoff titers, the endoscopic retrograde cholangiography
inclusion of alcohol and drug intake, and being reserved for cases where a dominant
their lack of utility in acute liver failure. A stricture requires intervention, or if there is
new scoring system has been proposed to need for tissue diagnosis to help exclude
address these concerns but this still requires cholangiocarcinoma.
prospective validation [71].
While variant syndromes, otherwise
referred to as “overlap syndromes,” are dis
cussed in more detail in Chapter 15, a few Treatment
points are worthy of mention. Approximately
10% of adult patients with AIH will present, Long‐term immunosuppression is the corner
usually consecutively and rarely simulta stone of therapy in pediatric and adult AIH
neously, with clinical features of coexistent and requires corticosteroid‐based induction
PBC or primary sclerosing cholangitis (PSC) followed by, ideally, steroid‐sparing mainte
[5]. While widely validated diagnostic cri nance of remission with azathioprine (Figure
teria for these syndromes are lacking, it is 6.2). Treatment is required for all patients
accepted that the revised and simplified with AIH who have symptoms, significant
IAIHG‐developed scores perform poorly in biochemical or histologic activity, and/or
this setting and are not recommended for advanced fibrosis or cirrhosis; the recommen
diagnosis of AIH in these patients. Patients dation for therapy in patients with mild
with AIH and features of PBC should meet activity without fibrosis and in elderly patients
diagnostic criteria for PBC and require a for whom prognostic impact of treatment is
biopsy to establish the presence of moderate less well defined remains unclear [5,72,73].
to severe interface hepatitis. In patients with Despite this, given the potential for unpredict
features of AIH and PSC, diagnostic features able severe fluctuations of disease and the risk
of AIH should be established and cholangio of progression, in the absence of contraindica
graphic features of biliary disease should be tions treatment for AIH is recommended in
demonstrated [5]. all patients. Lifelong immunosuppression is
generally the rule, with durable sustained
remission off therapy being achieved in the
Autoimmune Sclerosing Cholangitis
minority (<20%) of patients [74,75]. The goal
Unique to childhood is the particularly chal is to prevent or delay progressive liver disease
lenging issue of distinguishing between AIH by achieving biochemical, serologic, and his
and autoimmune sclerosing cholangitis tologic remission, defined as normalization of
(ASC). These patients present in a very sim alanine aminotransferase (ALT) and IgG and
ilar manner and have a similar immunologic resolution of interface activity.
Budesonide 9 mg daily
Predni(so)lone 0.5–1 mg/kg/day may be alternative in mild disease
(without advanced fibrosis)
Consider 2 week observation
Azathioprine 50 mg daily with up-

titration to target 1–2 mg/kg
ASSESS RESPONSE
TREATMENT RESPONDER
SUB-OPTIMAL RESPONSE
(normal ALT and IgG)
CONFIRM ADHERENCE
Taper prednisone and continue
long term azathioprine Verify absence of superimposed drug injury
maintenance Verify absence of superimposed liver disease (NASH,
viral hepatitis)
Assess for variant syndromes with PBC or PSC
Suboptimal response (elevated IgG or ALT Alternative etiology identified

OR
treatment intolerant Treat as appropriate
optimize adherence
withdraw offending drug
Consider second line therapy Rx for NAFLD, viral hepatitis
6-MP manage variant syndrome
Mycophenolate Mofetil
Calcineurin inhibitors
(cyclosporin, tacrolimus)
Salvage options: rituximab,
infliximab
Figure 6.2 Suggested treatment algorithm for management of autoimmune hepatitis in adults.
Therapeutic Approach in Adults

hepatotoxicity. In general, azathioprine
First‐line Therapy should be initiated when bilirubin level is
First‐line therapy relies on corticosteroid‐ below 100 μmol/L (5.8mg/dL) and can be
based induction usually at a daily dose of started at 50 mg daily up‐titrated to a target
0.5–1 mg/kg, though effective use of lower dose of 1–2 mg/kg daily. In non‐cirrhotic
doses has been suggested [5,76]. Prednis(ol) patients with mild disease, particularly in those
one with or without azathioprine in combi with relative contraindications to systemic cor
nation are equally effective in achieving remis ticosteroid use, budesonide is a reasonable
sion and improving outcome and superior to alternative and may circumvent steroid‐related
azathioprine alone for induction of remission toxicity [79]. In typical AIH, prompt improve
[77,78]. After starting corticosteroids, it is rea ment in liver tests is the rule after initiation of
sonable to delay initiation of azathioprine alone corticosteroids and absence of response should
for a few weeks to observe response and avoid prompt a search for alternative etiologies. With
the uncommon, albeit possible, confounding this approach, histologic remission can be
of suboptimal response due to azathrioprine achieved in 75% of patients after 18 months of
steroid‐based therapy though, importantly, or advanced fibrosis who wish for a trial off
histologic remission lags behind biochemical therapy, a pretreatment withdrawal biopsy
normalization by months [77]. should be performed; in patients with active
Once remission is achieved, the target dose interface hepatitis or an hepatic activity index
of azathioprine should be optimized to 2 mg/ (HAI) above 3, the risk of relapse exceeds 90%
kg and subsequently prednisone should be and in this case treatment should be continued
gradually tapered off. With this approach [55]. If histologic remission is confirmed
steroid‐free relapse is rare after 1 year, and and treatment is stopped, patients should be
greater than 80% remission rates can be closely monitored for relapse; this is most
expected after 67 months of follow‐up [80,81]. likely to occur within the first 12 months and
Long‐term maintenance with azathioprine warrants resumption of therapy which after
monotherapy is optimal, though low‐dose relapse should be lifelong [74].
prednisone therapy (<10 mg) may be consid
ered but is a less favorable option. Safety and Second‐line Options
efficacy data for maintenance budesonide Second‐line therapy may need to be consid
therapy in adults and children alike are lacking ered in 10–15% of patients as a result of
and this approach is discouraged. Tolerability either intolerance or suboptimal response to
to standard regimens remains problematic, standard regimens. After confirmation of
with 10–15% of patients discontinuing therapy adherence and exclusion of alternative etiol
due to side effects that can result in clinical ogies of liver injury, true treatment failure is
relapse warranting reinduction therapy. On‐ seen in less than 5% of patients and in this
therapy relapse warrants a detailed history of case second‐line therapy may need to be
adherence, which is the most likely reason for considered. The most frequently considered
loss of disease control. Cosmetic, psychologic, agents include 6‐mercaptopurine (6‐MP),
and metabolic toxicity of corticosteroids can mycophenolate mofetil (MMF), and calci
be particularly troublesome with long‐term neurin inhibitors such as ciclosporin or
use. Monitoring while on azathioprine is tacrolimus. Data for use of biologic agents
required, particularly for features of bone including rituximab and anti‐tumor necrosis
marrow suppression. An increased risk of factor (TNF) agents are limited to very small
malignancies has been reported, though this case series. In patients intolerant to azathio
remains controversial; however, this may be of prine, use of 6‐MP has been trialed with
relevance with lifelong use [82,83]. response rates of up to 75% from small series
being reported [84]. As 6‐MP is a downstream
Withdrawal of Therapy metabolite of azathioprine, manipulation of
Durable remission off therapy is unusual, the same biologic pathway is a rational option
with relapse rates of 50–90% widely reported. particularly when azathioprine use is limited
For all patients, maintenance treatment by gastrointestinal complaints or myelosup
should be continued for at least 3 years with pression. There are limited data of patients
at least 2 years of complete remission (as tolerating 6‐MP in the setting of azathio
evidenced by normal serum ALT and IgG) prine‐related hepatotoxicity and if this is
before a trial of withdrawal may be consid considered careful monitoring of liver tests
ered [78]. In patients with advanced fibrosis needs to be ensured with prompt discontinu
or cirrhosis, prior episodes of relapse, or ation if drug injury is noted. MMF is most
intolerance to induction therapy, continuing widely used as a second‐line agent at doses of
maintenance agents without a trial of discon 1.5–2 g daily in divided doses, with reported
tinuation is acceptable after discussion of rates of remission ranging between 31 and
risks and benefits, and this may be the 73% in both poor responders and those intol
preferred option for both the patient and erant to azathioprine [85,86]. Data on the
practitioner. In patients without prior relapse efficacy of ciclosporin in adults is limited
whereas good biochemical response has been and 6‐methylmercaptopurine levels can be
observed in children. Use of tacrolimus as a useful, especially where there is hypersplen
steroid‐sparing second‐line option has also ism with baseline neutropenia and thrombo
been variably reported; however, a recent ret cytopenia and in order to demonstrate
rospective study of 201 patients treated with compliance. An alternative strategy is to
second‐line therapy in AIH showed no reserve azathioprine for delayed steroid
difference in remission rates between MMF‐ response and/or steroid side effects.
treated and tacrolimus‐treated patients who Normal biochemistry can be expected to be
had failed azathioprine (69.4 vs. 72.5%) [87]. achieved in more than 80% of cases by 6
Doses of tacrolimus used generally range months of treatment and a complete response
between 1 and 8 mg daily with target trough is defined as normal biochemistry, normal
blood levels of 4–6 ng/ml being reasonable immunoglobulins, undetectable autoanti
but clearly empiric. For refractory disease, bodies, and absence of histologic activity.
salvage options of therapy with rituximab Maintenance therapy with low‐dose cortico
have been reported in small series demon steroids and azathioprine should be continued
strating biochemical improvement and for at least 2–3 years assuming a complete
improved histologic inflammation, though response. Repeat liver biopsy should be sched
only small numbers of patients have been uled after at least 2 years of biochemical
exposed in this setting [88]. response and only if it would be appropriate to
Use of infliximab has also been trialed with consider relaxing or withdrawing treatment.
moderate success in a series of 11 patients,
though this approach is controversial due to risk Withdrawal of Therapy
of infections noted and reports of anti‐TNF‐ Where there has been complete response the
related immune‐mediated hepatitis [89,90]. options of reduction/discontinuation of
treatment should be discussed with the young
person and their family. This should be con
Therapeutic Approach in Children
sidered at a time of physiologic stability, spe
First‐line Therapy cifically avoiding the rapid growth phase of
As with adults, where there is a possible drug early adolescence. The relapse rate is particu
trigger for AIH, this should be discontinued larly high in type 2 AIH. Many young people
immediately. Induction treatment starts with may consider continuing long‐term low‐dose
prednisolone 2 mg/kg daily up to a maximum monotherapy with azathioprine rather than
of 60 mg. This should be combined with an risk relapse with the subsequent necessity for
antacid in the first instance. This dose should reintroduction of steroids. During relaxation/
be maintained for at least 2 weeks and subse withdrawal of treatment, frequent monitoring
quently reduced over the next 4–12 weeks, should be undertaken including liver bio
according to response, to a maintenance dose chemistry and autoantibody levels.
of 0.1 mg/kg daily. This should be guided by Relapses are relatively common and gener
weekly, or at minimum fortnightly, labora ally associated with non‐compliance. They
tory monitoring. should be treated similarly to initial presenta
Having excluded thiopurine S‐methyl tion with sufficient steroids to induce remis
transferase (TPMT) deficiency, most centers sion and subsequent gradual relaxation to
now introduce azathioprine electively, maintenance treatment according to response.
unless there is severe cholestasis. Once there
has been an initial biochemical response, Second‐line Options
azathioprine should be introduced in a dose First‐line treatment fails or is associated with
of 1 mg/kg daily and if tolerated subse significant side effects in 10–15% of cases.
quently increased to 1.5–2 mg/kg daily over Subsequent choices would include either
the next month. Monitoring 6‐thioguanine MMF or one of ciclosporin/tacrolimus [87].
There is insufficient evidence to make a rec and chronic viral hepatitis in the era of such
ommendation between these agents. The studies. Suffice to say that untreated
choice will often depend on the experience moderate to severe AIH carries a very poor
and familiarity with these drugs of the treating prognosis [91,92]. The survival benefit
physician and family preference. The greatest associated with immunosuppression using
experience is with MMF, commencing at a prednisone with or without azathioprine
dose of 20 mg/kg daily increased as tolerated has been demonstrated in a number of
to 30–40 mg/kg daily. Remission rates of randomized controlled trials with four to
greater than 80% are reported. Particular side seven-fold reduction in mortality in treated
effects include abdominal pain and constipa patients, emphasizing the importance of
tion and neutropenia. A significant concern timely diagnosis and prompt institution of
with this drug is that it is teratogenic and is therapy [80]. Liver‐related morbidity and
contraindicated in pregnancy. transplantation may be avoided with early
Tacrolimus is a potent calcineurin inhib effective intervention and maintenance of
itor and is the default immunosuppressive remission, though in a minority of cases
agent following solid organ transplantation transplantation will be required [80,91]. In
in childhood. It is usually commenced twice North America and Europe, approximately
daily in a lower dose than used in transplan 4% of liver transplant activity is due to AIH
tation and guided by therapeutic monitoring, [21]. Treatment failure is more likely in
aiming for trough levels of approximately severe acute presentations, particularly in
5 ng/ml. Side effects include nephrotoxicity, sub‐fulminant disease; prognosis is particu
hypertension, headache, and diabetes. larly poor in this setting, with mortality
Ciclosporin has a similar mode of action and ranging from 19 to 45% and requirement for
side‐effect profile to tacrolimus. However, liver transplantation in 9–81% [93]. Liver
ciclosporin has proved to be particularly dif transplantation can be a highly effective sal
ficult to use in AIH due to toxicity, and vage option; however, recurrence of disease
increasingly pediatricians are unfamiliar has been reported in up to 36% of patients at
with using and monitoring this drug. 5 years, though late acute cellular rejection
For children resistant or intolerant to sec episodes are also more frequent in patients
ond‐line treatment, there are few proven with AIH and how clinically distinct these
options. There is anecdotal positive experi two entities are remains somewhat unclear
ence with rituximab, infliximab, and siroli [94,95]. Indeed some groups report a lower
mus. When embarking on such treatment, it frequency of recurrent AIH post transplan
should ideally be done as part of a multidisci tation with long‐term prolonged use of low‐
plinary team and following appropriate con dose prednisone post transplantation [96].
sultation with locally available and recognized
expertise. Randomized controlled trials of
Childhood
such treatment are probably impractical but
there is a need for at least a register of such Transplantation will be necessary in approxi
treatment protocols so that knowledge can mately 10–15% of young people with AIH,
be accumulated and shared. with indications falling into two broad groups:
(i) those presenting with acute liver failure
and encephalopathy who do not respond rap
idly to immunosuppressive treatment; and (ii)
Prognosis the more common scenario where transplant
is required for a combination of chronic liver
Adulthood
disease and intolerance of immunosuppres
The prognosis of untreated disease is diffi sive treatment. The hepatic indications for
cult to determine historically in part because transplantation are similar to any chronic
of difficulties with distinction between AIH liver disease. For some young people with
autoimmune liver disease, myalgia and aner Recurrent disease remains the commonest
gia is a persistent and dominant symptom. cause of graft loss in this group. Factors con
The prognosis after liver transplantation is tributing to recurrence include uncontrolled
excellent, with the expectation of in excess of inflammation at transplant, manifesting
90% high‐quality long‐term survival. The HLA‐DRB1*03, and discontinuing cortico
absolute rate of recurrence following liver steroids [97]. The impact of HLA matching
transplantation is high, with published rates between donor and recipient is unclear. In
ranging from 30 to 80% with increasing inci recognition of this, most programs issue
dence over time. specific immunosuppression regimens for
Severe recurrent disease requiring retrans AIH including long‐term low‐dose cortico
plantation is less common, although a steroids, maintenance azathioprine, and
significant burden at 15–25% overall. calcineurin inhibitors.
References
1 Manns, M.P., Czaja, A.J., Gorham, J.D. et al. 10 Liberal, R., Krawitt, E.L., Vierling, J.M. et al.
(2010). Diagnosis and management of (2016). Cutting edge issues in autoimmune
autoimmune hepatitis. Hepatology 51 (6): hepatitis. J. Autoimmun. 75: 6–19.
2193–2213. 11 Donaldson, P.T. (2004). Genetics of liver
2 van Gerven, N.M., Verwer, B.J., Witte, B.I. disease: immunogenetics and disease
et al. (2014). Epidemiology and clinical pathogenesis. Gut 53 (4): 599–608.
characteristics of autoimmune hepatitis in 12 Alvarez, F., Berg, P.A., Bianchi, F.B. et al.
the Netherlands. Scand. J. Gastroenterol. 49 (1999). International Autoimmune
(10): 1245–1254. Hepatitis Group Report: review of criteria
3 Czaja, A.J. (2016). Diagnosis and management for diagnosis of autoimmune hepatitis. J.
of autoimmune hepatitis: current status and Hepatol. 31 (5): 929–938.
future directions. Gut Liver 10 (2): 177–203. 13 de Boer, Y.S., van Gerven, N.M., Zwiers, A.
4 Mackay, I.R. (2008). Historical reflections on et al. (2014). Genome‐wide association
autoimmune hepatitis. World J. study identifies variants associated with
Gastroenterol. 14 (21): 3292–3300. autoimmune hepatitis type 1.
5 European Association for the Study of the Gastroenterology 147 (2): 443–452 e5.
Liver (2015). EASL clinical practice 14 Liu, J.Z., Almarri, M.A., Gaffney, D.J.
guidelines: autoimmune hepatitis. J. Hepatol. et al. (2012). Dense fine‐mapping study
63 (4): 971–1004. identifies new susceptibility loci for
6 Muratori, P., Lalanne, C., Fabbri, A. et al. primary biliary cirrhosis. Nat. Genet. 44
(2015). Type 1 and type 2 autoimmune (10): 1137–1141.
hepatitis in adults share the same clinical 15 Zhernakova, A., Stahl, E.A., Trynka, G.
phenotype. Aliment. Pharmacol. Ther. 41 et al. (2011). Meta‐analysis of genome‐wide
(12): 1281–1287. association studies in celiac disease and
7 Liberal, R., Longhi, M.S., Mieli‐Vergani, G., rheumatoid arthritis identifies fourteen
and Vergani, D. (2011). Pathogenesis of non‐HLA shared loci. PLoS Genet. 7 (2):
autoimmune hepatitis. Best Pract. Res. Clin. e1002004.
Gastroenterol. 25 (6): 653–664. 16 Migita, K., Jiuchi, Y., Furukawa, H. et al.
8 Doycheva, I., Watt, K.D., and Gulamhusein, (2015). Lack of association between the
A.F. (2019). Autoimmune hepatitis: current CARD10 rs6000782 polymorphism and
and future therapeutic options. Liver Int. 39 type 1 autoimmune hepatitis in a Japanese
(6): 1002–1013. population. BMC Res. Notes 8: 777.
9 van Gerven, N.M., de Boer, Y.S., Zwiers, A. 17 Manns, M.P., Johnson, E.F., Griffin, K.J.
et al. (2015). HLA‐DRB1*03:01 and HLA‐ et al. (1989). Major antigen of liver kidney
DRB1*04:01 modify the presentation and microsomal autoantibodies in idiopathic
outcome in autoimmune hepatitis type‐1. autoimmune hepatitis is cytochrome
Genes Immun. 16 (4): 247–252. P450db1. J. Clin. Invest. 83 (3): 1066–1072.
18 Wies, I., Brunner, S., Henninger, J. et al. T‐cells on CD8 T‐cell function in patients
(2000). Identification of target antigen for with autoimmune hepatitis. J. Autoimmun.
SLA/LP autoantibodies in autoimmune 25 (1): 63–71.
hepatitis. Lancet 355 (9214): 1510–1515. 30 Peiseler, M., Sebode, M., Franke, B. et al.
19 Lapierre, P., Hajoui, O., Homberg, J.C., and (2012). FOXP3+ regulatory T cells in
Alvarez, F. (1999). Formiminotransferase autoimmune hepatitis are fully functional
cyclodeaminase is an organ‐specific and not reduced in frequency. J. Hepatol.
autoantigen recognized by sera of patients 57 (1): 125–132.
with autoimmune hepatitis. 31 Grant, C.R., Liberal, R., Holder, B.S. et al.
Gastroenterology 116 (3): 643–649. (2014). Dysfunctional CD39(POS)
20 Hardtke‐Wolenski, M. and Jaeckel, E. regulatory T cells and aberrant control of
(2010). Mouse models for experimental T‐helper type 17 cells in autoimmune
autoimmune hepatitis: limits and chances. hepatitis. Hepatology 59 (3): 1007–1015.
Dig. Dis. 28 (1): 70–79. 32 Gronbaek, L., Vilstrup, H., and Jepsen, P.
21 Manns, M.P., Lohse, A.W., and Vergani, D. (2014). Autoimmune hepatitis in Denmark:
(2015). Autoimmune hepatitis: update incidence, prevalence, prognosis, and
2015. J. Hepatol. 62 (1 Suppl): S100–S111. causes of death. A nationwide registry‐
22 Longhi, M.S., Ma, Y., Bogdanos, D.P. et al. based cohort study. J. Hepatol. 60 (3):
(2004). Impairment of CD4(+)CD25(+) 612–617.
regulatory T‐cells in autoimmune liver 33 Jimenez‐Rivera, C., Ling, S.C., Ahmed, N.
disease. J. Hepatol. 41 (1): 31–37. et al. (2015). Incidence and characteristics
23 Ferri, S., Longhi, M.S., De Molo, C. et al. of autoimmune hepatitis. Pediatrics 136
(2010). A multifaceted imbalance of T cells (5): e1237–1248.
with regulatory function characterizes type 34 Schramm, C., Kanzler, S., zum
1 autoimmune hepatitis. Hepatology 52 (3): Buschenfelde, K.H. et al. (2001).
999–1007. Autoimmune hepatitis in the elderly. Am.
24 Taubert, R., Hardtke‐Wolenski, M., Noyan, J. Gastroenterol. 96 (5): 1587–1591.
F. et al. (2014). Intrahepatic regulatory T 35 Chen, J., Eslick, G.D., and Weltman, M.
cells in autoimmune hepatitis are (2014). Systematic review with meta‐
associated with treatment response and analysis: clinical manifestations and
depleted with current therapies. J. Hepatol. management of autoimmune hepatitis in
61 (5): 1106–1114. the elderly. Aliment. Pharmacol. Ther. 39
25 Longhi, M.S., Ma, Y., Mieli‐Vergani, G., and (2): 117–124.
Vergani, D. (2010). Aetiopathogenesis of 36 Mieli‐Vergani, G., Vergani, D., Baumann,
autoimmune hepatitis. J. Autoimmun. 34 U. et al. (2018). Diagnosis and management
(1): 7–14. of pediatric autoimmune liver disease:
26 Korn, T., Bettelli, E., Oukka, M., and ESPGHAN hepatology committee position
Kuchroo, V.K. (2009). IL‐17 and Th17 cells. statement. J. Pediatr. Gastroenterol. Nutr.
Annu. Rev. Immunol. 27: 485–517. 66 (2): 345–360.
27 Seki, E. and Schwabe, R.F. (2015). Hepatic 37 Rigopoulou, E.I., Dalekos, G., and
inflammation and fibrosis: functional links and Bogdanos, D.P. (2007). How common are
key pathways. Hepatology 61 (3): 1066–1079. connective tissue disorders in patients
28 Zhao, L., Tang, Y., You, Z. et al. (2011). with autoimmune hepatitis? Semin.
Interleukin‐17 contributes to the Arthritis Rheum. 36 (5): 332. author
pathogenesis of autoimmune hepatitis reply 333.
through inducing hepatic interleukin‐6 38 Teufel, A., Weinmann, A., Kahaly, G.J. et al.
expression. PLoS One 6 (4): e18909. (2010). Concurrent autoimmune diseases
29 Longhi, M.S., Ma, Y., Mitry, R.R. et al. in patients with autoimmune hepatitis.
(2005). Effect of CD4+ CD25+ regulatory J. Clin. Gastroenterol. 44 (3): 208–213.
39 Hurlburt, K.J., McMahon, B.J., Deubner, H. patients with autoimmune hepatitis. Exp.
et al. (2002). Prevalence of autoimmune Ther. Med. 7 (1): 145–148.
liver disease in Alaska Natives. Am. J. 50 Feld, J.J., Dinh, H., Arenovich, T. et al.
Gastroenterol. 97 (9): 2402–2407. (2005). Autoimmune hepatitis: effect of
40 Minuk, G.Y., Liu, S., Kaita, K. et al. (2008). symptoms and cirrhosis on natural history
Autoimmune hepatitis in a North and outcome. Hepatology 42 (1): 53–62.
American Aboriginal/First Nations 51 Miyake, Y., Iwasaki, Y., Terada, R. et al.
population. Can. J. Gastroenterol. 22 (10): (2006). Clinical characteristics of
829–834. fulminant‐type autoimmune hepatitis: an
41 Verma, S., Torbenson, M., and Thuluvath, analysis of eleven cases. Aliment.
P.J. (2007). The impact of ethnicity on the Pharmacol. Ther. 23 (9): 1347–1353.
natural history of autoimmune hepatitis. 52 Al‐Chalabi, T., Underhill, J.A., Portmann,
Hepatology 46 (6): 1828–1835. B.C. et al. (2008). Impact of gender on the
42 Lim, K.N., Casanova, R.L., Boyer, T.D., and long‐term outcome and survival of patients
Bruno, C.J. (2001). Autoimmune hepatitis with autoimmune hepatitis. J. Hepatol. 48
in African Americans: presenting features (1): 140–147.
and response to therapy. Am. J. 53 Floreani, A., Niro, G., Rosa Rizzotto, E.
Gastroenterol. 96 (12): 3390–3394. et al. (2006). Type I autoimmune hepatitis:
43 Wong, R.J., Gish, R., Frederick, T. et al. clinical course and outcome in an Italian
(2012). The impact of race/ethnicity on the multicentre study. Aliment. Pharmacol.
clinical epidemiology of autoimmune Ther. 24 (7): 1051–1057.
hepatitis. J. Clin. Gastroenterol. 46 (2): 54 Hennes, E.M., Zeniya, M., Czaja, A.J. et al.
155–161. (2008). Simplified criteria for the diagnosis
44 Levy, C., Naik, J., Giordano, C. et al. (2014). of autoimmune hepatitis. Hepatology 48
Hispanics with primary biliary cirrhosis are (1): 169–176.
more likely to have features of autoimmune 55 Luth, S., Herkel, J., Kanzler, S. et al. (2008).
hepatitis and reduced response to Serologic markers compared with liver
ursodeoxycholic acid than non‐Hispanics. biopsy for monitoring disease activity in
Clin. Gastroenterol. Hepatol. 12 (8): autoimmune hepatitis. J. Clin.
1398–1405. Gastroenterol. 42 (8): 926–930.
45 Zolfino, T., Heneghan, M.A., Norris, S. 56 Vergani, D., Alvarez, F., Bianchi, F.B. et al.
et al. (2002). Characteristics of (2004). Liver autoimmune serology: a
autoimmune hepatitis in patients who are consensus statement from the committee
not of European Caucasoid ethnic origin. for autoimmune serology of the
Gut 50 (5): 713–717. International Autoimmune Hepatitis
46 Czaja, A.J. (2011). Autoimmune hepatitis in Group. J. Hepatol. 41 (4): 677–683.
special patient populations. Best Pract. Res. 57 Bogdanos, D.P., Invernizzi, P., Mackay, I.R.,
Clin. Gastroenterol. 25 (6): 689–700. and Vergani, D. (2008). Autoimmune liver
47 Stravitz, R.T., Lefkowitch, J.H., Fontana, serology: current diagnostic and clinical
R.J. et al. (2011). Autoimmune acute liver challenges. World J. Gastroenterol. 14 (21):
failure: proposed clinical and histological 3374–3387.
criteria. Hepatology 53 (2): 517–526. 58 Muratori, P., Granito, A., Quarneti, C. et al.
48 Ferrari, R., Pappas, G., Agostinelli, D. et al. (2009). Autoimmune hepatitis in Italy: the
(2004). Type 1 autoimmune hepatitis: Bologna experience. J. Hepatol. 50 (6):
patterns of clinical presentation and 1210–1218.
differential diagnosis of the “acute” type. 59 Frenzel, C., Herkel, J., Luth, S. et al. (2006).
QJM 97 (7): 407–412. Evaluation of F‐actin ELISA for the
49 Peng, M., Li, Y., Zhang, M. et al. (2014). diagnosis of autoimmune hepatitis. Am. J.
Clinical features in different age groups of Gastroenterol. 101 (12): 2731–2736.
60 Dalekos, G.N., Makri, E., Loges, S. et al. diagnosis of autoimmune hepatitis in

(2002). Increased incidence of anti‐LKM Chinese patients. J. Hepatol. 54 (2):
autoantibodies in a consecutive cohort of 340–347.
hepatitis C patients from central Greece. 71 Mieli‐Vergani, G., Vergani, D., Baumann,
Eur. J. Gastroenterol. Hepatol. 14 (1): 35–42. U. et al. (2018). Diagnosis and management
61 Kerkar, N., Choudhuri, K., Ma, Y. et al. of paediatric autoimmune liver disease:
(2003). Cytochrome P4502D6(193–212): a ESPGHAN hepatology committee position
new immunodominant epitope and target statement. J. Pediatr. Gastroenterol. Nutr.
of virus/self cross‐reactivity in liver kidney 66: 345–360. https://doi.org/10.1097/
microsomal autoantibody type 1‐positive MPG.0000000000001801.
liver disease. J. Immunol. 170 (3): 72 De Groote, J., Fevery, J., and Lepoutre, L.
1481–1489. (1978). Long‐term follow‐up of chronic
62 Baeres, M., Herkel, J., Czaja, A.J. et al. active hepatitis of moderate severity. Gut
(2002). Establishment of standardised SLA/ 19 (6): 510–513.
LP immunoassays: specificity for 73 Czaja, A.J. (2009). Features and
autoimmune hepatitis, worldwide consequences of untreated type 1
occurrence, and clinical characteristics. autoimmune hepatitis. Liver Int. 29 (6):
Gut 51 (2): 259–264. 816–823.
63 Montano‐Loza, A.J., Shums, Z., Norman, 74 Czaja, A.J., Menon, K.V., and Carpenter,
G.L., and Czaja, A.J. (2012). Prognostic H.A. (2002). Sustained remission after
implications of antibodies to Ro/SSA and corticosteroid therapy for type 1
soluble liver antigen in type 1 autoimmune autoimmune hepatitis: a retrospective
hepatitis. Liver Int. 32 (1): 85–92. analysis. Hepatology 35 (4): 890–897.
64 Mieli‐Vergani, G., Heller, S., Jara, P. et al. 75 Montano‐Loza, A.J., Carpenter, H.A., and
(2009). Autoimmune hepatitis. J. Pediatr. Czaja, A.J. (2007). Consequences of
Gastroenterol. Nutr. 49 (2): 158–164. treatment withdrawal in type 1
65 O’Brien, C., Joshi, S., Feld, J.J. et al. (2008). autoimmune hepatitis. Liver Int. 27 (4):
Long‐term follow‐up of antimitochondrial 507–515.
antibody‐positive autoimmune hepatitis. 76 Pape, S., Gevers, T.J.G., Belias, M. et al.
Hepatology 48 (2): 550–556. (2019). Predniso(lo)ne dosage and chance
66 Czaja, A.J. and Carpenter, H.A. (1993). of remission in patients with autoimmune
Sensitivity, specificity, and predictability of hepatitis. Clin. Gastroenterol. Hepatol. 17
biopsy interpretations in chronic hepatitis. (10): 2068–2075.e2.
Gastroenterology 105 (6): 1824–1832. 77 Tan, P., Marotta, P., Ghent, C., and Adams,
67 Desmet, V.J., Gerber, M., Hoofnagle, J.H. P. (2005). Early treatment response predicts
et al. (1994). Classification of chronic the need for liver transplantation in
hepatitis: diagnosis, grading and staging. autoimmune hepatitis. Liver Int. 25 (4):
Hepatology 19 (6): 1513–1520. 728–733.
68 Kessler, W.R., Cummings, O.W., Eckert, G. 78 Murray‐Lyon, I.M., Stern, R.B., and
et al. (2004). Fulminant hepatic failure as Williams, R. (1973). Controlled trial of
the initial presentation of acute prednisone and azathioprine in active
autoimmune hepatitis. Clin. Gastroenterol. chronic hepatitis. Lancet 1 (7806):
Hepatol. 2 (7): 625–631. 735–737.
69 Czaja, A.J. (2008). Performance parameters 79 Manns, M.P., Woynarowski, M., Kreisel, W.
of the diagnostic scoring systems for et al. (2010). Budesonide induces remission
autoimmune hepatitis. Hepatology 48 (5): more effectively than prednisone in a
1540–1548. controlled trial of patients with
70 Qiu, D., Wang, Q., Wang, H. et al. (2011). autoimmune hepatitis. Gastroenterology
Validation of the simplified criteria for 139 (4): 1198–1206.
80 Lamers, M.M., van Oijen, M.G., Pronk, M., 89 Weiler‐Normann, C., Schramm, C., Quaas,
and Drenth, J.P. (2010). Treatment options A. et al. (2013). Infliximab as a rescue
for autoimmune hepatitis: a systematic treatment in difficult‐to‐treat autoimmune
review of randomized controlled trials. J. hepatitis. J. Hepatol. 58 (3): 529–534.
Hepatol. 53 (1): 191–198. 90 Rodrigues, S., Lopes, S., Magro, F. et al.
81 Stellon, A.J., Keating, J.J., Johnson, P.J. et al. (2015). Autoimmune hepatitis and anti‐
(1988). Maintenance of remission in tumor necrosis factor alpha therapy: a
autoimmune chronic active hepatitis with single center report of 8 cases. World J.
azathioprine after corticosteroid Gastroenterol. 21 (24): 7584–7588.
withdrawal. Hepatology 8 (4): 781–784. 91 Cook, G.C., Mulligan, R., and Sherlock, S.
82 Meyer zum Buschenfelde, K.H. and Lohse, (1971). Controlled prospective trial of
A.W. (1995). Autoimmune hepatitis. N. corticosteroid therapy in active chronic
Engl. J. Med. 333 (15): 1004–1005. hepatitis. Q. J. Med. 40 (158): 159–185.
83 Ngu, J.H., Gearry, R.B., Frampton, C.M., 92 Soloway, R.D., Summerskill, W.H.,
and Stedman, C.A. (2012). Mortality and Baggenstoss, A.H. et al. (1972). Clinical,
the risk of malignancy in autoimmune liver biochemical, and histological remission of
diseases: a population‐based study in severe chronic active liver disease: a
Canterbury, New Zealand. Hepatology 55 controlled study of treatments and early
(2): 522–529. prognosis. Gastroenterology 63 (5):
84 Hubener, S., Oo, Y.H., Than, N.N. et al. 820–833.
(2016). Efficacy of 6‐mercaptopurine as 93 Yeoman, A.D., Westbrook, R.H., Zen, Y.
second‐line treatment for patients with et al. (2014). Prognosis of acute severe
autoimmune hepatitis and azathioprine autoimmune hepatitis (AS‐AIH): the role
intolerance. Clin. Gastroenterol. Hepatol. of corticosteroids in modifying outcome. J.
14 (3): 445–453. Hepatol. 61 (4): 876–882.
85 Wolf, D.C., Bojito, L., Facciuto, M., and 94 Montano‐Loza, A.J., Mason, A.L., Ma, M.
Lebovics, E. (2009). Mycophenolate mofetil et al. (2009). Risk factors for recurrence of
for autoimmune hepatitis: a single practice autoimmune hepatitis after liver
experience. Dig. Dis. Sci. 54 (11): transplantation. Liver Transpl. 15 (10):
2519–2522. 1254–1261.
86 Jothimani, D., Cramp, M.E., and Cross, T.J. 95 Sebode, M., Hartl, J., Vergani, D., and
(2014). Role of mycophenolate mofetil for Lohse, A.W. (2018). International
the treatment of autoimmune hepatitis: an Autoimmune Hepatitis Group.
observational study. J. Clin. Exp. Hepatol. 4 Autoimmune hepatitis: from current
(3): 221–225. knowledge and clinical practice to future
87 Zizzo, A.N., Valentino, P.L., Shah, P.S., and research agenda. Liver Int. 38 (1): 15–22.
Kamath, B.M. (2017). Second‐line agents in 96 Krishnamoorthy, T.L., Miezynska‐Kurtycz,
pediatric patients with autoimmune J., Hodson, J. et al. (2016). Longterm
hepatitis: a systematic review and meta‐ corticosteroid use after liver
analysis. J. Pediatr. Gastroenterol. Nutr. 65 transplantation for autoimmune hepatitis is
(1): 6–15. safe and associated with a lower incidence
88 Burak, K.W., Swain, M.G., Santodomingo‐ of recurrent disease. Liver Transpl. 22 (1):
Garzon, T. et al. (2013). Rituximab for the 34–41.
treatment of patients with autoimmune 97 Stirnimann, G., Ebadi, M., Czaja, A.J., and
hepatitis who are refractory or intolerant to Montano‐Loza, A.J. (2019). Recurrent and
standard therapy. Can. J. Gastroenterol. 27 de novo autoimmune hepatitis. Liver
(5): 273–280. Transpl. 25 (1): 152–166.
123
Primary Biliary Cholangitis

Alessio Gerussi and Marco Carbone
Division of Gastroenterology, University of Milan Bicocca, Milan, Italy
Abstract
Primary biliary cholangitis (PBC) is a complex disease resulting from the combination of genetic
susceptibility, unknown environmental exposures and epigenetic alterations. The diagnosis is based
on persistently elevated cholestatic liver function tests and PBC-specific serology. The majority of
PBC patients are asymptomatic at diagnosis; when symptomatic, they usually report fatigue, pru-
ritus, and dry eyes and mouth. The introduction of ursodeoxycholic acid (UDCA) has dramatically
changed the disease course of individuals with PBC and reduced the rate of liver transplantation
(LT). PBC has a relatively predictable natural history and accurate prognostic models have been
developed. PBC is a good indication for LT considering the excellent long-term post-transplant
survival, better than for other indications, although the disease recurs almost universally.
Keywords cholestasis; liver transplantation; primary biliary cholangitis; risk assessment;

ursodeoxycholic acid
Key Points
●● Primary biliary cholangitis (PBC) is a ●● The introduction of ursodeoxycholic acid
complex disease resulting from the (UDCA) has dramatically changed the
combination of genetic susceptibility and disease course of individuals with PBC
as yet unknown environmental exposures and reduced the rate of liver transplanta-
and epigenetic alterations. tion (LT).
●● The diagnosis is based on persistently ele- ●● The majority of patients have improve-
vated cholestatic liver function tests and ment in liver biochemistry under UDCA,
PBC‐specific serology. In those seronega- which translates into good outcomes.
tive, a liver biopsy is usually required for Those with inadequate response to UDCA
diagnostic purposes. now have several licensed/evidence‐based
●● The majority of PBC patients are asymp- options that should be used as add‐on
tomatic at diagnosis; when symptomatic, therapies to UDCA.
they usually report fatigue, pruritus, and ●● PBC has a relatively predictable natural
dry eyes and mouth. history and accurate prognostic models
have been developed. These can aid clini- ritus; fatigue is not improved by LT and
cians in the decision‐making process therefore is not an indication.
regarding escalation of management and ●● PBC is a good indication for LT consid-
therapeutic approach. ering the excellent long‐term post‐
●● Indications for LT in PBC are similar to transplant survival, in general better than
those for other chronic liver diseases. An for other indications, although the disease
exceptional indication is intractable pru- recurs almost universally.
Introduction and Definition agreed to rename the disease “primary bil-

iary cholangitis” as it is known nowadays.
Primary biliary cholangitis (PBC) is an
autoimmune liver disease characterized by
destructive cholangitis affecting the small Epidemiology
intrahepatic bile ducts, leading to chronic
cholestasis and progressive fibrosis. The etio- Epidemiologic studies report heterogeneous
pathogenesis is not well understood. It is clin- incidence rates of 0.33–5.8 per 100 000 inhab-
ically characterized by chronic cholestasis, itants per year, and prevalence rates of 1.9–
serologic reactivity to anti‐mitochondrial 40.2 per 100 000 inhabitants. Such figures, in
antibodies (AMAs) or specific anti‐nuclear particular the prevalence, have increased in
antibody (ANA), and histologic evidence of the last few decades [2]. Improvements in
chronic non‐suppurative, granulomatous, diagnostic tools, increasing disease aware-
lymphocytic cholangitis. Many patients even- ness, and digitalized patient registration with
tually develop end‐stage liver disease with ease of case finding, along with improved
attendant need for liver transplantation (LT). survival, have likely contributed to the rising
The first report of the disease dates back to prevalence rates.
1851 by Addison and Gull who described a The disease has been described world-
clinical picture of progressive jaundice in the wide, even though North America and
absence of mechanical obstruction of the northern Europe have shown the highest
large bile ducts. Ahrens and colleagues in incidence and prevalence rates. It is not clear
1950 published the first detailed description whether there is a true variation in disease
of 17 patients with this condition and coined prevalence among populations of different
the term “primary biliary cirrhosis.” In 1959, geographic areas and of different ethnicity
Dame Sheila Sherlock reported a further or this is a consequence of a difference in
series of PBC patients and recognized that study quality.
the disease could be diagnosed in a pre‐ PBC is considered a prime example of the
cirrhotic stage and proposed the term female preponderance in autoimmunity,
“chronic intrahepatic cholestasis” as a more with a female to male ratio of up to 9 : 1, con-
appropriate description of this disease [1]. firmed by large cohort studies, although
However, this nomenclature failed to gain some recent data, using administrative reg-
acceptance and the term “primary biliary cir- istries, suggest an increasing male preva-
rhosis” lasted for decades. In 2014, to correct lence. Major defects of sex chromosomes,
the inaccuracy and remove the social stig- such as enhanced monosomy X in female
mata of cirrhosis as well as all the misunder- patients and enhanced Y chromosome loss
standing, disadvantages and discrimination in male patients, have been described and
emanating from this misnomer in daily life might well explain the greater female pre-
for patients, international liver associations disposition to develop PBC.
Chapter 7 Primary Biliary Cholangitis 125
Even though there are case reports of been highlighted, all of them located in
patients diagnosed at the age of 15 or 93, the immune‐related regions.
typical disease onset is between 30 and Environmental factors have been linked to
60 years. It is estimated that prevalence of immune tolerance breakdown in PBC. In a
PBC in women over the age of 45 years could genetically predisposed patient, an unidenti-
exceed 1 in 800 individuals. Women who fied environmental agent, such as bacterial
present at a younger age (<45 years) are often infection or xenobiotics, is probably able to
more symptomatic and more likely to fail trigger the pathogenetic cascade of events.
standard treatment with ursodeoxycholic PBC patients show higher prevalence of
acid (UDCA), which may translate into a urinary tract infections, mainly associated
higher risk of disease progression. Men are with Escherichia coli. Molecular mimicry
often diagnosed with PBC later in the disease between epitopes of E. coli and pyruvate
course compared to women, with more dehydrogenase E2 subunit (PDC‐E2), the
advanced disease at presentation, they show dominant mitochondrial autoantigen of
poorer biochemical response to UDCA PBC, has been described [4].
therapy and higher risk of developing hepa- Considering that the liver has a major role
tocellular carcinoma (HCC). in chemical detoxification, hepatocytes and
cholangiocytes are highly exposed to xenobi-
otics, which might affect liver pathobiology
Etiopathogenesis and trigger immune response. Although still
controversial, some epidemiologic studies
PBC is the archetypal autoimmune disease have linked the use of cosmetics and smoking
characterized by the presence of disease‐ with PBC. The 2‐octynoic acid, present in
specific autoantibodies toward a mitochon- cosmetic products, has been shown to induce
drial target [3], a genetic architecture antibodies to PDC‐E2 (AMA) and PBC‐like
pointing to immunoregulatory pathways, a lesions in non‐obese diabetic mice immu-
histologic pattern of chronic granulomatous nized with this substance.
lymphocytic cholangitis of the middle and Whatever the trigger is, it leads to loss of
small-size bile ducts, and a high proportion PDC‐E2 immune tolerance and self‐epitopes
of patients having at least one concurrent on the cholangiocytes are targeted by an
autoimmune condition. On the other hand, it autoimmune reaction. Even though the
is noteworthy that immunosuppressant PDC‐E2 protein is ubiquitous, only cholan-
drugs are not effective in PBC. giocytes are selectively damaged in PBC.
PBC is a complex trait deriving from the Defective clearance of apoptotic bodies
combination of genetic susceptibility, containing intact and antigenically reactive
environmental exposures, and epigenetic
PDC‐E2 released from dying cholangiocytes,
alterations. The prominence of a genetic but not other cell types, has been proposed
component in PBC has been highlighted by as a possible cause of the tissue specificity of
several studies. Twin studies have shown the disease.
higher PBC frequency in monozygotic twins PBC is considered a Th1‐dominant auto-
than in dizygotic ones. Genome‐wide immune response as Th1 effector cells are
association studies (GWAS) have identified critical for the development of cholangitis.
dozens of single‐nucleotide polymorphisms Also described is a skewing toward Th1/
(SNPs) associated with disease suscepti- Th17 activation, seen in the cytokine profile
bility. Among them, the strongest [interleukin (IL)‐12/Th1 and IL‐23/Th17] in
association has been found within the the liver of patients with PBC along with
human leukocyte antigen (HLA) class II downregulation of Th2 markers (e.g. IL‐10).
loci, but many non‐HLA SNPs have also While the IL‐12/Th1 pathway may be a
strong determinant in initiation of the dis- Table 7.1 Symptoms and signs in PBC.
ease process, IL‐23/Th17 pathway plays a
dominant role in the subsequent effector Symptoms Signs
stages of the disease. Cells of innate immu-
Pruritus Skin hyperpigmentation
nity are also involved, such as CD1d‐positive
Fatigue Scratching marks or
natural killer T (NKT) cells, whose fre- Sicca syndrome excoriations
quency in the liver of patients with PBC is (xerostomia and Xanthelasma/xanthoma
higher than normal, or NK cells, that are xerophthalmia) Kayser–Fleischer rings
upregulated as well. The association with Right upper Sclerodactyly
quadrant pain Teleangiectasia
granulomas is also indirect evidence of
Raynaud’s
the role of innate immunity, especially phenomenon
macrophages. Arthralgia
Recently, a novel pathogenic theory pro-
posed that the immune dysregulation in PBC
is a downstream process, which follows pri- strong correlation between severity of disease
mary biliary damage. Decreased expression and symptoms, although asymptomatic
and activity of the anion exchanger 2 (AE2) patients are more likely to have early disease.
on the cholangiocytes might lead to impaired Fatigue is a non‐specific symptom affecting
biliary bicarbonate secretion with conse- 50–78% of patients. It can be debilitating,
quent cholestasis and cholangiocyte injury, with a huge impact on quality of life. Its path-
followed by immune tolerance breakdown ogenesis is still unknown, and it is quite chal-
and autoimmune responses toward the sus- lenging to explore its specificity and to treat.
ceptible cholangiocytes, resulting in pro- Comorbidities that could contribute to, or
found biliary damage. Epigenetic studies are worsen fatigue, such as depression, hypothy-
currently ongoing to test this hypothesis. roidism, anemia, obesity, or medication-side
effects, should be promptly identified and
treated.
Clinical Presentation Pruritus can develop at any stage of the dis-
ease, does not correlate with progression of
Asymptomatic Patients liver disease, and may even improve or disap-
Most patients are currently diagnosed when pear as disease becomes more advanced. It is
asymptomatic, having been referred to the usually reported by over 70% of patients, and
hepatologist for abnormal liver function tests it is typically mild to moderate in intensity.
(LFTs), mostly raised gamma‐glutamyltrans- Given the impact on quality of life and night
peptidase (GGT) or alkaline phosphatase sleep, pruritus is correlated with fatigue.
(ALP), performed for annual screening blood It can rarely be severe, non‐responsive to
tests. Other frequent scenarios include medical therapy, and requiring LT. Pruritus
screening of patients with non‐liver auto- in PBC is characteristically intermittent,
immune diseases (e.g. rheumatoid arthritis) worse at night, and improves during summer.
or investigation of elevated cholesterol, Most patients with xerostomia and
evaluation of itch or unresolved cholestasis xerophthalmia have sicca syndrome rather
postpartum. than primary Sjögren syndrome. Other,
less often referred symptoms incluse non‐
progressive right upper quadrant pain,
Symptomatic Patients
Raynaud phenomenon, and arthralgia.
When symptomatic, PBC patients usually Although decompensated liver disease is
report fatigue, pruritus, and sicca complex. less common now, occasionally patients pre-
Symptoms and signs commonly found in sent with jaundice, ascites, or variceal hem-
PBC are summarized in Table 7.1. There is no orrhage at diagnosis. Jaundice is present only
in the later stage of the disease, unless the Triglycerides levels are normal or slightly
patient has a rare form of premature ducto- elevated. It is still unclear whether hyper-
penic variant of disease. lipidemia increases the cardiovascular risk
The physical examination is often unre- in PBC.
markable in early PBC. Signs that should be As disease progresses, hepatosplenomeg-
sought include: aly, spider nevi, palmar erythema, muscle
wasting, ascites, edema, and jaundice can be
●● skin hyperpigmentation, which reflects
found. Of note in PBC, cirrhosis is not neces-
deposition of melanin rather than bilirubin;
sarily required for portal hypertensive
●● scratching marks or excoriations from
complications to develop since this can be
intractable itch, which are more common on
secondary to a pre‐sinusoidal portal hyper-
the trunk and legs, and can lead to scarring;
tension caused by nodular regenerative
●● xanthelasma, subcutaneous deposits of
hyperplasia, although the majority of patients
cholesterol typically found around the eyes
with varices are in fact cirrhotic.
and with a yellow color;
●● more rarely xanthoma, cholesterol deposits
potentially found anywhere in the body,
but more often present around tendons Diagnosis
and bone prominences.
Diagnosing PBC is generally straightforward
Kayser–Fleischer rings are seen very occa- [5]. The foundations for a definite diagnosis
sionally in severe cholestasis and result from are persistently elevated cholestatic LFTs,
copper retention. Signs of associated auto- usually defined as elevated ALP for more
immune conditions such as scleroderma, than 6 months, but patients may just have
sclerodactyly or telangiectasia can be also elevated GGT; positive serology, i.e. AMAs
present. with a titer over 1 : 40 or specific ANAs such
It is reasonable that screening tests for as gp210 and sp100; and liver histology show-
celiac disease, thyroid disorders, and Sjogren ing florid bile duct lesions and, in advanced
syndrome are performed at baseline in all disease, ductopenia. It is necessary to obtain
PBC patients. a patient history excluding recent drug
When prolonged jaundice is present, fat‐ exposure, and cross‐sectional imaging ruling
soluble vitamin (A, D, E, K) deficiency may out bile duct obstruction.
occur, with the main problem being the Given the high specificity of serologic
development of osteoporosis, which is gen- markers, liver biopsy is not necessary for the
erally exacerbated by postmenopausal bone diagnosis of PBC; however, it is still required
loss. Night blindness is now exceptional. when PBC‐specific antibodies are absent or
Steatorrhea has been reported in PBC when coexistent autoimmune hepatitis (AIH)
patients with scleroderma, possibly due to or non‐alcoholic steatohepatitis (NASH) is
decreased bile acid delivery coupled with suspected. Liver biopsy can be useful for
exocrine pancreatic insufficiency, celiac dis- staging the disease for fibrosis and ductope-
ease, and bacterial overgrowth. nia. Finally, it may also be appropriate in the
Hyperlipidemia is another cholestasis‐ presence of other extrahepatic comorbidities.
driven complication, present in up to 85% of
patients at diagnosis. It is characterized by
Biochemical Tests
high levels of total cholesterol, mainly due to
elevation of high‐density lipoproteins (HDLs) The most common scenario is chronic chole-
in the early stages; as disease progresses, HDL stasis in the absence of jaundice, and with
levels fall and low‐density lipoproteins (LDL) transaminases less than three times the
levels rise, due to the pathologic loss of LDL upper limit of normal (ULN). The level of the
receptors and consequent LDL clearance. liver enzymes themselves does not readily
predict severity of the disease, although the techniques are used. Antibodies against the
subsequent biochemical response to therapy major M2 components, i.e. PDC‐E2, 2‐oxo-
is relevant to outcome. glutarate dehydrogenase complex (OGDC),
ALP is more specific for cholestasis than and branched‐chain 2‐oxoacid dehydrogenase
GGT, which can be raised secondary to drug, complex (BCOADC), may then be identified
alcohol or fat. However, when raised, GGT is using MIT3‐based enzyme‐linked immuno-
helpful for confirming that ALP is of biliary sorbent assay (ELISA).
origin (as ALP may also come from bone, Serial monitoring of AMA values is of no
gut, and placenta). clinical value, and AMA titers can fall on
Raised serum transaminases do not neces- treatment over time. AMA has been found in
sarily imply an overlap with AIH, and criteria less than 0.5% of healthy controls, but in up
to diagnose PBC–AIH variant syndrome to 20% of patients with AIH.
have been developed in order to avoid ANAs are present in approximately 30% of
over-diagnosis and over-treatment. Moreover, patients with PBC. The anti‐sp100 (nuclear
it is not rare to see normalization of transami- dots) and anti‐gp210 (perinuclear rims) are
nases together with cholestatic enzymes after specific for PBC (>95%) and are useful in
successful choleretic treatment, revealing how the diagnosis of AMA‐negative patients.
these abnormalities could be part of the Positivity for anti‐gp210 has been associated
biochemical picture of PBC. with aggressive disease and progression to
Another biochemical feature of PBC is liver failure in individuals with PBC.
increased serum IgM concentrations, prob- Other autoantibodies can be found in
ably driven by epigenetic changes, which is PBC, such as anti‐thyroid autoantibodies,
frequent but not specific to PBC. Serum rheumatoid factor, anti‐platelet antibodies,
levels of IgG may also be elevated, even in the anti‐histone antibodies, anti‐centromere anti-
absence of concomitant AIH, more fre- bodies, and anti‐smooth muscles antibodies.
quently in AMA‐negative cases and/or in Their presence is not necessarily related to a
individuals with advanced disease. coexisting autoimmune condition.
Attention should be given to the platelet
count, which is a useful, albeit rough, marker
Liver Biopsy
for advanced liver disease, and can be used to
identify patients who should be considered Liver biopsy for the diagnosis and staging
for esophageal varices screening. of PBC lost favor when it was appreciated
Serum bilirubin is a surrogate marker of that there was an uneven distribution of duct
advanced disease or severe ductopenia. lesions and fibrosis within the liver. The
Raised bilirubin and low albumin levels, widespread availability of non‐invasive mea-
together with prolongation of prothrombin sures of fibrosis means that liver biopsy for
time, are features of advanced cirrhosis. staging of PBC is somewhat obsolete.
Serum bile acids can be elevated but should However, liver biopsy does remain useful
not be determined on a routine basis since in certain settings. The main indications are
their levels do not provide additional value to to confirm the diagnosis of PBC when PBC‐
the diagnostic process. specific antibodies are absent and to confirm
a diagnosis of PBC with AIH features (i.e.
overlap PBC–AIH). Liver biopsy is also use-
Autoantibodies
ful for assessing the relative contribution of
AMA is the hallmark of PBC and is present in each liver injury when a comorbid liver dis-
up to 90–95% of patients. A titer of more than ease such as NAFLD is present. In patients
1 : 40 is considered positive. In up to 10% with inadequate response to UDCA, liver
of patients, AMAs are absent or present biopsy may provide the explanation and
only in low titer when immunofluorescence could undoubtedly inform risk stratification.
(a)
Lymphoplasmacytic
portal infiltrate
(b)
Ductular reaction
Vanishing of native duct

in the portal tract
Figure 7.1 Photomicrograph of stage 2 PBC. (a) Mononuclear inflammatory cells expand the portal tract with
some disruption of the limiting plates (interface hepatitis). The bile duct is surrounded by inflammatory cells.
No evidence of fibrosis (hematoxylin–eosin, ×100). (b) Loss of bile ducts and ductular reaction (cytokeratin 7
staining, ×100).
For example, it may identify a previously ductopenia and fibrosis (Figure 7.1). The
unsuspected variant syndrome, steatohepati- Ludwig and Scheuer scoring systems have
tis, or interface hepatitis of moderate or historically been used to stratify four (1–4)
greater severity. It is also useful in AMA‐ and “stages” of PBC, with stage 4 indicating the
ANA‐specific antibody‐negative cholestatic presence of cirrhosis. In the new system of
patients to indicate an alternative process, Nakanuma, the stage of disease is based on
such as sarcoidosis, small duct primary fibrosis, bile duct loss, and features of chole-
sclerosing cholangitis (PSC), or adult idio- stasis (i.e. deposition of orcein‐positive gran-
pathic ductopenia. ules), whereas the grade of necroinflammatory
PBC is characterized by chronic non‐sup- activity is based on cholangitis and inter-
purative inflammation, which surrounds and face hepatitis. The accumulation of orcein‐
destroys interlobular and septal bile ducts. positive granules occurs evenly across the
Progressive damage of the bile duct leads to PBC liver, which means that staging using
the Nakanuma system is more reliable rejection, ischemic cholangitis, and infectious
regarding sampling variability. cholangitis (e.g. AIDS) should be considered,
although these can be easily excluded based
on the clinical context.
Imaging
A diagnosis of PBC requires the exclusion

of extrahepatic causes of cholestasis or Natural History
liver neoplasms by abdominal ultrasound.
In seronegative patients, careful exclusion of AMA positivity can sometimes be found in
PSC by means of magnetic resonance chol- the presence of normal liver biochemistry.
angiography (MRCP) is essential. Abdominal This typically occurs after screening of family
ultrasound is also useful for identifying signs members of patients affected by PBC or when
and complication of advanced liver disease, AMA is detected following a wide panel of
such as focal liver lesions, splenomegaly, and autoantibodies performed for clinical suspi-
ascites. Detection of lymph nodes at hilum is cion of an autoimmune disease. Only one
frequent and does not necessarily suggest the patient in six will develop PBC within 5 years
presence of a malignancy. from the detection of AMA positivity. A sen-
In cross‐sectional studies, both enhanced sible approach would be to follow isolated
liver fibrosis test (ELF™) and transient elastog- AMA‐positive subjects at a primary‐care
raphy (e.g. FibroScan) have shown accuracy in level, with LFTs performed on an annual or
determining disease stage as confirmed by biannual basis. If LFTs become abnormal,
biopsy. There are no longitudinal data at the re-evaluation by a hepatologist is sensible, and
individual patient level regarding change of introduction of UDCA is suggested when the
these parameters over time; and it is not clear abnormality persists for more than 6 months.
whether they express disease severity rather In the era before UDCA therapy was intro-
than disease stage, and whether they add pre- duced, PBC was commonly diagnosed at an
dictive value to the liver biochemistry. While advanced stage with jaundice and liver
their use is increasing in clinical practice, decompensation. Pre‐UDCA era studies
particularly transient elastography, they are reported a median survival for symptomatic
not yet linked to paradigms for management patients between 8 and 10 years; conversely,
and intensity of patient follow‐up. asymptomatic patients developed symptoms
HCC surveillance should be performed within 5–7 years of follow‐up in 60% of cases
in those with cirrhosis as in other liver and had a longer median survival, close to
conditions. 15 years, which was however worse than
healthy controls matched for age and gender.
Based on cohort studies with follow‐up
Differential Diagnosis
biopsies, 31% and 50% of patients diagnosed
A careful evaluation is required when PBC‐ at stage 1 and 2, respectively, become cir-
specific autoantibodies are negative. Liver rhotic within 4 years. The rate of progression
biopsy might be useful. In case of normal to advanced fibrosis/cirrhosis in untreated
small bile ducts at histology assessment, patients is approximately 30% per year
cholestasis secondary to drug‐induced liver (Figure 7.2). The rate of histologic progres-
injury, systemic illnesses (e.g. virus), or sion, assessed in three large groups of
benign recurrent intrahepatic cholestasis untreated patients, showed that the median
should be sought. In case of abnormal small time to develop extensive fibrosis was 2 years
bile ducts, small‐duct PSC, sarcoidosis, idio- and the probability of remaining in the early
pathic adulthood ductopenia, and prolonged stage was 29% after 4 years of follow‐up. The
drug‐induced cholestasis are more likely; rate of development of esophageal varices,
also, graft‐versus‐host disease, liver allograft assessed in a prospective study of 256
1
Probability of remaining free of extensive
0.75
fibrosis or cirrhosis
0.5
0.25
0 2 4 6 8 10
Years
Figure 7.2 Probability of remaining free of extensive fibrosis or cirrhosis in UDCA‐ and placebo‐treated
patients with early‐stage PBC. At 4 years, the probability for UDCA‐treated patients to remain in the early stage
of disease is 76% (95% CI 58–88%), as compared with 29% (95% CI 15–52%) in placebo‐treated patients.
•
Vertical bars indicate 95% confidence interval (CI). (—O—), UDCA; (— —), placebo; (‐ ‐ ‐O‐ ‐ ‐), natural history
(which derives from the study of Locke et al. [6]). Source: Corpechot et al. [7]. Reproduced with permission of
John Wiley and Sons.
patients during a median follow‐up of Patients receiving UDCA therapy have a

5.6 years, was estimated to be 31%. These delayed rate of histologic progression to cir-
older studies should now be interpreted with rhosis. In a study from the USA, the rate of
caution, especially regarding the described progression to cirrhosis at 6 years was 13% in
differences between symptomatic and patients receiving UDCA and 49% in the
asymptomatic patients, as the definition of control group of untreated patients. In the
symptomatic included the presence of fea- French UDCA trial, the therapy was associ-
tures such as jaundice or ascites, which ated with a fivefold lower progression rate
would more accurately be regarded as fea- from early‐stage disease to extensive fibrosis
tures of advanced disease, the association of or cirrhosis, being 7% per year with UDCA
which with poor prognosis is unsurprising. compared with 34% per year under placebo.
Nonetheless, taken altogether, these data At 4 years, the probability of UDCA‐treated
clearly indicated that PBC, if untreated, is a patients remaining in early‐stage disease was
progressive disease [8]. 76% compared with 29% in the placebo
group. In a prospective study of 180 patients
followed up for 4 years the risk of developing
Disease Course in the UDCA Era varices was 16% for the UDCA‐treated
patients and 58% for those receiving the
The introduction of UDCA has dramatically placebo.
changed the pattern and course of the dis- Long‐term observational studies and prob-
ease. Numerous trials and observational abilistic models have been used to study the
studies have demonstrated its efficacy on effect of UDCA on survival. In 262 patients
liver biochemistry, histologic progression, who had received UDCA 13–15 mg/kg daily
and transplant‐free survival [9]. for a mean of 8 years, the LT‐free survival
rates were 84% and 66% at 10 and 20 years, Risk Stratification

respectively. In early‐stage patients, 6% were
predicted to progress to LT or death after All patients with PBC, who are generally on
10 years and 22% by 20 years. In contrast, the UDCA therapy, should be evaluated for their
probability of death or LT was significantly risk of developing complications of end‐stage
increased in UDCA‐treated patients in the disease and consequently their potential
advanced stages of the disease. need for additional therapies [10]. Several
Among the UDCA‐treated patients, the parameters can be used for risk stratifica-
degree of liver biochemistry improvement tion at presentation or at any time during
(i.e. the UDCA response) identifies patients treatment. They consist of standard labora-
with different long‐term prognoses. LT‐free tory investigations, demographics, serologic
survival of patients with normal or near‐ profiles, non‐invasive markers of fibrosis,
normal liver biochemistry on UDCA is sim- and histologic features.
ilar to that of the general population, whereas Historically, the Mayo PBC risk score,
it is significantly reduced in those with including age, serum bilirubin, albumin, pro-
abnormal liver biochemistry on treatment thrombin time, and the presence of fluid
(Figure 7.3). retention and/or use of diuretics, was devel-
HCC is infrequent in PBC. Recent large‐ oped in late‐stage PBC patients to predict
scale cohort studies highlighted that the lack outcome. Similarly, non‐PBC‐specific scor-
of UDCA response after 12 months of therapy ing systems such as the Model for End‐Stage
and male sex are associated with increased Liver Disease (MELD) are of value when the
future risk of developing HCC in PBC. disease is advanced.
The UDCA response is a major predictor
of long‐term outcome. Criteria of UDCA
response and sophisticated disease‐specific
risk scores assessing UDCA response and
Proportion surviving
0.8
disease stage have been recently developed
to predict prognosis in patients with PBC.
They have been reproduced widely across
0.4
large cohorts and their use is recommended
for all patients after 1 year of UDCA therapy.
However, it is still unclear how to implement
0.0
risk scores in clinical practice.
0 2000 6000 10000 The two most important parameters in eval-
Follow-up (days) uating response to UDCA are ALP and total
bilirubin. Qualitative and quantitative defini-
Figure 7.3 Kaplan–Meier plot for survival according
to Paris I criteria. Kaplan–Meier plots from time‐to‐
tions of UDCA response have been developed,
event analysis in which “failure” is defined as liver based on changes in bilirubin, transaminases
transplant (LT) for PBC, death from PBC‐related liver and ALP after 6–24 months of treatment with
failure, or (for surviving non‐LT recipients) a serum UDCA at 13–15 mg/kg daily. The first criteria
bilirubin level of 100 μmol/l or greater. In the time‐ of UDCA response, proposed in 2006, were
to‐event analysis, the start point is the date of first
presentation and the end point is the date of failure.
the Barcelona criteria, which are the only ones
Surviving non‐LT recipients with bilirubin levels less that consider the change in ALP over
than 100 μmol/l were censored at the date of their treatment. These define response as an ALP
most recent blood tests. The plot shows survival decrease greater than 40% from baseline values
curves for patients who met the Paris I criteria for or normal levels after 1 year of treatment. In
treatment response after 12 months of treatment
with UDCA (blue line) versus those who did not
the Spanish cohort, the observed LT‐free
meet criteria for treatment response (red line). survival was higher than that predicted by the
Source: adapted from Carbone et al. [9]. Mayo model and, for those treated at an early
stage of disease, similar to that estimated for ilirubin. A comprehensive list of all available
b
the control population. The Paris I criteria biochemical response criteria is outlined
defined response as an ALP <3 ULN, aspartate in Table 7.2.
aminotransferase (AST) <2 ULN, and normal Such dichotomous criteria are easy‐to‐use
bilirubin after 1 year of UDCA. In the French prognostic tools for clinical practice, but are
cohort, responders had a 10‐year LT‐free simplistic in identifying only two classes of risk
survival of 90% compared with 51% of non‐ (high and low). Furthermore, most of them do
responders. A combination of the Paris‐I cri- not take into account other prognostic vari-
teria and the AST‐to‐platelet ratio index ables, such as markers of disease stage.
(APRI), a surrogate marker of disease stage, A refinement of stratification tools has
applied after 1 year of UDCA improves the been possible thanks to the availability of
predictive power of the UDCA‐response cri- large cohorts, and this has led to the
teria alone. A further evolution were the Paris development of two important continuous
II criteria, which better fit early‐stage disease, scores: the Globe score and the UK‐PBC risk
and define response as ALP <1.5 ULN or score. These scoring systems derive from
AST <1.5 ULN or normal bilirubin. The large multicenter cohorts, are externally vali-
Toronto criteria, developed to predict histo- dated and convey probability of survival/risk
logic stage progression, stratify based on ALP of events on a continuous as opposed to
under treatment (<1.67 ULN), whereas the dichotomous scale, with better predictive
Rotterdam criteria are focused toward liver performances compared with the response
function/stage and include albumin and criteria.
Table 7.2 Biochemical response criteria for risk stratification in UDCA‐treated PBC patients.
Response definitions Definition and parameters Type of Number

and prognostic models evaluated prediction of patients Centers
Paris I, 2008 ALP <3× ULN Dichotomous 292 Single center

AST <2× ULN
Bilirubin ≤1 mg/dl after
1 year
Barcelona, 2006 More than 40% decrease of Dichotomous 192 Single center
ALP or normalization after
1 year
Toronto, 2010 ALP ≤1.67× ULN after Dichotomous 69 Single center
2 years
Paris II, 2011 ALP ≤1.5× ULN Dichotomous 165 Single center
AST ≤1.5× ULN
Bilirubin ≤1 mg/dl after
1 year
Rotterdam, 2009 Normalization of Dichotomous 375 Single center
abnormal bilirubin and/or
albumin after 1 year
GLOBE score, 2015 Age, bilirubin, albumin, Continuous 2488 15 tertiary centers
ALP, platelet count
UK‐PBC score, 2016 Bilirubin, ALT/AST, ALP, Continuous 1916 155 secondary and
platelets count, serum tertiary centers
albumin
ALT, alanine aminotransferase; ALP, alkaline phosphatase; AST, aspartate aminotransferase; ULN, upper limit of
normal.
A drawback of most of these prognostic have relevant prognostic value. The Nakanuma
tools is that they ignore the baseline (pre- system has been shown to reflect liver
treatment) liver biochemistry and what is dysfunction before UDCA treatment and to
defined as treatment response may actually correlate well with the future development of
be just disease severity, with many patients cirrhosis‐related complications. Liver biopsy
having an indolent disease course indepen- may therefore be valuable for risk stratifica-
dently of treatment. tion purposes in PBC. Considering the risk–
More recently, a joint UK/Italian cohort benefit ratio of this invasive procedure, its use
study identified baseline parameters, including may be advocated in patients with poor
younger age, elevated ALP, bilirubin and trans- biochemical response to UDCA.
aminases at diagnosis, the trend of ALP before
therapy, and a longer waiting time before
receiving UDCA, as major predictors of future Treatment
treatment failure to UDCA (Figure 7.4). These
variables were included in a predictive score Medical therapy of PBC targets disease pro-
(UDCA‐response score) that allows risk strati- gression and symptom control. The back-
fication of patients at an early stage, i.e. at bone of treatment for PBC is bile acid. UDCA
diagnosis and provides a basis for exploring has been the only drug available for two
alternative approaches to treatment stratifica- decades, although more recently obeticholic
tion in PBC patients, for example the early acid (OCA), a semisynthetic hydrophobic
identification of patients who would be eligible bile acid analog, has been licensed in patients
for second‐line therapies before waiting for with failure to, or intolerant of, UDCA
treatment failure under UDCA, with potential (Figure 7.5). Several other agents have been
impact on disease course. studied, including immunosuppressants, but
Liver biopsy for diagnostic purposes is no robust evidence of benefit is lacking.
longer recommended, and transient elastog-
raphy can now be used as an alternative option
for detecting severe fibrosis or cirrhosis. Ursodeoxycholic Acid
However, histologic parameters such as UDCA was the first drug to be approved for
fibrotic stage, the degree of lymphocytic inter- use in PBC and remains the first‐line
face hepatitis, and the stage of ductopenia therapy for all patients with PBC [12].
Figure 7.4 Relationship between age at diagnosis and the probability of UDCA response (left) and between
treatment time lag (interval from diagnosis to start of UDCA treatment) and the probability of UDCA response
(right). Source: adapted from Carbone et al. [11].
Figure 7.5 Flowchart showing the guideline‐ UDCA (13–15mg/kg)

supported, response‐guided approach of
treatment in PBC.
YES
Continue UDCA Biochemical reponse
20–40% NO
Add second-line therapy
Licensed Off-labels
OCA Bezafibrates
Budesonide
UDCA normally accounts for about 4% of Although nearly all patients with PBC have
bile acids, but with pharmacotherapy it some improvement in serum ALP, 20–40%
becomes the predominant bile acid. UDCA have an inadequate biochemical response to
exert its choleretic effect though complex UDCA and remain at risk of disease progres-
posttranscriptional molecular mechanisms sion. Definition of inadequate response to
and the stabilization of the “biliary HCO3– UDCA is based on the level of improvement
umbrella.” UDCA protects cholangiocytes of the LFTs, which can be assessed using
against the cytotoxicity of hydrophobic bile either qualitative definitions (e.g. Barcelona
acids, stimulates the hepatobiliary secretion and Paris criteria) or quantitative scoring
of bile acids, increases the hydrophilicity systems computed from continuous parame-
index of the circulating bile pool, and ters (i.e. Globe and UK‐PBC risk score).
may have immunomodulatory and anti‐ Patients with PBC who have an inadequate
inflammatory effects. UDCA has been response to UDCA or those few (<3%) who
shown to improve the liver biochemistry, are intolerant to UDCA should be candidates
slow down histologic progression, and for second‐line therapies.
improve LT‐free survival.
The optimum dose of UDCA is 13–15 mg/
Obeticholic Acid
kg, which can be given as a single oral daily
dose or in divided doses when tolerability is OCA is approved for patients with an inade-
an issue. Side effects are negligible, mainly an quate response to UDCA or for patients
increase in weight of a few kilograms within unable to tolerate UDCA. OCA is a derivative
the first year and, less frequently, some degree of chenodeoxycholic acid, the naturally
of flatulence, loose stools and hair loss. occurring ligand of the farnesoid X receptor
Concomitant use of bile acid‐binding seques- (FXR), which mediates the synthesis and
trants such as colestyramine or antacids can enterohepatic circulation of bile acids. OCA
alter absorption, and UDCA should be taken is 100 times more potent as an FXR ligand
1 hour before or 4 hours after these agents. compared with chenodeoxycholic acid. In
UDCA is not teratogenic, and even though the liver, FXR activation decreases the
evidence‐based studies of its safety during conversion of cholesterol to bile acids and
pregnancy are lacking, clinicians should not increases the transport of bile acids out of
be worried for patients to continue UDCA hepatocytes. Activation of FXR in the ileum
when used for intrahepatic cholestasis of decreases bile acid reabsorption and
pregnancy (although the drug is not licensed increases expression of fibroblast growth
in pregnancy). UDCA is not effective on factor (FGF)‐19, which acts in the liver to
fatigue, pruritus or bone disease. further decrease bile acid synthesis. OCA
may also have antifibrotic properties and 6 months if the response is judged inade-
may improve portal hypertension. quate by the clinician, provided that the drug
OCA has been evaluated in two Phase II is tolerated. An agreed definition of response
trials and one Phase III trial. The Phase III to OCA has not been produced yet and may
study [13] evaluated the effects of 12 months vary across different countries. The use of
of treatment with OCA in patients with PBC OCA in advanced liver disease needs cau-
who showed an inadequate response to tion. In patients with Child–Pugh score B
UDCA, defined by an ALP 1.67 ULN or and C, a careful risk–benefit evaluation
serum bilirubin above ULN but <2 ULN, or should be performed. The dose should be
were intolerant of UDCA. Patients were reduced to 5 mg per week, to a maximum of
randomized in a double‐blind fashion (1 : 1 : 1) 10 mg twice weekly. Caution is suggested also
to receive placebo daily, OCA 10 mg/day, or in patients with Child–Pugh score A and
OCA 5 mg/day with titration to 10 mg after 6 signs of portal hypertension.
months based on response and tolerability. Exacerbation of pruritus during OCA
The primary end point measured after therapy is dose‐dependent so dose titration is
12 months of treatment was the combination therefore suggested, and the pruritus usually
of ALP <1.67 ULN, with a reduction of at improves with time. A drug holiday for a few
least 15% from baseline, and a normal total weeks and restarting at a lower dose (e.g. 5 mg
bilirubin. More than 90% of the 216 patients every other day) can be helpful. Rifampicin is
with PBC received UDCA as background preferred to colestyramine to avoid reduced
therapy. The primary end point was met by absorption. Modifications in lipid profile are
47 and 46% of patients in the 10 mg and titra- also possible side effects, most frequently
tion groups, respectively, compared with 10% mild reductions in HDL levels, but no data
in the placebo group. There was no change in are available about whether this is associated
non‐invasive measures of fibrosis. Pruritus with a higher likelihood of future cardiovas-
was more common in the OCA group and cular events.
was reported in 68 and 56% of patients in the
10 mg and 5–10 mg OCA‐treatment arms,
Fibric Acid Derivatives
respectively, compared with 38% in the
placebo arm. Discontinuation owing to pru- Fibric acid derivatives, or fibrates, are ago-
ritus occurred in 10% of the 10 mg OCA‐ nists of the peroxisome proliferator‐activated
treated patients compared with 1% in the receptor (PPAR), a nuclear receptor involved
titration group and none in the placebo group. in several metabolic pathways. Fibrates are
After the initial 12 months of the trial, patients licensed for the treatment of hypertriglyceri-
were given the option to continue in an open‐ demia. They also exert potent anticholestatic
label long‐term extension starting at 5 mg effects. Among the fibrates, bezafibrate and
OCA and increasing to 10 mg OCA as toler- fenofibrate, both PPARa selective agonists,
ated. The majority of patients (91%) entered have been extensively studied as therapeutic
the extension phase and results to date agents because of their potential ability to
support ongoing efficacy of OCA through 2 decrease bile acid synthesis and bile acid‐
years. Longer‐term efficacy of OCA and gen- related hepatic inflammation. Numerous
eralisability to the patient population as a small pilot studies and case reports have
whole needs confirmation in prospective shown that fibrates, including fenofibrate in
follow‐up studies. Survival benefit has yet to the USA and bezafibrate in Europe and Japan,
be demonstrated and, for that purpose, a improve liver biochemistries, liver stiffness
long‐term Phase IV randomized trial is cur- measurements, and pruritus in patients with
rently ongoing. PBC. These data have been further sup-
OCA should be started at an initial dose of ported by the recent Phase III randomized
5 mg/day, and increased to 10 mg/day at placebo‐controlled Bezurso trial. Patients
with an inadequate response to UDCA by liver histology, the primary end point of the
Paris‐2 criteria were treated with bezafibrate study, were observed. Full trial data are
400 mg/day or placebo for 2 years. The pri- awaited. Caution must be applied when using
mary end point, i.e. normal total bilirubin, budesonide in patients with cirrhosis and
ALP, AST, alanine aminotransferase (ALT), portosystemic shunts. Evidence is lacking as
albumin, and prothrombin time at 2 years, to how to approach long‐term maintenance
was reached more frequently in the bezafi- therapy in overlap patients stepping down
brate group than in the placebo group (30 vs. from corticosteroids. A reasonable approach
0%, respectively). Whether biochemical is to extrapolate from management regimens
improvements translate to significant decrea for pure AIH with the use of antimetabolites.
ses in liver‐related death or need for LT has
not yet been shown.
Liver Transplantation
The safety of fibrates is a matter of con-
cern. Hepatotoxicity is a known side effect Rates of LT for PBC have declined over the
though reversible, reported in 6% of cases in last decades. Indications are similar to those
the Bezurso trial. Elevated creatinine, likely for other chronic liver diseases, i.e. MELD
related to increased release by muscle rather >15 or bilirubin values steadily rising above
than a real worsening in kidney function, is 3–5 mg/dl (40–85 μmol/l). An exceptional
also reversible. Up to 20% of patients com- indication is intractable pruritus, while
plain of cramps and musculoskeletal pain. fatigue is not improved by LT and therefore is
Taken together, these adverse effects repre- not an indication. Outcomes are favorable,
sent a concern for their use in some patients. with 5‐year patient survival rates better than
Although fibrates are currently still used off‐ for most other indications for LT (80–85%).
label as a second‐line treatment in PBC, PBC recurs after LT, with rates that
their efficacy and ability to improve pruritus increase the longer the follow‐up (20–30%
will likely lead to formal approval for this within 10 years, around 50% within 20 years)
indication. [14]. Since AMAs typically persist after LT,
liver biochemistry could be normal, and
abnormal ALP could be due to other post‐
Budesonide
transplant liver complications; the final
Budesonide is currently used as an off‐label diagnosis of PBC recurrence requires his-
medication in PBC. In PBC patients showing tology. However, recurrent PBC infrequently
interface hepatitis on liver biopsy, some leads to graft loss and current evidence
groups have demonstrated the efficacy of does not suggest an impact on graft or
budesonide in improving liver histology and patient survival after LT. The use of UDCA
biochemistry when used in combination with as a preventive measure in this setting is sup-
UDCA. In 1999 a randomized placebo‐con- ported by non‐randomized studies showing
trolled trial showed the efficacy of budesonide lower rate of PBC recurrence.
9 mg on liver biochemistry and liver histology
at 2 years follow‐up. These data have been
Symptom Management
recently challenged by a yet unpublished
Phase III, randomized, double‐blind, placebo‐ All patients should be evaluated for the
controlled trial (NCT00746486), which presence of symptoms, particularly fatigue
showed improved LFTs in the budesonide and pruritus, bearing in mind that the severity
arm (9 mg/day of budesonide as add‐on of symptoms does not correlate with stage of
therapy), with a mean change in ALP disease. Pruritus is a troublesome issue for
from baseline of −94.5% and −8.9% in the many but not all patients with PBC. If skin is
budesonide group and placebo group, respec- dry, moisturizing agents should be used, and
tively. However, no significant changes in patients should be counseled to avoid hot
baths or showers. Clinicians should exclude Others agents include the antihistamines,
secondary causes, like allergies, hypereosino- which sometimes have a non‐specific anti-
philic conditions, or bile duct obstruction. pruritic effect; this may be due to their seda-
First‐line treatment is colestyramine, a tive properties but they are not recommended
non‐absorbable resin and bile sequestrant. as specific therapy.
Even though evidence supporting its use is Physical approaches, such as nasobiliary
weak, clinical experience is solid. The starting drainage, molecular absorbance recirculating
dosage is usually one sachet (4 g) once or system (MARS), and ultraviolet light therapy
twice a day but a total daily dose of 16 g or are experimental approaches with small series
greater may be required. Clinicians must showing benefit. LT is a rare indication for
carefully respect the pharmacokinetics of the intractable cholestatic pruritus.
drug: patients should take colestyramine Fatigue is a non‐specific but often reported
2 hours after or 4 hours before other medica- symptom in PBC, and represents an unmet
tions to avoid preventing their absorption, need since there are no licensed therapies.
and 20 minutes before meals. Tolerability is Since fatigue is not related to the stage of
often an issue, with side effects including disease, and potentially can persist after LT,
bloating and constipation. Colestyramine is its management proceeds in parallel with
best taken with food, and is most effective management of the disease itself. A struc-
when taken with the first meal of the day. tured approach to management – quanti-
Second‐line treatment is rifampicin, which fying fatigue and its impacts (through the
probably acts through its pregnane X use of disease‐specific tools such as the
receptor (PXR) agonist function. There are PBC‐40 quality‐of‐life measures), address-
concerns over potential side effects, including ing contributing and exacerbating factors,
hepatotoxicity and hemolysis and increased and supporting patients to cope with its
international normalized ratio (INR). impact – is effective. Since fatigue in PBC is
Therefore, regular blood tests should be per- not specific, it is important to identify other
formed. Rifampicin is also a potent enzyme disease processes and therapies which may
inducer and caution is required when the be contributing to fatigue, such as hypothy-
drug is given with other medications (such as roidism, depression, anemia, sleep apnea,
warfarin). Also, appropriate consideration and overtreatment with antihypertensive
should be given to balancing the benefits drugs. Drugs such as coenzyme Q and ritux-
against the risks of antimicrobial resistance. imab have been shown to be ineffective. A
Third‐line treatment is oral naltrexone, an graded program of exercise helps some
opiate antagonist, which reduces the sensa- individuals.
tion of itching. It should be started at a Patients with PBC often complain of dry
low dose (12.5 mg/day) and titrated slowly to eyes (xerophthalmia) and/or dry mouth
avoid withdrawal‐like symptoms in the (xerostomia), and sometimes dysphagia and
first few days of treatment. Some patients vaginal dryness, but this is rarely associated
require an intravenous induction stage with with an actual diagnosis of primary Sjögren
naloxone. Long‐term tolerability can be an syndrome. When symptoms are mild,
issue, with many patients having ongoing artificial eye drops and saliva substitutes are
opiate withdrawal‐like reactions or reduced usually sufficient. Patients should be advised
threshold to pain. to undergo regular dental check‐ups, as
Fourth‐line treatment involves the selective they are at risk of caries and oral candidi-
serotonin reuptake inhibitors and gabapen- asis. If xerophthalmia is moderate to severe,
tin, which are used empirically in the an ophthalmologist’s review is suggested.
management of cholestatic itch, typically in Cholinergic agents such as pilocarpine are
patients with pruritus unresponsive to other effective in patients symptomatic despite
agents. first‐line treatments, but usually have
Symptom Treatment Notes

Itching Colestyramine 4 g q.d.up to a maximum of q.i.d.
More effective when taken in the morning
2 hours after or 4 hours before other drugs
Tolerability issues (constipation)
Rifampicin Start at 150 mg q.d. or b.i.d., titrate up to max 600 mg daily
Monitor liver function (every 2 weeks in the first month)
Caution in advanced disease
Naltrexone Start at 12.5 mg q.d.

Slow titration (up to 50 mg q.d.)
Tolerability issues (opiate withdrawal reactions)
Sertraline 100 mg q.d.
Possible pharmacokinetic interactions
Fatigue No drugs available Address possible contributing or exacerbating factors
Advise lifestyle changes
Support patients through counseling strategies
Sicca Artificial eye drops For mild complaints

syndrome Saliva substitutes
Pilocarpine/Civimeline For moderate-to-severe symptoms
Tolerability issues (gastrointestinal symptoms, sweating)
When symptoms are severe/refractory seek
ophthamologist/ENT consultation
Abbreviation: b.i.d., twice a day; q.d., once a day; q.i.d., 4 times a day
Figure 7.6 Practical approach to drug management of most relevant symptoms in PBC.
cholinergic side effects (e.g. nausea, diar- Complications of Liver Disease

rhea, sweating). A practical approach to the
management of these symptoms is summa- Osteoporosis
rized in Figure 7.6.
Up to 25% of patients with PBC exhibit All patients with PBC should have a risk
Raynaud’s phenomenon. Patients should assessment for osteoporosis, which affects
therefore be asked specifically about the up to 40% of patients. In addition to the
classical symptoms, i.e. extremities turning vitamin D deficiency secondary to chronic
white, then blue and finally red, often associ- cholestasis, osteopenic bone disease in PBC
ated with pain/burning/tingling when the can be related to female gender, menopausal
blood flow returns. Practical measures, such status, low body mass index, older age,
as wearing gloves, using hand warmers and and advanced disease. Patients should
avoiding cold environments, are often useful receive general advice to prevent bone loss
for mild symptoms. For more marked symp- (i.e. weight‐bearing exercise, cessation of
toms and signs, such as digital ulceration, smoking, minimizing alcohol intake).
specialist rheumatologic advice should be Vitamin D deficiency should be corrected
sought. Approximately 8% of patients with and an adequate dietary intake assured.
PBC have limited scleroderma (CREST syn- Dual‐energy X‐ray absorptiometry to assess
drome: calcinosis, Raynaud’s phenomenon, bone mineral density should be performed
esophageal dysmotility, sclerodactyly, telan- at presentation and at follow‐up where
giectasia). These symptoms should be indicated. Medical therapy with bisphos-

sought, and if present patients should be phonates is suggested when there is a sig-
referred to a rheumatologist. nificantly elevated fracture risk from
osteoporosis (generally with a T‐score lower portal hypertension can occur in PBC and
than −1.5). Caution is required when using the possibility of its presence should be
oral bisphosphonates in patients with var- considered in all PBC patients with a gas-
ices because of the risk of superficial erosion trointestinal bleed. The management of
and enhanced bleeding risk. Intravenous gastroesophageal varices and variceal hem-
bisphosphonates can be used if there is orrhage in patients with PBC follows the
clinical concern. Baveno‐VI guidelines.
HCC
Advanced Liver Disease
As with all forms of cirrhosis, patients with
Varices PBC may develop HCC and should therefore
All PBC patients with suspected portal undergo regular screening with cross‐sec-
hypertension should be screened for gas- tional imaging. Male sex and UDCA non‐
troesophageal varices. Patients with PBC response are risk factors for HCC development
can develop varices even in the absence of in PBC. Men without advanced disease who
established cirrhosis, in association with are UDCA non‐responders are at higher risk
nodular regenerative hyperplasia, although of developing HCC than women with cir-
this is relatively uncommon. Non‐cirrhotic rhosis who respond to UDCA.
Useful Websites
www.uk‐pbc.com
www.globalpbc.com
www.pbcfoundation.org.uk
References
1 Sherlock, S. (1959). Primary biliary 8 Metcalf, J.V., Mitchison, H.C., Palmer, J.M.
cirrhosis. Chronic intrahepatic obstructive et al. (1996). Natural history of early primary
jaundice. Gastroenterology 31: 574. biliary cirrhosis. Lancet 348: 1399–1402.
2 Boonstra, K., Beuers, U., and Ponsioen, C.Y. 9 Carbone, M., Mells, G.F., Pells, G. et al.
(2012). Epidemiology of primary sclerosing (2013). Sex and age are determinants of the
cholangitis and primary biliary cirrhosis: a clinical phenotype of primary biliary
systematic review. J. Hepatol. 56: cirrhosis and response to ursode
1181–1188. oxycholic acid. Gastroenterology 144 (3):
3 Gershwin, M.E., Mackay, I.R., Sturgess, A., 560–569.e7.
and Coppel, R.L. (1987). Identification and 10 Trivedi, P.J., Corpechot, C., Pares, A., and
specificity of a cDNA encoding the 70 kd Hirschfield, G.M. (2016). Risk stratification
mitochondrial antigen recognized in in autoimmune cholestatic liver diseases:
primary biliary cirrhosis. J. Immunol. 138 opportunities for clinicians and trialists.
(10): 3525–3531. Hepatology 63 (2): 644–659.
4 Joplin, R., Lindsay, J.G., Johnson, 11 Carbone, M., Nardi, A., Flack, S. et al.
G.D. et al. (1992). Membrane (2018). Pretreatment prediction of response
dihydrolipoamide acetyltransferase (E2) to ursodeoxycholic acid in primary biliary
on human biliary epithelial cells in cholangitis: development and validation of
primary biliary cirrhosis. Lancet 339 the UDCA Response Score. Lancet
(8785): 93–94. Gastroenterol Hepatol 3 (9): 626–634.
5 Hirschfield, G.M., Beuers, U., Corpechot, C. 12 Poupon, R.E., Poupon, R., and Balkau, B.
et al. (2017). The diagnosis and management (1994). Ursodiol for the long‐term
of patients with primary biliary cholangitis. treatment of primary biliary cirrhosis. The
J. Hepatol. 67 (1): 145–172. UDCA‐PBC Study Group. N. Engl. J. Med.
6 Locke, G.R. III, Therneau, T.M., Ludwig, J. 330 (19): 1342–1347.
et al. (1996). Time course of histological 13 Nevens, F., Andreone, P., Mazzella, G. et al.
progression in primary biliary cirrhosis. (2016). A placebo‐controlled trial of
Hepatology 23 (1): 52–56. obeticholic acid in primary biliary cholangitis.
7 Corpechot, C., Carrat, F., Bonnand, A.M. N. Engl. J. Med. 375 (7): 631–643.
et al. (2000). The effect of ursodeoxycholic 14 Neuberger, J., Portmann, B., Macdougall,
acid therapy on liver fibrosis progression in B.R. et al. (1982). Recurrence of primary
primary biliary cirrhosis. Hepatology 32 (6): biliary cirrhosis after liver transplantation.
1196–1199. N. Engl. J. Med. 306 (1): 1–4.
143
Primary Sclerosing Cholangitis

Mette Vesterhus1,2, Benedetta Terziroli Beretta‐Piccoli3, Kirsten Muri Boberg1,4,5,
and Giorgina Mieli‐Vergani6
1
Division of Surgery, Department of Transplantation Medicine, Norwegian PSC Research Center, Inflammatory Medicine and
Transplantation, Oslo University Hospital, Rikshospitalet, Oslo, Norway
2
Department of Clinical Science, University of Bergen, Bergen, Norway
3
Epatocentro Ticino, Lugano, Switzerland
4
Institute of Clinical Medicine, University of Oslo, Oslo, Norway
5
Division of Surgery, Department of Transplantation Medicine, Section for Gastroenterology, Inflammatory Medicine and
Transplantation, Oslo University Hospital, Rikshospitalet, Oslo, Norway
6
Faculty of Life Sciences and Medicine, Paediatric Liver, GI & Nutrition Centre, Mowat Labs, King’s College Hospital, London, UK
Abstract
This chapter provides guidance as regards patient management and clinical followup, emphasizing
the peculiarities of pediatric compared to adult presentations of the disease. Adult primary
sclerosing cholangitis (PSC) is a rare disease but occurs in 5–10% of patients with inflammatory
bowel disease (IBD). Cholangiocarcinoma and colorectal cancer are dreaded complications in PSC
and major causes of death; alertness to alarm symptoms and adherence to surveillance strategies
may improve survival. There is no medical therapy with proven effect to halt disease progression in
adult PSC, but ursodeoxycholic acid is often prescribed and immunosuppression may be indicated
in patients with additional features of autoimmune hepatitis. Full colonoscopy with biopsies is
recommended at diagnosis of PSC in all patients without known IBD in order to diagnose subclin
ical colitis. PSC has been denominated a premalignant condition by some, due to the firmly
established and extremely high risk of hepatobiliary as well as colorectal cancer.
Keywords autoimmune hepatitis; cholangiocarcinoma; colorectal cancer; inflammatory bowel
disease; medical therapy; primary sclerosing cholangitis; surveillance strategies
Key Points
●● Adult primary sclerosing cholangitis major causes of death; alertness to alarm
(PSC) is a rare disease but occurs in symptoms and adherence to surveillance
5–10% of patients with inflammatory strategies may improve survival.
bowel disease. ●● There is no medical therapy with proven
●● Cholangiocarcinoma and colorectal can effect to halt disease progression in adult
cer are dreaded complications in PSC and PSC, but ursodeoxycholic acid is often
prescribed and immunosuppression (cor may be challenging but has therapeutic

ticosteroids) may be indicated in patients implications.
with additional features of autoimmune ●● The most common form of juvenile
hepatitis (AIH). sclerosing cholangitis has strong autoim
●● Endoscopic therapy for clinically mune features and responds to immuno
significant biliary stenosis using balloon suppression similarly to AIH.
dilatation or stenting shows similar out ●● Adult‐type PSC is rare in children and
come, with reports of higher complication adolescents.
rates in stenting; if stenting is performed, ●● Prognostication in PSC will likely profit
short‐term (2–3 weeks) stenting is from the implementation of non‐invasive
recommended. tests for disease severity and stage, also
●● Demarcating IgG4‐associated cholangitis allowing risk stratification guiding differ
and features of AIH in patients with PSC entiated follow‐up.
Introduction children, the autoimmune and inflammatory

aspects of the disease are more pronounced
Primary sclerosing cholangitis (PSC) is a compared to adults. An overlap syndrome
severe chronic liver disease of unknown eti between autoimmune hepatitis (AIH) and

ology characterized by inflammation and autoimmune sclerosing cholangitis (ASC) is
fibrosis of the bile ducts leading to chole probably the most frequent cause of the
stasis and multifocal biliary strictures [1]. juvenile disease [3]. In pediatrics, it is impor
About two‐thirds of PSC patients in northern tant to identify well‐defined forms of sclerosing
Europe and the USA have concurrent cholangitis [4], including biliary atresia and
inflammatory bowel disease (IBD). Moreover, autosomal recessive neonatal sclerosing chol
the patients carry a strikingly high risk of angitis, other inherited conditions [e.g. mild
colorectal and hepatobiliary malignancy, to moderate defects in the ABCB4 (MDR3)
necessitating surveillance and vigilance. A gene] associated with small‐duct disease, and
plethora of intercurrent events may compli sclerosing cholangitis secondary to a wide
cate the condition, related to either the bil variety of disorders including primary and
iary disease itself (gallstones, dominant secondary immunodeficiencies, Langerhans
stenoses, bacterial cholangitis) or to any of cell histiocytosis, cystic fibrosis, and sickle cell
the associated diseases. anemia. The definition of “primary” should be
PSC progresses to end‐stage liver disease restricted to those children/adolescents in
within 15–20 years in the majority of patients whom sclerosing cholangitis occurs without
[2]. However, the variability in disease course any of these defining features. Adult‐type PSC
as well as transplant‐free survival is large and is rare in children, if strict diagnostic criteria
validated prognostic tools are not established. are followed [5]. In the largest pediatric series
The notoriously unpredictable nature of PSC published so far, 67% of 781 children did not
places a burden of uncertainty that is heavily have concomitant AIH or associated condi
felt by patients. With no established effective tions and were classified as PSC, but no
medical therapy in adult PSC, PSC is a major information was reported about IBD pharma
indication for liver transplantation in many cologic treatment before the appearance of
countries, with disease recurrence being a liver disease, which is relevant, as immunosup
major challenge. pression may mask AIH [6]. In the only pro
PSC is typically diagnosed in adults in their spective pediatric study, PSC was diagnosed
thirties and forties, but children down to the in only 18% of the children presenting with
age of three have been diagnosed; usually, in sclerosing cholangitis [3].
Chapter 8 Primary Sclerosing Cholangitis 145
Thus PSC is a complex and heterogeneous PSC is strongly associated with IBD; how
disease, with important knowledge gaps and ever, the frequency of IBD in PSC varies
unmet needs contributing to the challenges across the world, with a similar geographic
in patient care and follow‐up. There are gradient as for the disease itself. The high
important differences between adult PSC est prevalence rates are found in northern
and juvenile sclerosing cholangitis. This Europe and the USA where 60–80% of
chapter provides guidance as regards patient patients have concurrent IBD. Lower
management and clinical follow‐up, empha frequencies are observed in southern
sizing the peculiarities of pediatric compared Europe (40–60%) and Asia (reports of 34%
to adult presentations of the disease. in Japan, 50% in India), possibly due to a
larger fraction of IBD‐unrelated sclerosing
cholangitis (e.g. IgG4‐associated sclerosing
Definition cholangitis) in these areas. In support of the
latter, current data from Japan suggest that
PSC is a chronic and progressive immune‐ among patients diagnosed with PSC there
associated cholestatic liver disease of is a group of elderly patients without IBD
unknown etiology characterized by inflam that is clinically distinct from younger
mation, fibrosis, and multifocal strictures of patients (which show an IBD prevalence
the bile ducts. rate of 60%).
The reported prevalence for PSC among
patients with IBD varies and is probably less
than 10%. PSC was seen more frequently in
Adult PSC extensive rather than distal colitis in a large
population of ulcerative colitis (UC) patients
Epidemiology
from Sweden. There is a seemingly stronger
PSC is a rare disease according to formal def association for PSC with UC compared with
initions, with the highest prevalence and Crohn’s colitis (80% of IBD in PSC is classi
incidence rates in northern Europe and the fied as UC). However, PSC patients with
USA. Population‐based studies are scarce, Crohn’s disease are more likely to be female
but studies from northern Europe and the with small‐duct PSC and a milder (hepato
USA report prevalence rates of about 1 in biliary) disease course and this group may
10 000 and incidence rates of 0.4–2.0 per be underdiagnosed at present. This is sup
100 000 per year. There is a geographic gra ported by a study of a large cohort of IBD
dient with 10–100‐fold lower prevalence patients from Norway where MRC surveil
rates in southern Europe and Asia. Several lance revealed findings consistent with PSC
reports support a rising incidence of PSC. in 7.8%, of which only one‐third were recog
This rise is thought to be real rather than an nized clinically and with similar frequencies
effect of improved diagnostics alone as the of PSC in Crohn’s disease and UC [7].
clinical features of newly diagnosed patients Current data indicate that PSC may be
have not changed over time. Earlier diagnosis occurring equally frequently in men and
of PSC as magnetic resonance cholangiog women but runs a clinically more quiescent
raphy (MRC) has improved and replaced course in females, explaining the tradition
invasive diagnostic methods does not seem ally reported 2 : 1 male to female ratio. Data
to explain fully the increased incidence. from a large multicenter study of 7119 PSC
The parallel increase in incidence in other patients support this, showing significantly
autoimmune and idiopathic inflammatory better outcomes for female patients [8]. In
diseases has fueled hypotheses of a common the largest pediatric series to date, the male
environmental cause driving the increased to female ratio was 1.5 : 2, and disease pro
incidences. gression was similar in boys and girls [6].
Diagnosis
contrast‐enhanced liver magnetic resonance
The diagnosis is made based on characteristic imaging (MRI) within 6 months of diagnosis if
cholangiographic findings, featuring irregular not undertaken as part of the initial work‐up,
bile ducts with multiple alternating strictures with the aim of early detection of cholangio
and dilatations in combination with elevated carcinoma; however, any effect on outcome
cholestatic liver enzymes, i.e. alkaline phos has not been proven and the risk of side
phatase (ALP) and/or gamma‐glutamyltrans effects is debated. ERC is generally reserved
peptidase (GGT). The biliary changes are for cases where intervention (stenting, dilata
most often found both intrahepatically and tion) is required [9].
extrahepatically (50–80%), but may be intra Known causes of secondary cholangitis
hepatic only (<25%) or, more rarely, only should be excluded, for example IgG4‐asso
extrahepatic (<5%) (Figure 8.1). MRC is the ciated cholangitis (IAC), and comorbid liver
imaging modality of choice, with diagnostic diseases identified (viral or AIH, alcohol)
accuracy reaching that of endoscopic retro based on detailed history, extensive blood
grade cholangiography (ERC), with sensi work‐up, and ultrasound. Liver biopsy is not
tivity and specificity of 0.86 and 0.94, part of standard diagnostic work‐up but
respectively, at expert centers. If the initial may be useful in suspected variant PSC with
MRC is inconclusive, showing equivocal concomitant features of autoimmune
minimal changes of the biliary tree, second hepatitis (PSC/AIH; previously denomi
opinion evaluation by an expert center either nated “AIH overlap syndrome” [10]; suspect
of the original images or a repeat MRC may if elevated transaminases five times upper
yield a definitive diagnosis. Some experts in limit of normal, ULN) or small‐duct disease
the field recommend MRC combined with (when a strong suspicion of PSC remains
(a)
(b)
Figure 8.1 (a) Endoscopic retrograde cholangiography (ERC) in a patient with primary sclerosing cholangitis
(PSC) showing multifocal intrahepatic bile duct strictures with intervening dilatations and narrowing of the
entire extrahepatic bile duct (only catheter visible). (b) Magnetic resonance cholangiography (MRC) in a PSC
patient featuring multifocal intrahepatic bile duct strictures with intervening dilatations and a dilated irregular
extrahepatic bile duct. Source: courtesy of Gunter Kemmerich.
iagnosed when diagnostic criteria for both

d
PSC and AIH, respectively, are met. PSC/
AIH affects 7–14% of adult PSC patients and
should be suspected in cases of considerably
elevated transaminases. Liver biopsy is
mandatory to make the diagnosis. There are
no established formal criteria for a diagnosis
of PSC with features of AIH, but commonly
the condition is diagnosed in a patient diag
nosed with PSC based on:
●● elevation of transaminases at least five
times ULN;
●● IgG at least two times ULN; and
●● typical or compatible histologic findings
dependent on other factors (including the
Figure 8.2 Liver histology in PSC showing a portal presence of autoantibodies).
field with bile duct changes. There are infiltrating
lymphocytes in the epithelium and slight periductal Some patients have a clinical presentation
fibrosis. Source: courtesy of Krzysztof Grzyb. leading to a primary diagnosis of AIH, and
the diagnosis of PSC may then be missed.
despite repeated negative cholangiograms)
(Figure 8.2). Elevated IgG4
Full colonoscopy with biopsies is recom Moderately elevated IgG4 is observed in
mended at diagnosis of PSC in all patients 10–20% of classical PSC patients. Most com
without known IBD in order to diagnose monly IgG4 is less than 2.8 g/l (2× ULN) in
subclinical colitis. This has implications for these patients, but higher levels may be seen.
prognosis (risk of colorectal cancer) and It is important to distinguish this group from
surveillance. patients with IAC, a secondary sclerosing
cholangitis which mimics the cholangio
Diagnosis of Variant Phenotypes graphic findings typical for PSC, because
Small‐duct PSC IAC responds to steroid treatment and even
When clinical suspicion of PSC remains severe biliary changes may resolve com
despite repeated cholangiograms lacking pletely on treatment. IAC should be diag
diagnostic cholangiographic changes, patients nosed according to modified HISORt
should undergo liver biopsy to identify small‐ (histology, imaging, serology, other organ
duct PSC, which represents 6–16% of PSC involvement, response to therapy) criteria. It
patients. Histologic criteria for this depend on has been suggested that IAC may be indi
IBD status; more distinct histologic features cated by an IgG4/IgG1 ratio above 0.24 and
are required in non‐IBD patients in whom the condition is more frequent in older men
differential diagnoses of secondary sclerosing without a diagnosis of IBD. In the absence of
cholangitis should be particularly considered other imaging features raising concern about
(e.g. genetic cholestasis resulting from ABCB4 IgG4 disease, IAC is very unlikely in patients
mutations). with IgG4 levels below 5 g/l (4× ULN).
Features of AIH (see also Chapter 6)

Presentation
There is a continuous spectrum of autoim
mune features associated with PSC, probably Typically, PSC is diagnosed in a 30–40‐year‐
springing from a pool of shared pathogenetic old male with a diagnosis of IBD colitis (most
factors for PSC and AIH. PSC with features often UC) and elevated cholestatic liver
of autoimmune hepatitis (PSC/AIH) is enzymes (ALP and GGT). Reflecting the
g enetic autoimmune disposition, one in four Bacterial cholangitis is a frequent compli

patients has other autoimmune diseases (e.g. cation but symptoms may be atypical (i.e. the
type 1 diabetes, rheumatoid arthritis and classical triad of fever, jaundice, and upper
autoimmune thyroid disease). In the majority right quadrant pain may be lacking) and
of patients, the diagnosis of IBD precedes standard definitions for cholangitis are not
that of PSC, but IBD may present at diag applicable in PSC. Patients developing bacte
nosis of PSC or any time later, even after liver rial cholangitis should undergo MRC to iden
transplantation. Similarly, PSC may present tify any flow‐limiting biliary strictures and
in an IBD patient after colectomy. Genetically ERC with intervention if indicated. Some
and phenotypically, the IBD in PSC is dis patients experience recurrent bacterial
tinct from IBD without PSC. Typically, IBD cholangitis and may benefit from easy access
in PSC is mild or asymptomatic with endo to antibiotics (i.e. prescriptions for self‐
scopic pancolitis (often more pronounced on administration of ciprofloxacin at first symptom
the right side), backwash ileitis and rectal recurrence); schemes where antibiotics are
sparing; however, severe colitis requiring rotated lack evidence and risk promoting
biological treatment or colectomy is also antibiotic resistance, so this cannot be rec
observed. Despite a considerable heritability, ommended. Recurrent bacterial cholangitis
with an 11‐fold risk of developing PSC in may constitute an indication for liver trans
siblings of patients, the rarity of the disease plantation in severe cases.
leads to a low absolute risk of PSC in children Dominant stricture, reported in up to 45%
of PSC patients. of patients, is defined as a clinically significant
Approximately 50% of adult PSC patients stenosis within the extrahepatic biliary tree
are asymptomatic at diagnosis. Often, the in PSC (≤1.5 mm in the common bile duct or
disease is suspected on incidental findings of ≤1 mm in the hepatic duct within 2 cm of the
abnormal liver enzymes in an IBD patient. If hilum at ERC) and may cause jaundice. These
present, symptoms typically include pruritus, diagnostic definitions are not applicable for
fatigue, upper right quadrant abdominal dis MRC because of the lack of spatial resolution
comfort, jaundice, and sometimes weight and hydrostatic pressure which is present at
loss. The patients are prone to developing ERC. In fact, a complete occlusion cholan
cholelithiasis and may report episodes of giogram at ERC may sometimes reveal that a
typical biliary colic or cholecystitis. Some
suspected dominant stricture on MRC is not
patients report episodes of fever and chills, a stricture at all. Moreover, focus has increas
typically resolving within 24 hours, represent ingly shifted from diameter definitions to
ing cholangitis. Longer‐lasting fever particu assessment of clinical significance of stric
larly with concurrent jaundice could indicate tures and the potential benefit from endo
bacterial cholangitis necessitating antibiotic scopic intervention. Thus, the diameter
therapy; symptoms may be atypical and stan criterion is not applied strictly; rather, the
dard definitions are not applicable. Rapid decision for intervention results from a
deterioration should always raise suspicion of compound clinical evaluation focusing on
cholangiocarcinoma, the development of the clinical relevance of the stricture (ele
which may precipitate the symptoms leading vated serum bilirubin, significantly increasing
to the diagnosis. cholestatic liver enzymes, repeated episodes
Cholelithiasis is a common finding in PSC of bacterial cholangitis or pruritis). Limited
(25%). The gallbladder is frequently found to be data exist as to the natural history of
large, but this has no known clinical implica untreated dominant strictures. Although
tions. The incidence of cholecystitis is increased. endoscopic intervention is advocated if there
Gallbladder polyps or mass lesions are observed is associated jaundice or cholestatic liver
in 4–6.5% of PSC patients and harbor a reported enzymes, one study observed similar changes
high malignant potential (see below). in liver biochemistry irrespective of whether
a dominant stenosis was present or not, silent jaundice or as a recent‐onset dominant

questioning the effect of dominant stenoses stricture, whereas intrahepatic cholangiocar
on short‐term outcome. cinoma typically presents as a mass lesion
with ALP elevation but no jaundice. Abrupt
presentation with jaundice and weight loss or
Malignancy
rapid clinical deterioration (weight loss, jaun
Malignancy is a major cause of death in dice, fatigue, abdominal pain, cholestatic liver
PSC. PSC has been denominated a “prema tests) may be merely a reflection of disease
lignant condition” by some, due to the progression (development of benign stenosis
firmly established and extremely high risk or end‐stage liver disease) but should always
of hepatobiliary as well as colorectal cancer. precipitate swift and adequate work‐up aimed
PSC is associated with cholangiocarci at diagnosing potential cholangiocarcinoma.
noma, gallbladder cancer and colorectal If cholangiocarcinoma is suspected, combined
cancer (in patients with colitis), and contrast MRI/MRC and serum carbohydrate
patients with cirrhosis are at risk of hepato antigen 19‐9 (CA19‐9) testing should be per
cellular carcinoma. formed, followed by ERC with brush cytology.
Notably, the sensitivity of brush cytology is
Cholangiocarcinoma dependent on operator variability, access to
Patients with PSC carry a strikingly high risk of the stricture, and the experience of the pathol
cholangiocarcinoma, with a lifetime risk of 7% ogist. Repeated brushings and additional anal
in population‐based cohorts and more than ysis using fluorescence in situ hybridization
10–15% in series from transplant centers [2,11]. (FISH) may improve sensitivity if available.
Older age at PSC diagnosis was associated with In patients with negative brush cytology
increased risk of cholangiocarcinoma in both a and no mass on cross‐sectional imaging,
large international multicenter study and the endoscopic ultrasound with fine needle aspi
large Dutch population‐based study, with a ration may be of use in the establishment of a
median age at diagnosis of cholangiocarcinoma cholangiocarcinoma diagnosis; however, con
of 47 years in the latter [2,8]. Between one‐third cerns have been raised regarding the potential
and half of the cholangiocarcinomas are for needle track seeding of tumor cells and
detected at PSC diagnosis or within a year after practice varies. While cholangioscopy (e.g.
diagnosis of PSC, likely because cholangiocar SpyGlass, Boston Scientific, USA) has shown
cinoma precipitates the diagnosis of PSC in some promise in the diagnosis of sporadic
these cases. Thereafter estimated yearly inci cholangiocarcinoma, only limited evidence
dence is 0.5–1.5%; late presentations also occur, (and limited success) exists for any utility of
even more than 10 years after the PSC diag cholangioscopy for the detection of cholan
nosis. After 30 years, the cumulative risk of giocarcinoma in PSC [9]. The value of posi
cholangiocarcinoma was reported to reach tron emission tomography lies in the further
20%. Cholangiocarcinoma is the most common staging of cholangiocarcinoma diagnosed by
PSC‐related cause of death (32%). another modality. Computed tomography
Early detection of cholangiocarcinoma is (CT) or MRI alone may show lesions but lack
severely hampered by the scarcity of symp diagnostic accuracy in early cholangiocarci
toms and the poor performance of current noma due to difficulties in distinguishing
diagnostic tools. Early stages are often asymp inflammatory, benign and malignant lesions.
tomatic; symptomatic cholangiocarcinoma It should be noted that the utility of serum
most often reflects advanced incurable cancer. CA19‐9 alone is limited; inactivating poly
The clinical presentation of cholangiocarci morphisms in the FUT3 gene may lead to
noma may depend on the phenotype. Perihilar CA19‐9‐negative cholangiocarcinoma in 7%
(Klatskin tumor; most common) and extrahe of cases and elevated levels may be observed
patic cholangiocarcinoma often present with with cholangitis or other malignancies.
Surveillance for cholangiocarcinoma has no transplantation or death over time. Mortality

documented effect on survival or clinical out in PSC is fourfold that of the general
come but is increasingly being performed. population. Malignancy makes a major con
Annual or regular surveys with combined tribution to the high mortality. In a large
MRI/MRC (sensitivity 89%, specificity 75% for study, cholangiocarcinoma and colorectal
cholangiocarcinoma), preferably with con cancer accounted for 32% and 8% of the PSC‐
trast enhancement for increased accuracy, related mortality, respectively, as compared
combined with serum CA19‐9 may be the to liver failure (15%) and liver transplant‐
best choice for surveillance, increasing sensi related complications (9%).
tivity at the cost of reduced specificity [12]. Studies report widely variable median liver
transplantation‐free survival times. In a
Gallbladder Cancer large international cohort study and a large
The frequency of gallbladder carcinoma in Dutch population‐based study, respectively,
the overall PSC population is reported to the reported median liver transplantation‐
2.5–3.5%. Gallbladder polyps in patients with free survival was 14.5 years and 21.5 years
PSC have a high malignant potential; as many [2,8]. In the Dutch study, comprising all hos
as 55–75% are malignant. Detection of gall pitals in half of the Netherlands, longer
bladder polyps in PSC should lead to chole median transplant‐free survival was
cystectomy; general cutoffs regarding size do observed in general hospitals compared with
not apply as neoplasia is observed also in transplant centers (21.3 vs. 13.2 years), sug
polyps smaller than 0.8 cm. gesting that referral bias may confound
studies of natural history in PSC. Despite the
Colorectal Cancer overall grim prognosis, a proportion of
The risk of colorectal cancer is fivefold higher patients may never need a liver transplant.
in PSC than in IBD without PSC and ranks There are currently no established prog
among the top four causes of death in PSC. nostic tools that reliably estimate prognosis
Patients with PSC and UC seem to be at at the individual level. The severity or stage
particular risk. Colorectal cancer may occur of disease in PSC is not captured by any
at any point. A large population‐based study single biochemical test or radiologic feature.
from the Netherlands demonstrated that The gauging of disease stage and prognosis is
patients with PSC were on average 20 years complicated by the remarkably variable
younger compared to patients without PSC natural history of PSC, with its complex
when they were diagnosed with colorectal interactions between chronic inflammation,
cancer [2]. In this study, even patients with fibrosis and dysplasia of the bile ducts as well
PSC in their early twenties developed colo as intercurrent events (e.g. bacterial cholan
rectal cancer, justifying surveillance from the gitis, gallstones) that contribute to the natural
time of diagnosis of PSC. After 20 and fluctuations in ALP, bilirubin and symptoms
30 years, the cumulative risk of colorectal which are, at least in part, dissociated from
cancer reached 6 and 13%, respectively. the severity of the underlying liver disease.
It is debated whether pancreatic cancer In the following sections we discuss historic
occurs at increased frequencies or whether and current attempts to capture disease
misclassification of cholangiocarcinoma may severity and stage in PSC.
have led to such conclusions.
Risk Stratification: Clinical
Characteristics
Prognosis
Patients who are asymptomatic at diagnosis
PSC is a progressive disorder, almost univer show better survival compared to patients
sally leading to cirrhosis and complications presenting with symptoms. However, the
related to end‐stage liver disease and liver severity of symptoms such as fatigue and
ruritus correlate poorly with the severity of

p revised Mayo Risk Score is most widely used
the liver disease in PSC. Similarly, patients [13]. However, the influence from parame
without concurrent IBD (although ascertain ters reflecting late‐stage disease [variables
ment is difficult as IBD may present later) or include age, bilirubin, albumin, aspartate
having a diagnosis of Crohn’s disease are at aminotransferase (AST), variceal bleeding]
lower risk of progression to death or liver and its relatively short horizon (4 years)
transplantation compared to UC patients. On render the model less useful in early disease.
the other hand, there is poor correlation bet Moreover, its failure to predict adverse out
ween colonic disease severity and progression comes in high‐dose ursodeoxycholic acid
or outcome of the liver disorder. Female (UDCA) treatment trials raised concerns and
patients seem to fare better than males; how the model is not validated for use at an
ever, no differences were demonstrated bet individual level.
ween male and female PSC patients when
matched for IBD phenotype in a large multi ALP and Bilirubin
center study, indicating that UC in PSC con Elevated bilirubin is consistently associated
fers a strong increased risk which may override with outcome in PSC; however, temporary
the protective effect of female sex. elevations may be caused by intercurrent
Typically, small‐duct PSC runs a quiescent events (gallstones, bacterial cholangitis, bil
course with significantly improved long‐term iary stenosis) and may later resolve. In pri
survival compared to large‐duct PSC. mary biliary cholangitis (PBC), ALP response
Cholangiocarcinoma is a rare complication to UDCA therapy is an established part of
but does occur. Colonoscopic surveillance is the prognostic evaluation. The ALP level is
indicated in patients with concurrent IBD as associated with outcome in PSC as well, and
an increased risk of colorectal cancer seems studies have reported associations of ALP
to persist. In about 20% of patients, transition reduction with clinical outcome after UDCA
to large‐duct PSC occurs over a median treatment as a parallel to scores used in PBC.
period of 7.4 years, and cholangiography However, study designs vary and all attempts
should be repeated on clinical deterioration. to cross‐validate suggested criteria have hith
Adult PSC patients with features of AIH erto failed. The naturally fluctuating nature
seem to have a prognosis similar to that of of ALP in PSC is very different from the
other patients with large‐duct PSC. PSC relentless increase over time in PBC for
patients with elevated IgG4 often have a example and may, on the individual level,
more aggressive disease course and reduced limit the value of single measurements at any
transplant‐free survival has been observed. point in time for patient follow‐up. Further
studies are warranted.
Natural History Models
A number of prognostic models have been Non‐invasive Evaluation of Fibrosis
proposed, typically combining laboratory Over recent years, there has been a strong
results (e.g. ALP and bilirubin) and clinical interest in the identification of prognostic
features of cirrhosis (e.g. ascites, variceal biomarkers and non‐invasive evaluation of
bleeding or spleen size) [1]. Histology allows liver fibrosis in PSC. Markers of liver fibrosis
assessment of disease stage and several histo appear to be of particular interest for prog
logic scores show robust association with nostic purposes.
outcome on group level. However, while The enhanced liver fibrosis (ELF™) test is a
early models included histology, this was direct serum marker panel of liver fibrosis
later abandoned due to its invasive nature reflecting three components of fibrogenesis
and the patchy distribution of pathology in and matrix remodeling, and has demon
PSC, limiting the diagnostic value of liver strated a strong ability to predict clinical out
biopsy. Regarding natural history models, the come [14] independently of the Mayo Risk
Score in several independent PSC panels, prevents progression to liver transplantation

including a large international multicenter or death in PSC patients. Even the largest
study. clinical trial of UDCA in PSC to date was
Liver stiffness measurement using transient underpowered and could not demonstrate an
elastography as a proxy for liver fibrosis eval effect on outcome. High‐dose UDCA (28–
uation has been successfully established in 30 mg/kg daily) was surprisingly associated
the management of patients with viral with a significant increase in clinical end
hepatitis and is recommended by interna points in a well‐designed placebo‐controlled
tional guidelines in the follow‐up of patients clinical trial and hence should be avoided.
with PBC. In PSC, baseline liver stiffness as International guidelines offer diverging
well as the change in liver stiffness over time recommendations regarding the use of
were associated with clinical outcome in a low‐ to medium‐dose UDCA in PSC. The
French monocenter study [15] and the find American Association for the Study of Liver
ings were validated in an independent patient Diseases (AASLD) advises against the use of
panel. Notes of caution have been raised by UDCA. A Cochrane review stated that evi
case series demonstrating the influence of dence is lacking for its use. The European
dominant strictures and cholestasis on liver Association for the Study of the Liver (EASL)
stiffness measurements in PSC, causing high concludes that high‐level evidence is lacking
values which normalize following biliary for an effect of UDCA on clinical outcomes
intervention, underscoring the importance of while noting an effect of UDCA (13–15 mg/
interpreting results as part of a full clinical kg daily) on surrogate markers such as ALP.
picture. An ongoing, large, prospective, inter The American College of Gastroenterology
national multicenter study will further inves has taken an intermediate position, recom
tigate the prognostic value of transient mending against high‐dose UDCA while
elastography in PSC. acknowledging prescription of low to
medium doses. Some centers choose a
pragmatic approach, prescribing a 6‐month
Treatment
trial period of low‐dose UDCA (13–15 mg/
There is no established medical therapy with kg daily) followed by liver enzyme assessment
a proven effect to halt disease progression in and continuation of UDCA only in patients
PSC. Liver transplantation remains the only showing significant ALP reduction (or effect
curative therapeutic option at present, but on pruritus). The strategy is hardly harmful
PSC may recur in the liver transplant. but its benefit is unclear. There is no good
Symptomatic treatment for pruritus and evidence to support the use of UDCA to pre
patient management is described Chapter 17. vent the development of colorectal cancer;
However, in the following section we discuss such practice is based on retrospective and in
some aspects related to current practice part inconsistent data and cannot be advised.
regarding medical and endoscopic treatment
in PSC and briefly mention potential future Antibiotics
avenues. Antibiotic therapy is indicated in bacterial
cholangitis. Prophylaxis with broad‐spec
Medical Treatment trum antibiotics is mandatory at biliary
Ursodeoxycholic Acid intervention. Biliary infections promote dis
UDCA in low to medium dose (10–20 mg/kg ease progression in PSC and prevention of
daily) is widely prescribed based on clinical repeated attacks is thought to be important
experience but practice varies between cen in order to avoid further scarring of the bile
ters and countries. However, while UDCA ducts. Some patients experiencing recurrent
has demonstrated effect in reducing ALP and bacterial cholangitis may need easy access to
other hepatic biochemistries in PSC, there is a supply of antibiotics (e.g. ciprofloxacin),
insufficient evidence to claim that UDCA with recommendations to start treatment at
the first sign of infection. Schemes involving the various aspects of the current model for
prescription of rotating antibiotics lack evi PSC pathogenesis such as bile acids (nor‐
dence and risk promoting antibiotic resis UDCA, obeticholic acid), the gut microbiota
tance and cannot be recommended. A and gut–liver axis (vancomycin in particular;
different gut microbiota has been demon fecal transplantation), inflammation (vedoli
strated in PSC compared to healthy controls zumab), and fibrosis (simtuzumab/anti‐
or UC patients and opens up intriguing pos LOXL2; bezafibrate/PPAR agonists). At the
sibilities of using antibiotics, probiotics or moment, drugs aiming at modulating chole
fecal transplantation to manipulate the gut stasis and antibiotics seem most promising.
microbiota with the aim of treating the bil
iary disease. Vancomycin has attracted Bile Acid Therapy
particular attention; however, data are cur Nor‐UDCA is a side‐chain‐shortened
rently insufficient for any form of manipula derivative of UDCA which unlike UDCA is
tion of the gut microbiota in the treatment of secreted unconjugated into the bile and
PSC and further studies are awaited. inhibits toxic bile formation; Phase II trial
results were promising (showing dose‐
Immunosuppression responsive ALP reduction) and a Phase III
There is no indication for the use of study is ongoing. The farnesoid X receptor
immunosuppressant agents to treat regular
(FXR) is one of the master regulators of the
adult PSC. Studies have shown no effect of protective pathways in hepatocyte chole
immunosuppression on transplant‐free sur
stasis, partly inducing its effects through the
vival whereas serious side effects appeared. downregulation of bile salt production.
Preliminary results from retrospective studies Obeticholic acid is an FXR agonist that has
investigating the effect of biological treatment been conditionally approved in PBC with
[tumor necrosis factor (TNF)‐α inhibitors] on promising results from a Phase II study in
PSC are so far disappointing, but data are PSC (safe and showed dose–response effect
scarce. Given the fact that PSC is an autoim on ALP reduction); however, concerns
mune disease as defined genetically and by remain in PSC for potential deleterious
other features, this is somewhat of a paradox. effects of FXR agonism in the presence of
However, in patients with features of AIH (i.e. cholestasis and increased pruritus has been
transaminases >5× ULN, IgG >2× ULN, and reported as a side effect in some patients.
typical or compatible histologic findings Recently, a small retrospective study reported
dependent on the presence or absence of auto biochemical improvement and reduced pru
antibodies, as defined in the Diagnosis section), ritus in patients treated by a combination of
immunosuppressive therapy may be beneficial UDCA and fenofibrate or bezafibrate, sug
and is recommended, with prescription follow gesting a role for PPAR agonists. Like
ing standard guidelines for the treatment of obeticholic acid, fibrates downregulate bile
AIH. The evidence for such treatment is weak acid synthesis, stimulate canalicular secre
and the response to treatment is variable and tion, and upregulate bile acid detoxification.
generally less pronounced than in AIH without A Phase II clinical trial is being planned.
a diagnosis of PSC; hence, the risk of side effects
(e.g. osteopenia) should be recognized and Microbiota Modulators
weighed against treatment effect during close Given the increasing evidence for a role of the
monitoring. intestinal microbiota on PSC pathogenesis,
antibiotic therapy appears attractive. Vancom
Future Drugs ycin is perhaps the best‐studied antibiotic in
Based on new insights into the pathogenesis PSC, with promising results in smaller studies
of PSC, there has been a surge of interest in in adults and children; however, the evidence is
clinical trials in PSC. Several new drugs are currently insufficient to support treatment
being investigated. Approaches vary, t argeting recommendations. A larger clinical trial is
ongoing. Another potential antibiotic therapy MRI/MRC should be performed prior to ERC
for PSC, although with insufficient evidence, is to confirm the indication and identify any
minocycline, whereas rifaximin did not show focal lesions. As further discussed in the EASL
any effect in a clinical trial. Clinical Practice Guidelines describing the
role of endoscopy in PSC [9], the optimal
Anti‐inflammatory Treatment interventional approach has not been resolved
Paradoxically, although PSC is an immune‐ and the choice between balloon dilatation
associated disease, immunosuppression has with or without short‐term stenting for 1–2
not proved effective in the past. The focus weeks, or short‐term stenting only is left to
now is mainly on preventing the potential the operator. Studies including the interna
aberrant “homing” of lymphocytes from the tional multicenter prospective DILSTENT
gut to the liver. Timolumab (BTT1023), an study have shown similar outcome for the two
anti‐VAP‐1 antibody, reduces the homing of alternative approaches, although some data
gut‐activated lymphocytes to the biliary tree suggest a higher rate of procedure‐related
and is currently being investigated in PSC. complications associated with stenting [16].
Current evidence indicates that biological Different stenting durations have not been
therapies used successfully for the treatment formally compared in PSC but short‐term
of IBD are safe but not effective treatments stenting is favored because stents tend to clog
for PSC (including adalimumab, infliximab rapidly in PSC and the efficacy reported in
and vedolizumab). studies of short‐term (1–2 weeks) versus stan
dard (8–12 weeks) stenting was similar.
Antibiotic Therapy Sphincterotomy, although not generally
Results from a Phase II clinical trial investi advised, is recommended following difficult
gating simtuzumab, an anti‐LOXL2 agent tar cannulation or if repeated procedures are
geting fibrosis, disappointingly did not show anticipated to reduce the risk of complications
any effect on event‐free survival, histologic in future procedures. PSC patients carry a
fibrosis scores, or hepatic collagen content. slightly higher risk of complications (cholangi
tis, post‐ERC pancreatitis) related to ERC
Better tools for risk stratification, improving compared to other patients and preprocedure
patient selection for clinical trials, and administration of prophylactic parenteral
surrogate end points to evaluate treatment antibiotics and rectal non‐steroidal anti‐
effect would facilitate the establishment of inflammatory drugs (e.g. 100 mg diclofenac or
effective therapy. It is possible that indometacin) is recommended. In case of a
combination therapy, attacking two or several high risk of post‐ERC pancreatitis, a prophy
components of pathogenesis, may yield lactic pancreatic stent should be considered.
improved therapeutic effect, or that better Endoscopic treatment of clinically significant
characterization or stratification of patients strictures in PSC has been demonstrated to
may direct tailored therapy targeting the dis improve symptoms but the evidence for an
ease‐driving factor in the individual patient. impact on prognosis is less clear, although
some studies have shown increased trans
Endoscopic Treatment plant‐free survival. ERC with brush sampling
Endoscopic treatment plays an important role plays an important role in the surveillance and
in PSC and should be considered whenever a diagnosis of cholangiocarcinoma in PSC.
patient with PSC shows increasing symptoms
(jaundice, pruritus, cholangitis), rapid increase
Surveillance for Malignancy
of cholestatic liver enzymes, or a new or pro
gressing dominant stricture accompanied by Malignancy is a major cause of death in PSC.
appropriate clinical findings [9]. Concomitant Several recommendations regarding surveil
ductal sampling with brush cytology should lance in PSC target the increased risk of hep
always be performed. Most often, combined atobiliary and colorectal cancer and are
aimed at the early detection of relevant can sensitivity as well as specificity. Combined
cers, focusing on measures with proof of MRI/MRC has the highest sensitivity (89%)
effect on outcome. Complications of chole and specificity (75%) for detection of particu
stasis justify other forms of surveillance. larly small cholangiocarcinomas. Cases with
suspicious findings should be followed up by
Colorectal Cancer: Colonoscopy endoscopic retrograde cholangiopancreatog
The risk of colorectal cancer is fivefold higher raphy (ERCP) with brush cytology. In a posi
than in IBD without PSC. Colonoscopy with tion paper, an international expert panel of
biopsies should be performed at diagnosis of hepatologists and radiologists suggest that
PSC in all patients without a former diagnosis contrast MRI/MRC should be performed
of IBD in order to diagnose subclinical colitis. within 6 months of diagnosis of PSC if not
Surveillance by colonoscopy should be per taken at the time of diagnosis as surveillance
formed regularly (annually or biannually) from for cholangiocarcinoma, as the risk is high in
the time of diagnosis of PSC, including in the first year following diagnosis. ERC with
young adults. Adherence to surveillance has brush cytology should be performed if a new
been demonstrated to increase survival. or progressing dominant stenosis is detected.
Gallbladder Cancer: Ultrasound Bone Mineral Density

Annual ultrasound is recommended to iden Cholestasis may cause steatorrhea, reduced
tify gallbladder polyps, which have a more vitamin D uptake, and osteoporosis. Bone
than 50% risk of neoplasia. Cholecystectomy mineral density should be assessed at diag
is advised if a polyp is found: the absolute risk nosis in adult patients and at regular inter
of malignancy is low but not absent in polyps vals depending on findings to identify
less than 0.8 cm in size and cholecystectomy patients at risk, and vitamin D supplements
should be considered in polyps 5–8 mm in may be recommended.
size in these patients [17].
Hepatocellular Cancer: Ultrasound Juvenile Sclerosing

Hepatocellular carcinoma occurs in patients Cholangitis
with PSC and cirrhosis, although possibly at
slightly lower rates compared to patients With the advent of non‐invasive biliary
with cirrhosis of other etiologies. Standard imaging, sclerosing cholangitis, hitherto con
surveillance (liver ultrasound every 6 months) sidered rare in children, is diagnosed with
is recommended in patients with PSC and increasing frequency in the pediatric age
cirrhosis; any added value of alpha‐fetopro group. Untreated sclerosing cholangitis prog
tein is debated. resses to liver cirrhosis and end‐stage liver
disease and is an important cause of mor
Cholangiocarcinoma bidity and mortality, accounting for approxi
There has been no evidence‐based support mately 2% of pediatric liver transplants in the
for surveillance of cholangiocarcinoma in USA between 1988 and 2008 (United Network
PSC, and international guidelines have not for Organ Sharing Data Report, October
recommended specific surveillance strat 2009. http://www.unos.org/data).
egies. However, a recent report supports a
beneficial effect of surveillance for cholan
Epidemiology in Pediatric Disease
giocarcinoma on survival [17]. The optimal
surveillance modality and frequency have There is paucity of data on the epidemiology
not been settled. Combining annual imaging of juvenile sclerosing cholangitis. It is a rare
(ultrasound, CT or MRI/MRC) with serum condition, with a reported incidence in
CA19‐9 is currently practiced by many cen North America of 0.1–0.2 cases per 100 000
ters. Annual CA19‐9 alone has suboptimal children and a reported prevalence of 1.5 per
100 000 children [4]. Epidemiologic data are Approximately 50% of the patients enrolled
hampered by the limited number of pub had alterations of the bile ducts characteristic
lished series of pediatric sclerosing cholangi of sclerosing cholangitis at presentation and
tis, most being retrospective, single center were diagnosed as having ASC. One‐quarter
and including small patient numbers [4]. of these had no histologic features suggesting
Reported incidences of the various clinical bile duct involvement despite abnormal chol
forms of sclerosing cholangitis differ, angiograms. Serologically, ASC patients were
reflecting different study designs, patterns of indistinguishable from AIH type 1. In contrast
referral and diagnostic protocols. In all retro to AIH, which had a clear female preponder
spective series cholangiographic studies were ance, ASC affected a similar proportion of
prompted by biochemical and/or histologic boys and girls. Surveillance endoscopy to
features of cholestatic disease. Boys are investigate possible IBD was performed in all
reported to be more affected than girls, cases, showing IBD in 45% of ASC and 20% of
20–40% having only intrahepatic cholangi AIH type 1. At presentation, immunologic
opathy. As for adults, there is a strong IBD and biochemical indices did not help in dis
association (60–90% of cases), with over two‐ criminating between AIH and ASC, although
thirds of patients having UC, the others the ALP/AST ratio was significantly higher in
Crohn’s disease or indeterminate colitis. IBD ASC. Biochemical response to predniso
can precede the diagnosis of liver disease by lone with or without azathioprine was similar
many years, be diagnosed at the same time, in AIH and ASC, leading to normalization of
or develop during follow‐up. transaminase levels, although bile duct
damage progressed over the years in half of
the ASC cases.
Autoimmune Sclerosing Cholangitis
In view of the postulated autoimmune
All retrospective studies recognize a form of pathogenesis of adult PSC, the question
sclerosing cholangitis with strong autoim arises as to whether ASC is the early manifes
mune features. However, in some this tation of what will eventually become adult
condition is reported to respond favorably to PSC. Similarities and differences between
immunosuppression, having a better prog PSC and ASC are summarized in Table 8.1.
nosis than PSC; in others the prognosis of
AIH/PSC overlap syndrome is reported to be
Diagnosis in Children
severe and not ameliorated by immunosup
pressive treatment or is similar to that of PSC The diagnosis of sclerosing cholangitis in
irrespective of treatment [4]. Major limita children/adolescents is based on the finding
tions of retrospective studies are uneven of abnormal bile ducts on cholangiography
diagnostic protocols and lack of accurate often, but not always, in association with a
information on IBD treatment before the cholestatic liver profile. Conditions associ
diagnosis of sclerosing cholangitis, as immu ated with bile duct disease or secondary
nosuppression for IBD might affect the liver sclerosing cholangitis should be excluded by
disease. appropriate investigations. Strictly “primary”
Only one study has been conducted pro sclerosing cholangitis is rare [3].
spectively with the aim of establishing the The diagnosis of ASC is based on abnormal
relative incidence of AIH and ASC among cholangiographic findings by MRC at pre
children with liver disease and positive auto sentation in children that fulfill the diagnosis
immune serology, by performing cholangio of AIH. Both intrahepatic and extrahepatic
grams at disease onset, irrespective of bile ducts are most often involved, with iso
biochemical or histologic evidence of chole lated involvement of the intrahepatic bile
static disease. Patients were recruited over ducts in one‐third of the cases. A liver biopsy
16 years [3] and are currently being followed. is mandatory in the diagnostic work‐up of
Table 8.1 Differences between adult primary sclerosing cholangitis (PSC) and juvenile autoimmune sclerosing
cholangitis (ASC).
Adult PSC Pediatric ASC
Response to No Biochemical in 80%

immunosuppressive No progression of
treatment disease in 50%
Male to female ratio 2 : 1 1 : 1
Anti‐neutrophil 26–94% 75%
cytoplasmic antibody
SMA/ANA positivity Up to 83% 96%
Association with IBD 60–80% 50%
Biochemical cholestatic Part of the diagnostic criteria May be missing at
profile presentation
Liver biopsy part of the No Yes
diagnostic criteria
Outcome Median transplantation‐free 10‐year transplant‐free
survival: 13–21 years survival: 65%
ANA, anti‐nuclear antibody; IBD, inflammatory bowel disease; SMA, smooth muscle antibody.
Figure 8.3 Dense portal tract inflammatory infiltrate comprising lymphocyte and plasma cells and invading
the surrounding parenchyma (interface hepatitis) in a child with autoimmune sclerosing cholangitis. The arrow
shows bile duct damage. Hematoxylin and eosin stain.
ASC. The classical histologic changes of hepatitis, similar to what is described in AIH
periductular concentric fibrosis (often (Figure 8.3). Other reported histologic
referred to as “onion skin fibrosis”) are only features include fibro‐obliterative cholangitis
rarely seen in children, probably because and primary ductular involvement [6].
they result from long‐standing disease. Most The International Autoimmune Hepatitis
commonly the histology shows dense mono Group scoring systems, devised for the diag
nuclear inflammatory infiltrate of the portal nosis of AIH in adults, do not discriminate
tract with several plasma cells and interface between AIH and ASC, as they do not include
cholangiographic studies at disease onset. A is higher than in AIH. The immune serologic
position paper of the European Society of profile is most often indistinguishable from
Paediatric Gastroenterology, Hepatology AIH type 1; the great majority of patients are
and Nutrition (ESPGHAN) Hepatology positive for anti‐nuclear and/or anti‐smooth
Committee has recently proposed a diag muscle antibodies, whereas the presence of
nostic scoring system for juvenile autoim anti‐liver kidney microsomal type 1 antibody
mune liver diseases that awaits validation is rare [3]. Perinuclear anti‐neutrophil
[18]. Current EASL guidelines recommend nuclear antibody is more frequently positive
that every child or adolescent diagnosed with in ASC than in AIH, being found in three‐
AIH must undergo MRC in order to actively quarters of cases [3]. Data from the UK show
screen for ASC [19]. All children diagnosed that ASC is associated with HLA‐DRB1*1301
with ASC or PSC should undergo upper gas genotype, at variance with AIH type 1, which
trointestinal endoscopy and full colonoscopy is associated with DRB1*0301, and with AIH
in order to screen for concomitant IBD. type 2, associated with DRB1*0701 [4]. One‐
Juvenile small‐duct sclerosing cholangitis quarter of patients with abnormal cholangio
is a poorly characterized entity. Reliable diag gram do not have histologic evidence of bile
nostic criteria for this condition are lacking. duct injury. Inflammation of the portal tracts
Moreover, genetic defects such as MDR3 is frequent, and interface hepatitis, the histo
deficiency are associated with small‐duct logic hallmark of AIH, is as frequent in ASC
sclerosing cholangitis, and the role of auto as in AIH. Histologically, AIH is associated
immune small‐duct disease is still unknown. with more active inflammation and a higher
Therefore, this condition in children cannot frequency of cholangitis than ASC, but
be directly compared with adult small‐duct overall the histologic picture is similar in the
PSC and deserves further research. two conditions and the final diagnosis is
based on cholangiographic findings. AIH can
evolve to ASC, i.e. AIH patients with normal
Clinical Features in Pediatric Disease
baseline cholangiogram can develop bile
In retrospective studies including various duct disease during follow‐up. This observa
forms of juvenile sclerosing cholangitis most tion raises the question as to whether AIH
patients are males. Symptoms at presenta and ASC belong to the same disease spec
tion depend on the cause underlying the bile trum or are different conditions.
duct damage. ASC affects males and females
equally, the median age at presentation being
Treatment of Pediatric Disease (ASC
12 years. Symptoms and clinical signs at dis
and Juvenile PSC)
ease onset are indistinguishable from those
of AIH, the majority presenting with jaun Treatment of sclerosing cholangitis varies
dice, hepatosplenomegaly and intestinal according to the underlying cause of the bile
symptoms. Extrahepatic autoimmune dis duct injury. For ASC and PSC, the treatment
eases, including IBD, are present in about schedule varies between centers. Randomized
half of the patients. IBD may be asymptom controlled trials are missing. Published ret
atic at presentation and diagnosed at endos rospective series have used uneven diag
copy. Autoimmune disorders in first‐degree nostic criteria and data on treatment
relatives are found in one‐third of the chil regimens are often scarce. Recommended
dren with either ASC or AIH. treatment for ASC is the same as for AIH:
In ASC, the biochemical profile is often steroids with or without azathioprine.
similar to that of AIH, and at least one‐ Additional UDCA (15 mg/kg daily) is widely
quarter of the patients have normal ALP and prescribed even though data on its impact on
GGT levels. Transaminase levels tend to be disease progression and survival are lacking.
lower than in AIH, while the ALP/AST ratio Most ASC patients have an excellent
biochemical response to immunosuppres cancer surveillance are lacking, and those

sion. However, in half of the cases the bile already described for adult PSC should be
duct disease progresses over time, with a 10‐ implemented.
year transplant‐free survival of 65% in ASC Small‐duct sclerosing cholangitis has been
compared to 100% in AIH. reported to have a more favorable course in
Oral vancomycin has shown promising terms of biliary complications, but not in
results in a small study including 14 children terms of transplant‐free survival [6].
with sclerosing cholangitis and IBD, but Liver transplant is offered to patients who
further high‐quality data are required before progress to end‐stage liver disease despite
any recommendation can be made. For ASC treatment. However, post‐transplant recur
patients with intolerance or insufficient rence is frequent and increases over time.
response to azathioprine, mycophenolate Reported frequencies range from 10 to 50%
mofetil may be effective [4]. For those with dif depending on the time from transplantation
ficult‐to‐control IBD, anti‐TNF‐α treatment and the criteria employed to diagnose recur
has been reported to be beneficial to both liver rence. The prognosis of recurrent ASC is
and gut [20]. particularly poor, the majority of patients
Liver transplantation is needed for patients requiring retransplantation [4]. There is no
progressing to end‐stage liver disease, who standard treatment for ASC recurrence,
often have a difficult‐to‐control concomitant though increases in the steroid dose
IBD. combined with azathioprine or mycopheno
late mofetil are advisable. UDCA has no
proven benefit. Both progression of liver dis
Prognosis in Pediatric Disease
ease before and recurrence of sclerosing
Pediatric disease appears to be marginally cholangitis after transplantation are usually
less aggressive than the adult disease, with a associated with poorly controlled IBD,
10‐year transplant‐free survival of 70% as underscoring the importance of controlling
compared to 61–64% in adults, possibly the intestinal disease in order to avoid pro
because children are diagnosed at an earlier gression of liver damage.
stage [6]. However, about half of the children
develop portal hypertension or biliary com
plications in the first decade following the Conclusion
diagnosis [6].
Immunosuppressive treatment is effective Juvenile sclerosing cholangitis is the result of
in controlling both parenchymal and biliary different pathologies, with a high proportion
disease long term in 50% of ASC cases. The of patients having associated IBD and/or auto
real prognosis of ASC compared to PSC immune features. In contrast to adult PSC,
cannot be adequately established in retro ASC responds satisfactorily long term to
spective cohorts with variable diagnostic immunosuppressive treatment in at least 50%
approaches and treatment protocols. In the of cases. The question as to whether ASC is
large series from Canada, there was no the pediatric manifestation of adult PSC
difference in the 5‐year transplant‐free remains open, and needs to be answered by
survival rate between PSC and ASC [6]. Liver long‐term longitudinal interdisciplinary
malignancies are rarer in juvenile sclerosing studies. Multicenter prospective studies are
cholangitis than in adult PSC, even over long needed to better define the juvenile sclerosing
periods of follow‐up. Cholangiocarcinoma cholangitis phenotype and diagnostic criteria
has been reported in only a few patients, and for exploring pathogenic mechanisms
mostly males, all during their second decade with the aim of devising more effective
of life [6]. Specific pediatric guidelines on treatment for this often unrelenting condition.
References
1 Karlsen, T.H., Folseraas, T., Thorburn, D., 9 Aabakken, L., Karlsen, T.H., Albert, J. et al.
and Vesterhus, M. (2017). Primary (2017). Role of endoscopy in primary
sclerosing cholangitis: a comprehensive sclerosing cholangitis: European Society of
review. J. Hepatol. 67: 1298–1323. Gastrointestinal Endoscopy (ESGE) and
2 Boonstra, K., Weersma, R.K., van Erpecum, European Association for the Study of the
K.J. et al. (2013). Population‐based Liver (EASL) Clinical Guideline. Endoscopy
epidemiology, malignancy risk, and outcome 49: 588–608.
of primary sclerosing cholangitis. 10 Boberg, K.M., Chapman, R.W., Hirschfield,
Hepatology 58: 2045–2055. G.M. et al. (2011). Overlap syndromes: the
3 Gregorio, G.V., Portmann, B., Karani, J. et al. International Autoimmune Hepatitis
(2001). Autoimmune hepatitis/sclerosing Group (IAIHG) position statement on a
cholangitis overlap syndrome in childhood: controversial issue. J. Hepatol. 54: 374–385.
a 16‐year prospective study. Hepatology 33: 11 Boberg, K.M., Bergquist, A., Mitchell, S.
544–553. et al. (2002). Cholangiocarcinoma in
4 Mieli‐Vergani, G. and Vergani, D. (2016). primary sclerosing cholangitis: risk factors
Sclerosing cholangitis in children and and clinical presentation. Scand. J.
adolescents. Clin. Liver Dis. 20: 99–111. Gastroenterol. 37: 1205–1211.
5 Terziroli Beretta‐Piccoli, B., Vergani, D., 12 Schramm, C., Eaton, J., Ringe, K.I. et al.
and Mieli‐Vergani, G. (2018). (2017). Recommendations on the use of
Autoimmune sclerosing cholangitis: magnetic resonance imaging in PSC. A
evidence and open questions. J. position statement from the International
Autoimmun. 95: 15–25. PSC Study Group. Hepatology 66:
6 Deneau, M.R., El‐Matary, W., Valentino, P.L. 1675–1688.
et al. (2017). The natural history of primary 13 Kim, W.R., Therneau, T.M., Wiesner, R.H.
sclerosing cholangitis in 781 children: a et al. (2000). A revised natural history
multicenter, international collaboration. model for primary sclerosing cholangitis.
Hepatology 66: 518–527. Mayo Clin. Proc. 75: 688–694.
7 Lunder, A.K., Hov, J.R., Borthne, A. et al. 14 Vesterhus, M., Hov, J.R., Holm, A. et al.
(2016). Prevalence of sclerosing cholangitis (2015). Enhanced liver fibrosis score
detected by magnetic resonance predicts transplant‐free survival in primary
cholangiography in patients with long‐term sclerosing cholangitis. Hepatology 62:
inflammatory bowel disease. 188–197.
Gastroenterology 151: 660–669.e4. 15 Corpechot, C., Gaouar, F., El Naggar, A.
8 Weismuller, T.J., Trivedi, P.J., Bergquist, A. et al. et al. (2014). Baseline values and changes in
(2017). Patient age, sex, and inflammatory liver stiffness measured by transient
bowel disease phenotype associated with elastography are associated with severity of
course of primary sclerosing cholangitis. fibrosis and outcomes of patients with
Gastroenterology 152: 1975–1984.e8. primary sclerosing cholangitis.
Gastroenterology 146: 970–979. quiz ESPGHAN Hepatology Committee

e15–16. Position Statement. J. Pediatr.
16 Ponsioen, C.Y., Arnelo, U., Bergquist, A. Gastroenterol. Nutr. 66: 345–360.
et al. (2018). No superiority of stents vs 19 European Association for the Study of the
balloon dilatation for dominant strictures in Liver (2015). EASL clinical practice
patients with primary sclerosing cholangitis. guidelines: autoimmune hepatitis. J.
Gastroenterology 155: 752–759.e5. Hepatol. 63: 971–1004.
7 Ali, A.H., Tabibian, J.H., Nasser‐Ghodsi, N.
1 0 Nedelkopoulou, N., Vadamalayan, B.,
2
et al. (2018). Surveillance for hepatobiliary Vergani, D., and Mieli‐Vergani, G. (2018).
cancers in patients with primary sclerosing Anti‐TNFalpha treatment in children and
cholangitis. Hepatology 67: 2338–2351. adolescents with combined inflammatory
8 Mieli‐Vergani, G., Vergani, D., Baumann,
1 bowel disease and autoimmune liver
U. et al. (2018). Diagnosis and management disease. J. Pediatr. Gastroenterol. Nutr. 66:
of pediatric autoimmune liver disease: 100–105.
163
IgG4‐Related Liver and Biliary Disease

Eleanor Barnes1,2 and Emma L. Culver1,2
1
Translational Gastroenterology Unit, Nuffield Department of Medicine, University of Oxford, Oxford, UK
2
Oxford NIHR Biomedical Research Centre, Oxford, UK
Abstract
IgG4-related disease (IgG4-RD) is a multisystem disorder associated with inflammation and fibrosis
in affected organs. Almost any organ may be affected, forming part of the IgG4-RD spectrum.
Patients with IgG4-RD may present to hepatologists with symptoms related to focal masses or dif-
fuse swelling of the pancreas, with intrahepatic and/or extrahepatic biliary disease, as inflammatory
pseudotumors of the liver and biliary tree, and as a variant of autoimmune hepatitis. IgG4-RD
involving the pancreas, bile ducts and/or liver may mimic pancreatic cancer, cholangiocarcinoma,
and primary sclerosing cholangitis. Distinguishing IgG4-RD from malignancy is critical, often
requiring biopsy and histologic analysis. IgG4-RD may be associated with progressive organ
dysfunction, organ failure and malignancy. IgG4-RD is usually treated with oral corticosteroids as
first line therapy, but relapse rates are high once steroid treatment is tapered or stopped. Secondline
and steroidsparing agents are commonly used, but the optimal use of these is not well defined in
head to head clinical studies.
Keywords biliary tree; first line therapy; histologic analysis; IgG4-related disease; inflammatory
pseudotumors; pancreatic cancer; primary sclerosing cholangitis; secondline therapy; steroid sparing
agents
Key Points
●● IgG4-related disease (IgG4‐RD) is a multifocal masses or diffuse swelling of the
system disorder associated with inflam- pancreas (autoimmune pancreatitis), with
mation and fibrosis in affected organs. intrahepatic and/or extrahepatic biliary
Almost any organ may be affected, form- disease (IgG4‐related sclerosing cholan-
ing part of the IgG4‐RD spectrum. It is gitis), as inflammatory pseudotumors of
increasingly diagnosed globally. the liver and biliary tree (IgG4‐related
●● Patients with IgG4‐RD may present to hepatopathy), and as a variant of autoim-
hepatologists with symptoms related to mune hepatitis.
●● IgG4‐RD involving the pancreas, bile steroid‐sparing agents are commonly

ducts and/or liver may mimic pancreatic used, but the optimal use of these is not
cancer, cholangiocarcinoma, and pri- well defined in head-to-head clinical
mary sclerosing cholangitis. Distinguishing studies.
IgG4‐RD from malignancy is critical, often ●● Rituximab is known to be effective in
requiring biopsy and histologic analysis. inducing remission and is generally used
●● IgG4‐RD may be associated with progres- in patients who are intolerant or relapse
sive organ dysfunction, organ failure and after steroid therapy. Relapse still occurs
malignancy. and maintenance rituximab therapy may
●● The histologic hallmarks of the condition be considered in these patients.
are a lymphoplasmacytic infiltrate com- ●● Disease pathogenesis is not fully under-
posed of CD4+ T cells, B cells and plasma stood, but immune analysis suggests
cells that produce IgG4 antibodies in that the IgG4 B‐cell response is associ-
affected organs, a storiform fibrosis and ated with T follicular helper 2 and cyto-
an obliterative phlebitis, with variable toxic T effector cells. There is a strong
presence of eosinophils. An IgG4:IgG association with allergy and T helper 2
plasma cell ratio above 40% is supportive pathways. Causal factors that drive the
in the histologic diagnosis of IgG4‐RD. disease have not been reproducibly
●● Serum IgG4 may be elevated in up to two‐ identified.
thirds of patients, but can be within ●● It is not known if the full clinical spec-
normal limits, and is not a good biomarker trum of IgG4‐RD truly represents a
of disease. single disease entity. However, there
●● Specific criteria for disease diagnosis appears to be sufficient commonality in
have been established and recently mod- underlying pathogenesis and treatment
ified, based on a combination of clinical, pathways to categorize diverse clinical
laboratory, radiologic, and histologic phenotypes as IgG4‐RD. Large cohorts of
characteristics. carefully phenotyped IgG4‐RD patients,
●● IgG4‐RD is usually treated with oral corti- stratified by subtypes and followed pro-
costeroids as first-line therapy, but relapse spectively, are required to better under-
rates are high once steroid treatment is stand long‐term outcomes and response
tapered or stopped. Second‐line and to therapies.
Introduction and Historical The very earliest reports of IgG4‐RD were

Perspective reported in the literature in the early 1960s,
when the pancreatic form of the disease was
IgG4‐related disease (IgG4‐RD) is a first described as “chronic inflammatory
relatively newly described disease, clinically sclerosis of the pancreas” and the biliary
complex and although treatable is fre- form as “sclerosing cholangitis” associated
quently not recognized or misdiagnosed. with retroperitoneal fibrosis. In the 1990s,
Malignancy is often high on the list of reports continued to emerge from the
differential diagnoses. Therefore, failing to Japanese and English literature of cases of
make a diagnosis or miscalling the diagnosis pancreatitis that were responsive to immu-
will have serious adverse clinical conse- nosuppressive therapy and the term “auto
quences for patients, and for this reason the immune pancreatitis” (AIP), reflecting the
topic is important. presumed pathogenesis, was adopted.
Chapter 9 IgG4-Related Liver and Biliary Disease 165
Table 9.1 Disease nomenclature that forms part of the IgG4‐RD

spectrum and target organ.
IgG4‐RD disease Target organ
Mikulicz disease Salivary and lacrimal glands

Kuttner tumors Submandibular gland
Reidel thyroiditis Thyroid
Chronic sclerosing aortitis Aorta
Abdominal aortitis Aorta
Retroperitoneal fibrosis Retroperitoneum
Sclerosing cholangitis Biliary tree
Orbital pseudotumors Orbital annexa
Eosinophilic angiocentric fibrosis Nasal cavities and sinuses
Multifocal fibrosclerosis Multiple and various organs
However, it is currently not clear if AIP is Etiology and Pathogenesis

truly an autoimmune condition. It was not of IgG4‐RD
until 2001 that elevated levels of serum IgG4
levels were recognized as an important dis- Development and Characteristics
ease biomarker for AIP. of an IgG4 Antibody Response
Over the last 15 years, our conceptual
understanding of IgG4‐RD has progressed The etiology of IgG4‐RD is currently
rapidly; patients initially diagnosed with AIP unknown. However, understanding disease
were found to have, or to develop, multi pathogenesis may give insights into etiology
organ disease such that AIP is now clearly and many aspects of IgG4‐RD disease biology
recognized to be part of an IgG4‐RD
are currently under investigation. The IgG4
spectrum that may affect almost any organ. antibody, in health, is usually a minor compo-
Many diseases historically named and nent of total IgG (<6%), but can be enhanced
thought to be organ specific have now been by chronic exposure to antigens. For example,
reclassified and are thought to belong to the immunotherapy used for the treatment of
IgG4‐RD spectrum (Table 9.1). Reflecting allergy with repeat antigen stimulation is
the new recognition of IgG4‐RD as a multi- associated with a decline in IgE levels and an
system disease, a condition that was initially increase in IgG4 levels and beekeepers
the domain of specialists within hepatobili- repeatedly exposed to bee venom develop
ary medicine is now frequently managed by elevated IgG4 responses. An IgG4 response
rheumatologists and clinicians from other may also be generated in response to repeat
disciplines. Our understanding has pro- exposure to therapeutic monoclonal anti-
gressed further in the last 5 years, with new bodies such as adalimumab in patients with
studies giving insight into disease pathogen- rheumatoid arthritis. In general, IgG4 anti-
esis [1], the development of new clinical cri- bodies are considered to be “regulatory,” lim-
teria for diagnosis, new therapeutic treatment iting effector B‐cell responses, both in the
options [2,3], and a realization that IgG4‐RD context of inflammation and also in cancer.
may be associated with organ failure, mor- The biological properties of IgG4 are in keep-
bidity and mortality, and probably malig- ing with this regulatory role; in health, IgG4
nancy [4]. is the only IgG antibody that is unable to
a ctivate complement and has a low affinity to fore related to IgG4‐RD underlying processes
Fcγ receptors. Uniquely, IgG4 antibodies are rather than allergy per se [7]. The human
able to exchange half molecules resulting in response to allergens drives both an IgE and
bispecific antibodies that are unable to cross‐ an IgG4 response and both are dependent on
link and therefore result in the formation of interleukin (IL)‐4 as a “class switching factor”
small immune complexes. Because of these but the kinetics and drivers of the responses
properties, IgG4 antibodies are unable to are different, with IgE and a modified Th2
trigger immunologic effector functions and response dominated by IgG1 developing
have a limited role in the clearance of antigens early. Only after repeat antigen exposure
and in inflammation and may actively inhibit does IgG4 start to dominate and it is likely
IgG1 effector functions. IgG4‐positive B cells that additional immune factors, possibly
may be generated in IgG4‐RD as a regulatory those involved in immune tolerance in
response to an underlying inflammatory, response to chronic antigen stimulation such
antigen‐driven process that has yet to be as IL‐10 or transforming growth factor
clearly defined. (TGF)‐β, are required to drive the IgG4
However, studies of IgG4 biology in patients response. These same factors may also drive
with IgG4‐RD have shown that IgG4 anti- the fibrotic responses that ultimately lead to
bodies can form immune complexes via C1q organ failure in IgG4‐RD. Rituximab therapy
binding via both the classical and mannose lec- leads to a major decline in both IgG4 and IgE
tin‐binding pathway. This, together with the levels in IgG4‐RD patients, suggesting that
hypocomplementemia that is found in more both isotypes are produced, at least in part by
than 30% of patients with AIP implicates short‐lived antibody‐secreting B cells [2].
complement activation in the pathogenesis [5]. Some studies suggest that patients with
In addition, immune complex formation (C3c IgG4‐RD and allergy are a specific patient
and IgG4) has been demonstrated in the tissue subset characterized by circulating Th2
in pancreatic and bile ducts in AIP and IgG4‐ memory cells [8] and targeting distinct auto-
related sclerosing cholangitis (IgG4‐SC), and antigens [9]. The Th2 response in IgG4‐RD is
also to colocalize with collagen IV, and we polyclonal, supporting the notion that mul-
have recently shown decreased Fc antibody tiple antigens drive disease [10,11].
galactosylation in AIP/IgG4‐SC patients com-
pared with healthy controls, a post-translational
Antigens That May Drive an IgG4‐RD
modification important in complement
Response
activation [6]. Overall, this may support a
novel role of IgG4 antibodies in IgG4‐RD path- Since IgG4 antibodies can be generated in
ogenesis, and a potential role in fibrosis. response to chronic antigen exposure, efforts
An IgG4 antibody response is also related, have been made to identify self or exogenous
at least in part, to allergic pathways, although antigens that may trigger IgG4‐RD. However,
the role of allergy in IgG4‐RD is currently the IgG4 response in IgG4‐RD is oligoclonal
debated. High levels of polyclonal IgE, eosin- and reactive against a range of antigens,
ophilia and an increase in allergy and/or including food antigens [11], suggesting that
atopy is described in some IgG4‐RD cohorts no single antigen is driving the IgG4 anti-
[1]. One hypothesis is that IgG4‐RD is a two‐ body response seen in IgG4‐RD. IgG1 anti-
hit disease process: the first hit is the bodies against a range of self antigens have
development of allergy and the second hit been associated with IgG4‐RD, including
redirects the immune response to an IgG4‐ lactoferrin, carbonic anhydrase, pancreatic
RD phenotype. Others have suggested that secretory trypsin inhibitor, and heat shock
the increase in IgE and eosinophilia occurs in protein but none are specific for the disease.
the absence of atopy in patients with IgG4‐ Antibodies to galectin‐3 [12], annexin 11
RD and that this immune phenotype is there- [13], and laminin 511 [9] have more recently
been associated with IgG4‐RD in major histopathologic severity within affected

cohort subsets (18% galectin‐3, 28% annexin organs. These IgG4 antibodies were then
11, and 51% laminin 511). All are ubiquitous transferred to pancreatic tissue in patients
proteins that vary in abundance in organs, with AIP, and IgG4 antibodies bound to col-
and findings need to be confirmed in addi- lagen IV‐expressing extracellular matrix.
tional patient cohorts. The observation that Together these data suggest that although
antibodies against laminin 511 were specifi- usually “regulatory,” in IgG4‐RD both IgG1
cally detected in IgG4‐RD patients without and IgG4 may directly contribute to disease
allergy suggests that substratification of pathogenesis [15]. Additional studies have
patients depending on clinical phenotype recently implicated blood and tissue cyto-
will be required to enhance the specificity of toxic CD4+ T cells, through RNA‐seq anal-
putative autoantibodies for diagnostic utility. ysis of tissue T‐cell subsets [10].
To date there is no convincing evidence that In summary, IgG4‐RD appears to reflect an
IgG4‐RD is an autoimmune disease, leading aberrant immune process associated with
investigators to seek exogenous antigens as inflammation, fibrosis and tissue destruc-
potential triggers; gastric Helicobacter pylori tion. No specific self or exogenous antigens
has been proposed to cause IgG4‐RD through have been systematically linked to the IgG4
molecular mimicry, but we find no increased response, though specific patient subsets
risk of H. pylori infection, and no evidence of may be associated with immune responses to
B‐cell memory to H. pylori antigens in a some self antigens. The question is often
cohort of IgG4‐RD patients [14]. Whilst some raised as to whether IgG4‐RD really repre-
have proposed that occupational exposure sents a single disease. There appears to be
may be associated with IgG4‐RD, this needs to sufficient commonality in underlying patho-
be evaluated in larger cohorts with detailed genesis and treatment pathways to catego-
occupational exposure histories. rize diverse clinical phenotypes as IgG4‐RD
but large cohort studies with prospective
sampling, and immune analysis will be
IgG4 Antibodies and Pathogenesis
required to address this question.
There is some evidence to support the notion
that IgG4‐positive B cells or antibodies are
driving immune pathology and organ Incidence and Prevalence
damage. B cells may be linked to pathogen- of IgG4‐RD
esis since rituximab therapy leads to a decline
in circulating IgG4‐positive plasmablasts in As IgG4‐RD has been only recently described,
associated with disease resolution. IgG4 anti- is difficult to diagnose, and presents to mul-
bodies from patients with AIP may bind to tiple clinical specialities, the true prevalence
normal pancreatic tissue, but IgG1 has also and incidence of IgG4‐RD is not known.
been shown to bind, and it is not clear if these Furthermore, the diagnostic criteria for
antibodies arise as a consequence of, rather IgG4‐RD have evolved over time, compli-
than being causally linked to, pancreatic dis- cating both the inclusion criteria within and
ease. Interestingly, the passive transfer of comparisons between prevalence studies.
IgG1 and IgG4 antibodies purified from the Current prevalence data are largely confined
serum of patients with active AIP and of con- to Japan and include only AIP, where an
trols into mice leads to pathologic changes in overall prevalence rate of 4.6 and an annual
both pancreatic and salivary exocrine organs. incidence rate of 1.4 per 100 000 population
Although transfer of either purified IgG1 or is described [16]. However, since this study
IgG4 alone produced the disease phenotype, assessed patients in a hospital setting, with
the transfer of both antibodies together led an established diagnosis of AIP, the true
to reduced IgG1 deposition and decrease in prevalence rate is likely to be much higher.
linical Characteristics of IgG4‐

C salivary glands, whilst orbital disease pres-
related Hepatobiliary Disease ents with pain in the orbit or proptosis and
most commonly includes the lacrimal gland
Hepatobiliary Disease and soft tissue, although extraocular muscles
and palpebral and optic nerves may also be
IgG4‐RD typically affects males (>70%) in involved. Eye involvement is very frequently
the fifth and sixth decades of life [17]. Clinical bilateral (>50%) [19].
presentation is dependent on the organs Thoracic manifestations of IgG4‐RD are
affected. Hepatobiliary and pancreatic dis- numerous and can mimic many other
ease commonly present with abdominal pain, respiratory conditions. In a prospective cohort,
weight loss and jaundice, closely mimicking over 40% of IgG4‐RD patients had radiologic
malignancy, whilst pancreatic disease may and/or histologic evidence of thoracic involve-
present with endocrine or exocrine insuffi- ment, predominantly mediastinal lymph-
ciency. Two types of AIP are recognized: adenopathy, the majority associated with
●● type 1 is part of the IgG4‐RD spectrum; multisystem disease outside the chest [20].
●● type 2 (also called idiopathic duct‐centric Recent efforts using large cross‐sectional
pancreatitis) affects younger patients, is international patient cohorts (765 cases) have
not associated with systemic disease, and identified four major patient groups, including
is not classified as IgG4‐RD. pancreatic–hepatobiliary disease (31%), retro-
peritoneal fibrosis and/or aortitis (24%), head
IgG4‐RD may also affect the liver with and neck‐limited disease (24%), and classic
inflammatory pseudotumors. An IgG4‐ Mikulicz syndrome with systemic involvement
related autoimmune hepatitis has been (22%). Interestingly, being Asian or female
described but it is uncertain to what extent predisposed individuals to head and neck‐
it is part of the IgG4‐RD spectrum. Hepato limited disease [21]. However, almost any
biliary disease may arise in the absence of any organ may be affected by IgG4‐RD, high-
other organ involvement at presentation, lighting the need to promote awareness to
though other organs may be affected subse- physicians of all specialties if the diagnosis
quently. Acute pancreatitis may exhibit mini- rates are to increase. To complicate diagnosis
mally elevated serum IgG4 levels, typically further, coexistent autoimmune disease (e.g.
caused by alcohol and gallstones, but is not a thyroid and celiac disease) has been described
manifestation of IgG4‐RD. in 10% of patients with IgG4‐RD [4]. IgG4‐RD
is likely to present to the hepatobiliary spe-
Disease Outside the Hepatobiliary System cialist, but knowledge of the disease pheno-
type outside the hepatobiliary system is
IgG4‐related hepatobiliary disease is often essential for effective patient management.
found in association with additional organ
involvement.
Renal disease is frequently detected on iagnostic Criteria, Histologic
D
imaging in patients with other symptoms, and
seems to present relatively infrequently in the
Features, and Approach
absence of other organ involvement. Tubulo to Diagnosis of IgG4‐related
interstitial nephritis with abundant IgG4‐ Hepatobiliary Disease
positive plasma cells is the most common
finding in biopsy (90%) with a minority having Diagnostic Criteria for AIP and IgG4‐RD
membranous nephropathy [18]. Specific criteria have been established by a
Retroperitoneal fibrosis and aortitis are also consensus of specialists to support the diag-
well‐recognized manifestations of IgG4-RD. nosis of IgG4‐RD. The first guidelines were
Head and neck disease frequently presents created specifically for AIP in both eastern
with swelling of mandibular and/or parotid Asia (the Japan–Korea consensus criteria)
and in the Mayo clinic, USA (the HISORt cri- be evaluated and applied within an appro-
teria) [22] (Table 9.2). More recently, the priate clinical context. In general, the more
Comprehensive Diagnostic Criteria have features of IgG4‐RD that meet the criteria,
been established that can be usefully applied the more secure the diagnosis. Even with
to all manifestations of IgG4‐RD, rather than these criteria, diagnosis is often complex and
AIP alone [23] (Table 9.3). These guidelines is best confirmed in a multidisciplinary team
each focus on a combination of environment, with specialists familiar with
histopathologic findings, imaging, serum
the disease.
IgG4 levels, and response to therapy. No Making the correct diagnosis in IgG4‐RD
single criterion can be used to make the is imperative since the disease requires
diagnosis and the diagnostic criteria need to lifelong follow‐up and/or long‐term

Table 9.2 HISORta criteria for AIP.
1 Histopathology: one or both Characteristic appearances within biopsy or resection materialb

criteria required At least 10 IgG4‐positive plasma cells per high power field within areas
of lymphoplasmacytic infiltrate
2 Imaging and serology: all Diffusely enlarged pancreas with delayed and ‘rim’ enhancement
three criteria required Irregular pancreatic duct
Increased serum IgG4 concentration
3 Response to steroid therapy: Unexplained pancreatic disease after a full clinical work‐up, including
all three criteria required exclusion of cancer
Raised serum IgG4 concentration and/or extrapancreatic organ
involvement with increased numbers of tissue IgG4‐positive plasma cells
Resolution or marked improvement in disease with steroid therapy
a
HISORt: histology, imaging, serology, other organ involvement, response to therapy.
b
This includes a lymphoplasmacytic infiltrate, “storiform” fibrosis and obliterative phlebitis; the inflammatory cell
infiltrate alone is not sufficient to meet this criterion.
Source: Chari [22]. Reproduced with permission of Springer.
Table 9.3 The Comprehensive Diagnostic Criteria for all manifestations of IgG4‐RD.
1 Clinical examination showing characteristic diffuse/localized swelling or masses in single or multiple

organs
2 Blood tests shows elevated serum IgG4 concentrations (>135 mg/dl)
3 Histopathologic examination shows:
●● Marked lymphocyte and plasmacytic infiltration and fibrosis
●● Infiltration of IgG4‐positive plasma cells: ratio of IgG4+/IgG+ cells >40% and >10 IgG4‐positive
plasma cells per high power field
Definite diagnosis: meets criteria 1 + 2 + 3
Probable: diagnosis: meets criteria 1 + 3
Possible diagnosis: meets criteria 1 + 2
Note that it is important to differentiate IgG4‐RD from malignant tumors of each organ (e.g. cancer, lymphoma) and
similar diseases (e.g. Sjogren syndrome, primary sclerosing cholangitis, Castleman disease, secondary retroperitoneal
fibrosis, Wegener granulomatosis, sarcoidosis, Churg–Strauss syndrome) by additional histopathologic examination.
Even when patients cannot be diagnosed using these criteria, they may be diagnosed using organ‐specific
diagnostic criteria for IgG4‐RD (see Table 9.4).
Source: Deshpande et al. [23]. Reproduced with permission of Springer Nature.
immunosuppressive therapy and obtaining 1) an IgG4+/IgG+ plasma cell ratio of over
tissue at the point of diagnosis is one of the 40% is required to make a histologic diag-
most important components of this. nosis of IgG4‐RD; and
2) the absolute numbers of IgG4‐positive
Histologic Diagnosis of IgG4‐related plasma cells that support an IgG4‐RD
Hepatobiliary Disease diagnosis depends on the organ involved.
Numbers of IgG4‐positive plasma cells
In general, malignancy is high on the list of required to make the diagnosis for each
differential diagnoses and should be actively organ are given within the consensus state-
excluded. Histologic evaluation is often key ment [23]; in brief, for liver and p ancreas
in making the correct diagnosis of IgG4‐RD with at least two classical histopathologic
in excluding malignancy and therefore every features of disease, this is >10/high‐power
effort should be made to obtain biopsy spec- field (HPF) for a biopsy specimen and >50/
imens before treatment is commenced to HPF for a resection specimen compared
establish the diagnosis. As a broad principle, with >20/HPF for lung and >50/HPF for
fine needle aspiration (FNA) of tissue may be lacrimal gland involvement.
useful to exclude malignancy, but is not
sufficient to make a diagnosis of IgG4‐RD For suspected IgG4‐related hepatobiliary
where tissue with intact architecture is disease, ultrasound‐guided FNA or cytology
required. from biliary brushings may be used to diag-
The classic histopathologic features nose malignancy, but intrabiliary, liver and/or
of IgG4‐RD include storiform (swirling) ampullary biopsies will be required to make a
fibrosis, obliterative phlebitis, and a lympho- histologic diagnosis of IgG4‐RD. Even then,
plasmacytic infiltrate with predominance of the quantification of IgG4‐positive plasma
IgG4‐positive plasma cells. However, in cells will support the diagnosis, but additional
some organs affected by IgG4‐RD and in histologic features may well be absent. In
small biopsy samples, this triad of features is patients with advanced fibrotic disease, IgG4‐
frequently not found (Figure 9.1). positive plasma cells may be reduced in num-
Organ‐specific criteria have been established bers though the IgG4/IgG1 ratio may remain
to aid the diagnosis of IgG4‐RD [23] (Table 9.4). elevated. In these cases, the diagnosis will
Two important principles are that in those need to take into account the clinical features,
with at least one classical histopathologic fea- imaging and ultimately response to therapy.
ture of disease: IgG4‐SC usually affects the extrahepatic and
Figure 9.1 IgG4‐related AIP histology at low power (left) showing an IgG immunostain, and at high
power (right) showing plasma cells positive for IgG4 immunostain (>40%). IgG4 cells are positive for
3,3′‐diaminobenzidine (DAB) and therefore appear dark brown.
Table 9.4 The Boston consensus criteria for the histologic diagnosis (generic and organ‐specific) of IgG4‐RD.
A Key morphologic features

Dense lymphoplasmacytic inflammation
Fibrosis that is at least focally storiform in nature
Obliterative phlebitis
B Other morphologic features that may be present
Phlebitis without obliteration of the lumen
Increased numbers of eosinophils
C Raised numbers of IgG4‐positive plasma cells
Measured as the mean IgG4‐positive plasma cell count per high‐power microscope field (HPF)
within the three HPFs containing the greatest number of these cells
The minimum mean number of IgG4‐positive plasma cells per HPF required varies between
anatomic sites and whether biopsy or surgical material is available for examination
The IgG4+/IgG+ plasma cell ratio must be >40%
Overall assessment of histologic features
Histologically highly suggestive Two or more of the key morphologic features and the required
of IgG4‐RD number of IgG4‐positive plasma cells per HPF for that site
Probable histologic features of One of the key morphologic features and the required number of
IgG4‐RD IgG4‐positive plasma cells per HPF for that site
Insufficient histopathologic Cases with appearances falling outside the above two categories
features of IgG4‐RD
Source: Deshpande et al. [23]. Reproduced with permission of Springer Nature.
hilar bile ducts, but small intrahepatic ducts found in up to 80% of IgG4‐RD patients, but
may be affected and observed on liver biopsy. an elevated IgG4 serum antibody is neither
IgG4‐related liver disease may show portal sensitive nor specific for IgG4‐RD. In general,
inflammation, large bile duct obstructive fea- in our experience, the higher the IgG4 serum
tures, portal sclerosis, lobular hepatitis, and antibody level, the more likely the patient is to
evidence of cholestasis. IgG4‐RD in the liver have IgG4‐RD, have multiorgan disease, and
may be considered a variant of AIH, with be at risk for disease relapse. IgG4 serum
infiltration of IgG4‐positive plasma cells and levels more than four times upper limits have
without other classical features of AIH [24], a specificity and positive value of almost 100%
or as part of the IgG4‐RD spectrum especially for distinguishing IgG4‐SC from primary
if additional organs are involved. For sclerosing cholangitis (PSC), but the s ensitivity
pancreatic disease, biopsy can be obtained by remains low at 42%. Many patients with histo-
core biopsy percutaneously, but recently logically confirmed IgG4‐RD that fulfills diag-
biopsy needles (SharkCore) have been nostic criteria have normal serum IgG4
adapted to reliably obtain core biopsies using antibody levels, whilst IgG4 antibody levels
endoscopic ultrasound (EUS). Biopsy at EUS are elevated in up to 25% of people with other
is probably a more comfortable and safer inflammatory, malignant and autoimmune
alternative way to obtain tissue [25]. diagnoses and in a small subset (5%) of healthy
people. This has led investigators to seek more
specific criteria that utilize serum IgG4 anti-
Utility of Blood Tests, Including Serum IgG4
body but in a ratio with IgG1, whereby an
Levels in IgG4‐RD Diagnostics
IgG1/IgG4 ratio above 0.24 enhances sensi-
An elevated serum IgG4 antibody (>1.4 g/l) tivity for an IgG4‐SC diagnosis compared to
has utility in the diagnosis in IgG4‐RD and is PSC with elevated serum IgG4 levels (>80%).
Detection of circulating plasmablasts, even actively excluded. Common disease mimics

in those with normal serum IgG4 levels, in our experience include:
and elevated serum IgE levels (>408 kU/l) are
●● PSC;
additional biomarkers of IgG4‐RD that may in
●● cholangiocarcinoma;
the future be incorporated into algorithms
●● pancreatic cancer;
to enhance diagnostic accuracy. Other blood
●● atypical presentations of anti‐neutrophil
tests must be interpreted with caution:
cytoplasmic antibody (ANCA)‐positive
inflammatory markers may be non‐specifically
vasculitis;
elevated in IgG4‐RD, serum electrophoresis
●● B‐cell/plasma cell lymphomas;
typically shows a polyclonal hypergamma-
●● histiocytosis.
globulinemia, and anti‐nuclear antibody
(ANA) and rheumatoid factor are elevated in Some but not all inflammatory pseudotu-
50 and 20% of patients, respectively [26]. mors are part of the IgG4‐RD spectrum; it
may be difficult to distinguish IgG4‐RD
pseudotumors from myofibroblastic tumors
Common Alternative Diagnoses to Consider
by the level of IgG4‐positive plasma cells,
Recognizing the difficulties in diagnosing and the absence of anaplastic lymphoma
IgG4‐RD, alternative diagnoses should be kinase (ALK) gene rearrangement by
Pancreas Pancreas
Figure 9.2 Imaging changes typical in IgG4‐related hepatobiliary disease showing IgG4‐related AIP. (Top left)
Avid uptake in the pancreas on PET‐CT and (top right) “sausage”‐shaped pancreas typical of IgG4‐RD using
conventional CT imaging. (Bottom) Multiple and long extrahepatic and intrahepatic biliary strictures at ERCP.
immunostaining and cytogenetics supports months) but may fall out s pontaneously. EUS
the diagnosis of IgG4‐RD. may be used to evaluate mass lesions, to
further characterize duct involvement, and to
perform needle aspiration and tissue biopsies.
Radiologic Characteristics
of Hepatobiliary IgG4‐RD
IgG4‐RD and Relationship
Abdominal ultrasound may demonstrate with Malignancy
biliary obstruction and pancreatic enlarge-
ment suggestive of AIP and exclude other Several groups have shown that IgG4‐RD
pathologies such as gallstones. Cross‐sec- may be associated with an increased risk of
tional imaging, usually computed tomog- malignancy in a variety of organs, and not
raphy (CT) as first line, is essential to define just those affected by the IgG4‐RD itself. A
organ involvement (Figure 9.2). CT abdominal history of malignancy has been observed to
imaging in IgG4‐RD may also reveal biliary be as much as threefold higher (16%) in
strictures, thickened bile duct walls and IgG4‐RD patients compared with matched
liver, hilar or pancreatic masses, and a controls [28]. However, other studies that
swollen or atrophic pancreas. We advocate have evaluated similar‐sized cohorts have
a baseline CT chest, abdomen and pelvis at not shown a significantly increased risk of
diagnosis to define subclinical organ involve- prior malignancy. An overall increased can-
ment. Magnetic resonance cholangiopan- cer risk when patients are followed prospec-
creatography (MRCP) will be indicated in tively [4], during initial diagnosis or
IgG4‐related hepatobiliary disease; contin- particularly in the first year after diagnosis
uous bile duct involvement, bile duct walls of IgG4‐RD [29], has also been reported,
thicker than 2.5 mm, and the presence of though again others have failed to confirm
other organ involvement are all characteristic these observations. Though debated, there is
of IgG4‐related hepatobiliary disease [27]. no conclusive evidence that AIP is associ-
18
F‐fluorodeoxyglucose positron emission ated with an increased risk of pancreatic
tomography (PET) is useful for demon- cancer.
strating clinical involvement in other sites. The reasons for the probable association
The role of whole‐body diffusion‐weighted between IgG4‐RD and malignancy are not
magnetic resonance imaging (MRI) without resolved; it may be that the state of chronic
contrast is of particular interest given the inflammation seen in IgG4‐RD is genuinely
radiation dose from repeat CT and PET‐CT. associated with the development of malig-
Endoscopic retrograde cholangio nancy, as has been observed in other settings
pancreatography (ERCP) will be indicated for (e.g. obesity). However, it is also possible that
patients with possible IgG4‐SC in order to malignancy (or the treatment of malignancy)
identify features that may distinguish this may promote an IgG4 response (a paraneo-
from cholangiocarcinoma and PSC, and to plastic syndrome) and is causally linked to
take biliary biopsies and brushings for the development of IgG4‐RD. Alternative
cytology and the evaluation of malignant explanations are that patients with IgG4‐RD
cells. Typical findings in IgG4‐SC include undergo cross‐sectional whole‐body imaging
long and multifocal strictures, diffuse that may lead to an increase in cancer detec-
pancreatic swelling with a thin narrowed tion, or that risk factors predispose to both
pancreatic duct, and mild biliary dilatation. conditions. Careful, prospective studies of
Temporary biliary stenting may be required large cohorts will ultimately be needed to
to relieve biliary obstruction; the stent should resolve these possibilities. Irrespective, the
be removed once the disease has responded importance of differentiating IgG4‐RD from
to therapy (typically in approximately 2 malignant tumors of each organ needs to be
highlighted. Comprehensive cross‐sectional complete disease remission 12 months after

imaging alongside biopsy and histologic eval- starting steroid therapy, disease relapse rates
uation has a distinct role in both mapping are high and occur in at least 50% of patients,
involved organs and for excluding malignancy often shortly after steroid discontinuation.
and should be performed whenever possible Some groups suggest this may be related to
at diagnosis. the speed of corticosteroid taper. Relapse
rates are likely to be even higher during
long‐term follow‐up, but currently large
Management and Treatment prospective cohorts followed over decades
of Patients with IgG4‐RD post treatment are lacking. We have obser-
vational experience of patients relapsing up
IgG4‐RD is a potentially treatable condition to 8 years after discontinuation of therapy.
and immunosuppression is the cornerstone As the long‐term relapse rates are not
of therapy. IgG4‐RD is not a benign condition known, patients should not be discharged
in that the development of fibrosis and organ from clinic. Risk factors for disease relapse
failure is well recognized [4]. Treatment include high levels of serum IgG4 (>2.8 g/l),
should therefore aim to prevent organ serum IgE and multiorgan involvement and
damage, improve organ function, and relieve especially in those with proximal biliary
patient symptoms. Cross‐sectional imaging strictures [4,17,33,34].
of at least chest and abdomen with or without There is currently significant regional vari-
PET‐CT will be required at initial presenta- ation about how and when to introduce ste-
tion to stage the extent of disease. Optimal roid‐sparing agents; the reason for this is that
regimens for follow‐up imaging will depend the evidence base around optimal second‐
on the organs involved at presentation. line therapies assessed head to head and the
Consensus guidelines on the treatment of long‐term relapse rates and clinical outcome
IgG4‐RD have been developed following an associated with these therapies is currently
international meeting of over 40 investiga- lacking. Some investigators promote the
tors from Asia, North America, and Europe introduction of steroid‐sparing agents at the
[30]. Broad agreement was reached in most start of primary treatment, or introduce
key areas of IgG4‐RD treatment and these as steroids taper in patients at high risk
management, although when and how to of relapse, analogous to the treatment for
incorporate second‐line immunosuppres- autoimmune liver disease. This has been
sive agents into treatment strategies was not shown to reduce the risk of relapse. However,
fully agreed. In general, corticosteroids are there are no clinical trials comparing this
recommended as primary treatment for approach with that of incorporating second‐
IgG4‐RD [30,31]. The recommended line therapies when and if relapse occurs.
starting dose is 30–40 mg daily for 4 weeks, There are also some data to show that main-
aiming to reduce by 5 mg every 2 weeks, but taining low‐dose steroids long term will
titrated depending on response. Initial reduce the risk of relapse. However, main-
response rates to steroids is very high (>90%) taining corticosteroids long term clearly
and typically response occurs in the first 4–6 carries health risks, including the
weeks after starting therapy [17]. Smaller development of diabetes and bone loss. Our
doses, 10–20 mg daily, have been shown in a own practice in those with AIP alone is gen-
retrospective analysis to induce remission in erally to discontinue steroids and monitor
patients with AIP and this dose may be used closely for disease relapse, although our
if there are concerns about steroid tolerance threshold for adding second‐line agents after
[32]. If a non‐response to primary therapy is steroid induction in those with IgG4‐SC and
observed, the diagnosis of IgG4‐RD should multiorgan disease at presentation is much
be questioned. Although up to 66% achieve lower.
There is currently no evidence comparing when clinically indicated. These patients

one second‐line immunosuppressive agent should be maintained under follow‐up to
with another in head‐to‐head studies, but look for new disease in other organs. The
most physicians treating IgG4‐related hepa- answer to the question “What am I aiming to
topancreatobiliary disease use azathioprine achieve with therapy?” should be carefully
at a dose of 2 mg/kg daily. Other second‐line considered in all patients, and therapy
agents that have been used include mycophe- directed accordingly.
nolate, mercaptopurine, methotrexate, tacro-
limus, and cyclosphosphamide. Rituximab, a Monitoring and Follow‐up
monoclonal antibody that depletes CD20+ B
cells, is gaining traction as a therapy for Response to therapy can be monitored by
IgG4‐RD and has been shown to be highly assessing serum IgG4 levels (when elevated),
effective in treating in IgG4‐RD [2,35]. imaging, and the recovery of organ function.
Rituximab is generally administered by two Patient symptomology must also be taken
intravenous infusions of 1000 mg, 2 weeks into account, with caution about interpreta-
apart. Again relapse is well described after tion of the response to corticosteroids, which
rituximab therapy. In one single‐center are well known to improve the feeling of
analysis of 60 patients treated with ritux-
general well‐being. An IgG4‐RD “responder
imab, 37% experienced relapses with a index” has been developed primarily as a
median time from the first infusion to relapse research tool, but is also useful for the
of 244 days. Baseline concentrations of serum objective documentation of response to
IgG4, IgE and circulating eosinophils pre- therapy. Establishing a clear baseline of dis-
dicted disease relapses [36]. Additional pro- ease, including histology, imaging and
spective studies of patient follow‐up are serology, is imperative before committing
required to understand the long‐term relapse patients to a therapeutic pathway that may be
rate, which is likely to be higher. Rituximab lifelong. Once remission is induced, patients
maintenance therapy has been shown to should be followed up indefinitely. Repeat
reduce the risk of relapse (to 11%) but carries cross‐sectional imaging will be required to
the risk of infectious complications. Newer monitor organ recovery and to assess new
agents such as abatacept have been shown to symptomology. As previously discussed,
be efficacious in patients who fail to sustain a where possible MRI should be used rather
response with rituximab, and a Phase II than CT to reduce radiation doses long term.
clinical trial is currently enrolling in the USA.
For organ/life‐threatening disease (e.g.
orbital, sight‐threatening eye disease) addi- Conclusions and Future
tional rapidly acting therapies may be con- Directions
sidered (e.g. cyclophosphamide). At the
opposite end of the spectrum, a diagnosis of IgG4‐RD is a clinically complex and multifac-
IgG4‐RD may be made, but organ function eted disease. Diagnosis may be challenging
may have already been lost and the patient and require knowledge of disease classification
asymptomatic; this clinical scenario is criteria, specific histologic and radiologic
common in “burnt‐out” AIP. When the expertise, and clinical expertise in immuno-
patient is asymptomatic and there is no suppressive therapies that include biologics
organ‐threatening disease, there may be little such as rituximab. Clinicians from several
benefit to starting treatment with immuno- clinical specialties may need to contribute to
suppressive agents, all of which carry a patient care. In our experience, these patients
significant side‐effect profile; however, are best managed in a multidisciplinary team
exocrine replacement therapy (Creon) and meeting and in our institution we have
diabetes management should be instigated established this to take referrals nationally.
The pathogenesis of IgG4‐RD is now evaluate existing therapies, including the

under investigation by several established optimal second‐line immune therapies and
research groups in the USA, Asia and steroid dosing. Future studies should also
Europe. Current therapies are associated evaluate the use of rituximab as first‐line
with a high response rate, but relapse rates therapy to see if a “top‐down” approach can
are very high. This is perhaps not surprising more effectively prevent progression to
since these agents do not modify the under- meaningful clinical end points such as organ
lying drivers of disease, but rather have a failure. Large cohorts of IgG4‐RD patients
temporary effect on immune effector popu- and international registries with prospective
lations. Understanding disease pathogenesis follow‐up are required to understand dis-
may lead to new biomarkers for diagnosis ease outcomes and to stratify patients by
and new therapeutics that may facilitate an clinical subtypes. This will clearly require
IgG4‐RD cure, including those that target international cooperation. Such collabora-
allergic/Th2 pathways. Clinical trials or tions would also facilitate large genetic
careful retrospective analysis of treated studies of IgG4‐RD, which may give new
patient cohorts are required to properly insights into pathogenesis.
References
1 Culver, E.L., Sadler, R., Bateman, A.C. et al. eosinophilia, and IgE elevation in IgG4‐
(2017). Increases in IgE, eosinophils, and related disease. Allergy 69 (2): 269–272.
mast cells can be used in diagnosis and 8 Mattoo, H., Della‐Torre, E., Mahajan, V.S.
to predict relapse of IgG4‐related disease. et al. (2014). Circulating Th2 memory cells
Clin. Gastroenterol. Hepatol. 15 (9): 1444– in IgG4‐related disease are restricted to a
1452.e6. defined subset of subjects with atopy.
2 Carruthers, M.N., Topazian, M.D., Allergy 69 (3): 399–402.
Khosroshahi, A. et al. (2015). Rituximab for 9 Shiokawa, M., Kodama, Y., Sekiguchi, K.
IgG4‐related disease: a prospective, open‐ et al. (2018). Laminin 511 is a target
label trial. Ann. Rheum. Dis. 74 (6): antigen in autoimmune pancreatitis. Sci.
1171–1177. Transl. Med. 10 (453): eaaq0997.
3 Khosroshahi, A., Carruthers, M.N., 10 Mattoo, H., Mahajan, V.S., Maehara, T.
Deshpande, V. et al. (2012). Rituximab for et al. (2016). Clonal expansion of CD4(+)
the treatment of IgG4‐related disease: cytotoxic T lymphocytes in patients with
lessons from 10 consecutive patients. IgG4‐related disease. J. Allergy Clin.
Medicine (Baltimore) 91 (1): 57–66. Immunol. 138 (3): 825–838.
4 Huggett, M.T., Culver, E.L., Kumar, M. et al. 11 Culver, E.L., Vermeulen, E., Makuch, M.
(2014). Type 1 autoimmune pancreatitis and et al. (2015). Increased IgG4 responses to
IgG4‐related sclerosing cholangitis is multiple food and animal antigens indicate
associated with extrapancreatic organ a polyclonal expansion and differentiation
failure, malignancy, and mortality in a of pre‐existing B cells in IgG4‐related
prospective UK cohort. Am. J. Gastroenterol. disease. Ann. Rheum. Dis. 74 (5): 944–947.
109 (10): 1675–1683. 12 Perugino, C.A., AlSalem, S.B., Mattoo, H.
5 Kawa, S. (2017). The immunobiology of et al. (2019). Identification of galectin‐3 as
immunoglobulin G4 and complement an autoantigen in patients with IgG4‐
activation pathways in IgG4‐related related disease. J. Allergy Clin. Immunol.
disease. Curr. Top. Microbiol. Immunol. 143 (2): 736–745.e6.
401: 61–73. 13 Hubers, L.M., Vos, H., Schuurman, A.R.
6 Culver, E.L., van de Bovenkamp, F.S., et al. (2018). Annexin A11 is targeted by
Derksen, N.I.L. et al. (2018). Unique patterns IgG4 and IgG1 autoantibodies in IgG4‐
of glycosylation in immunoglobulin subclass related disease. Gut 67 (4): 728–735.
G4‐related disease and primary sclerosing 14 Culver, E.L., Smit, W.L., Evans, C. et al.
cholangitis. J. Gastroenterol. Hepatol. (2017). No evidence to support a role for
https://doi.org/10.1111/jgh.14512. Helicobacter pylori infection and
7 Della Torre, E., Mattoo, H., Mahajan, V.S. plasminogen binding protein in
et al. (2014). Prevalence of atopy, autoimmune pancreatitis and IgG4‐related
disease in a UK cohort. Pancreatology 25 Detlefsen, S. and Kloppel, G. (2018).

17 (3): 395–402. IgG4‐related disease: with emphasis on the
15 Shiokawa, M., Kodama, Y., Kuriyama, K. biopsy diagnosis of autoimmune
et al. (2016). Pathogenicity of IgG in pancreatitis and sclerosing cholangitis.
patients with IgG4‐related disease. Gut Virchows Arch. 472 (4): 545–556.
65 (8): 1322–1332. 26 Sah, R.P. and Chari, S.T. (2011). Serologic
16 Kanno, A., Masamune, A., Okazaki, K. issues in IgG4‐related systemic disease and
et al. (2015). Nationwide epidemiological autoimmune pancreatitis. Curr. Opin.
survey of autoimmune pancreatitis in Japan Rheumatol. 23 (1): 108–113.
in 2011. Pancreas 44 (4): 535–539. 27 Ohara, H., Okazaki, K., Tsubouchi, H. et al.
17 Ghazale, A., Chari, S.T., Zhang, L. et al. (2012). Clinical diagnostic criteria of
(2008). Immunoglobulin G4‐associated IgG4‐related sclerosing cholangitis 2012.
cholangitis: clinical profile and response to J. Hepatobiliary Pancreat. Sci. 19 (5):
therapy. Gastroenterology 134 (3): 706–715. 536–542.
18 Evans, R.D.R., Cargill, T., Goodchild, G. 28 Wallace, Z.S., Wallace, C.J., Lu, N. et al.
et al. (2019). Clinical manifestations and (2016). Association of IgG4‐related disease
long‐term outcomes of IgG4‐related kidney with history of malignancy. Arthritis
and retroperitoneal involvement in a Rheumatol. 68 (9): 2283–2289.
United Kingdom IgG4‐related disease 29 Asano, J., Watanabe, T., Oguchi, T. et al.
cohort. Kidney Int. Rep. 4 (1): 48–58. (2015). Association between
19 Ebbo, M., Patient, M., Grados, A. et al. immunoglobulin G4‐related disease and
(2017). Ophthalmic manifestations in malignancy within 12 years after diagnosis:
IgG4‐related disease: clinical presentation an analysis after longterm followup.
and response to treatment in a French J. Rheumatol. 42 (11): 2135–2142.
case‐series. Medicine (Baltimore) 96 30 Khosroshahi, A., Wallace, Z.S., Crowe,
(10): e6205. J.L. et al. (2015). International consensus
20 Corcoran, J.P., Culver, E.L., Anstey, R.M. guidance statement on the management
et al. (2017). Thoracic involvement in and treatment of IgG4‐related disease.
IgG4‐related disease in a UK‐based patient Arthritis Rheumatol. 67 (7): 1688–1699.
cohort. Respir. Med. 132: 117–121. 31 Masaki, Y., Matsui, S., Saeki, T. et al.
21 Wallace, Z.S., Zhang, Y., Perugino, C.A. (2017). A multicenter phase II prospective
et al. (2019). Clinical phenotypes of IgG4‐ clinical trial of glucocorticoid for patients
related disease: an analysis of two with untreated IgG4‐related disease. Mod.
international cross‐sectional cohorts. Ann. Rheumatol. 27 (5): 849–854.
Rheum. Dis. 78 (3): 406–412. 32 Buijs, J., van Heerde, M.J., Rauws, E.A.
22 Chari, S.T. (2007). Diagnosis of et al. (2014). Comparable efficacy of low‐
autoimmune pancreatitis using its five versus high‐dose induction corticosteroid
cardinal features: introducing the Mayo treatment in autoimmune pancreatitis.
Clinic’s HISORt criteria. J. Gastroenterol. Pancreas 43 (2): 261–267.
42 (Suppl 18): 39–41. 33 Sandanayake, N.S., Church, N.I.,
23 Deshpande, V., Zen, Y., Chan, J.K. et al. Chapman, M.H. et al. (2009). Presentation
(2012). Consensus statement on the and management of post‐treatment
pathology of IgG4‐related disease. Mod. relapse in autoimmune pancreatitis/
Pathol. 25 (9): 1181–1192. immunoglobulin G4‐associated
24 Chung, H., Watanabe, T., Kudo, M. et al. cholangitis. Clin. Gastroenterol. Hepatol.
(2010). Identification and characterization 7 (10): 1089–1096.
of IgG4‐associated autoimmune hepatitis. 34 Culver, E.L., Sadler, R., Simpson, D. et al.
Liver Int. 30 (2): 222–231. (2016). Elevated serum IgG4 levels in
diagnosis, treatment response, organ rituximab: the Mayo Clinic experience. Gut
involvement, and relapse in a prospective 62 (11): 1607–1615.
IgG4‐related disease UK cohort. Am. J. 6 Wallace, Z.S., Mattoo, H., Mahajan, V.S.
3
Gastroenterol. 111 (5): 733–743. et al. (2016). Predictors of disease relapse
35 Hart, P.A., Topazian, M.D., Witzig, T.E. et al. in IgG4‐related disease following
(2013). Treatment of relapsing autoimmune rituximab. Rheumatology (Oxford) 55 (6):
pancreatitis with immunomodulators and 1000–1008.
181
Section III
Specific Clinical Challenges

183
10
Managing Acute and Chronic Seronegative Liver Disease

Marcus C. Robertson1,2 and Peter C. Hayes3
1
Department of Gastroenterology and Hepatology, Monash Health, Melbourne, Victoria, Australia
2
Department of Medicine, University of Melbourne, Austin Hospital, Heidelberg, Victoria, Australia
3
Department of Hepatology, Royal Infirmary of Edinburgh, Edinburgh, UK
Abstract
Patients with seronegative or cryptogenic liver disease represent a heterogeneous and challenging
group. Cryptogenic cirrhosis represents the end stage of a chronic liver disease process where
the underlying etiology remains unknown after comprehensive investigation. The incidence
and prevalence of cryptogenic cirrhosis has decreased with recognition of conditions such
as nonalcoholic fatty liver disease and atypical presentations of autoimmune liver disease.
Autoimmune hepatitis has diverse clinical phenotypes, with patient presentations ranging from
fulminant liver failure to asymptomatic, which makes diagnosis and management challenging.
Up to 30% of patients with autoimmune hepatitis present without significant titers of antinuclear
antibodies and smooth muscle antibodies and up to 20% of patients may have no detectable
autoantibodies. Acute liver failure (ALF) is an uncommon but life threatening condition occurring
in patients with no pre-existing liver disease. The use of corticosteroid therapy in ALF remains
controversial and should be decided on a case-by-case basis in conjunction with a transplant center.
Keywords acute liver failure; antinuclear antibodies; autoimmune hepatitis; corticosteroid
therapy; lifethreatening condition; seronegative chronic liver disease; smooth muscle antibodies
Key Points
●● Acute seronegative or “indeterminate” ●● N‐Acetylcysteine is now recommended by
hepatitis refers to a cryptogenic hepatitis the European Association for the Study of
resulting in acute liver failure. It is the the Liver in all cases of acute liver failure,
second most common cause of acute
regardless of etiology, and may improve
liver failure after acetaminophen overdose transplant‐free survival. It should be given
in developed countries and occurs most for a period up to 5 days at the standard
commonly in young females. It is consis- dose for a acetaminophen overdose.
tently associated with a low chance of ●● The administration of corticosteroids
spontaneous recovery. remains controversial in acute liver
184 Section III Specific Clinical Challenges
failure, even if an autoimmune etiology ●● Autoimmune hepatitis has diverse clinical

is suspected. Patients with a MELD score phenotypes, with patient presentations
above 35 and/or high‐grade encephalo ranging from fulminant liver failure to
pathy are unlikely to gain benefit from asymptomatic, which makes diagnosis and
corticosteroids. management challenging. Up to 30% of
●● Cryptogenic cirrhosis represents the end patients with autoimmune hepatitis present
stage of a chronic liver disease process without significant titers of anti‐nuclear
where the underlying etiology remains antibodies and smooth muscle antibodies
unknown after comprehensive i nvestigation. (≤1 : 40 titer) and up to 20% of patients may
The incidence and prevalence of crypto- have no detectable autoantibodies.
genic cirrhosis has decreased with recogni- ●● In all cases of seronegative liver disease,
tion of conditions such as non‐alcoholic a comprehensive and multidisciplinary
fatty liver disease and atypical presentations approach is recommended with regular
of autoimmune liver disease. reevaluation of possible etiologies.
Introduction Autoimmune hepatitis (AIH) was first

described as an important cause of chronic
Cryptogenic or seronegative liver disease is liver disease in 1950, although its clinical diag-
a challenging and heterogeneous condition nosis remained erratic. The first diagnostic cri-
in which the underlying etiology remains teria from an international working group were
unknown after extensive clinical, serologic, not proposed until 1998 and then modified in
and pathologic evaluations have been per- 2008; this has undoubtedly led to improved
formed [1]. There is a wide spectrum of diagnosis. Seronegative autoimmune liver dis-
clinical presentation, ranging from acute or eases are a challenging clinical entity and likely
fulminant liver failure to chronic liver disease represent a subset of cryptogenic liver disease.
or cirrhosis, in all cases from an undeter-
mined disease process. Cryptogenic cirrhosis
(CC) constituted approximately 50% of all n Approach to Seronegative
A
cases of cirrhosis in the 1960s, but the preva- Acute Liver Failure
lence has dramatically decreased following
advances in the field and the discovery of Acute liver failure (ALF) is an uncommon
hepatitis B virus (HBV) (1965), hepatitis D but life‐threatening condition occurring in
virus (1977), hepatitis C virus (1989) [2] patients with no preexisting liver disease.
and hepatitis E virus (1990). In addition, the There is considerable heterogeneity in
development of viral hepatitis testing, auto- the clinical presentation of patients, ranging
immune serology, and recognition of con from rapidly progressive to insidious, with
ditions such as non‐alcoholic fatty liver evidence of jaundice, coagulopathy, deranged
disease (NAFLD) have further decreased the liver function tests, and development of mul-
proportion of cirrhosis with no clear etiology tiorgan failure. The onset of clinically appre-
[1,3]. Currently, CC accounts for 5–10% of all ciable hepatic encephalopathy heralds liver
cases of cirrhosis in specialist centers and as failure and differentiates it from patients
the diagnosis of liver disorders continues to with a severe acute liver injury [4].
evolve this proportion is likely to further The most common cause of ALF in the UK,
decline in the future. Seronegative liver USA, Australia and some parts of Europe
disease contributes to the morbidity and
remains acetaminophen overdose, typically
mortality associated with liver disease world- representing approximately 50% of presenta-
wide and represents a significant indication tions [4,5]. Other common causes include
for liver transplantation. acute viral hepatitis‐induced ALF (especially
Chapter 10 Managing Acute and Chronic Seronegative Liver Disease 185
hepatitis A and B), drug‐induced liver injury In 1992, O’Grady et al. [12] modified the
(DILI), and AIH. Despite the varying etiologies, definition of ALF based on the time between
a number of epidemiologic observations have jaundice and the development of hepatic
been made about ALF. Firstly, there is a pre- encephalopathy following observations from a
ponderance among females (67% of cases in a large series of patients with ALF at King’s
large series of 1147 patients) and relatively College. “Hyperacute liver failure” was defined
young patients (mean age 38 years) [5,6]. as the development of encephalopathy within
Transplant‐free survival in ALF was tradition- 7 days of the onset of jaundice. Hyperacute
ally quoted as below 20% [7,8], but more recent liver failure represents the most common
series have shown significantly improved out- form of ALF and is also the most likely group
comes; a recent large series of over 2000 to have an identifiable cause, frequently from
patients with ALF across 31 centers in the USA acetaminophen hepatotoxicity. It is associated
reported an overall survival rate of 71.4%, with with a high incidence of cerebral edema, but
22.3% of patients receiving liver transplantation paradoxically has the best outcomes and
[9]. There is a profound and consistently patients in this group are more likely to sur-
reported disparity in spontaneous recovery vive without liver transplantation. “Acute liver
rates between ALF due to acetaminophen (70% failure” is defined as the development of
transplant‐free survival) and ALF due to non‐ encephalopathy 8–28 days from the onset of
acetaminophen causes (34.5% transplant‐free jaundice; this group also has a high incidence
survival) [9], and acetaminophen‐related ALF of cerebral edema but has a much poorer
is now considered a favorable etiology. prognosis without transplantation. Finally,
Outcomes for ALF have significantly improved “subacute liver failure” pertains to individuals
compared with the pre‐transplant era where with a more insidious onset of encephalop-
mortality rates were at least 80%. athy, which develops 5–12 weeks after the
Acute seronegative (also known as “non‐A, onset of jaundice; in this group the incidence
non‐B, non‐C,” “non A‐E” or “indeterminate”) of cerebral edema is low but prognosis is very
hepatitis refers to a cryptogenic hepatitis poor without transplantation.
resulting in ALF where no identifiable cause Patients with indeterminate or seronega-
is found with currently available diagnostic tive hepatitis most commonly present with
techniques [10]. Historically it was assumed subacute liver failure, which can be difficult
that unidentified pathogenic viruses or hepa- to differentiate clinically from decompen-
totoxic drugs were the most likely cause sated liver cirrhosis or acute‐on‐chronic liver
for the majority of these cases; the term failure. Such cases have poor spontaneous
“seronegative” was introduced more recently, recovery and survival rates and represent a
which recognized the possibility of an unchar- common indication for emergency liver
acterized autoimmune or immune‐mediated transplantation [4,12].
liver injury. A recent systematic review
by Brennan et al. [4] identified 5027 cases
of ALF, of which 20% were estimated to
represent seronegative or indeterminate
otential Causes of Seronegative
P
liver failure, and in the USA it represents ALF and Features that Suggest
the second most common cause of ALF after an Autoimmune Pathogenesis
acetaminophen [11]. Large case series of
seronegative ALF have also consistently dem- Potentially implicated etiologies for seroneg-
onstrated a strong female preponderance ative ALF include viruses, unrecognized or
(59–71%) who present at a young age (37– seronegative autoimmune disease, unidenti-
39 years). In addition, seronegative ALF has fied or missed acetaminophen overdose,
consistently been associated with a poor rate unrecognized drugs, toxins or idiosyncratic
of spontaneous recovery and a poor prognosis drug reactions, and occult metabolic or ge-
in the absence of liver transplantation. netic disorders (Table 10.1).
Table 10.1 Possible causes of seronegative acute liver failure.
Category Example causes Comments
Infections Parvovirus B19 Recognized as a cause of self‐limiting acute hepatitis;

can be associated with aplastic anemia
SEN virus Recently characterized single‐stranded DNA virus;
postulated as a potential cause of indeterminate ALF
but no strong evidence of this
Herpes simplex virus HSV is a documented but uncommon cause of ALF
(HSV) that is thought to represent <1% of ALF cases. Most
common in pregnant or immunosuppressed patients
and may not be associated with mucocutaneous
lesions; empiric aciclovir can be considered in these
patients. HSV typically results in massive
hepatocellular necrosis and patients typically do not
become jaundiced
Togavirus‐like Detected in the native livers of seven patients
particles transplanted for indeterminate hepatitis in one UK
center
Herpes zoster virus
Dengue fever
Undetected APAP protein adducts can be measured in serum to
acetaminophen detect concealed or missed acetaminophen overdose
overdose
Seronegative Seronegative
autoimmune autoimmune hepatitis
disorders
Immune disorders Hemophagocytic Several case reports document HLH as the cause of
lymphohistiocytosis ALF in patients with indeterminate ALF. HLH is often
(HLH) associated with high fever, hepatosplenomegaly,
cytopenias, coagulopathy, multiorgan failure, and
histopathologic evidence of hemophagocytosis
Unrecognized drug Mushroom ingestion Usually preceded by profuse vomiting and diarrhea
reactions or toxins (Amanita phalloides) resulting in acute kidney injury; no laboratory test
available to identify the toxin
Non‐acetaminophen Less than 10% of patients progress to ALF, but in those
drug‐induced liver who do, 80% die or require transplantation
injury
Pregnancy‐related Acute fatty liver Presentation may occur after delivery of the fetus;
liver conditions disease of pregnancy transaminases are relatively low; hypoglycemia is
(AFLP) common
Hemolysis, elevated Presentation may occur after delivery of the fetus; must
liver enzymes and be differentiated from hemolytic uremic syndrome and
low platelets (HELLP) thrombotic thrombocytopenic purpura
Hyperthermic injury May be seen in recreational drug use such as ecstasy
from heat shock
Ischemic hepatitis Rarely considered an indication for transplantation
ALF, acute liver failure.

A thorough patient history is critical in Based on histology, 58% of patients were

determining etiology, as offending drugs or considered to have probable autoimmune
toxins may have been ingested weeks or ALF, and this group also demonstrated
months before presentation. Unintentional higher serum globulin levels and a higher
acetaminophen overdoses are also common, prevalence of anti‐nuclear antibodies (ANAs)
with patients inadvertently consuming toxic and/or anti‐smooth muscle antibodies
levels of acetaminophen from multiple differ- (SMAs) (73 vs. 48%; p = 0.034) [11]; these
ent products or when mentation is impaired serologic features in themselves may have
with drug and/or alcohol ingestion. It is often pointed toward a diagnosis of AIH.
necessary to take repeat histories from In addition, Bernal et al. [14] investigated
patients and gain collateral history from rela- the presence of autoimmune features in a
tives or friends. cohort of 73 ALF patients of varying etiol-
Undiagnosed autoimmune disease is thought ogies, including immunoglobulins, non‐
to be a significant cause of seronegative or organ‐specific antibodies, and antibodies to
indeterminate ALF. It is known that up to soluble liver antigen (SLA). Autoantibodies
25% of patients with AIH present with were present in 23 patients, comprising
acute hepatitis and a small minority of anti‐SLA (n = 16), ANA (n = 6), SMA (n = 4),
these progress to autoimmune ALF [11,13]. and anti‐mitochondrial antibody (AMA)
Around 20% of patients presenting with ALF (n = 1). Autoantibodies were absent in acet-
are cryptogenic, and more than 80% of these aminophen‐related cases but present in 23
remain undiagnosed even after retrospective of 53 (43%) of non‐acetaminophen cases. Of
analysis of stored specimens for occult viral 16 cryptogenic cases in the cohort, 5 (31%)
infections and testing for detectable protein– had positive anti‐SLA. Overall, using the
acetaminophen adducts, which are sugges- revised system with diagnosis of AIH, 50%
tive of distant acetaminophen ingestion. In of patients with cryptogenic ALF were
this group, some patients are suspected of deemed to have “probable autoimmune
having an autoimmune pathogenesis on the hepatitis” [14].
basis of characteristics such as being a young Finally, Ganger et al. [15] recently pub-
female, a history of other autoimmune dis- lished a large case series from the ALFSG
eases, hyperglobulinemia, and positive auto- that highlights the importance of reevaluat-
antibodies. There may also be suggestive ing cases of seronegative ALF in an effort to
features on histologic analysis, although in determine an identifiable etiology. A total
the setting of ALF and massive hepatic necrosis of 314 patients with a diagnosis of indeter-
these are usually not specific. The American minate ALF were included and patient
Acute Liver Failure Study Group (ALFSG) samples were retrospectively tested for

examined 72 patients with indeterminate acetaminophen adducts, hepatitis E
ALF who had liver biopsy or explant tissue serology, and occult viral sequences by
available for review. The reviewing patholo- microarray analysis and deep sequencing;
gist was blinded to all clinical information, if available, liver histology was also re‐
and diagnosed probable autoimmune ALF reviewed. Following these investigations,
on the basis of four histologic features that 46.9% of patients were rediagnosed with a
suggested an autoimmune pathogenesis: single defined etiology, which was rated as
highly likely or probable. Among the rede-
●● distinctive patterns of massive hepatic fined etiologies, 45 patients were found to
necrosis with centrilobular lesions (pre- have previously unrecognized paracetamol
sent in 42% of sections); overdose, 34 AIH, 24 DILI, 13 viral causes,
●● presence of lymphoid follicles (32%); and 12 ischemia. The remaining 150
●● a plasma cell‐enriched inflammatory infil- patients, representing only 5.5% of the total
trate (63%); and ALF cohort, were deemed to be true inde-
●● central perivenulitis (65%). terminate ALF [15].
linical Features of Seronegative

C that 57% of patients with indeterminate
ALF hepatitis were listed for transplantation, with
a transplant‐free survival rate of 27.5% [9].
Indeterminate hepatitis most commonly Studies from England, Scotland and Australia
manifests as subacute liver failure, with have also demonstrated a transplant‐free
symptoms present for more than 4 weeks survival rate of 20–25% for indeterminate
before the development of encephalopathy. ALF [5,16,17]. By comparison, transplant‐
There is a female preponderance reported in free survival in patients with acetamino-
all cases series. Prodromal symptoms are phen‐induced ALF is approximately 65%.
typically non‐specific and include fatigue, A number of risk stratification algorithms
myalgia and nausea, and can often be mis- are used throughout the world; common
taken for a viral prodrome. elements among these algorithms include

Biochemically, patients with seronegative patient age, the degree of coagulopathy, and
ALF typically demonstrate lower transami- jaundice. One of the most widely used risk
nases and higher bilirubin levels than patients stratification algorithms is the King’s College
presenting with hyperacute liver failure from criteria, which have traditionally been used
etiologies such as acetaminophen toxicity. in the UK and Australia to predict poor out-
Other clinical features include a lesser degree comes in patients with ALF. The King’s
of coagulopathy, and less renal failure and College criteria use different parameters
acidosis when compared with acetamino- depending on whether the etiology of ALF
phen toxicity. is acetaminophen or non‐acetaminophen in
The leading causes of death in ALF are origin (Table 10.2). Multiple studies,
cerebral edema and sepsis. Acute renal including syste matic reviews and meta‐
failure, bacterial and fungal infections, and analyses, have consistently shown these cri-
respiratory distress requiring ventilation are teria to have high specificity but lower
also common manifestations [6]. sensitivity and negative predictive value. For
example, a meta‐ analysis examining the
performance of the King’s College criteria in
predicting the need for emergency liver
isk Stratification in Seronegative
R transplantation in non‐acetaminophen ALF
ALF revealed a specificity of 0.82, sensitivity of
0.68, and diagnostic odds ratio of 12.6 [18];
Emergency liver transplantation remains the specificity appeared to be highest in patients
only therapy to date that substantially alters with high‐grade hepatic encephalopathy.
the survival outcomes in patients with ALF. The Clichy criteria (Table 10.3) are widely
Thus, early identification of patients with used to risk stratify patients with ALF in
ALF who have a poor prognosis with current France and were derived from a prospective
medical and supportive therapy and a low study by Bernuau et al. [19] of patients
chance of spontaneous survival is crucial to presenting with hepatitis B‐associated ALF.
facilitate timely listing for emergency liver Patients identified as having the lowest
transplantation. In ALF, it is imperative that transplant‐free survival included those with
this prognostication occurs as early in the hepatic encephalopathy and low factor V
disease course as possible and prior to the levels and these criteria predicted mortality
onset of common complications such as with a positive predictive value of 82% and
increased intracranial pressure and sepsis. negative predictive value of 98% [20]. These
Indeterminate ALF has consistently been data indicated that the Clichy criteria could
associated with poor outcomes and low rates accurately select ALF patients who required
of transplant‐free survival. The ALFSG Adult liver transplantation. Subsequent studies,
Registry of 2070 patients presenting with however, have reported lower predictive
ALF between 1998 and 2013 documented values that were inferior to the King’s College
Table 10.2 King’s College criteria for liver transplantation in ALF.
Acetaminophen‐associated ALF Non‐acetaminophen causes of ALF
pH < 7.3 or INR >6.5 (PT >100 seconds) or

INR > 6.5 (PT >100 seconds), serum creatinine Any three of the following variables:
>300 μmol/l, and grade III–IV encephalopathy 1) Age <10 or >40 years
2) Unfavorable etiology: non‐A, non‐B hepatitis
or idiosyncratic drug reaction
3) Time from onset of jaundice to the
development of encephalopathy >7 days
4) INR > 3.5 (PT >50 seconds)
5) Serum bilirubin >300 μmol/l
ALF, acute liver failure; INR, international normalized ratio; PT, prothrombin time.
Table 10.3 Clichy criteria for selection of ALF patients patients with non‐acetaminophen ALF,
for liver transplantation. whereas the King’s College criteria outper-
formed MELD in acetaminophen‐induced
Presence of hepatic encephalopathy and ALF. The significant heterogeneity between
Factor V level <20% of normal (in patients <30 years of studies included in the meta‐analysis made
age), or <30% of normal (in patients >30 years of age) conclusions difficult and the authors suggested
that neither scoring system is optimal for all
patients [24].
criteria in other populations, including both
acetaminophen‐ and non‐acetaminophen‐
related ALF [21,22]. anagement of Seronegative
M
Since 2006, the Model for End‐stage Liver ALF
Disease (MELD) score has been proposed as
an alternative risk stratification algorithm to Management of ALF requires an expert mul-
the King’s College criteria. Katoonizadeh et al. tidisciplinary treating team incorporating
[23] investigated the utility of the MELD score hepatologists, transplant surgeons, and
to predict outcomes in 99 patients with non‐ intensive care physicians. All patients with
acetaminophen ALF. A MELD score above 30 ALF should be managed in an experienced
had a similar negative predictive value to the center, which is usually a transplant center;
King’s College criteria (92 and 91%, respec- intensive care unit (ICU) monitoring is
tively), and the optimal cutoff point for differ- mandatory once hepatic encephalopathy is
entiating between survivors and non‐survivors present. The primary aim of management in
was a MELD score above 35, with a sensitivity ALF, irrespective of etiology, is to achieve
and specificity of 86% and 75%, respectively. A metabolic and hemodynamic stability to pro-
recent meta‐analysis by McPhail et al. [24] vide optimal conditions for hepatic regenera-
quantitatively assessed and compared the tion and recovery. Patients often develop
prognostic accuracy of the King’s College hypoglycemia which should be monitored
criteria and MELD scores in patients with ALF. and glucose infusions commenced as appro-
In non‐acetaminophen causes, the King’s priate. Coagulopathy should not be corrected
College criteria demonstrated a sensitivity of unless there is evidence of active bleeding
58%, specificity of 74%, and diagnostic odds and stress ulcer prophylaxis should be
ratio of 4.16; MELD demonstrated a sensitivity provided. Patients should ideally be in a quiet
of 76%, specificity of 73%, and diagnostic odds environment with the head of the bed raised
ratio of 8.42. This study suggested that MELD to at least 30°; sedation, intubation and venti-
may be a better predictor of mortality in lation is recommended if encephalopathy
progresses to grade 3 or higher. Some survival was significantly better in patients

etiologies of ALF, such as acetaminophen tox- receiving NAC (40%) compared with those
icity, have a specific treatment antidote and given placebo (27%; p = 0.043); however, the
this should be instituted as soon as possible. survival benefit was limited to patients with
Unfortunately this is not the case for seroneg- grade 1 or 2 hepatic encephalopathy. Stravitz
ative ALF and thus management is supportive et al. [28] subsequently demonstrated that
in nature. The only proven effective therapy NAC may improve outcomes in non‐acet-
for ALF remains liver transplantation. aminophen ALF by reducing expression of
Patients should undergo rapid assessment proinflammatory cytokines, such as inter-
to determine the etiology and severity of the leukin (IL)‐17, and decreasing hepatocyte
liver injury, along with the suitability for liver necrosis. A meta‐analysis of prospective
transplantation. It is important to exclude clinical trials investigating administration of
the presence of underlying cirrhosis (which NAC in non‐acetaminophen ALF was con-
is critical to making the diagnosis of ALF), ducted by Hu et al. in 2015, including 331
acute alcoholic hepatitis (which still pre- patients receiving NAC and 285 controls. The
cludes transplantation in many centers), and authors concluded that administration of NAC
malignant infiltration of the liver. Early in non‐acetaminophen ALF is safe with no
identification of patients with a significantly hepatotoxic effects. While overall survival was
reduced chance of spontaneous survival is similar between groups, patients receiving
imperative as this increases the chance for NAC had a significantly higher transplant‐free
successful liver transplantation. survival (41 vs. 30%, p = 0.01) and post‐trans-
Both the American Academy for the Study plant survival (85.7 vs. 71.4%, p = 0.03) [29].
of Liver Disease (AASLD) [25] and the More recently, Darweesh et al. [30] demon-
European Association for the Study of the strated in a prospective multicenter observa-
Liver (EASL) [26] have released practice tional study that administration of NAC was
guidelines advising on the general and criti- associated with a significantly higher trans-
cal care management of patients with ALF. plant‐free survival rate in non‐acetaminophen
ALF. Administration of NAC was also associ-
ated with decreased encephalopathy (33 vs.
N‐Acetylcysteine
63%, p = 0.02), hospital stay, ICU admission,
N‐Acetylcysteine (NAC) is recommended and failure of other organs. Of note, this
in all cases of ALF due to acetaminophen cohort did not include patients with indeter-
overdose, regardless of the dose or timing of minate liver failure (the majority of patients
the acetaminophen ingestion. There is emerg- had ALF secondary to viral hepatitis, DILI or
ing evidence that NAC has complex antioxi- pregnancy‐related complications) and thus
dant and immunologic effects that may benefit the results may not be generalisable to patients
individuals with non‐acetaminophen‐related with indeterminate ALF [30].
ALF and administration is now recommended
as early as possible in ALF regardless of eti-
Corticosteroids
ology [26], typically at the same dose as for
acetaminophen overdose. In all cases, it is The use of corticosteroid therapy in ALF
advisable to limit the clinical use of NAC to a remains controversial and should be decided
maximum of 5 days’ duration due to a possible on a case‐by‐case basis in conjunction with a
increase in the risk of nosocomial sepsis as a transplant center. Karkhanis et al. [31] investi-
result of its anti‐inflammatory effects [26]. gated the use of corticosteroids in 361 patients
In 2009, Lee et al. [27] conducted a pro- with AIH, drug‐induced or indeterminate
spective randomized clinical trial comparing ALF, finding that corticosteroids did not
NAC with placebo in 173 patients with improve either overall or transplant‐free
non‐acetaminophen ALF. Transplant‐free survival. In addition, administration of
corticosteroids was associated with lower deteriorate. On a practical level, corticoste-

survival in patients with the highest MELD roids are not uncommonly trialed in patients
scores. More recently, Zhao et al. [32] with fulminant AIH or indeterminate ALF.
performed a retrospective study analyzing the Once instituted, this treatment should not
efficacy of corticosteroids in improving trans- delay transfer to a transplant center, should
plant‐free survival in patients with both acute be given for a defined period which should be
and subacute liver failure in Chongqing, less than 2 weeks, and patients must be inten-
China. Administration of corticosteroids in sively monitored for signs of deterioration.
this study was associated with a significantly Multiple studies have suggested that there is
increased rate of transplant‐free survival in a point at which AIH‐related ALF may be too
the ALF cohort (29.4 vs. 5.1%, p = 0.013). advanced to permit rescue with corticoste-
Patients with significantly elevated alanine roids, and thus patients with a MELD
aminotransferase (ALT) levels greater than 30 score above 35 and/or high‐grade encepha-
times the upper limit of normal, a coma grade lopathy are unlikely to gain benefit.
less than four, and a MELD score less than
35 appeared to particularly benefit from
Emergency Liver Transplantation
corticosteroids (transplant‐free survival 65%
vs. 17.4%, p = 0.002), as did patients with an ill- Despite current standards of care, patients
ness duration of less than 2 weeks (transplant‐ with indeterminate ALF continue to have a
free survival 51.4% vs. 15%, p = 0.010) [32]. reduced spontaneous survival compared
One limitation of this study is that outcomes with other etiologies and, as such, emergency
were not specified for different etiologies of liver transplantation remains the only effica-
ALF and thus it is difficult to evaluate any cious treatment. The use of liver transplanta-
specific benefit in indeterminate hepatitis. tion has undoubtedly been the most
In relation to patients presenting with significant development in the treatment of
acute severe or fulminant AIH, Ichai et al. ALF and has transformed survival.
[33] examined the efficacy of corticosteroids In the King’s College Hospital experience of
in 16 patients who required ICU admission, over 3300 patients with ALF treated between
62.5% of whom were encephalopathic on 1973 and 2008, the hospital survival rate
presentation. Twelve patients received corti- increased from 16.7 to 62.2%. Survival after
costeroids for a median duration of 7 days; transplantation improved from 66 to 86% and
median MELD score was 37. The authors overall transplant‐free survival also increased
found little benefit of corticosteroids in to 48%. However, improvements in trans-
fulminant acute AIH: all treated patients
plant‐free survival were not consistent across
experienced a deterioration in hepatic all etiologies of ALF, being most apparent in
encephalopathy and MELD scores prior to acetaminophen‐related ALF, intermediate in
transplantation and septic complications other drug‐induced cases, and absent in cases
may be higher in treated patients. of indeterminate ALF [16]. In addition, the
Given the lack of robust evidence in inde- use of transplantation for non‐acetamino-
terminate ALF, corticosteroid therapy cannot phen ALF increased over the study period,
be universally recommended. In cases of ALF with the proportion of patients undergoing
secondary to AIH or suspected autoanti- transplantation being higher in non‐acet-
body‐negative AIH, the utility of corticoste- aminophen cases compared with acetamino-
roids needs to be evaluated on a case‐by‐case phen cases [16]. Donnelly et al. [17] also
basis. In these situations, clinicians are faced recently reported an increased post‐trans-
with the conundrum of whether to deny cor- plantation survival rate for patients with non‐
ticosteroids to patients who may respond but acetaminophen ALF, with a trend toward
otherwise die or to introduce corticosteroids improvement in transplant‐free survival also
to patients who will not improve and may observed.
In evaluating the need for transplantation, hepatic encephalopathy is often a very late
multiple factors need to be taken into development in the clinical course of patients
consideration in any patient with ALF, with indeterminate ALF and patients may be
including the following: too sick to survive transplantation by the
time it manifests (particularly if encephalop-
●● Accurate prediction of the likelihood of
athy has been precipitated by infection).
survival without transplantation, thus
Conversely, the improving survival rates for
identifying patients with a poor prognosis
ALF secondary to acetaminophen means that
at an early time point. This minimizes the
a benefit for transplantation is not always pre-
chance the patient will become too unwell
sent and transplantation in this cohort should
to transplant and maximizes the time to
generally be reserved for patients with high‐
find a suitable donor organ.
grade encephalopathy unless there is severe
●● Consideration of survival potential follow-
acidosis that does not correct rapidly after
ing transplantation, including factors such
fluid resuscitation [34]. The current UK list-
as age, comorbidities, and sepsis.
ing criteria are shown in Table 10.4.
●● Psychological assessment of the patient to
assess expected compliance, social supports,
past history of suicide attempts, and high‐
risk substance and alcohol abuse. This is a
n Approach to Cryptogenic
A
complex area and collateral history should Chronic Liver Disease
be obtained from family and friends.
Typically, some of the classical contraindica- Cryptogenic cirrhosis (CC) represents the end
tions to liver transplantation in patients with stage of a chronic liver disease process where
chronic liver disease may be overlooked in the underlying etiology remains unknown even
ALF and thus the impact of psychosocial after comprehensive clinical, serologic, and
factors in influencing the decision to trans- pathologic investigations have been performed
plant requires careful consideration, consis- [1]. This is a heterogeneous condition that may
tency, clear documentation, and involvement be multifactorial in origin. CC traditionally
of all members of the multidisciplinary team. accounted for 5–30% of cases of cirrhosis [35]
but with improvements in viral hepatitis testing,
The decision to list an appropriate patient for the development of serologic markers for auto-
liver transplantation involves determination immune diseases and recognition of NAFLD,
of the etiology of ALF and dynamic applica- the prevalence has decreased to approximately
tion of a prognostic scoring system to identify 5%. Similar to indeterminate ALF, CC is more
patients with a low probability of spontaneous likely to affect females, although the average age
recovery. The presence of encephalopathy is higher at around 60 years [35]. Cryptogenic
has traditionally been an integral component chronic hepatitis (CCH) is the term given to
of both the definition of ALF and the decision patients with persistently elevated serum ami-
to transplant, although it is no longer such a notransferases of unknown etiology [36].
black‐or‐white scenario. In the UK, the cri-
teria used to list patients for emergency liver
transplantation have recently been modified Potential Causes of Seronegative
Chronic Liver Disease and Features that
to reflect the importance of distinguishing
Suggest an Autoimmune Pathogenesis
between favorable and unfavorable etiologies
in patients with non‐acetaminophen ALF. Every patient with CC requires a detailed
Indeterminate ALF is categorized as an unfa- and systematic assessment to exclude
vorable etiology and the new criteria now common and uncommon recognizable dis-
allow emergency liver transplant listing of orders. Possible causes of CC are shown in
such patients in the absence of hepatic Table 10.5. A comprehensive patient history
encephalopathy. This change recognizes that is mandatory, including:
Table 10.4 Refined UK listing criteria for super‐urgent liver transplantation.
Category Etiology Criteria
1 Acetaminophen pH <7.25 more than 24 hours after overdose and after fluid
resuscitation
2 Acetaminophen Coexisting prothrombin time >100 seconds or INR >6.5, and serum
creatinine >300 μmol/l or anuria, and grade III–IV encephalopathy
3 Acetaminophen Significant liver injury and coagulopathy following exclusion of
other causes of hyperlactatemia after adequate fluid resuscitation:
arterial lactate >5 mmol/l on admission and >4 mmol/l 24 hours
later in the presence of clinical hepatic encephalopthy
4 Acetaminophen Two of the three criteria from category 2 with clinical evidence of
deterioration (e.g. increased ICP, Fio2 >50%, increasing inotrope
requirements) in the absence of clinical sepsis
5 Favorable non‐ The presence of clinical hepatic encephalopathy is mandatory and
acetaminophen (e.g. prothrombin time >100 seconds or INR >6.5, or any three from the
acute viral hepatitis or following: age 40 years, prothrombin time >50 seconds or INR >3.5,
ecstasy/cocaine‐induced any grade of hepatic encephalopathy with jaundice to
ALF) encephalopathy time >7 days, serum bilirubin >300 μmol/l
6 Unfavorable non‐ a. Prothrombin time >100 seconds or INR >6.5
acetaminophen (e.g. b. In the absence of clinical hepatic encephalopathy, an INR >2
seronegative hepatitis or after vitamin K repletion is mandatory and any two from the
idiosyncratic drug following: age >40 years, prothrombin time >50 seconds or INR
reactions) > 3.5; serum bilirubin >300 μmol/l, and if encephalopathy is
present then jaundice to encephalopathy time >7 days
7 Acute presentation of A combination of coagulopathy and any grade of encephalopathy
Wilson disease or
Budd–Chiari syndrome
ICP, intracranial pressure; INR, international normalized ratio.
Table 10.5 Potential causes of cryptogenic cirrhosis. ●● risk factors for viral hepatitis (e.g. prior drug
use, previous travel, blood transfusions);
Established Non‐alcoholic ●● alcohol exposure (including details of
associations steatohepatitis
Seronegative autoimmune previous and cumulative intake), previous
hepatitis blood alcohol levels;
Occult viral hepatitis ●● family history of liver disease;
Concealed ethanol ●● prior obesity, diabetes, dyslipidemia, bariat-
Sarcoidosis ric surgery or previous evidence of fatty liver;
Wilson disease
●● personal or family history of any autoim-
Less established Occult biliary disease mune disorders;
Hepatic vascular disease
Celiac disease ●● occupational history;
●● migratory history (country of birth,
Other Mitochondriopathies
associations Familial Mediterranean fever countries lived in, time in refugee camps,
Systemic lupus etc.);
erythematosus ●● detailed medication history (for drugs such
Alstrom syndrome as methotrexate);
Keratin 18 mutations ●● past unexplained liver function test
Telomerase gene mutations
abnormalities.
The patient history may need to be taken on Table 10.6 Features on liver biology/histology that
multiple occasions and collaborative history may suggest an underlying etiology.
is often required from family members.
A full liver screen should be undertaken in Etiology Suggestive features
all patients and previous laboratory tests and
Non‐alcoholic Foci of macrosteatosis
imaging (from the patient’s general practi-
fatty liver Cellular ballooning
tioner or other hospitals) should be sourced disease Glycogenated nuclei
for review. In true CC, the diagnostic yield
Autoimmune Interface hepatitis (with or
from a liver screen is not high, and isolated hepatitis without bile duct injury)
results can be difficult to interpret. For Plasma cell infiltrate
example, ANA can be detected in around one‐ Portal inflammation
third of patients with non‐alcoholic steato- Centrilobular zone 3 necrosis
hepatitis (NASH), generally at a low titer, and Primary biliary Portal stage: infiltration of portal
hyperferritinemia is common in both NAFLD cholangitis tracts by lymphocytes,
neutrophils and eosinophils;
and alcohol‐related liver disease. Quantitative
possible granuloma formation
immunoglobulins may provide helpful clues: Periportal stage: lymphocytic
isolated elevation of IgG is suggestive of auto- cholangitis, mild interface
immune liver disease, isolated IgM is often hepatitis, portal/periportal
present in primary biliary cholangitis (PBC), ductular proliferation
Septal stage: presence of bridging
and isolated elevation of IgA can suggest ste-
fibrosis
atohepatitis (from alcohol and/or NAFLD). Cirrhotic stage: biliary cirrhosis
Other investigations such as IgG4 levels and with nodular regeneration of
celiac serology should also be considered. hepatic parenchyma and
A liver biopsy and histologic assessment prominent ductopenia
Small‐duct PSC
should be undertaken wherever possible in
patients with CC and may be helpful in estab-
lishing the etiology of disease. Often, interpre- diagnosis of NAFLD include foci of macros-
tation of biopsies necessitates close cooperation teatosis, cellular ballooning, and glycoge-
between the hepatologist and pathologist, nated nuclei [35].
looking for residual “histologic footprints” or
pre-existing disorders (Table 10.6) [35].
Wilson Disease
Wilson’s disease (WD) is an autosomal reces-

Non‐alcoholic Fatty Liver Disease
sive inherited disorder of hepatic copper meta
It was first proposed in 1990 by Powell et al. bolism resulting from mutations in an
[37] that NASH could be a cause of histologi- intracellular copper transporter ATPase,
cally bland CC due to paradoxical loss of ATP7B, which is primarily expressed in hepa-
macrosteatosis in late stages of the NAFLD tocytes. Loss of ATP7B function results in
disease process. Subsequent research demon- reduced hepatic biliary copper excretion,
strated a consistent relationship between reduced incorporation of copper into cerulo-
obesity, diabetes and cirrhosis and confirmed plasmin, and the accumulation of copper in
NASH as a cause of CC [38–40]. It is now rec- organs and tissues [41]. WD has a wide spec-
ognized that NAFLD/NASH likely represents trum of clinical presentations, with liver dis-
the etiology of cirrhosis in a significant ease and neuropsychiatric disturbances being
proportion of patients previously labeled with the most common. The diagnosis of WD is
CC. Because of the loss of macrosteatosis in notoriously challenging and relies on detection
advanced cirrhosis, the liver biopsy may of Kayser–Fleischer rings, low ceruloplasmin
appear quite bland; residual signs of prior ste- levels, elevated urinary and hepatic copper
atohepatitis that suggest an underlying levels, signs of liver and/or neurologic disease,
and histologic changes in the liver such as perinuclear anti‐neutrophil cytoplasmic

macrosteatosis and glycogenated nuclei [41]. antibody (pANCA). The diagnosis of AIH,

Other laboratory markers, such as an aspartate which requires the presence of characteristic
aminotransferase (AST)/ALT ratio above 2.2 clinical features and the exclusion of other
or an alkaline phosphatase (ALP) to total bili- conditions that resemble AIH, has been for-
rubin ratio less than 4, have also been sug- malized by the International Autoimmune
gested as screening tools to predict the Hepatitis Group, which characterizes patients
presence of WD [42]. The diagnosis of WD into “definite AIH” or “probable AIH.”
should be considered in any patient with acute AIH has diverse clinical phenotypes and
or chronic liver disease under the age of 40; patient presentations can range from fulminant
untreated, WD invariably results in severe liver failure to asymptomatic; this diversity,
disability and death. coupled with the lack of pathognomonic
Complicating the diagnosis of WD is the biochemical or histologic features, can compli-
fact that low ceruloplasmin levels are not cate both diagnosis and management. It has
adequately sensitive or specific to act as a been recognized that 20–30% of patients with
screening tool and can be commonly present AIH present without significant titers of ANA
in patients with chronic liver disease. The his- and SMA (≤1 : 40 titer) and up to 20% of
tologic changes most frequently associated patients may have no detectable autoantibodies
with WD are also non‐specific and can be [45,46]. In addition, although antibodies play a
seen in common conditions such as NAFLD. pivotal role in the diagnosis of AIH, they are
In addition, although the majority of patients neither pathogenic nor disease‐specific.
with WD are diagnosed prior to the age of 40, Autoantibody titers at presentation also corre-
multiple case reports and case series exist of late poorly with disease severity, degree of liver
confirmed WD in patients who become injury, and response to treatment [47].
symptomatic after the age of 40 and even in Seronegative or autoantibody‐negative
their sixties [42]. The largest case series by AIH occurs when patients with typical
Ferenci et al. [43] documents 46 patients who clinical, laboratory and/or histologic find-
became symptomatic with WD after 40 years ings of AIH have no detectable autoanti-
of age. Thus, the diagnosis of WD should be bodies. It is important to emphasize that the
considered in any patient presenting with absence of autoantibodies does not preclude
unclear hepatic or neurologic disease. a diagnosis of AIH or a benefit from cortico-
steroid therapy [44]. In such cases, the diag-
nosis is made typically on the basis of a
Seronegative Autoimmune Hepatitis
transaminitis, elevated serum IgG levels,
AIH is a complex autoimmune liver disease of typical histologic findings, and careful
unknown etiology that results in interface exclusion of other causes of liver disease
hepatitis and chronic necroinflammation that [48]. In patients with negative autoanti-
may progress to cirrhosis and can also present bodies, or indeed any patient with CC, care-
acutely. AIH may occur at any age in children ful review of liver histology is essential for
or adults and affects all ethnicities, although the diagnosis. Interface hepatitis is the hall-
there is a consistent female preponderance mark histologic feature of AIH and sugges-
and it most commonly presents b etween the tive features on biopsy may include portal
ages of 40 and 60 [44]. AIH is typically diag- and/or lobular lymphoplasmacytic necroin-
nosed by the presence of circulating auto flammatory infiltrates (Table 10.6). It is now
antibodies (ANA and SMA) along with recognized that a spectrum of other histo-
hypergammaglobulinemia; those with nega- logic findings may coexist with interface
tive ANA and SMA can also be tested for hepatitis without negating the diagnosis
serum anti‐liver kidney microsomal (LKM) [44,49]. Up to 30% of patients may have cen-
antibodies, anti‐SLA antibodies, and atypical trilobular zone 3 necrosis (with or without
cirrhosis), which can resolve in subsequent while bile duct injury manifested by
histologic examinations; this may represent uctopenia, portal fibrosis and portal edema
d
acute or severe forms of AIH or an acute suggests an overlap syndrome with primary
exacerbation of chronic disease [44]. The sclerosing cholangitis (PSC) [44]. A pro-
presence of centrilobular necrosis may posed algorithm to diagnose seronegative
confer a good response to corticosteroid AIH is shown in Figure 10.1.
therapy [50]. The presence of bile duct injury It is likely that a significant proportion of
in conjunction with interface hepatitis and patients previously labeled with CC actually
otherwise classical features of AIH may also had seronegative AIH and careful reevalua-
suggest an overlap syndrome. Bile duct tion of CC cases should be considered. For
injury including destructive cholangitis, in example, a study in Germany by Heringlake
conjunction with AMA, may represent an et al. [51] of 126 patients with presumed CCH
overlap syndrome between AIH and PBC, found that 34% of patients could be reclassi-
Liver disease of unknown cause
Full liver screen, appropriate liver imaging
ANA +, SMA +, Liver screen

Cause found AMA +
LKM + negative
AIH PBC
Other antibodies: F-Actin,

Liver biopsy
SLA, p-ANCA, PDA-E2
Changes
Antibodies Interface No Interface
PDA-E2 + F-Actin +, SLA + p-ANCA + suggestive
negative hepatitis hepatitis
of PBC
AMA
Consider Seronegative
PBC AIH negative
AIH or PSC AIH
PBC
Cryptogenic
Probable Consider
chronic
AIH PSC
hepatitis
Figure 10.1 Proposed algorithm to diagnose autoimmune liver diseases in patients with cryptogenic chronic
hepatitis or cryptogenic cirrhosis. AIH, autoimmune hepatitis, PBC, primary biliary cholangitis; PSC, primary
sclerosing cholangitis; ANA, anti‐nuclear antibody; AMA, anti‐mitchondrial antibody; LKM, anti‐liver kidney
microsomal antibody; SLA, soluble liver antigen; p‐ANCA, atypical perinuclear anti‐neutrophil cytoplasmic
antibody; PDA‐E2, PBC‐specific anti‐nuclear antibody.
fied as either definite or probable AIH using being reclassified as AMA positive. In addition,
the revised International Autoimmune the majority of truly AMA‐negative patients
Hepatitis (IAH) score. Similarly, two North will test positive to PBC‐specific ANAs such as
American studies concluded that 19–22% of gp210 and sp100 [55]. Such antibodies are
patients with cryptogenic hepatitis could be equivalent to AMA in terms of diagnostic
reclassified as AIH by the revised IAH score accuracy for PBC, and AMA‐negative ANA‐
[46] or by clinical judgment [52]. Lower positive patients do not require biopsy for
frequencies of seronegative AIH have been diagnostic confirmation [54]. Autoantibody‐
reported in other studies and there is likely to negative PBC cannot be diagnosed without a
be significant regional variation. liver biopsy (Table 10.6).
It remains controversial whether seronega- Patients with AMA‐negative PBC should
tive AIH has a different disease course com- be managed in the same way as those with
pared with seropositive cases and few studies AMA‐positive disease. There are conflicting
have examined this. Sonthalia et al. [53] data relating to the clinical course of AMA‐
published an abstract that examined 101
negative patients compared to those with
patients with seropositive or seronegative AMA‐positive PBC. Mendes and Lindor [56]
AIH. The authors found that patients with reviewed five studies that compared AMA‐
seronegative AIH were significantly more negative and AMA‐positive PBC, and
likely to present with severe AIH (defined as reported a similar clinical course, response to
ALF, acute‐on‐chronic liver failure or severe treatment, and prognosis. In addition,
acute hepatitis) (50% vs. 20.27%, p = 0.022), Invernizzi et al. [57] reviewed a historical
although response to treatment and proba- cohort of 297 patients with AMA‐negative
bility of remission was similar between and ‐positive PBC, finding no substantial dif-
groups. Conversely, Mehendiratta et al. [46] ferences in the clinical spectrum or course of
found that the prevalence of ANA or SMA the disease. More recently, some data suggest
did not correlate with the clinical or histo- that patients with PBC‐specific ANAs may
logic severity of AIH at diagnosis when have worse outcomes, with faster progression
studying 52 patients with seropositive or to jaundice, cirrhosis and liver failure [58].
seronegative AIH and treatment response
was similar in both groups.
Conclusion
Seronegative Primary Biliary
Cholangitis
Patients with seronegative or cryptogenic liver
PBC represents an autoimmune liver disease disease represent a heterogeneous and chal-
of immune‐mediated biliary epithelial cell lenging group. A systematic, comprehensive
injury, cholestasis and progressive fibrosis that and multidisciplinary approach is essential in
may progress to end‐stage biliary cirrhosis. It is evaluating such patients and this, coupled
most commonly diagnosed through the with ongoing improvements in diagnostic
combination of cholestatic liver function tests technology, is likely to result in fewer and
and the presence of AMAs [54]. Approximately fewer patients diagnosed with true crypto-
5% of patients with PBC do not have detectable genic liver disease. Atypical presentations of
AMA, a condition termed AMA‐negative autoimmune liver disease undoubtedly repre-
PBC or autoimmune cholangitis. The develop sent the underlying etiology in a significant
ment of complementary assays to detect AMA, proportion of seronegative patients present-
including immunofluorescence, ELISA and ing with both acute and chronic liver disease.
immunoblotting, has resulted in many patients Timely diagnosis is essential and may result in
originally considered to be AMA negative improved outcomes for the patient.
References
1 Mercado‐Irizarry, A. and Torres, E.A. observational cohort study. Ann. Intern.

(2016). Cryptogenic cirrhosis: current Med. 164: 724–732.
knowledge and future directions. Clin. 10 Wigg, A.J., Gunson, B.K., and Mutimer, D.J.
Liver Dis. 7: 69–72. (2005). Outcomes following liver
2 Desai, H.G. (2009). Cryptogenic cirrhosis: a transplantation for seronegative acute liver
vanishing entity. J. Assoc. Physicians India failure: experience during a 12‐year period
57: 751–754, 759. with more than 100 patients. Liver Transpl.
3 Charlton, M.R., Kondo, M., Roberts, S.K. 11: 27–34.
et al. (1997). Liver transplantation for 11 Stravitz, R.T., Lefkowitch, J.H., Fontana,
cryptogenic cirrhosis. Liver Transpl. Surg. R.J. et al. (2011). Autoimmune acute liver
3: 359–364. failure: proposed clinical and histological
4 Brennan, P.N., Donnelly, M.C., and criteria. Hepatology 53: 517–526.
Simpson, K.J. (2018). Systematic review: 12 O’Grady, J.G., Schalm, S.W., and Williams,
non A–E, seronegative or indeterminate R. (1993). Acute liver failure: redefining the
hepatitis; what is this deadly disease? syndromes. Lancet 342: 273–275.
Aliment. Pharmacol. Ther. 47: 1079–1091. 13 Manns, M.P. and Strassburg, C.P. (2001).
5 Gow, P.J., Jones, R.M., Dobson, J.L., and Autoimmune hepatitis: clinical challenges.
Angus, P.W. (2004). Etiology and outcome Gastroenterology 120: 1502–1517.
of fulminant hepatic failure managed at an 14 Bernal, W., Ma, Y., Smith, H.M. et al.
Australian liver transplant unit. J. (2007). The significance of autoantibodies
Gastroenterol. Hepatol. 19: 154–159. and immunoglobulins in acute liver
6 Lee, W.M., Squires, R.H., Nyberg, S.L. failure: a cohort study. J. Hepatol. 47:
et al. (2007). Acute liver failure: 664–670.
summary of a workshop. Hepatology 15 Ganger, D.R., Rule, J., Rakela, J. et al. (2018).
47: 1401–1415. Acute liver failure of indeterminate etiology:
7 Gimson, A.E., O’Grady, J., Ede, R.J. et al. a comprehensive systematic approach by an
(1986). Late onset hepatic failure: clinical, expert committee to establish causality. Am.
serological and histological features. J. Gastroenterol. 113 (9): 1319.
Hepatology 6: 288–294. 16 Bernal, W., Hyyrylainen, A., Gera, A. et al.
8 O’Grady, J.G., Alexander, G.J., Hayllar, (2013). Lessons from look‐back in acute
K.M., and Williams, R. (1989). Early liver failure? A single centre experience of
indicators of prognosis in fulminant 3300 patients. J. Hepatol. 59: 74–80.
hepatic failure. Gastroenterology 97: 17 Donnelly, M.C., Davidson, J.S., Martin, K.
439–445. et al. (2017). Acute liver failure in Scotland:
9 Reuben, A., Tillman, H., Fontana, R.J. et al. changes in aetiology and outcomes over
(2016). Outcomes in adults with acute liver time (the Scottish Look‐Back study).
failure between 1998 and 2013: an Aliment. Pharmacol. Ther. 45: 833–843.
18 McPhail, M.J.W., Wendon, J.A., and Bernal, improves transplant‐free survival in early
W. (2010). Meta‐analysis of performance of stage non‐acetaminophen acute liver
Kings’s College Hospital criteria in failure. Gastroenterology 137: 856–864.e1.
prediction of outcome in non‐paracetamol‐ 28 Stravitz, R.T., Sanyal, A.J., Reisch, J. et al.
induced acute liver failure. J. Hepatol. 53: (2013). Effects of N ‐acetylcysteine on
492–499. cytokines in non‐acetaminophen acute
19 Bernuau, J., Samuel, D., Durand, F. et al. liver failure: potential mechanism of
(1991). Criteria for emergency liver improvement in transplant‐free survival.
transplantation in patients with acute viral Liver Int. 33: 1324–1331.
hepatitis and factor V (FV) below 50% of 29 Hu, J., Zhang, Q., Ren, X. et al. (2015).
normal: a prospective study. Hepatology 14 Efficacy and safety of acetylcysteine in
(4): Abstract 49A. “non‐acetaminophen” acute liver failure: a
20 Bismuth, H., Samuel, D., Castaing, D. et al. meta‐analysis of prospective clinical trials.
(1996). Liver transplantation in Europe for Clin. Res. Hepatol. Gastroenterol. 39:
patients with acute liver failure. Semin. 594–599.
Liver Dis. 16: 415–425. 30 Darweesh, S.K., Ibrahim, M.F., and El‐
21 Izumi, S., Langley, P.G., Wendon, J. et al. Tahawy, M.A. (2017). Effect of N‐
(1996). Coagulation factor V levels as a acetylcysteine on mortality and liver
prognostic indicator in fulminant hepatic transplantation rate in non‐acetaminophen‐
failure. Hepatology 23: 1507–1511. induced acute liver failure: a multicenter
22 Ichai, P., Legeai, C., Francoz, C. et al. study. Clin. Drug Investig. 37: 473–482.
(2015). Patients with acute liver failure 31 Karkhanis, J., Verna, E.C., Chang, M.S.
listed for superurgent liver transplantation et al. (2014). Steroid use in acute liver
in France: reevaluation of the Clichy‐ failure. Hepatology 59: 612–621.
Villejuif criteria. Liver Transpl. 21: 32 Zhao, B., Zhang, H.‐Y., Xie, G.‐J. et al.
512–523. (2016). Evaluation of the efficacy of steroid
23 Katoonizadeh, A., Decaestecker, J., Wilmer, therapy on acute liver failure. Exp. Ther.
A. et al. (2007). MELD score to predict Med. 12: 3121–3129.
outcome in adult patients with non‐ 33 Ichai, P., Duclos‐Vallée, J.‐C., Guettier, C.
acetaminophen‐induced acute liver failure. et al. (2007). Usefulness of corticosteroids
Liver Int. 27: 329–334. for the treatment of severe and fulminant
24 McPhail, M.J.W., Farne, H., Senvar, N. et al. forms of autoimmune hepatitis. Liver
(2016). Ability of King’s College criteria Transpl. 13: 996–1003.
and model for end‐stage liver disease 34 O’Grady, J. (2014). Timing and benefit of
scores to predict mortality of patients with liver transplantation in acute liver failure.
acute liver failure: a meta‐analysis. Clin. J. Hepatol. 60: 663–670.
Gastroenterol. Hepatol. 14: 516–525.e5. 35 Caldwell, S. (2010). Cryptogenic cirrhosis:
25 Lee, W.M., Stravitz, R.T., and Larson, A.M. what are we missing? Curr. Gastroenterol.
(2011). Introduction to the Revised Rep. 12: 40–48.
American Association for the Study of 36 Gassert, D.J., Garcia, H., Tanaka, K., and
Liver Diseases Position Paper on Acute Reinus, J.F. (2007). Corticosteroid‐
Liver Failure 2011. Hepatology 55: responsive cryptogenic chronic hepatitis:
965–967. evidence for seronegative autoimmune
26 Wendon, J., Cordoba, J., Dhawan, A. et al. hepatitis. Dig. Dis. Sci. 52: 2433–2437.
(2017). EASL clinical practical guidelines 37 Powell, E.E., Cooksley, W.G., Hanson, R.
on the management of acute (fulminant) et al. (1990). The natural history of
liver failure. J. Hepatol. 66: 1047–1081. nonalcoholic steatohepatitis: a follow‐up
27 Lee, W.M., Hynan, L.S., Rossaro, L. et al. study of forty‐two patients for up to 21
(2009). Intravenous N‐acetylcysteine years. Hepatology 11: 74–80.
38 Maheshwari, A. and Thuluvath, P.J. (2006). autoimmune hepatitis. Hepatology 51:

Cryptogenic cirrhosis and NAFLD: are 2193–2213.
they related? Am. J. Gastroenterol. 101: 50 Miyake, Y., Iwasaki, Y., Terada, R. et al.
664–668. (2007). Clinical features of Japanese type 1
39 Duseja, A., Nanda, M., Das, A. et al. (2004). autoimmune hepatitis patients with zone
Prevalence of obesity, diabetes mellitus and III necrosis. Hepatol. Res. 37: 801–805.
hyperlipidaemia in patients with 51 Heringlake, S., Schütte, A., Flemming, P.
cryptogenic liver cirrhosis. Trop. et al. (2009). Presumed cryptogenic liver
Gastroenterol. 25: 15–17. disease in Germany: high prevalence of
40 Caldwell, S.H., Oelsner, D.H., Iezzoni, J.C. autoantibody‐negative autoimmune
et al. (1999). Cryptogenic cirrhosis: clinical hepatitis, low prevalence of NASH, no
characterization and risk factors for evidence for occult viral etiology. Z.
underlying disease. Hepatology 29: 664–669. Gastroenterol. 47: 417–423.
41 Rosencrantz, R. and Schilsky, M. (2011). 52 Seo, S., Toutounjian, R., Conrad, A. et al.
Wilson disease: pathogenesis and clinical (2008). Favorable outcomes of autoimmune
considerations in diagnosis and treatment. hepatitis in a community clinic setting.
Semin. Liver Dis. 31: 245–259. J. Gastroenterol. Hepatol. 23: 1410–1414.
42 Weitzman, E., Pappo, O., Weiss, P. et al. 53 Sonthalia, N., Rathi, P., Jain, S. et al. (2016).
(2014). Late onset fulminant Wilson’s disease: Seronegative autoimmune hepatitis
a case report and review of the literature. presents with more severe disease but has
World J. Gastroenterol. 20: 17656–17660. similar outcomes as compared to
43 Ferenci, P., Członkowska, A., Merle, U. seropositive patients. J. Clin. Exp. Hepatol.
et al. (2007). Late‐onset Wilson’s disease. 6 ((Suppl 1)): S81–S82.
Gastroenterology 132: 1294–1298. 54 Hirschfield, G.M., Dyson, J.K., Alexander,
44 Czaja, A.J. (2016). Diagnosis and G.J.M. et al. (2018). The British Society of
management of autoimmune hepatitis: Gastroenterology/UK‐PBC primary biliary
current status and future directions. Gut cholangitis treatment and management
Liver 10: 177–203. guidelines. Gut 67: 1568–1594.
45 Czaja, A.J. (1996). The variant forms of 55 Ozaslan, E., Efe, C., and Gokbulut Ozaslan,
autoimmune hepatitis. Ann. Intern. Med. N. (2016). The diagnosis of
125: 588–598. antimitochondrial antibody‐negative
46 Mehendiratta, V., Mitroo, P., Bombonati, A. primary biliary cholangitis. Clin. Res.
et al. (2009). Serologic markers do not predict Hepatol. Gastroenterol. 40: 553–561.
histologic severity or response to treatment 56 Mendes, F. and Lindor, K.D. (2008).
in patients with autoimmune hepatitis. Clin. Antimitochondrial antibody‐negative
Gastroenterol. Hepatol. 7: 98–103. primary biliary cirrhosis. Gastroenterol.
47 Czaja, A.J. (1999). Behavior and significance Clin. North Am. 37: 479–484.
of autoantibodies in type 1 autoimmune 57 Invernizzi, P., Crosignani, A., Battezzati,
hepatitis. J. Hepatol. 30: 394–401. P.M. et al. (1997). Comparison of the
48 Alvarez, F., Berg, P.A., Bianchi, F.B. et al. clinical features and clinical course of
(1999). International Autoimmune antimitochondrial antibody‐positive and ‐
Hepatitis Group Report: review of criteria negative primary biliary cirrhosis.
for diagnosis of autoimmune hepatitis. J. Hepatology 25: 1090–1095.
Hepatol. 31: 929–938. 58 Nakamura, M. (2014). Clinical significance
49 Manns, M.P., Czaja, A.J., Gorham, J.D. et al. of autoantibodies in primary biliary
(2010). Diagnosis and management of cirrhosis. Semin. Liver Dis. 34: 334–340.
201
11
Managing Pregnant Women with Autoimmune Liver Disease

Eleni Theocharidou and Michael A. Heneghan
Institute of Liver Studies, King’s College Hospital NHS Foundation Trust, London, UK
Abstract
Menstruation and ovulation is usually impaired in women with autoimmune hepatitis (AIH) but
returns to normal when disease activity is controlled. Pregnancy in women with AIH has been
associated with increased risk of preterm birth and small-for-gestational age neonates, and likely
increased risk of miscarriage in those with cirrhosis. The risk of maternal complications is closely
linked to the presence of cirrhosis and disease activity, and can be reduced with optimization of
medical management. Women with cirrhosis are at risk of decompensation, in particular those
with higher model for end-stage liver disease score at conception. Flare-up of disease activity is
more common in the postpartum period but can also occur during pregnancy, and can precipitate
liver-related complications and adverse maternal and fetal outcomes. The widespread use of
azathioprine in rheumatologic disorders and inflammatory bowel diseases has provided increasing
evidence on the safety of azathioprine in pregnancy.
Keywords autoimmune hepatitis; azathioprine; disease activity; end-stage liver disease;
inflammatory bowel diseases; medical management; preterm birth; small-for-geastational age
neonates
Key Points
●● Menstruation and ovulation is usually ●● The risk of maternal complications is
impaired in women with autoimmune closely linked to the presence of cir-
hepatitis (AIH) but returns to normal rhosis and disease activity, and can be
when disease activity is controlled. reduced with optimization of medical
●● Pregnancy in women with AIH has been management.
associated with increased risk of preterm ●● Flare‐up of disease activity is more
birth and small‐for‐gestational age neo- common in the postpartum period but
nates, and likely increased risk of miscar- can also occur during pregnancy, and can
riage in those with cirrhosis. precipitate liver‐related
complications
and adverse maternal and fetal women should undergo surveillance and
outcomes. therapy of gastroesophageal varices usu-
●● Corticosteroids and azathioprine therapy ally prior to conception as well as in the
is considered safe and should not be dis- second trimester of gestation.
continued but mycophenolate and mTOR ●● The management of women with AIH can
inhibitors should be avoided because of be challenging and should involve a multi-
teratogenicity. disciplinary team of experienced hepatol-
●● Women with cirrhosis are at risk of ogists and obstetricians.
decompensation, in particular those with ●● Pre‐conception counseling should be
higher Model for End‐stage Liver Disease offered to women considering pregnancy
(MELD) score at conception. These and can improve outcomes.
Introduction seven had “chronic active hepatitis” as the

etiology of their chronic liver disease, and
Autoimmune hepatitis (AIH) is a rare chronic were on treatment with steroids and/or
liver disease that predominantly affects women, azathioprine, likely corresponding to cases
including those of reproductive age. AIH can of AIH. Women with AIH often exhibit
have an impact on fertility and pregnancy out- amenorrhea in the context of their initial
comes. Its activity can be altered during preg- presentation. Menstruation and ovulation
nancy and in the postpartum period, and usually return to normal after optimal con-
women with cirrhosis are at risk of developing trol of their disease has been achieved [2].
liver‐related complications. The safety of med- Conception is therefore feasible in women
ications in pregnancy has been a significant with stable AIH.
concern among women and obstetricians.
Pregnancy Outcomes in AIH
Fertility in AIH Initial reports of increased risk of compli-
cations and suboptimal pregnancy out-
Gonadal dysfunction was initially described comes among women with AIH, alongside
in male patients with alcohol‐related liver concerns regarding the safety of medica-
disease. A significant proportion of female tion, in particular azathioprine, led clini-
patients with end‐stage liver disease experi- cians to advise women with AIH against
ence amenorrhea and anovulation, the pregnancy and even recommend termina-
etiology of which is presumably dysfunction tion of pregnancy. The landscape has
of the hypothalamic–pituitary axis. A study changed dramatically in the last one to two
of 12 young women (age 19–33 years) with decades with the emergence of increasing
cirrhosis and amenorrhea of at least evidence showing encouraging fetal and
3 months demonstrated that seven women maternal outcomes.
were hypogonadotrophic and five were Several cohort studies and case reports
normogonadotrophic based on luteinizing have reported on outcomes of pregnancy in
hormone levels [1]. The former had signifi- women with AIH. Their main limitations are
cantly lower testosterone and estradiol the relatively small number of patients/preg-
levels, were thinner, and exhibited lower nancies, their retrospective nature, the lack
breast development stage. It was therefore of control group, and in some cohorts the
hypothesized that undernutrition, which is methodology used to record outcomes (self‐
common in advanced liver disease, could reported outcomes via response to question-
potentially be the cause of hypothalamic‐ naires). The results of the largest cohort
mediated amenorrhea. Of the 12 women, studies are summarized in Table 11.1.
Table 11.1 Pregnancy outcomes in women with autoimmune hepatitis.
Preterm
birth
No. of patients Miscarriage (<36–37 Cesarean Congenital
Study or pregnancies Cirrhosis (<20–22 weeks) Stillbirth weeks) section malformations Maternal complications Outcome
Danielsson 71/131 33 (46.5%) Non‐cirrhotic 1 Non‐ Non‐ 3 (urethral aplasia, 5 NICU

Borssen et al. [3] of 43 10 (26.3%) cirrhotic 12 cirrhotic hand and forearm admissions
Werner et al. [4] pregnancies Cirrhotic 14 (21%) 6 (11%) defect, pre‐auricular 1 infant death at
(42.4%) Cirrhotic 10 Cirrhotic fibroma) 3 months
Terminations (23%) 11 (26%) (Self‐reported
13 Total 22 Total 17% outcomes)
(22%)
Terrabuio et al. 39/54 (51 26 (68.4%) 13 (24.1%) 1 6 (11.8%) 2 (anencephaly, 3 (7.8%): 1 NICU
[5] deliveries urinary tract 1 VB in second trimester admission
reported) malformation) 1 vaginal infection/
bleeding postpartum
1 ascites postpartum
Nomuras et al. 10/13 10 (100%) 1 0 5 (42%) 7 (58%) 0 3 (25%):
[6] 2 perineal hematoma
1 uterine atony
Westbrook et al. 53/81 21 (39.6%) 8 (10%) 1 Non‐ 2 (cerebral palsy, 9 (11%): 11 NICU
[7] of 33 Terminations cirrhotic Perthe’s disease of hip) 1 fetal and maternal admissions
pregnancies 12 7/40 (17.5%) death (pulmonary 1 neonatal death
Cirrhotic hypertension/embolism) at 10 weeks
5/19 (26.3%) 6 decompensation (sepsis)
Total 12/59 episodes
(20%) 3 maternal deaths within
12 months
(Continued )
Preterm
birth
No. of patients Miscarriage (<36–37 Cesarean Congenital
Study or pregnancies Cirrhosis (<20–22 weeks) Stillbirth weeks) section malformations Maternal complications Outcome
Schramm 22/42 4 (18%) 7 (17%) 2 7 (17%) 15 2 (Smith–Lemli–Opitz 3 (13.6%): 3 infant deaths

et al. [8] (45.5%) syndrome, spastic 1 maternal death (necrotizing
tetraparesis) secondary to sepsis enterocolitis,
1 fulminant hepatic congenital heart
failure and LT block, Edwards
1 decompensation syndrome)
(ascites)
Candia 58/101 12 (11.9%) 5/95 (7%) 11 2 (2%) Fetal loss 19
et al. [9] (12.4%) (19%)
Perinatal
mortality 4%
Buchel et al. [10] 6/14 2 (33.3%) 1 (7%) 0 1 (7.7%) 4 (30.8%) 0
Heneghan 18/35 7 (38.9%) 2 (5.7%) 2 2 (6%) 1 (3%) 2 (1 Perthe’s disease of 2 maternal deaths (11%): Fetal loss 5
et al. [2] Termination 1 hip, 1 cerebral palsy) 1 fetal and maternal (14.3%)
death due to pulmonary
hypertension/embolism;
1 maternal death
6 months postpartum
(VB)
1 VB in third trimester
1 decompensation/liver
failure (de novo AIH)
2 toxemia
Steven et al. [11] 16/30 3 (10%) 2 (9%) 7 (30.4%) 6 (26%) 1 (pyloric stenosis) Perinatal fetal
Termination 4 loss 4 (33%)
AIH, autoimmune hepatitis; LT, liver transplantation; NICU, neonatal intensive care unit; VB, variceal bleeding.
Chapter 11 Managing Pregnant Women with Autoimmune Liver Disease 205
In these studies the miscarriage rate, described by Borssen et al. [3], 26% of cir-
defined as fetal loss before weeks 20–22 of rhotic women delivered via CS as opposed to
gestation, ranged between 5.7 and 42.4%. In 11% of non‐cirrhotic women. In the study
the study by Heneghan et al. [2], 18 women that reported the highest CS rate of 58%, all
with AIH had 35 pregnancies; of these preg- 10 women had cirrhosis with thrombocyto-
nancies, there were only two (5.7%) miscar- penia (platelets <100 × 109/l) and evidence of
riages and one termination. The highest portal hypertension (gastroesophageal var-
miscarriage rate was reported among women ices) in 92% [6]. With the increasing screen-
with cirrhosis in the study by Borssen et al. ing for varices prior to conception and during
[3]. In this study, there were 131 pregnancies pregnancy, the practice of CS in cirrhosis is
in 71 women, including 43 pregnancies in 33 anticipated to subside, and CS will be con-
women with cirrhosis. The (self‐reported) fined mainly to women with obstetric indica-
miscarriage rate in cirrhotic women was tions for CS.
42.4%, but was significantly lower (26.3%) in The reported congenital abnormality and
non‐cirrhotic women. There were 13 termi- neonatal death rate was low in the majority
nations of pregnancy in this cohort, several of studies, and did not seem to be associated
of which occurred following clinician advice. with fetal exposure to azathioprine. The
The number of stillbirths, defined as fetal requirement for admission to a neonatal
loss beyond weeks 20–22 of gestation, in all intensive care unit (NICU) is often linked to
studies was very low (0–2). Cirrhosis seems prematurity and perinatal complications. In
to have an impact on the total live birth rate. the study by Westbrook et al. [7], 11 neonates
Westbrook et al. [7] reviewed the outcomes required admission to NICU. The risk of
of 81 pregnancies in 53 women, among which NICU admission was higher in neonates
33 pregnancies occurred in 21 women with born to cirrhotic women (21.1% vs. 5%).
cirrhosis. The live birth rate was 57.6% in cir- With regard to maternal complications,
rhotic and 83.3% in non‐cirrhotic women. Westbrook et al. reported four maternal
The rate of preterm birth (delivery before deaths (one secondary to pulmonary hyper-
weeks 36–37 of gestation) was on average tension/embolism, and three secondary to
20% in the majority of studies [3,4,7,8]. Some liver‐related complications within 12 months
studies reported lower risk of preterm birth postpartum) [7], and Schramm et al. [8]
(6–11.8%) [2,5,9,10], whereas in a single reported one death secondary to sepsis.
study with a limited number of patients pre- Hepatic decompensation, usually in the form
term birth occurred in 5 of 12 pregnancies of VB or ascites, can occur during pregnancy
(42%) [6]. The risk of preterm birth does not and in the postpartum period in women with
seem to be increased in women with cir- cirrhosis and portal hypertension [2,5,7,8].
rhosis. The risk was 21% in non‐cirrhotic and Hepatic decompensation or even acute/
23% in cirrhotic women in the study by acute‐on‐chronic liver failure can be precipi-
Borssen et al. [3], and 17.5 and 26.3%, respec- tated by an AIH flare or can occur in the con-
tively, in the study by Westbrook et al. [7]. text of de novo AIH presentation. Westbrook
Variable rates of Cesarean section (CS) et al. reported six decompensation episodes
have been reported in the literature (6–58%) during pregnancy, five of which occurred in
that likely reflect the variation in practices the context of AIH flares [7]. They concluded
and expertise among different centers in that the risk of serious maternal outcomes
managing women with chronic liver disease. was increased with maternal age, presence of
The concern regarding increased risk of vari- cirrhosis, absence of treatment for AIH, and
ceal bleeding (VB) during vaginal delivery in active disease in the last 12 months prior to
the context of the Valsava maneuver may conception. In the study by Schramm et al.,
provide an explanation for higher rates of CS one woman presented with de novo AIH and
in women with cirrhosis. In the cohort acute liver failure during pregnancy, and
required urgent hysterectomy and liver trans- and can be reduced with optimization of
plantation [8]. The implications of AIH flares medical management.
and de novo AIH are discussed in detail later.
All previous studies included small num-
bers of patients given the rarity of the dis- Liver‐related Outcomes
ease, and more importantly did not compare in Pregnancy
pregnancy outcomes to those in the general
population. Two registry studies aimed to AIH is characterized by loss of immunologic
overcome the latter limitation. In a Swedish tolerance to liver autoantigens. The loss of self‐
Patient Register study, 140 women with AIH tolerance is the result of impaired balance bet-
had 171 singleton deliveries between 2006 ween liver‐specific T‐regulatory cells and
and 2011 [12]. Compared to the general effector cells of liver damage [17]. Under
population, they found an increased relative normal circumstances, T‐regulatory cells exert
risk (RR) of preterm birth (RR 3.72), low control over the effector cells ensuring immune
birthweight (RR 2.51), and gestational tolerance. Impaired T‐regulatory cell control
diabetes mellitus (RR 4.35). There was no function or impaired effector cell responsive-
difference in the risk of preeclampsia, CS and ness results in recognition of liver‐specific
congenital malformation. The rate of still- autoantigens, cytokine release and cytotoxic-
birth in this cohort was 1.2%. Of the women ity, leading to liver damage [18]. In active AIH,
with AIH, 48% were on no treatment. T helper 1 (Th1) immune responses predomi-
A Danish nationwide registry‐based cohort nate, with release of tumor necrosis factor
included 721 women with AIH aged under (TNF)‐α and TNF‐β, interleukin (IL)‐1, IL‐2,
50 years who had their first singleton delivery IL‐12, and interferon (IFN)‐γ, whereas in
between 1994 and 2015, and 7210 healthy remission there is a shift toward anti‐
controls [13]. Of the women with AIH, 41% inflammatory Th2 responses, with release of
had cirrhosis and 51% were on immunosup- IL‐4, IL‐5, IL‐10, and IL‐13 [19].
pressive therapy. The risk of preterm birth A number of immunologic changes occur
(RR 3.19) and small‐for‐gestational age neo- in the advent of conception in an effort to
nate (RR 3.2) was increased in women with ensure immune tolerance to paternally
AIH. The risk of miscarriage (RR 1.17) and derived fetal antigens. The immune system
congenital malformations (RR 1.27) was sim- reacts to the recognition of fetal antigens with
ilar to that of the general population. There the development of protective mecha nisms.
were no stillbirths and no maternal deaths up Pregnancy is characterized by a shift from
to 6 months postpartum. Women with AIH proinflammatory Th1 to anti‐inflammatory
had a higher rate of hospital admissions post- Th2 immune responses and cytokine balance.
partum (10% vs. 3%). The risk of intrahepatic Increased levels of cortisol, estrogen and
cholestasis of pregnancy was also higher in progesterone likely play a role in this shift.
the same group compared to the general High corticosteroid levels inhibit IL‐2 syn-
population (11.4% vs. 0.9%). thesis, and progesterone enhances IL‐4 pro-
Pregnancy is feasible in women with AIH duction [19]. Progesterone also inhibits
even in the presence of cirrhosis and portal natural killer (NK) cell activity and prevents
hypertension. Pregnancy in this population abortion [20]. Of particular interest is the
is associated with increased risk of preterm induction of T‐regulatory cells during preg-
birth and small‐for‐gestational age neonates, nancy that facilitate tolerance to fetal
but not congenital malformations. The risk antigens, but which also maintain tolerance
of miscarriage is likely increased mainly in to self‐antigens [21].
women with cirrhosis [6,14–16]. The risk of The immunologic shift in Th1/Th2 balance
maternal complications is closely linked to and the induction of T‐regulatory cells pro-
the presence of cirrhosis and disease activity, vide a potential explanation for the AIH
remission that is observed in a significant highlight the importance of adequate

proportion of patients during pregnancy. immunosuppression even in the immune‐
Delivery and the associated changes in hor- tolerant state of pregnancy.
mone levels result in reactivation of the The immune reconstitution that follows
immune system that can precipitate AIH delivery increases the risk of postpartum
exacerbations postpartum. However, the AIH flares. Initial case reports described
immunologic changes in pregnancy are very the characteristic pattern of remission of
complex, and AIH flares have also been disease activity during pregnancy even
described during pregnancy (Table 11.2). without treatment, followed by a relapse
AIH can present for the first time (de novo) postpartum [19,26,27]. In the largest
during pregnancy or in the postpartum previous cohorts, the risk of postpartum
period (Table 11.3). flares was 11.9–52% in the first 3 months
Despite the favorable immunologic envi- after delivery. In the relatively small cohort
ronment, AIH flares have been reported described by Nomuras et al. [6], there were
with variable frequency (0–34.7%) during no flares during pregnancy or postpartum,
pregnancy, but in most cohorts did not whereas in the study by Buchel et al. [10]
exceed 15% [3–5,7,10]. This variation in postpartum flares were observed in 12 of 13
different cohorts may reflect the different pregnancies (92.3%). Postpartum flares are
immunosuppression strategies, with some associated with the same risk of decompen-
centers electing to withdraw azathioprine sation [5]. In view of these risks, some
prior to (planned) conception or in the first experts recommend increasing the baseline
trimester [6,10]. A proportion of women immunosuppression immediately after
might have elected to stop azathioprine delivery. Although this practice might not
and/or steroids due to concerns regarding be endorsed universally, increased vigilance
the impact on the fetus. AIH flares typically is required with close monitoring at least in
respond to courses of high‐dose steroids or the first 3 months after delivery to ensure
to intensification of the baseline immuno- that AIH flares are diagnosed and treated
suppression. AIH flares can have delete- promptly [28].
rious sequelae, particularly in women with De novo AIH during pregnancy and in the
cirrhosis, such as adverse pregnancy out- postpartum period, although a rare event,
comes or decompensation episodes. In the may pose an even greater risk for the mother
study by Buchel et al. [10], an AIH flare and the fetus. De novo AIH develops in
resulted in miscarriage. Aggarwal et al. [25] women with no preexisting liver disease or
reported AIH flares precipitating decom- with previously undiagnosed AIH. A
pensation in two patients that resulted in proportion of the latter might have
preterm birth in one case and liver failure established cirrhosis. Eight cases of de novo
and death secondary to postpartum AIH during pregnancy [2,3,5,7,8,22], and 14
bleeding in the other. Terrabuio et al. [5] in the first 3 months after delivery have been
reported an episode of VB that occurred in reported in the literature [7,8,23,24]. In most
the context of an AIH flare in the second cases the initial presentation was with acute
trimester, and in the study by Westbrook hepatitis with optimal response to steroids,
et al. [7], five of six AIH flares precipitated but there have been cases of decompensa-
a decompensation episode during preg- tion of preexisting cirrhosis or frank acute/
nancy. The latter investigators demon- subacute liver failure. A patient who pre-
strated that increased maternal age, sented with de novo AIH in the first tri-
presence of cirrhosis, absence of treatment, mester developed decompensation in the
and disease activity in the year prior to con- form of spontaneous bacterial peritonitis,
ception were associated with adverse and had a miscarriage [5]. A woman who
maternal outcomes. These observations presented with subacute liver injury at
Table 11.2 Impact of pregnancy on liver disease.
No. of patients AIH flare during AIH flare

Study or pregnancies Cirrhosis Treatment pregnancy postpartum Liver outcome
Danielsson Borssen 71/131 33 (46.5%) of 84/100 on treatment Non‐cirrhotic Non‐cirrhotic

et al. [3] 43 pregnancies 31/100 on 13 (23%) 24 (42%)
Werner et al. [4] azathioprine Cirrhotic 2 (5%) Cirrhotic 8
Total 15 (15.1%) (19%)
Total 32 (32.3%)
Terrabuio et al. [5] 39/54 26 (68.4%) 43/54 on treatment 4 (7.9%) 23 (45.1%) AIH flare/VB in second trimester
36/54 on azathioprine Ascites/renal failure postpartum
Nomuras et al. [6] 10/13 10 (100%) 11/12 on treatment 0 0 No decompensation
Azathioprine stopped
first trimester
Westbrook et al. [7] 53/81 21 (39.6%) of 61/81 on treatment 6 (7.4%) 20 (24.7%) 3 maternal deaths within 12 months
33 pregnancies 32/81 on azathioprine 6 decompensation episodes (5 in the
context of AIH flares)
Schramm et al. [8] 22/42 4 (18%) 14/42 on azathioprine 9 (21%) 22 (52%) 1 maternal death secondary to sepsis
1 fulminant hepatic failure resulting
in fetal loss/mother LT (de novo AIH)
1 decompensation (ascites)
Candia et al. [9] 58/101 35 (34.7%) 12 (11.9%)
Buchel et al. [10] 6/14 2 (33.3%) Azathioprine stopped 1 (7%) 12 (92.3%) Miscarriage in the context of AIH
first trimester flare
Heneghan et al. [2] 18/35 7 (38.9%) 26/35 on treatment 4 (12.5%) 4 (12.5%) 1 maternal death 6 months
15/35 on azathioprine postpartum (VB)
1 VB in third trimester
1 decompensation/liver failure
(de novo AIH)
2 toxemia
AIH, autoimmune hepatitis; LT, liver transplantation; VB, variceal bleeding.

Table 11.3 Cases of de novo autoimmune hepatitis during pregnancy or postpartum.
De novo De novo
No. of during postpartum
Author patients pregnancy (1–3 months) Presentation Outcome
Danielsson Borssen 2 2
et al. [3]
Terrabuio et al. [5] 1 1 Decompensation Miscarriage
(SBP)
Sato et al. [22] 1 1 Subacute liver CS at 29 weeks
injury (mother and infant
survived)
Westbrook et al. [7] 2 1 1 Good response to
steroids
Efe et al. [23] 3 3 Good response to
steroids
Schramm et al. [8] 6 1 5 1 liver failure 1 hysterectomy and
LT at 18 weeks
Heneghan et al. [2] 2 2 1 CS at 28 weeks: child
decompensation/ with developmental
liver failure problems
Samuel et al. [24] 5 5 Good response to
steroids
CS, Cesarean section; LT, liver transplantation; SBP, spontaneous bacterial peritonitis.
16 weeks required CS at 29 weeks [22]. maternal and fetal mortality [30]. Worsening
Another woman who presented with liver of portal hypertension may occur during
failure required emergency hysterectomy pregnancy as a result of physiologic volume
and liver transplantation at 18 weeks [8]. A expansion and the pressure of the expand-
woman who presented with decompensa- ing uterus on the inferior vena cava and col-
tion/liver failure had CS at 28 weeks and the lateral circulation [29]. These hemodynamic
child had severe developmental prob- changes are more prominent in the second/
lems [2]. De novo AIH is usually responsive thrird trimester when the risk of decompen-
to conventional induction therapy, but occa- sation becomes higher. Initial observations
sionally can be associated with liver failure/ yielded a high rate of VB (18–32%) during
decompensation with deleterious sequelae pregnancy among women with cirrhosis,
for the mother and the fetus. and a rate of up to 78% in those with preex-
isting varices with high associated mortality
[32]. Subsequent studies demonstrated a
Pregnancy in Cirrhosis significantly lower VB rate that likely reflects
improvements in variceal surveillance and
The presence of cirrhosis increases the risk optimal management of gastroesophageal
of both fetal and maternal complications varices. A large study that reviewed the
[29]. Cirrhosis is associated with increased complications of 339 pregnancies in women
risk of miscarriage, preterm birth and small‐ with cirrhosis reported decompensation
for‐gestational age neonate [3,30,31]. The episodes in 15% (ascites 11%, VB 5%). The
primary concern in this population is the maternal mortality rate was 1.8% in cir-
risk of decompensation that increases both rhotic and 0% in non‐cirrhotic women,
while the fetal mortality rate was 5.2 and may impair uterine blood flow [38].
2.1%, respectively [30]. The maternal Propranolol is the beta‐blocker of choice and
mortality rate increased to 6% and the fetal is considered generally safe in pregnancy
mortality rate to 12% in cirrhotic women despite concerns regarding decreased pla-
who decompensated during pregnancy. Two cental perfusion, fetal and neonatal brady-
smaller studies in women with cirrhosis cardia, and hypoglycemia (FDA category C).
showed a similar decompensation rate of In the event of VB, EVL, vasoactive agents
10% [33], and 16% (VB 4%) with associated that reduce portal pressure, and broad‐spec-
maternal mortality of 4% [34]. Patients with trum antibiotics are the mainstay of
cirrhosis are at increased risk of infections, treatment. Although terlipressin is the vaso-
which are associated with poor outcomes. A active agent of choice in non‐pregnant
single study in 187 women with cirrhosis patients, octreotide (FDA category B) might
demonstrated an increased risk of peripar- be preferred in pregnancy due to the lack of
tum infections in this population [35]. adequate evidence with regard to safety of
Women with cirrhosis and portal hyperten- the former (potential risk of uterine contrac-
sion are also more likely to develop post- tions and decreased uterine blood flow).
partum bleeding and other complications Transjugular intrahepatic portosystemic
that may require hospital admission [31]. shunt (TIPS) insertion is limited in preg-
The main predictor of pregnancy outcomes nancy due to fetal radiation exposure, but
and liver‐related complications is the severity good outcomes have been reported in a small
of liver disease at conception. Westbrook number of pregnant women with recurrent/
et al. [33] reviewed outcomes of 62 preg- refractory VB [39–41]. Ascites and hepatic
nancies in 29 women with cirrhosis, and encephalopathy are managed with diuretics
found that higher Model for End‐stage Liver and lactulose, respectively, as in non‐preg-
Disease (MELD) and Child–Pugh scores were nant patients.
associated with increased risk of preterm Another consideration in women with cir-
birth. A MELD score of 10 or above was an rhosis and portal hypetension is the optimal
accurate predictor of decompensation, mode of delivery. The increase in intra‐
whereas a MELD score of 6 or less was associ- abdominal pressure during vaginal delivery
ated with no liver‐related complications. A may precipitate rupture of gastroesophageal
further study of 165 pregnancies in 100 varices, which provides an explanation for
women with chronic liver disease (80 preg- the increased frequency of CS in women
nancies in 48 women with cirrhosis) showed with cirrhosis [30,35]. This risk has dramati-
that the albumin/bilirubin ratio is an accurate cally decreased with the implementation of
predictor of live births, and the aspartate ami- variceal screening, and vaginal delivery is
notransferase (AST)/platelet ratio is a good likely safe in women with no or small varices.
predictor of preterm birth [36]. CS should be reserved for women with severe
Screening for varices with esophagogastro- portal hypertension and varices refractory to
duodenoscopy (EGD) prior to (planned) con- medical/endoscopic therapy. Otherwise, the
ception or in the second trimester of decision for CS should be driven by obstetric
pregnancy is indicated in all women with cir- indications.
rhosis [37]. Esophageal varices should be
managed according to their grade and the
presence of high‐risk endoscopic stigmata Safety of Medication
with non‐selective beta‐blockers and/or in Pregnancy
endoscopic variceal ligation (EVL). EGD is
considered safe during pregnancy, with the Prednisolone at high doses is used for
exception of a small risk of fetal hypoxia due induction of remission in AIH, and can also
to sedation and maternal positioning that be used at low doses for maintenance of
remission usually in combination with aza- terminate pregnancy, stemmed mainly from
thioprine. Despite the lack of well‐controlled animal studies that showed risk of teratoge-
data in humans, prednisolone is considered nicity with azathioprine and the lack of con-
safe during pregnancy and lactation as the trolled data in humans. The widespread use
transplacental passage is limited (Table 11.4) of azathioprine in rheumatologic disorders
[42]. There is a slightly increased risk of cleft and inflammatory bowel diseases has
lip/palate in neonates born to women provided increasing evidence on the safety of
exposed to steroids in the first trimester of azathioprine in pregnancy. Despite initial
gestation [43]. Neonates might need to be reports of congenital anomalies and adverse
monitored for adrenal insufficiency when the pregnancy outcomes, these have not been
mother has been exposed to long‐term ste- confirmed in large‐scale studies [45–47].
roid therapy, although this event is rare [44]. None of the previous cohort studies in AIH
Azathioprine and 6‐mercaptopurine have demonstrated increased risk of adverse preg-
been the source of great concern among phy- nancy outcomes or neonatal complications
sicians and women with AIH in the past in women exposed to azathioprine during
because of the equivocal evidence with pregnancy [7]. Rare cases of leukopenia in
regard to their safety in pregnancy. These the neonate have been associated with high
concerns, which led women previously to levels of active metabolites (6‐thioguanine
avoid pregnancy, stop therapy or even nucleotide, 6‐TGN) secondary to low
Table 11.4 Drug safety in pregnancy.
FDA
pregnancy
Drug category Evidence in pregnancy Evidence in lactation
Azathioprine D Evidence of teratogenicity in animals Excreted in human milk

No controlled data in humans Low or undetectable active
Reports of congenital anomalies, metabolites in milk and
leukopenia/thrombocytopenia of neonatal serum
the neonate
Reports of miscarriage, preterm
birth, low birthweight
Prednisolone C/D No adequate data in humans Excreted in human milk
Slightly increased risk of cleft lip/ Avoid breastfeeding for
palate 4 hours after dose
Low birthweight, adrenal Monitor neonate for
suppression of the neonate adrenal suppression if
maternal dose >40 mg daily
Mycophenolate D Contraindicated Unknown
mofetil Risk of miscarriage and congenital
anomalies
Tacrolimus C No controlled data in humans Excreted in human milk
Risk of preterm birth, neonatal Unknown effect
hyperkalemia and renal dysfunction
Ciclosporin C Evidence of adverse effect on fetus Excreted in human milk
in animals Undetectable levels in most
No adequate data in humans neonates
Risk of preterm birth and low
birthweight
FDA, Food and Drug Administration.

thiopurine S‐methyltransferase (TPMT) contraception should be used for 3 months

enzyme activity in the mother and fetus [48]. after stopping MMF.
6‐TGN (associated with myelotoxicity) and Ciclosporin and tacrolimus have been used
6‐methylmercaptopurine (6‐MMP, associ- as second‐line therapies in AIH. The data on
ated with hepatotoxicity) have been mea- safety in pregnancy derive mainly from
sured in both mother and fetus (umbilical studies in post‐transplant populations.
cord blood) [49]. 6‐TGN levels were signifi- Despite the concerns regarding risks of pre-
cantly lower in the fetus compared to term birth, low birthweight, neonatal hyper-
maternal blood, while 6‐MMP levels were kalemia, and renal dysfunction, these have
essentially undetectable in the fetus. This not been confirmed in large‐scale studies
indicates that the placenta forms a relative [51,52], and calcineurin inhibitors can be
barrier to active metabolites of azathioprine. continued during pregnancy.
Similarly, although azathioprine is excreted When balancing the risks and benefits of
in milk, the levels of active metabolites were AIH therapy during pregnancy, it is evident
low or undetectable in both milk and neo- that uncontrolled or suboptimally controlled
natal serum. In light of the rare risks, disease poses a greater risk for maternal
measurement of TPMT activity and active and fetal outcomes than the medication per se.
metabolites might be prudent in pregnant A pragmatic approach is to continue the
women on azathioprine to minimize the risk regimen that has achieved optimal disease
of toxicity. control throughout pregnancy (Table 11.5).
Mycophenolate mofetil (MMF) has been This approach is not only very rarely associ-
associated with increased risk of miscarriage ated with adverse events, but more impor-
and congenital anomalies and is therefore tantly has been shown to improve fertility and
contraindicated in pregnancy [50]. MMF pregnancy outcomes in women with autoim-
should be discontinued at least 3 months mune disorders [53]. Women on MMF wish-
prior to conception, and effective barrier ing to pursue pregnancy should be switched
Table 11.5 Practical management of autoimmune hepatitis (AIH) in pregnancy.
Pre‐conception/ ●● Amenorrhea and anovulation are not uncommon at the initial presentation of AIH,
conception and usually reverse with initiation of treatment and optimal control of disease
activity
Pregnancy ●● Pregnancy in AIH is associated with increased risk of preterm birth and small‐for‐
outcomes gestational age neonates, but not congenital malformations
●● The risk of miscarriage is increased mainly in women with cirrhosis
Maternal ●● The risk of maternal complications is closely linked to the presence of cirrhosis/
complications severity of liver disease and AIH activity
●● Optimal control of disease activity is an important goal in pregnancy: Prednisolone
and azathioprine are considered safe and should not be discontinued. MMF is
contraindicated in pregnancy
●● Screening for varices prior to conception or in the second trimester is indicated in
all women with cirrhosis
●● Cesarean section should be reserved for women with severe portal hypertension, or
otherwise should be driven by obstetric indications
Postpartum ●● AIH flares can occur in 15% of pregnancies, and more commonly postpartum, and
can be associated with adverse pregnancy outcomes or decompensation episodes
●● Close monitoring is required in the first 3 months postpartum
MMF, mycophenolate mofetil.

to a different regimen at least 3 months prior and fetal outcomes. Optimal control of dis-
to conception. ease activity is thus an important goal in
pregnancy. First‐ and second‐line AIH
therapy is considered safe and should not be
Summary discontinued, with the exception of MMF.
Women with cirrhosis are at risk of decom-
Pregnancy in women with AIH has been pensation, in particular those with higher
associated with increased risk of preterm MELD score at conception. Women with cir-
birth and small‐for‐gestational age neonates, rhosis should undergo surveillance and
and likely increased risk of miscarriage in therapy of gastroesophageal varices prior to
women with cirrhosis. The risk of maternal conception and/or in the second trimester of
complications is closely linked with the gestation. The management of women with
presence of cirrhosis and disease activity, and AIH can be challenging and should involve a
can be reduced with optimization of medical multidisciplinary team of experienced hepa-
management. AIH flares are more common tologists and obstetricians. Preconception
in the postpartum period, but can also occur counseling should be offered to women con-
during pregnancy, and can precipitate liver‐ sidering pregnancy and can improve
related complications and adverse maternal outcomes.
References
1 Cundy, T.F., Butler, J., Pope, R.M. et al. pregnancy in women with autoimmune
(1991). Amenorrhoea in women with hepatitis. J. Autoimmun. 38: J239–J244.
non‐alcoholic chronic liver disease. Gut 32: https://doi.org/10.1016/j.jaut.2011.1012.
202–206. 1002.
2 Heneghan, M.A., Norris, S.M., O’Grady, J.G. 8 Schramm, C., Herkel, J., Beuers, U. et al.
et al. (2001). Management and outcome of (2006). Pregnancy in autoimmune
pregnancy in autoimmune hepatitis. Gut 48: hepatitis: outcome and risk factors. Am. J.
97–102. Gastroenterol. 101: 556–560. https://doi.
3 Danielsson Borssen, A., Wallerstedt, S., org/10.1111/j.1572‐0241.2006.00479.x.
Nyhlin, N. et al. (2016). Pregnancy and 9 Candia, L., Marquez, J., and Espinoza, L.R.
childbirth in women with autoimmune (2005). Autoimmune hepatitis and
hepatitis is safe, even in compensated pregnancy: a rheumatologist’s dilemma.
cirrhosis. Scand. J. Gastroenterol. 51: Semin. Arthritis Rheum. 35: 49–56. https://
479–485. https://doi.org/10.3109/00365521. doi.org/10.1016/j.semarthrit.2005.1003.
00362015.01115893. 1002.
4 Werner, M., Bjornsson, E., Prytz, H. et al. 10 Buchel, E., Van Steenbergen, W., Nevens, F.,
(2007). Autoimmune hepatitis among fertile and Fevery, J. (2002). Improvement of
women: strategies during pregnancy and autoimmune hepatitis during pregnancy
breastfeeding? Scand. J. Gastroenterol. 42: followed by flare‐up after delivery. Am. J.
986–991. https://doi.org/10.1080/0036552 Gastroenterol. 97: 3160–3165. https://doi.
0601155266. org/10.1111/j.1572‐0241.2002.07124.x.
5 Terrabuio, D.R., Abrantes‐Lemos, C.P., 11 Steven, M.M., Buckley, J.D., and Mackay,
Carrilho, F.J., and Cancado, E.L. (2009). I.R. (1979). Pregnancy in chronic active
Follow‐up of pregnant women with hepatitis. Q. J. Med. 48: 519–531.
autoimmune hepatitis: the disease 12 Stokkeland, K., Ludvigsson, J.F., Hultcrantz,
behavior along with maternal and fetal R. et al. (2016). Increased risk of preterm
outcomes. J. Clin. Gastroenterol. 43: birth in women with autoimmune
350–356. https://doi.org/10.1097/MCG. hepatitis: a nationwide cohort study. Liver
1090b1013e318176b318178c318175. Int. 36: 76–83. https://doi.org/10.1111/
6 Nomuras, R.M., Kleine, R.T., Igai, A.M. et al. liv.12901.
(2013). Clinical and obstetrical management 13 Gronbaek, L., Vilstrup, H., and Jepsen, P.
of pregnant women with autoimmune (2018). Pregnancy and birth outcomes in a
hepatitis complicated by moderate or severe Danish nationwide cohort of women with
thrombocytopenia. Rev. Assoc. Med. Bras. autoimmune hepatitis and matched
(1992) 59: 28–34. population controls. Aliment. Pharmacol.
7 Westbrook, R.H., Yeoman, A.D., Kriese, S., Ther. 48: 655–663. https://doi.org/10.1111/
and Heneghan, M.A. (2012). Outcomes of apt.14925.
14 Szczepanska, M., Horst, N., Mikolajczyk, cirrhosis overlap syndrome: a report of two
M. et al. (2018). The course and delivery of cases. Clin. Res. Hepatol. Gastroenterol.
a pregnancy in a patient with autoimmune 35: 687–689. https://doi.org/10.1016/j.
hepatitis complicated by cirrhosis of the clinre.2011.1006.1003.
liver. Ginekol. Pol. 89: 339–340. https://doi. 24 Samuel, D., Riordan, S., Strasser, S. et al.
org/10.5603/GP.a2018.0058. (2004). Severe autoimmune hepatitis first
15 Orgul, G., Ozkan, E.U., Celik, H.T., and presenting in the early post partum period.
Beksac, M.S. (2017). Autoimmune hepatitis Clin. Gastroenterol. Hepatol. 2: 622–624.
and pregnancy: report of two cases with 25 Aggarwal, N., Chopra, S., Suri, V. et al.
different maternal outcomes. Clin. Exp. (2011). Pregnancy outcome in women with
Hepatol. 3: 212–214. https://doi.org/10. autoimmune hepatitis. Arch. Gynecol.
5114/ceh.2017.71445. Obstet. 284: 19–23. https://doi.
16 Braga, A. and Braga, J. (2016). Successful org/10.1007/s00404‐00010‐01540‐z.
pregnancy with autoimmune cirrhosis. 26 Malhotra, B., Malhotra, N., Deka, D., and
BMJ Case Rep. 2016:bcr2015212501. Takkar, D. (2002). Immunosuppressive
doi: https://doi.org/10.1136/bcr‐2015‐ effect of pregnancy on autoimmune
212501. hepatitis: a case report and review of
17 Wen, L., Peakman, M., Lobo‐Yeo, A. et al. literature. Eur. J. Obstet. Gynecol. Reprod.
(1990). T‐cell‐directed hepatocyte damage Biol. 101: 91–92.
in autoimmune chronic active hepatitis. 27 Colle, I. and Hautekeete, M. (1999).
Lancet 336: 1527–1530. Remission of autoimmune hepatitis during
18 Ferri, S., Longhi, M.S., De Molo, C. et al. pregnancy: a report of two cases. Liver 19:
(2010). A multifaceted imbalance of T cells 55–57.
with regulatory function characterizes type 28 Riely, C.A. (2002). Meet Sarah: a vignette of
1 autoimmune hepatitis. Hepatology 52: autoimmune hepatitis and pregnancy. Am.
999–1007. https://doi.org/10.1002/ J. Gastroenterol. 97: 2945. https://doi.
hep.23792. org/10.1111/j.1572‐0241.2002.07089.x.
19 Muratori, P., Loffreda, S., Muratori, L. et al. 29 Tan, J., Surti, B., and Saab, S. (2008).
(2002). Spontaneous remission of Pregnancy and cirrhosis. Liver Transpl. 14:
autoimmune hepatitis during pregnancy. 1081–1091. https://doi.org/10.1002/
Dig. Liver Dis. 34: 608–609. lt.21572.
20 Szekeres‐Bartho, J. (2002). Immunological 30 Shaheen, A.A. and Myers, R.P. (2010). The
relationship between the mother and the outcomes of pregnancy in patients with
fetus. Int. Rev. Immunol. 21: 471–495. cirrhosis: a population‐based study. Liver
21 Aluvihare, V.R., Kallikourdis, M., and Betz, Int. 30: 275–283. https://doi.
A.G. (2004). Regulatory T cells mediate org/10.1111/j.1478‐3231.2009.02153.x.
maternal tolerance to the fetus. Nat. 31 Puljic, A., Salati, J., Doss, A., and Caughey,
Immunol. 5: 266–271. https://doi. A.B. (2016). Outcomes of pregnancies
org/10.1038/ni1037. complicated by liver cirrhosis, portal
22 Sato, H., Tomita, K., Yasue, C. et al. (2015). hypertension, or esophageal varices.
Pregnant woman with non‐comatose J. Matern. Fetal Neonatal Med. 29:
autoimmune acute liver failure in the 506–509. https://doi.org/10.3109/14767058
second trimester rescued using medical .14762015.11009438.
therapy: a case report. Hepatol. Res. 45: 32 Russell, M.A. and Craigo, S.D. (1998).
349–355. https://doi.org/10.1111/ Cirrhosis and portal hypertension in
hepr.12344. pregnancy. Semin. Perinatol. 22: 156–165.
23 Efe, C., Ozaslan, E., and Purnak, T. (2011). 33 Westbrook, R.H., Yeoman, A.D., O’Grady,
Outcome of pregnancy in patients with J.G. et al. (2011). Model for end‐stage
autoimmune hepatitis/primary biliary liver disease score predicts outcome in
cirrhotic patients during pregnancy. Dig. Liver Dis. 40: 387–390. https://doi.
Clin. Gastroenterol. Hepatol. 9: 694–699. org/10.1016/j.dld.2007.1002.1013.
https://doi.org/10.1016/j.cgh.2011.1003. 42 van Runnard Heimel, P.J., Franx, A.,
1036. Schobben, A.F. et al. (2005).
34 Jena, P., Sheela, C.N., Venkatachala, R.P., Corticosteroids, pregnancy, and HELLP
and Devarbhavi, H. (2017). Obstetric syndrome: a review. Obstet. Gynecol. Surv.
outcome in women with chronic liver 60: 57–70; quiz 73–74.
disease. J. Obstet. Gynaecol. India 67: 43 Bermas, B.L. (2014). Non‐steroidal anti
263–269. https://doi.org/10.1007/ inflammatory drugs, glucocorticoids and
s13224‐13016‐10959‐y. disease modifying anti‐rheumatic drugs for
35 Murthy, S.K., Heathcote, E.J., and Nguyen, the management of rheumatoid arthritis
G.C. (2009). Impact of cirrhosis and liver before and during pregnancy. Curr. Opin.
transplant on maternal health during labor Rheumatol. 26: 334–340. https://doi.
and delivery. Clin. Gastroenterol. Hepatol. org/10.1097/BOR.0000000000000054.
7: 1367–1372. 44 Ostesen, M. (1994). Optimisation of
36 Gonsalkorala, E.S., Cannon, M.D., Lim, T.Y. antirheumatic drug treatment in pregnancy.
et al. (2019). Non‐invasive markers (ALBI Clin. Pharmacokinet. 27: 486–503.
and APRI) predict pregnancy outcomes in 45 Francella, A., Dyan, A., Bodian, C. et al.
women with chronic liver disease. Am. J. (2003). The safety of 6‐mercaptopurine for
Gastroenterol. 114: 267–275. childbearing patients with inflammatory
37 Shergill, A.K., Ben‐Menachem, T., bowel disease: a retrospective cohort study.
Chandrasekhara, V. et al. (2012). Gastroenterology 124: 9–17. https://doi.
Guidelines for endoscopy in pregnant and org/10.1053/gast.2003.50014.
lactating women. Gastrointest. Endosc. 76: 46 Alami, Z., Agier, M.S., Ahid, S. et al. (2018).
18–24. https://doi.org/10.1016/j. Pregnancy outcome following in utero
gie.2012.1002.1029. exposure to azathioprine: a French
38 O’Mahony, S. (2007). Endoscopy in comparative observational study. Therapie
pregnancy. Best Pract. Res. Clin. 73: 199–207. https://doi.org/10.1016/j.
Gastroenterol. 21: 893–899. https://doi. therap.2017.1006.1006.
org/10.1016/j.bpg.2007.1005.1007. 47 Saavedra, M.A., Sanchez, A., Morales, S.
39 Ingraham, C.R., Padia, S.A., Johnson, G.E. et al. (2015). Azathioprine during
et al. (2015). Transjugular intrahepatic pregnancy in systemic lupus erythematosus
portosystemic shunt placement during patients is not associated with poor fetal
pregnancy: a case series of five patients. outcome. Clin. Rheumatol. 34: 1211–1216.
Cardiovasc. Intervent. Radiol. 38: https://doi.org/10.1007/s10067‐10015‐
1205–1210. https://doi.org/10.1007/s00270‐ 12987‐x.
00015‐01053‐z. 48 Thomas, C., Monteil‐Ganiere, C., Mirallie,
40 Wood, A.M., Grotegut, C.A., Ronald, J. S. et al. (2018). A severe neonatal
et al. (2018). Identification and lymphopenia associated with
management of abdominal wall varices in administration of azathioprine to the
pregnancy. Obstet. Gynecol. 132: 882–887. mother in a context of Crohn’s disease.
https://doi.org/10.1097/AOG.000000000 J. Crohns Colitis. 12: 258–261. https://doi.
0002805. org/10.1093/ecco‐jcc/jjx1123.
41 Lodato, F., Cappelli, A., Montagnani, M. 49 de Boer, N.K., Jarbandhan, S.V., de Graaf,
et al. (2008). Transjugular intrahepatic P. et al. (2006). Azathioprine use during
portosystemic shunt: a case report of pregnancy: unexpected intrauterine
rescue management of unrestrainable exposure to metabolites. Am. J.
variceal bleeding in a pregnant woman. Gastroenterol. 101: 1390–1392.
https://doi.org/10.1111/j.1572‐0241.2006. J. Nephrol. 2016: 4539342. https://doi.org/

00538.x. 10.1155/2016/4539342.
50 Alsebayel, M.M., Abaalkhail, F.A., 52 Durst, J.K. and Rampersad, R.M. (2015).
Alsebayel, F.M. et al. (2018). Congenital Pregnancy in women with solid‐organ
esophageal atresia and microtia in a transplants: a review. Obstet. Gynecol. Surv.
newborn secondary to mycophenolate 70: 408–418. https://doi.org/10.1097/OGX.
mofetil exposure during pregnancy: a case 0000000000000194.
report and review of the literature. Am. J. 3 Buchanan, N.M., Khamashta, M.A.,
5
Case Rep. 19: 523–526. https://doi. Morton, K.E. et al. (1992). A study of 100
org/10.12659/AJCR.908433. high risk lupus pregnancies. Am. J. Reprod.
1 Shah, S. and Verma, P. (2016). Overview of
5 Immunol. 28: 192–194.
pregnancy in renal transplant patients. Int.
219
12
Bone Health in Patients with Autoimmune Liver Diseases

Albert Parés1,3 and Núria Guañabens2,3
1
Liver Unit, University of Barcelona, Barcelona, Spain
2
Department of Rheumatology, Metabolic Bone Diseases Unit, University of Barcelona, Barcelona, Spain
3
Hospital Clínic, Institut d’Investigacions Biomèdiques August Pi Sunyer (IDIBAPS), Centro de Investigación Biomédica en Red de
Enfermedades Hepáticas y Digestivas (CIBERehd), Barcelona, Spain
Abstract
Autoimmune liver diseases include primary biliary cholangitis (PBC), primary sclerosing cholangi
tis, and autoimmune hepatitis (AIH). Osteoporosis is frequently observed in patients with autoim
mune liver diseases, particularly PBC. The pathogenesis of osteoporosis is mainly characterized by
low bone formation, although increased bone resorption has been described in patients with severe
cholestasis and advanced liver disease. Osteomalacia, characterized by poor bone mineralization,
is very uncommon and has only been reported in a few patients with advanced PBC and severe
intestinal malabsorption, and in geographic areas with decreased exposure to sunlight. The preva
lence of bone disease in patients with AIH is uncertain, although it is assumed that osteoporosis is
one of the most frequent complications of corticosteroid treatment. For the prevention and
treatment of osteoporosis, good nutrition, avoidance of smoking and excess alcohol, weightbearing
exercise, treatment of risk factors, and prescription of calcium and vitamin D supplements appear
helpful.
Keywords autoimmune hepatitis; bone formation; bone mineralization; corticosteroid treatment;

intestinal malabsorption; osteoporosis pathogenesis; primary biliary cholangitis; primary sclerosing
cholangitis; risk factors
Key Points
●● Osteoporosis is frequently observed in ●● The pathogenesis of osteoporosis is mainly
patients with autoimmune liver diseases, characterized by low bone formation,
particularly primary biliary cholangitis although increased bone resorption has also
(PBC). been described in patients with severe cho
●● Osteoporosis is associated with the severity lestasis and advanced liver disease.
of liver disease, older age, use of corticoste ●● Osteomalacia, characterized by poor bone
roids, and the duration of cholestasis mineralization, is very uncommon and
has only been reported in a few patients ●● For the prevention and treatment of oste
with advanced PBC and severe intestinal oporosis, good nutrition, avoidance of
malabsorption, and in geographic areas smoking and excess alcohol, weight‐
with decreased exposure to sunlight. bearing exercise, treatment of risk factors,
●● The prevalence of bone disease in patients and prescription of calcium and vitamin
with autoimmune hepatitis is uncertain, D supplements appear helpful.
although it is assumed that osteoporosis is ●● Oral bisphosphonates, including weekly
one of the most frequent complications of alendronate and monthly ibandronate,
corticosteroid treatment. can be effective for increasing bone min
●● Bone density measurement is the usual eral density.
diagnostic procedure. Lateral X‐ray of the ●● The efficacy of parenteral bisphospho
dorsal and lumbar spine should also be nates such as zoledronic acid, as well
carried out to identify vertebral fractures, as other newer agents, in patients at high
and laboratory measurements of risk for developing osteoporotic frac
circulating levels of calcium, phosphorus, tures needs ongoing evaluation in
25‐hydroxyvitamin D, and parathyroid patients with autoimmune liver
hormone can be assessed at diagnosis. diseases.
Introduction of other immunomodulators in PBC. In

c ontrast, the standard therapy of PSC is con
Autoimmune liver diseases include primary troversial since no effective treatment has
biliary cholangitis (PBC, formerly termed been clearly demonstrated, although UDCA
primary biliary cirrhosis), primary scler is commonly used based on improvement in
osing cholangitis (PSC), and autoimmune liver biochemistries.
hepatitis (AIH). The cholestatic component It has been known for decades that bone
predominates in the first two diseases, while metabolic disorders occur in patients with
hepatocytes are the main target in AIH. In chronic cholestatic diseases, specifically PBC
addition, there are variants or overlap syn [1–4]. This is likely due to the lack of bone
dromes of these diseases, PBC coexisting mineralization as a consequence of the
with AIH, and PSC with AIH. The variant reduced intestinal absorption of vitamin D
PBC/PSC is controversial, possibly due to resulting from the severity and duration of
misinterpretations. the cholestasis. Thus it was considered that
All these diseases have an autoimmune osteomalacia, characterized by diminished
basis and are considered as such. bone mineralization, would be the metabolic
Consequently, the therapeutic strategy is bone disease associated with PBC. However,
based on the use of immunomodulatory subsequent studies have indicated that oste
agents (mostly corticosteroids and azathio omalacia is almost non‐existent in this dis
prine) in AIH, and on agents that modify or ease and that osteoporosis, characterized by
improve cholestasis, such as ursodeoxycholic systemic impairment of bone mass and
acid (UDCA) and other new drugs such as microarchitecture resulting in fragility frac
obeticholic acid and bezafibrate combined tures, is the metabolic bone disease present
with UDCA, in PBC. There is a favorable in PBC. There is less evidence regarding PSC,
therapeutic effect of the combination of and because it is predominantly a chronic
UDCA and budesonide in non‐cirrhotic PBC cholestatic disease [5], it is assumed that
and these medications remain under evalua bone pathology would be similar to that in
tion. There is little evidence for the efficacy PBC. Neither is there much information
Chapter 12 Bone Health in Patients with Autoimmune Liver Diseases 221
about the presence of bone metabolic diseases. Developments in imaging methods

pathology in AIH. However, AIH is treated such as bone densitometry have improved
with corticosteroids and it is essentially the the diagnosis of osteoporosis. At present, the
action of these steroids that are responsible gold standard for the diagnosis of osteopo
for the development of the osteoporosis and rosis consists of assessing bone mineral
bone fractures seen in cases with intense and density (BMD) and the recognition of
long‐term treatment. fractures (Figure 12.1). Subsequently, the
For many years, up until the 1990s, the diagnosis of osteoporosis is mainly based on
diagnosis of bone disease was based on bone BMD with a T‐score equal or below −2.5.
X‐ray and histomorphometry of bone biopsy. Osteopenia is diagnosed when the T‐score is
However, bone biopsy has only been used in between −1 and −2.5. Severe or “established”
a few cases, as it is an invasive method and osteoporosis refers to individuals who meet
also because of the difficulties in obtaining densitometric criteria and have one or more
the specimens at the transilial site and fragility fractures.
performing the analysis of the undecalcified This chapter describes the prevalence of
tissue, in addition to the potential complica metabolic bone disease in these three auto
tions. These problems may explain the immune liver conditions, and the preva
reduced information available from this lence and risk for developing fractures.
diagnostic tool. However, using bone histo Likewise, the pathogenesis of the associ
morphometry it was clarified that osteopo ated bone disease is reported, as well as an
rosis but not osteomalacia is the common update of the recommendations for diag
metabolic bone disease associated with liver nosis and treatment.
(a) (b)
Figure 12.1 (a) The diagnosis of osteoporosis is based on densitometric criteria (T‐score equal or below –2.5).
(b) Severe osteoporosis is diagnosed when the densitometric criteria are accompanied by one or more fragility
fractures.
Prevalence of Osteoporosis
while the prevalence is higher in advanced
and Fractures stages and before transplantation [13].
The prevalence of fractures ranges from 2
Primary Biliary Cholangitis
to 23% (Table 12.1), being about 22% in
Over the years the prevalence of bone dis patients with advanced disease. In a series of
ease in PBC has changed because patients 185 patients, fractures were detected in 21%
are now diagnosed in the early stages of the of the cases (11% vertebral and 12% peripheral
disease and are mostly asymptomatic [3]. fractures) and associated with osteoporosis
Moreover, in the earlier studies bone disease and the severity of cholestasis. A BMD
was assessed by X‐ray of the spine or by bone threshold that captures most vertebral
biopsy and histomorphometry. Bone mass is fractures has been identified. Thus, the
now best assessed by bone densitometry of patients with a lumbar or femoral BMD
the lumbar spine and proximal femur. T‐score below −1.5 are those with a high risk
The reported prevalence of osteoporosis in for fractures. The clear‐cut correlation bet
patients with PBC varies considerably, from ween vertebral fracture and a T‐score less
20 to 90%. Using bone histomorphometry, than −1.5 observed in these patients may
the prevalence of osteoporosis ranges even indicate that this densitometric measurement
more markedly, from 9 to 90% [1,4,6–15]. In is a useful guide for considering therapy.
the most recent and largest series of patients Data regarding incident fractures are fairly
with PBC, the prevalence of osteoporosis, uncertain. One study reported 5% of incident
assessed by densitometry, is approximately fractures over 2 years, while another reported
31% (Table 12.1), and is significantly higher that 13% of patients with PBC had new radio
than in the age‐matched population. Age graphic vertebral deformities over the same
above 56 years, duration of PBC longer than period. Recent reports are contradictory
4 years, and advanced histologic stage are the with respect to the increased fracture risk in
variables associated with osteoporosis [12]. PBC. Thus, one report did not show an
A lower prevalence of osteoporosis is increase in the fracture risk, while another
observed in patients with less severe disease, population‐based cohort study including
Table 12.1 Prevalence of osteoporosis and fractures according to the severity and stage of primary biliary
cholangitis.
No. of Advanced
Reference cases Females (%) disease (%) Osteoporosis (%) Fractures (%)
Guañabens et al. (1994) [6] 38 100 94 45 13

Springer et al. (2000) [7] 72 100 11 24
Menon et al. (2001) [8] 176 83 59 20
Newton et al. (2001) [9] 272 94 54 31
Parés et al. (2001) [10] 61 100 26 21 13
Solerio et al. (2003) [11] 133 100 39 35
Guañabens et al. (2005) [12] 142 100 26 31 14
Guichelaar et al. (2006) [13] 156 86 100 44 22
Guañabens et al. (2010) [14] 185 100 23 37 21
Seki et al. (2017) [15] 128 100 0 26
Average 136 96 43 31 17
930 people with PBC found modest increases a ge‐ and sex‐matched population. About half
in both the absolute and relative fracture risk of patients had early disease, and 22% had
compared with the general population. This cirrhosis. Deficiency of vitamin D and
last study observed an approximately twofold calcium was found in 23% and 4% of patients,
increased risk of any fracture, hip fracture, respectively. The multivariate analyses iden
and ulna/radius fracture among people with tified age older than 54 years, body mass
PBC compared with the general population. index (BMI) below 24.1 kg/m2, and associ
ated IBD for more than 19 years as the
variables associated with osteoporosis.

Primary Sclerosing Cholangitis
Interestingly, osteoporosis was present in
There are fewer reports regarding bone 75% of the patients with these three factors,
health in patients with PSC as compared but in only 3% of patients with none of them.
with PBC. BMD in PSC is lower than There was no significant difference between
expected for age‐ and sex‐matched controls. patients with and without IBD in terms of
Thus, in a cohort of 81 patients included in a baseline bone density or in the rate of bone
randomized trial of UDCA, BMD of the loss of 1% per year, which was significantly
lumbar spine was lower than expected when higher than the bone loss rate expected in a
compared to normal values adjusted for age, matched population. Duration of IBD,
sex and ethnic group at entry and after alkaline phosphatase, aspartate aminotrans
5 years of follow‐up. In this study, 8.6% of ferase, and triglycerides levels, as well as the
patients had lumbar BMD below the fracture Mayo risk score were significantly associated
threshold at entry, and these patients were with a greater rate of bone loss. However, in
older, had a longer duration of inflammatory the multivariate analysis the unique variable
bowel disease (IBD), and more advanced dis that correlated with the rate of bone loss was
ease. Age was the only variable inversely the longer duration of IBD. Another inter
related with baseline lumbar BMD. None of esting finding was the fact that all fragility
the variables predicted progression of the fractures occurred with a BMD T‐score
bone disease. Previous steroid therapy, given lower than −1.5, as previously described for
in 35% of the patients, apparently did not PBC patients. This observation further sup
affect baseline BMD measurements and the ports consideration of treatment in patients
rate of bone loss during the follow‐up. with BMD below this measurement, particu
Moreover, no significant difference in the larly if they have risk factors for osteoporosis
rate of change of BMD was found between or if they are on the waiting list for liver
men and women, neither with respect to the transplantation, since a significant increment
association with IBD. Incident factures were in the risk of fragility fractures is expected
observed in two patients with BMD below after transplantation [13]. Thus, in patients
the fracture threshold. with advanced PSC, the prevalence of osteo
In another more recent bone health evalu porosis was 32.5%, and fractures were present
ation which included 237 patients (42% in 16% of the cases.
female and 58% male) with PSC [5], osteopo
rosis assessed by a BMD T‐score below −2.5
Autoimmune Hepatitis
was found in 15% of patients and occurred
about 25‐fold more frequently than expected There are no confident and conclusive data
for the matched population; 97 patients regarding the prevalence of bone disease in
(41%) had osteopenia (T‐score between −1 patients with AIH, although it is assumed
and −2.5). Severe osteoporosis, defined by that osteoporosis is one of the most frequent
the authors as a Z‐score less than −2, was complications resulting from long‐lasting
found in 33 patients, a prevalence that was corticosteroid therapy. It is estimated that
6.1 times higher than the expected in an osteoporosis and vertebral fracture related
to osteoporosis occur in 5–10% of patients The available bone histomorphometric

receiving immunosuppressant therapy with studies in patients with PBC and PSC support
steroids. In an early series of 39 patients with the fact that decreased bone formation medi
AIH, BMD measured using single‐photon ated by osteoblasts is the main mechanism
absorptiometry of the distal forearm was for osteoporosis, although increased bone
inversely related to the duration of cortico resorption, mediated by osteoclasts, may
steroid treatment and to age at menarche. participate in the end‐stage phase of chole
Thus, in patients with advanced AIH it is static diseases. Thus, a decreased rate of
critical to consider the amount of corticoste bone formation and impaired osteoblast
roids received. Accordingly, patients can be function resulting in lower mean wall thick
stratified to receive low (≤7.5 mg/day) or ness and a defect in matrix synthesis has
high (>7.5 mg/day) prednisone doses, with a been observed. Similar data are present in
clear‐cut association since patients receiving patients with advanced cholestatic liver dis
high doses have a substantially increased ease before transplantation. These morpho
relative risk for fractures. This is particularly metric data are consistent with decreased
relevant for patients receiving 30 mg/day, serum levels of osteocalcin, a bone turnover
with a cumulative dose higher than 5 g in a marker of bone formation.
year. In an analysis of the fracture risk of oral Diverse mechanisms may be implicated in
corticosteroid therapy, it was concluded that osteoblast dysfunction, involving trophic
intermittent use of high‐dose oral cortico agents such as insulin growth factor (IGF)‐1
steroids (daily dose >15 mg, cumulative or the retained substances of cholestasis,
exposure <1 g) may result in a small increased including bilirubin and bile acids
risk of osteoporotic fracture. Conversely, (Figure 12.2). Thus, serum IGF‐1 levels are
patients who receive several courses of high‐ decreased in cirrhotic patients and some
dose corticosteroids (daily dose >15 mg, experiments have demonstrated that IGF‐1
cumulative exposure >1 g) have a consider enhances bone mass and bone density in
ably increased risk of fracture. The conse cirrhotic rats. Increased production of a
quences on bone health of budesonide fibronectin isoform during liver disease may
treatment, an alternative to prednisone in also participate in the decrease in bone
non‐cirrhotic patients with AIH, are still formation. In cholestatic diseases, the con
unknown, although recent data from tribution of bilirubin and bile acids may be
patients with PBC receiving a combination essential. Both bilirubin and bile acids
of UDCA and budesonide revealed a decrease proliferation, differentiation, and
decrease in bone mass after 3 years. mineralization of osteoblasts. Moreover,
sera from jaundiced patients impair osteo
blast differentiation, and bilirubin and lith
Pathogenesis ocholic acid induce apoptosis in osteoblastic
cells. Interestingly, UDCA counteracts these
The mechanisms resulting in impaired bone damaging effects on osteoblasts. The neu
health in patients with autoimmune liver dis tralizing effects of UDCA might explain the
eases may be different, given that in PBC and favorable consequences of the drug, which
PSC the cholestatic component prevails, while not only improves liver damage in PBC but
in AIH, apart from the severity and duration also the course and consequences of low
of liver damage, the consequences of cortico bone formation, thus preventing, or at least
steroid therapy predominate. Nevertheless, delaying, the development of osteopenia
the pathogenesis is mainly related to osteopo and osteoporosis in these patients. However,
rosis, since osteomalacia is a very infrequent this benefit of UDCA on bone health has
condition that is only present in cases with not been demonstrated in the clinical
extremely severe and long‐lasting cholestasis. setting.
Mechanisms What’s involved

(a)
Sclerostin
Low bone Retained substances of cholestasis:
formation • bilirubin and bileacids
Decreased trophic factors (IGF1)
(b)
HPT secondary to vitamin D deficiency
Transient
Bilirubin; RANKL/OPG system?
increased bone
resorption Unknown
(c) Bone matrix protein alterations Low vitamin K

Genetic susceptibility COLIA1, VDR, IGF1, CLDN14
Poor nutrition
Figure 12.2 (a) Pathogenesis of osteoporosis in primary biliary cholangitis. Low bone formation is the main
mechanism for osteoporosis, related to the detrimental effects of retained cholestatic substances and
sclerostin. Other factors such as IGF‐1 may influence bone formation. (b) Transient increased bone resorption is
the other mechanism, with vitamin D deficiency and alterations in the RANKL/OPG system. HPT,
hyperparathyroidism. (c) Bone matrix protein alterations related to low vitamin K and some gene
polymorphisms also contribute to osteoporosis.
Recent data have proposed that increased the cause of increased bone turnover in some
sclerostin, a key regulator of the Wnt/β‐ patients. Osteoprotegerin (OPG) and
catenin signaling pathway mainly produced receptor activator of nuclear factor kappa‐B
by osteocytes and which therefore regulates ligand (RANKL) can be involved in the
bone formation, may be involved in the path increased bone resorption, although at pre
ogenesis of osteoporosis in PBC. Vitamin K sent their role remains uncertain.
deficiency may be another additional factor Nevertheless, an increased RANKL/OPG
for osteoporosis based on some studies pro ratio, resulting in high bone resorption and
posing that vitamin K deficiency causes ultimately bone loss, has been observed in
reductions in BMD and increases the frac cirrhotic patients. In addition, serum from
ture risk. Likewise, vitamin K supplementa jaundiced patients upregulates RANKL/
tion has been associated with improvements OPG gene expression, a system intimately
in BMD. The influence of some gene poly involved in osteoblast‐induced osteoclasto
morphisms has also been assessed in patients genesis. The role of some cytokines such as
with PBC. However, these polymorphisms interleukin (IL)‐1 and IL‐6 and tumor
either do not take part in or have a minor necrosis factor (TNF)‐α, which directly or
effect on the development of osteoporosis in indirectly activates the osteoclasts, and the
this cholestatic disease. role of specific adipokines in the pathogen
Elevated bone resorption may also be esis of osteoporosis in patients with PBC and
involved in the pathogenesis of osteoporosis PSC remains to be established, although a
in end‐stage cholestatic diseases. Accor negative correlation between adiponectin
dingly, reduced trabecular thickness and and BMD has been reported.
increased bone turnover proportional to the The pathogenesis of osteoporosis induced
severity of hepatic dysfunction and chole by corticosteroids is mostly relevant for
stasis have been reported. Calcium and patients with AIH. Reduced bone formation
vitamin D deficiencies leading to secondary is the critical process. Even at low doses, cor
hyperparathyroidism have been proposed as ticosteroids have been shown to rapidly
s uppress several indices of osteoblast activity, Table 12.2 Risk factors for bone loss.
including serum markers of bone formation
Alcohol abuse
such as serum procollagen type 1 N‐terminal
propeptide (P1NP) and osteocalcin. More Smoking
over, corticosteroid excess inhibits both oste Body mass index <19 kg/m2
oblast differentiation and function while at Male hypogonadism
the same time inducing osteoblast apoptosis, Early menopause
resulting in rapid and deep suppression of
Secondary amenorrhea of more than 6 months
bone formation. Corticosteroids have early
Family history of osteoporotic fracture
effects on bone resorption as well and there
fore they not only compromise the ability of Treatment with corticosteroids (5 mg/day or more
of prednisone for 3 months or longer)
the skeleton to form new bone but also dis
turb the regulatory balance between bone Advanced age
formation and resorption, which finally leads
to a reduction in bone mass and an increase
in fracture risk. In addition to the direct history of bone fragility fractures, and
effects of corticosteroids on osteoblasts, treatment with glucocorticoids (>5 mg of
osteoclasts and osteocytes, they may indi prednisone over more than 3 months)
rectly affect bone metabolism and fracture (Table 12.2).
risk through effects on gonadal hormones Blood levels of calcium, phosphorus,
and the neuromuscular system (as indicated 25‐hydroxyvitamin D and parathyroid hor
by an increased propensity for falls in corti mone should be measured at diagnosis and at
costeroid‐treated patients). The severity and 1‐ or 2‐year intervals. This is important since
advanced stage of AIH can also contribute to vitamin D (25‐hydroxyvitamin D) levels
decreasing bone mass and increasing frac below 20 ng/ml have been reported in
ture risk. 64–92% of patients with liver diseases, pre
dominantly in chronic cholestatic condi
tions, and which are usually inversely
Assessment of Bone Disease correlated with more advanced disease.
Consequently, these measurements should
The gold standard for the diagnosis of osteo be evaluated more frequently in patients with
porosis involves evaluation of BMD by dual‐ severe cholestasis. The biochemical markers
energy X‐ray absorptiometry (DXA) and the of bone turnover can be determined, but
assessment of fractures. A BMD T‐score below they are mainly helpful for monitoring the
−2.5 is the criteria for osteoporosis. Osteopenia response to specific therapy for bone disease.
is diagnosed when the T‐score is between −1 Disturbances in thyroid and gonadal function
and −2.5. Severe or “established” osteoporosis should be ruled out in particular cases
refers to individuals who meet densitometric depending on the medical history.
criteria and have one or more fragility fractures. BMD should be repeated after 2–3 years to
BMD should be performed in all patients with detect bone loss in patients with baseline
autoimmune liver diseases. Lateral X‐rays of normal or nearly normal range. However,
the dorsal and lumbar spine should also be car evaluation should be performed at a shorter
ried out to reveal vertebral fractures. interval of approximately 1–2 years in patients
In all the patients with autoimmune liver with more than one additional risk factor for
diseases, risk factors for low bone mass and osteoporosis in patients with deep chole
fractures should be assessed, including high stasis, and in those treated with high doses of
alcohol intake and tobacco abuse, BMI lower corticosteroids. This strategy is also recom
than 19 kg/m2, early menopause, secondary mended for patients with late‐stage disease
amenorrhea of more than 6 months, family and those eligible for transplantation.
Prevention and Treatment women with PBC the mean annual change in

of Bone Loss BMD was 0.1% in the patients treated with
vitamin D and −3.1% in the control group,
General Measures suggesting that vitamin D prevents bone loss
in this disease. However, other studies did
Factors contributing to bone loss should be not support the claim that vitamin D supple
reduced to a minimum by stopping alcohol mentation avoids bone loss.
intake and smoking [3]. Physical activity and
exercises aimed at improving the mechanics
Pharmacologic Agents
of the spine are recommended. A balanced
diet should be prescribed, especially in Bisphosphonates
patients with advanced disease because they There is no agreement concerning the appro
may have little appetite and are malnourished. priate time to start treatment of bone disease
Supplements of calcium (1000–1500 mg/day) in patients with autoimmune liver illnesses.
and vitamin D (800–1000 IU/day or 260 μg At present, patients with established osteo
of 25‐hydroxyvitamin D every 2 weeks) or porosis, and therefore with fragility fractures,
the dose required to maintain normal levels should be treated to reduce the risk of further
should be provided (Figure 12.3). Patients fractures. On the other hand, and consid
receiving colestyramine should be closely ering that patients with a lumbar spine or
assessed since resins may reduce the proximal femur T‐score lower than −1.5 have
absorption of vitamin D. a high risk for fracture, it seems reasonable to
There are no explicit data confirming the treat patients with a bone mass below this
efficacy of calcium and vitamin D supple densitometric value, particularly if other risk
ments for preventing bone loss in patients factors are present.
with liver disease. Thus there have been few Bisphosphonates are anti‐resorptive drugs
clinical trials of the effects of vitamin D sup that increase bone mass and reduce the
plementation alone, the optimal dose, and incidence of fractures in postmenopausal

the formulation in patients with cirrhosis. In osteoporosis. Their effects in PBC are weak,
Diagnosis Management
DXA
lumbar spine and hip DXA Treatment
Repeat in 2–3 yrs
Normal
Ca + Vitamin Da
T-score > –1.5
Osteopenia Repeat in 1–2 yrsb Ca + Vitamin Da

T-score ≤ –1.5 > –2.5 Bisphosphonatesc
Ca + Vitamin Da
Repeat in 1–2 yrsb
Osteoporosis Bisphosphonates
T-score ≤ –2.5 New agents
Figure 12.3 Diagnosis and management of osteoporosis in primary biliary cholangitis. a Calcium (1000–
1500 mg/day) and vitamin D (400–800 IU/day or 260 μg 25‐hydroxyvitamin D every 2 weeks) to preserve
normal levels. b According to the severity of cholestasis. c Depending on additional risk factors. DXA,
dual‐energy X‐ray absorptiometry.
basically because of the low number of trials administration was associated with higher
and the few patients treated (Table 12.3). BMD values in patients with higher indices
Nonetheless, it has been demonstrated that of cholestasis. Ibandronate is also effective at
cyclic administration of etidronate is able to increasing bone mass. Accordingly, in a
prevent bone loss after 2 years of treatment randomized trial comparing monthly iban
and that alendronate increases bone mass in dronate and weekly alendronate for osteopo
patients with PBC, effects which are rosis in PBC, both treatments resulted in a
comparable to those reported in osteopo significant increase in BMD at the lumbar
rosis due to other causes. A placebo‐con spine after 2 years of therapy (Figure 12.4)
trolled trial of alendronate (70 mg/week) [21]. The mean percentage change was 4.5%
indicates that this bisphosphonate is able to and 5.7% at the lumbar spine for alendro
increase bone mass after one year with no or nate and ibandronate, respectively. Changes
minor adverse effects. The BMD of patients in bone markers were similar in both groups
treated with alendronate increased signifi and one patient receiving alendronate devel
cantly at both the lumbar spine and proximal oped a new vertebral fracture. Adherence to
femur compared with placebo. The effects of therapy was higher with ibandronate. Neither
weekly alendronate on BMD were more treatment impaired liver function or chole
patent at the lumbar spine, and greater than stasis. Serious adverse effects were not
with the daily alendronate regimen. Our observed and potential unfavorable effects of
results of once‐weekly alendronate 70 mg in oral bisphophonates, such as esophagitis,
PBC indicate that this regimen is more effec were not reported, likely because the patients
tive and has a better tolerability profile than were carefully instructed to take the medica
daily dosing in the treatment of low bone tion on an empty stomach, swallowing an
mass in patients treated for 1 year. An inter adequate amount of water and remaining in
esting observation was that alendronate an upright position for at least 30 minutes.
Table 12.3 Effect of different oral bisphosphonates on lumbar spine and femoral T‐scores after 1 or 2 years
of therapy in patients with primary biliary cholangitis.
Percent Percent
lumbar femoral
No. of
Dose patients 1 year 2 years 1 year 2 years
Etidronate
Guañabens et al. (1997) [16] 400 mg/day, 13 0.5 1.3
Wolfhagen et al. (1997) [17] cycles 3 months 6 1
Lindor et al. (2000) [18] 29 0.7 1 1.4 0.5
Guañabens et al. (2003) [19] 13 1.9 0.4
Alendronate
Guañabens et al. (2003) [19] 10 mg/day 13 5.8 3.9
Guañabens et al. (2005) [3] 70 mg/week 16 3.3 1.2
Guañabens et al. (2005) [3] 10 mg/day 10 1.2 −0.3
Zein et al. (2005) [20] 70 mg/week 15 10.4 0.4
Guañabens et al. (2013) [21] 70 mg/week 19 4.5 2.7
Ibandronate
Guañabens et al. (2013) [21] 150 mg/month 14 5.7 1.2
5
% change
0
Basal 6 12 18 24
Month
Figure 12.4 Percentage changes in lumbar bone mineral density (BMD) with respect to baseline values for
patients with primary biliary cholangitis treated with ibandronate (broken line) or alendronate (solid line).
No significant differences in BMD change were observed between treatments for each time period, but the
increase in BMD was significant in each arm from 6 months.
There is little information on the effect of patients included in one trial suggest that
parenteral bisphosphonates in PBC. One hormone therapy may worsen cholestasis.
retrospective study assessed the effect of
Despite these fairly positive results, currently
zoledronic acid, pamidronate disodium, or hormone therapy is not considered the most
ibandronate sodium in PBC patients with appropriate treatment for osteoporosis as
osteoporosis. The percentage change of there are other efficacious non‐hormonal
lumbar spine and hip BMD was 2.9% and agents such as bisphosphonates.
0.4%, respectively. No serious adverse events The efficacy and safety of raloxifene, a
were found. Although the number of patients selective estrogen receptor modulator, has
treated was very small and with a diversity of been assessed in a pilot study including
regimens, parenteral bisphosphonates stabi nine postmenopausal women with PBC.
lized BMD in about half of the patients. More Compared to baseline, lumbar spine BMD
prospective studies are needed to evaluate improved significantly after 1 year of
the efficacy of specific parenteral bisphos therapy but not in matched controls, and
phonates in patients with PBC and osteopo no significant effect was observed in femoral
rosis. No information is available regarding neck.
bone disease therapy in patients with PSC Preliminary results with denosumab, a fully
and AIH. human IgG2 monoclonal antibody directed
against RANKL, in 13 patients with PBC indi
Other Agents cate that lumbar spine T‐score significantly
There have been some studies indicating that improved after 12 months, along with changes
hormonal treatment prevents bone loss or in bone alkaline phosphatase and tartrate‐
even increases BMD in patients with PBC. resistant acid phosphatase 5b, as markers of
Other studies have also confirmed that trans bone formation and osteoclast number,
dermal estradiol and medroxyprogesterone respectively. During the treatment period,
plus calcium and vitamin D supplements humeral fracture at 11 months and latent
increase both lumbar spine and femoral hypocalcemia was observed in two patients.
BMD after 2 years, with no severe side Data are very limited for definite recommen
effects. However, the individual data of the dations. The effect of other agents, such as
anabolic therapy with teriparatide that is used ndetermined, although it is assumed that
u
in postmenopausal osteoporosis, also war osteoporosis is one of the most frequent
rants additional investigation in patients with complications of corticosteroid treatment.
autoimmune cholestatic disease. DXA is the usual diagnostic procedure
and should be performed in all patients.
Lateral X‐ray of the dorsal and lumbar
Summary spine should also be carried out to disclose
vertebral fractures, and laboratory
Osteoporosis is a complication frequently measurements of circulating levels of calc
observed in patients with autoimmune liver ium, phosphorus, 25‐hydroxyvitamin D
diseases, particularly PBC. The pathogenesis and parathyroid hormone should be
of osteoporosis is mainly characterized by assessed at diagnosis.
low bone formation, although increased For the prevention and treatment of osteo
bone resorption has also been described in porosis good nutrition is required, as are the
patients with severe cholestasis and advanced suppression of risk factors for o steoporosis
liver disease. Osteoporosis is associated with and the administration of calcium and vitamin
the severity of liver disease, older age and the D supplements. There is no specific treatment
duration of cholestasis. Patients with osteo for bone loss, although it has been demon
porosis and those with a lumbar spine T‐score strated that oral bisphosphonates, especially
below −1.5 have a high risk for fractures, par weekly alendronate and monthly ibandronate,
ticularly of the spine. Osteomalacia, charac are effective at increasing BMD. The potential
terized by poor bone mineralization, is very of parenteral bisphosphonates such as zole
uncommon and has only been reported in a dronic acid or the efficacy of newer agents
few patients with advanced PBC and severe that are now prescribed in patients with post
intestinal malabsorption, and in geographic menopausal osteoporosis and at high risk for
areas with decreased sunlight expo sure. developing fractures needs further evaluation
Information regarding the prevalence in patients with autoimmune liver diseases
of bone disease in patients with AIH is and low bone mass.
References
1 Guañabens, N. and Parés, A. (2010). Liver D receptor gene polymorphisms and bone
and bone. Arch. Biochem. Biophys. 503: mass in primary biliary cirrhosis.
84–94. Hepatology 33 (3): 554–560.
2 Leslie, W.D., Bernstein, C.N., Leboff, M.S. 11 Solerio, E., Isaia, G., Innarella, R. et al.
et al. (2003). AGA technical review on (2003). Osteoporosis: still a typical
osteoporosis in hepatic disorders. complication of primary biliary cirrhosis?
Gastroenterology 125: 941–966. Dig. Liver Dis. 35 (5): 339–346.
3 Pares, A. and Guañabens, N. (2018). 12 Guañabens, N., Parés, A., Ros, I. et al.
Primary biliary cholangitis and bone (2005). Severity of cholestasis and
disease. Best Pract. Res. Clin. Gastroenterol. advanced histological stage but not
34–35: 63–70. menopausal status are the major risk
4 Collier, J. (2007). Bone disorders in chronic factors for osteoporosis in primary biliary
liver disease. Hepatology 46: 1271–1278. cirrhosis. J. Hepatol. 42: 573–577.
5 Angulo, P., Grandison, G.A., Fong, D.G. 13 Guichelaar, M.M., Kendall, R., Malinchoc,
et al. (2011). Bone disease in patients with M., and Hay, J.E. (2006). Bone mineral
primary sclerosing cholangitis. density before and after OLT: long‐term
Gastroenterology 140: 180–188. follow‐up and predictive factors. Liver
6 Guañabens, N., Parés, A., Navasa, M. et al. Transpl. 12: 1390–1402.
(1994). Cyclosporin A increases the 14 Guañabens, N., Cerdá, D., Monegal, A.
biochemical markers of bone remodeling in et al. (2010). Low bone mass and severity of
primary biliary cirrhosis. J. Hepatol. 21 (1): cholestasis affect fracture risk in patients
24–28. with primary biliary cirrhosis.
7 Springer, J.E., Cole, D.E., Rubin, L.A. et al. Gastroenterology 138 (7): 2348–2356.
(2000). Vitamin D‐receptor genotypes as 15 Seki, A., Ikeda, F., Miyatake, H. et al.
independent genetic predictors of (2017). Risk of secondary osteoporosis
decreased bone mineral density in primary due to lobular cholestasis in non‐cirrhotic
biliary cirrhosis. Gastroenterology 118 (1): primary biliary cholangitis.
145–151. J. Gastroenterol. Hepatol. 32 (9):
8 Menon, K.V., Angulo, P., Weston, S. et al. 1611–1616.
(2001). Bone disease in primary biliary 16 Guañabens, N., Parés, A., Monegal, A.
cirrhosis: independent indicators and rate et al. (1997). Etidronate versus fluoride
of progression. J. Hepatol. 35 (3): 316–323. for treatment ofosteopenia in primary
9 Newton, J., Francis, R., Prince, M. et al. biliary cirrhosis: preliminary results
(2001). Osteoporosis in primary biliary after 2 years. Gastroenterology 113 (1):
cirrhosis revisited. Gut 49 (2): 282–287. 219–24.
10 Parés, A., Guañabens, N., Alvarez, L. et al. 17 Wolfhagen, F.H., van Buuren, H.R., den
(2001). Collagen type Ialpha1 and vitamin Ouden, J.W. et al. (1997). Cyclical
etidronate in the prevention of bone loss cirrhosis. Am. J. Gastroenterol. 98 (10):

in corticosteroid‐treated primary biliary 2268–2274.
cirrhosis. A prospective, controlled pilot 0 Zein, C.O., Jorgensen, R.A., Clarke, B. et al.
2
study. J. Hepatol. 26 (2): 325–330. (2005). Alendronate improves bone
8 Lindor, K.D., Jorgensen, R.A., Tiegs, R.D.
1 mineral density in primary biliary cirrhosis:
et al. (2000). Etidronate for osteoporosis in a randomized placebo‐controlled trial.
primary biliary cirrhosis: a randomized Hepatology 42 (4): 762–771.
trial. J. Hepatol. 33 (6): 878–882. 21 Guañabens, N., Monegal, A., Cerdá, D. et al.
19 Guañabens, N., Parés, A., Ros, I. et al. (2013). Randomized trial comparing monthly
(2003). Alendronate is more effective than ibandronate and weekly alendronate for
etidronate for increasing bone mass in osteoporosis in patients with primary biliary
osteopenic patients with primary biliary cirrhosis. Hepatology 58: 2070–2078.
233
Section IV
Transplantation and Its Role in Autoimmune Liver Disease

235
13
Recurrent Autoimmune Liver Disease and Its Impact on

Clinical Practice
Carlos Moctezuma‐Velázquez1,2 and Aldo J. Montano‐Loza2
1
Mount Sinai Liver Cancer Program, Division of Liver Diseases, Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai,
New York, NY, USA
2
Division of Gastroenterology and Liver Unit, University of Alberta Hospital, Edmonton, Alberta, Canada
Abstract
Autoimmune liver diseases (ALIDs) primary biliary cholangitis (PBC), primary sclerosing cholan-
gitis (PSC) and autoimmune hepatitis (AIH) recur after liver transplantation (LT). This chapter
reviews indications for LT in patients with AILD, and the frequency and clinical impact of recur-
rent AILD after LT. Recurrent PBC is diagnosed on the basis of liver histology and may be seen with
normal liver tests; antimitochondrial antibodies recur irrespective of histologic recurrence.
Recurrent PSC is usually diagnosed on the basis of the demonstration of multiple nonanastomotic
strictures in the absence of other risk factors. Recurrent AIH may occur and is diagnosed on the
basis of characteristic autoantibodies and liver histology. Recurrence of AILDs reduces odds of
graft and patient survival, and management strategies are needed to prevent AILD recurrence or
reduce its negative effects.
Keywords antimitochondrial antibodies; autoimmune hepatitis; clinical impact; liver histology;

liver transplantation; primary biliary cholangitis; primary sclerosing cholangitis; recurrent disease;
risk factors
Key Points
●● Autoimmune liver diseases – primary recurrence. Graft loss is unusual and
biliary cholangitis (PBC), primary sclerosing treatment with ursodeoxycholic acid may
cholangitis (PSC) and autoimmune hepatitis slow progression. Recurrence appears
(AIH) – recur after liver transplantation. more frequent for those on tacrolimus‐
●● The reported prevalence of recurrence var- based immunosuppression compared
ies between series, in part because of varia- with ciclosporin.
tion in detection and diagnostic criteria. ●● Recurrent PSC is usually diagnosed on the
●● Recurrent PBC is diagnosed on the basis basis of the demonstration of multiple
of liver histology and may be seen with non‐anastomotic strictures in the absence
normal liver tests; anti‐mitochondrial of other risk factors (such as any alternate
antibodies recur irrespective of histologic cause of ischemic cholangiopathy).
236 Section IV Transplantation and Its Role in Autoimmune Liver Disease
Although controversial, recurrent PSC is ●● Recurrent AIH may occur and is diag-
less likely in those with a colectomy either nosed on the basis of characteristic auto-
before or at the time of transplant. There antibodies and liver histology. The natural
is no effective treatment other than history may be improved by optimisation
retransplantation. of immunosuppression.
Introduction PBC in recent years. Interestingly, the female

to male ratio has also halved, and this may be
Autoimmune liver disease (AILD), com- due to the differential response to UDCA [4].
prising primary biliary cholangitis (PBC), The most common indication for LT in
primary sclerosing cholangitis (PSC), and PBC is end‐stage liver disease with life‐threat-
autoimmune hepatitis (AIH), can progress to ening complications of portal hypertension.
end‐stage liver disease and constitutes the Referral for LT assessment according to
third most common indication for liver international guidelines should be considered
transplantation (LT) [1]. Overall, patient and early, when bilirubin approaches above 6 mg/
graft survival after LT of these patients are dl, the Mayo Risk Score is 7.8 or more, and/or
acceptable, with 5‐year survival rates of Model for End‐stage Liver Disease (MELD)
around 77, 73, and 74% for PBC, PSC, and score is above 12. Other accepted indications
AIH, respectively [2]. for LT in PBC include hepatocellular carci-
LT for AILDs shares some peculiarities, for noma (HCC) and intractable pruritus.
example these patients are more prone to Patients with PBC have a good outcome
have rejection episodes and the diseases can after LT when compared to other etiologies.
recur in the new graft, compromising quality Most cohorts report 1‐, 5‐, and 10‐year
of life and patient and graft function. patient/graft survival of around 90–94%/85–
Recurrence rates vary across the different 92%, 83–98%/78–87%, and 76–94%/72–84%,
studies depending on the length of follow‐up, respectively [1,3,5,6].
whether biopsies were done by protocol or
per clinical indication, and the criteria used Recurrence of PBC After LT
to define recurrence; because of this, factors
The rate of recurrent PBC (rPBC) increases as
associated with recurrent disease are incon-
time passes after LT; median time to recur-
sistent across the different studies.
rence is usually between 3.5 and 5 years. There
In this chapter, we review current indica-
are variable rates of rPBC across the studies,
tions for LT in patients with AILD, and the
depending on the criteria used to define rPBC
frequency and clinical impact of recurrent
and on whether protocol or clinically driven
AILD after LT.
biopsies were made. For example, in a previous
study from our group with clinically driven
biopsies, we calculated 5‐ and 10‐year
Primary Biliary Cholangitis recurrence rate of 13 and 29% [5], respectively,
whereas in a study with protocol‐guided
Currently PBC accounts for roughly 4% of all biopsies, the recurrence rate was higher, as
LTs performed in North America [3]. The expected, with 5‐, 10‐, and 15‐year rates of 27,
use of ursodeoxycholic acid (UDCA) in 47, and 61%, respectively [6].
patients with PBC, and more recently the
use of obeticholic acid (OCA), has resulted in
Diagnosis of PBC After LT
increased LT‐free survival and a reduction
in the number of patients in need of LT. Diagnosis of rPBC has some particular-
Therefore, less LTs have been performed for ities and it is not appropriate to use the
Chapter 13 Recurrent Autoimmune Liver Disease and Its Impact on Clinical Practice 237
pre‐LT criteria. Currently, diagnosis is lymphoid aggregates, epithelioid granulomas,

essentially based on histopathologic find- or florid bile duct lesions. Two of the
ings on liver biopsy. The following criteria previous features in the absence of endo-
are considered when making a diagnosis theliitis strongly suggests rPBC.
of rPBC. ●● Diagnosis of exclusion: it is necessary to
rule out other etiologies of biliary dis-
●● Confirmed diagnosis of PBC before LT. ease that are common in the LT setting
●● Positive anti‐mitochondrial antibodies: on and which may be confused with rPBC
their own, antibodies are not considered a such as hepatic artery thrombosis, anas-
reliable marker of recurrent disease tomotic stricture, rejection, viral
because they may persist after LT, or disap- hepatitis, or drug‐induced liver injury
pear and sometimes reappear later, without (Table 13.1).
any clinical manifestation.
●● Graft dysfunction: around 40% patients
Risk Factors Associated with PBC
may have histologically proven rPBC with
Recurrence After LT
normal liver tests [6].
●● Histopathology findings: because of points Factors associated with rPBC vary across
1–3, diagnosis of rPBC hardly depends on the different studies. In many of the studies,
pathology findings. Compatible histologic no multivariable analysis was made to
findings include lymphocytic cell infiltrate, adjust for potential confounders, so the
Table 13.1 Diagnostic criteria for recurrence autoimmune liver diseases after liver transplantation.
Disease Diagnostic criteria
Recurrent ●● History of liver transplantation related to PBC

primary biliary ●● Confirmed diagnosis of PBC in the explant histology with characteristic histologic
cholangitis features (lymphoplasmacytic portal infiltrate, lymphoid aggregates, epithelioid
granulomas, evidence of bile duct injury)
●● Persistence of AMA or AMA‐M2
●● Exclusion of other causes of graft dysfunction (acute and chronic rejection, graft‐
versus‐host disease, bile flow impairment or cholangitis, vascular complications,
viral hepatitis, drug‐induced hepatitis)
Recurrent ●● Confirmed diagnosis of PSC before LT
primary ●● Cholangiography showing non‐anastomotic intrahepatic and/or extrahepatic biliary
sclerosing strictures with beading and irregularities of bile ducts at least 90 days after LT
cholangitis and/or histopathologic findings of fibrous cholangitis and/or fibro‐obliterative lesions
●● Exclusion of other causes of graft dysfunction (hepatic artery thrombosis or
stenosis, chronic ductopenic rejection, anastomotic and non‐anastomotic strictures
before day 90 after LT, ABO‐incompatible LT)
Recurrent ●● History of liver transplantation related to AIH
autoimmune ●● Hypergammaglobulinemia, increased serum IgG levels, and ANA, SMA, or both
hepatitis ANA and SMA
●● Histologic aspectsa (prominent lymphocytic interface activity with or without
plasma cell infiltration, acute lobular hepatitis with focal hepatocyte necrosis,
acidophil bodies with lymphoplasmacytic cells, pseudo‐rosetting of hepatocytes,
and perivenular lymphoplasmacytic inflammation, confluent and bridging necrosis
with lymphoplasmacytic infiltration indicating severe inflammatory activity)
a
Features may be less pronounced, absent, or otherwise atypical in part because of concurrent anti‐rejection therapy.
ANA, anti‐nuclear antibody; AMA, anti‐mitochondrial antibody; SMA, smooth muscle antibody.
presumed associations are of questionable an earlier and more aggressive course [5].
significance. Azathioprine was found to protect against
Younger age at diagnosis and at the time of rPBC in some studies, though one of them did
LT is associated with higher risk of rPBC [7]. not adjust for confounding variables. There is
Male recipient was significant in one study, but less compelling evidence that steroids may also
this finding could not be replicated in others play a role in preventing rPBC, but his
studies [5,6]. Biochemical markers of chole- association is usually lost in multivariable anal-
stasis within the first year after LT are also ysis after adjusting for ciclosporin (Table 13.2).
associated with rPBC [7]. Patients with overlap
syndrome (PBC/AIH) on the native liver as an
Treatment of PBC Recurrence After LT
indication for LT are also at risk of recurrence.
Regarding human leukocyte antigen (HLA) The standard of treatment for rPBC is
status, there is conflicting evidence as some UDCA. However, unlike in pre‐LT patients,
studies have found an association between the use of UDCA has not been shown to
HLA mismatching or specific HLA and rPBC, modify the natural history of rPBC in the
and others have not. The polymorphism short to middle term, even after improving
rs62270414 at the IL12A locus has been asso- alkaline phosphatase (ALP) levels. Long‐
ciated with a higher risk of rPBC [6]. term studies showing a benefit on progres-
Tacrolimus when compared to ciclosporin sion of rPBC are needed. In addition, we
has been associated with rPBC, and also with lack data regarding the use of second‐line
Table 13.2 Frequency and risk factors for recurrence of autoimmune liver disease after liver transplantation.
Frequency of
Disease recurrence Risk factors for recurrence
Primary biliary 10–20% at 5 years ●● Younger age at diagnosis and at LT

cholangitis 10–35% at 10 years ●● HLA‐DR locus mismatching
●● Use of tacrolimus
●● Biochemical cholestasis within the first year after LTa
●● High IgM at the time of LT
●● Gender mismatch
Primary sclerosing 12–50% at 5 years ●● Male gender
cholangitis 20–60% at 10 years ●● Intact colon
●● Donor–recipient gender mismatch
●● Ulcerative colitis requiring maintenance steroids
●● HLA‐A locus mismatching
●● Cholestasis in the third month post‐LTb
●● IBD, older donor age, and higher INR at the time of LT
Autoimmune 18–32% at 5 years ●● Discontinuation of steroid therapy
hepatitis 36–68% at 10 years ●● HLA matching between donor and recipient for
HLA‐DR3 or ‐DR4
●● HLA‐DR locus mismatching
●● Moderate‐to‐severe hepatic inflammation in the explant
●● High IgG before transplant
a
Serum total bilirubin >100 μmol (>5.9 mg/dl), an increase in alkaline phosphatase (ALP) more than three times
upper limit of normal or a combination of the two.
b
ALP level more than twice upper limit of normal or a combined elevation of both bilirubin and ALP levels.
HLA, human leukocyte antigen; IBD, inflammatory bowel disease; INR, international normalized ratio.
Primary Biliary Cholangitis Primary Sclerosing Cholangitis Autoimmune Hepatitis
Pre-Transplantation Strategies
• Use of UDCA • Appropriate control of IBD before • Treatment of active cirrhosis

transplant • Treatment to normalization
• Consider pre-transplant colectomy of AST/ALT//IgG before
in patients with compensated liver transplant
disease and difficult to control IBD
Peri-post Transplantation Strategies
• Use of preventive UDCA • Maintaining remission IBD after • Long-term corticosteroid

Use ciclosporin as main transplant (low dose) use after liver
Immunosuppression • Consider post-transplant transplantation
colectomy (within 6 months)
in patients with difficult-to-control IBD
• Avoid EDC, ABO-incompatible and
DCD donors
Figure 13.1 Strategies to reduce the risk of autoimmune liver disease recurrence after liver transplantation.
ALT, alanine aminotransferase; AST, aspartate aminotransferase; UDCA, ursodeoxycholic acid;
IBD, inflammatory bowel disease; EDC, extended donor criteria; DCD, donation after cardiac death.
Source: Montano‐Loza et al. [2]. Reproduced with permission of John Wiley and Sons.
t reatments, such as OCA or b ezafibrate for compared to those with no recurrent

non‐responders to UDCA in the post‐LT disease [7].
setting. The preventive administration of
UDCA started within the first 2 years after
LT was associated with lower risk of rPBC [6] Primary Sclerosing
(Figure 13.1). Cholangitis
Contrary to what has happened in PBC, due
Prognostic Impact of PBC Recurrence
to the lack of effective therapeutics, the
After LT
proportion of LT due to PSC has remained
Prior reports suggested that the development relatively stable over time (4–5%) and is
of rPBC had no significant impact on long‐ currently the most common cause of LT

term patient survival or need for retransplan- secondary to AILD [1,4]. The usual time
tation [5,6], and thus its clinical impact had frame between the diagnosis of PSC and the
been questioned. However, the fact that those need for LT in patients outside of referral
observations were inaccurate may be due to centers is 12–18 years.
short follow‐up and a limited number of LT is indicated mostly in cases of end‐
patients. A recent multicenter study found stage liver disease with portal hypertension‐
rPBC to be associated with worse graft and related complications as in other etiologies
patient survival after LT. As expected, the of cirrhosis. HCC may be an indication for
majority of patients with rPBC lost their graft LT, but the risk is lower than with other
because of disease progression to cirrhosis etiologies of end‐stage liver disease such

due to recurrent disease. Additionally, overall as viral hepatitis. As in PBC patients,
survival was lower in patients with rPBC refractor y pruritus and non‐cirrhotic portal
ypertension may also be indications for LT.

h Risk Factors Associated with
A particular scenario when LT should be PSC Recurrence After LT
considered in PSC patients is the presence of The risk factors that have been identified
severe and/or recurrent episodes of cholan- among the different studies are not uniform,
gitis, which is considered a non‐standard though inflammatory bowel disease (IBD)
MELD exception policy. and the use of extended criteria donors seem
Some PSC patients with unresectable to play an important role. Colectomy before
hilar cholangiocarcinoma in an early stage or at the time of LT was found to be protective
might benefit from LT under a strict pro- against recurrence. More recently, the
tocol that includes the use of neoadjuvant performance of colectomy plus end ileos-
therapy, followed by staging surgery, with tomy was associated with more favorable
5‐year recurrence‐free survival of approxi- graft outcomes, compared to colectomy plus
mately 65%. Patients transplanted for PSC restorative surgery via ileal pouch–anal anas-
have good outcomes, with 1‐ and 5‐year tomosis [9].
patient/graft survival of 90/83% and 82/72%, Likewise, an intact colon before LT, and
respectively [1]. the presence of ulcerative colitis (UC) at the
time of post‐LT [8,10], were found to
Recurrence of PSC After LT increase the risk of rPSC. These factors
underscore the interplay between the bowel
Recurrence rate is variable across the differ- and the liver in PSC patients. Other risk
ent studies depending on the criteria used to factors associated with rPSC include young
define recurrence and the median follow‐up. age at LT [10], HLA‐DRB1*08 in the
The 1‐, 5‐, and 10‐year recurrence rates are recipient, male recipient, cytomegalovirus
2–5%, 12–60%, and 20–50% [8], respectively, hepatitis, cholangiocarcinoma in the native
with a median time to recurrence of 4.6 years liver, overlap syndrome, high MELD score
(6 months to 5 years). at the time of LT, and the need for mainte-
nance (>3 months) steroids for UC post‐LT
Diagnosis of Recurrence of (Table 13.2). In terms of immunosuppres-
PSC After LT sion, the use of OKT3 and tacrolimus has
been associated with a higher risk of rPBC.
The diagnosis of recurrent PSC (rPSC) is fre-
Finally, the occurrence of at least one epi-
quently done on imaging, as in the pre‐LT
sode of acute cellular rejection and in
setting. Diagnosis is based on the following.
particular of steroid‐resistant rejection
●● Confirmed diagnosis of PSC before LT. have been associated with rPSC, though the
●● Compatible cholangiographic findings latter scenario suggests that some cases of
identified 90 days or more after LT (intra- ductopenic rejection were probably misdi-
hepatic and/or extrahepatic stricturing, agnosed as rPSC.
beading, or irregularity) or compatible his- About 75–90% of PSC patients have con-
tologic findings (fibrous cholangitis and/or comitant IBD, mostly UC. It is desirable
fibro‐obliterative lesions with or without that their IBD is under complete remission
ductopenia, biliary fibrosis, or biliary at the time of LT because active disease
cirrhosis). increases the risk of graft failure post LT
●● Exclusion of secondary causes, including [11]. In patients with refractory colitis or
ABO mismatch, established ductopenic significant dysplasia, a colectomy can be
rejection, hepatic artery thrombosis/ste- performed at the time of LT with no evi-
nosis, anastomotic strictures alone, and dence of increased postoperative morbidity
non‐anastomotic strictures within the first or mortality when compared to LT‐only
90 days after LT (Table 13.1). patients.
Treatment of PSC Recurrence After LT utcomes of LT improve, rPSC may become

o
a more important issue [10].
As in the pre‐LT setting, there is no effective
Donation after circulatory death (DCD)
treatment for rPSC. UDCA is prescribed in
has helped expand the donor pool to solve
some transplant programs even when there
part of the graft shortage. However, when
is no solid evidence to support this recom-
considering DCD grafts in PSC patients one
mendation. In addition, some experts argue
must be cautious as some data suggest
in favor of the use of UDCA in patients with
inferior outcomes. A retrospective analysis
concomitant UC and rPSC to theoretically
of the UNOS database from 2002 to 2012
diminish the risk of colorectal cancer. Other
found that the use of DCD grafts in patients
than that, therapy is focused on addressing
with PSC was independently associated with
biliary complications when present,
lower graft survival; patients with PSC
managing portal hypertension‐related com-
receiving DCD grafts had a higher rate of bil-
plications if there is progression to cirrhosis,
iary complications and lower survival when
and considering retransplant in a timely
compared with conventional grafts. However,
fashion. Of note, there are case reports of
a more recent study that included the anal-
rPSC where the use of vancomycin led to
ysis of PSC patients undergoing LT from
complete normalization of liver tests, sug-
2006 to 2016 could not corroborate this
gesting more research is needed in this area.
association [13].
Currently there is no effective way to pre-
vent rPSC. Despite the fact that there is a
clear association between IBD and recur-
rence, the role of prophylactic colectomy Autoimmune Hepatitis
needs to be further assessed. A recent Nordic
Multicentre Study found tacrolimus to be Patients with AIH represent approximately
associated with recurrence; if this is con- 4% of those selected for LT [1] and this
firmed, using an alternative drug could pro- number has remained constant across the
vide an effective means to prevent rPSC [12] years. The most common indication is end‐
(Figure 13.1). stage liver disease with portal hypertension‐
derived complications. Approximately 10%
of AIH patients will progress to end‐stage
Prognostic Impact of PSC Recurrence
liver disease and require LT. The median age
After LT
at the time of listing in AIH is younger than
Patients who develop rPSC have decreased the median age at diagnosis, suggesting that
graft, re‐LT free, and patient survival when diagnosis of AIH at an early age increases the
compared to patients with no recurrence risk of needing LT [4].
[8,10]. There seems to be an interaction bet- HCC can also be an indication for LT in
ween age and rPSC because patients younger patients with AIH, though the risk of HCC is
than 40 years at the time of LT are at increased lower than in other liver diseases. Acute liver
risk of graft failure. Graft loss due to rPSC is failure and acute‐on‐chronic liver failure can
variable depending on the follow‐up time but be indications for LT in these patients as well.
roughly 25–50% of the retransplants cur- AIH comprises 15% of the patients that are
rently done in PSC are due to rPSC, and in listed as UNOS status 1, and this number has
long‐term follow‐up studies 30–50% of increased when compared to previous years,
patients who developed rPSC had graft but this may be the result of actually consid-
failure because of the rPSC. Hence, impact of ering AIH as a potential etiology of acute
rPSC on survival seems to be significant not liver failure in recent years. Up to 80% of
in the short term, but in the middle to long these patients will need LT. The role of the
term (>7 years), which suggests that as use of corticosteroids in this setting remains
controversial; if used, failure to improve ●● Elevated aminotransferases.

should prompt urgent assessment and listing ●● High IgG or hypergammaglobulinemia.
for LT. ●● Autoantibodies, which might not be
The outcomes after LT for AIH are accept- specific in the post‐LT setting, but high
able, with 1‐, 5‐, and 10‐year patient survivals titers have been associated with rAIH.
of around 84–94%, 73–86%, and 65–86%, ●● Compatible histopathology (lympho-
respectively; and 1‐, 5‐, and 10‐year graft plasmacytic infiltrate with interface
survivals of 84–91%, 65–85%, and 61–74%, hepatitis), with no endotheliitis or ductu-
respectively [1,14,15]. However, when com- litis. As mentioned earlier, histopatho-
pared to the other etiologies of AILD, survival logic changes may occur well in advance
has been shown to be lower in AIH, with of clinical and/or biochemical activity
graft failure being responsible for a significant and some authors consider there is a
proportion of deaths. In a recent analysis place for protocol liver biopsies in these
from the UNOS database, patients with AIH patients.
had higher retransplantation rates (6.1%) ●● Response to steroids (Table 13.1).
when compared with non‐alcoholic steato-
hepatitis, cryptogenic cirrhosis, or alcohol
cirrhosis. In the setting of acute liver failure, Risk Factors Associated with AIH
Recurrence After LT
they achieve similar survival rates to those
patients with non‐AIH etiologies. Factors The role of corticosteroid discontinuation/
that are implicated in patient/graft survival tapering in precipitating rejection episodes
include MELD score at time of LT, recipient’s is clear. However, their role in rAIH is con-
age, and recurrent AIH (rAIH) [16]. troversial, with most cases showing success-
ful withdrawal of steroids. The role of the
Recurrence of AIH After LT type of calcineurin inhibitor used is also
controversial. Tacrolimus has been sug-
Recurrence rates increase with time after LT, gested as a risk factor in one study, but there
with 1‐ and 5‐year rates of 8–12% and was no multivariable analysis to control for
36–68%, respectively [14,15]. A systematic confounders. A systematic review con-
review calculated a 23% recurrence rate after cluded there was no association between
a median of 26.4 months (14.4–55.2 months). the type of calcineurin inhibitor used and
Notably, patients may have rAIH with normal rAIH, and our group concluded the same
liver tests. In fact, histopathologic changes [14]. In a study of 42 patients of whom 15
can precede biochemical recurrence by as developed rAIH, on multivariate analysis
much as 9 years. the use of antimetabolites was found to
There are no specific criteria to diagnose increase the risk of rAIH, but this finding
rAIH, and the scoring system of the has not been replicated. The number of epi-
International AIH Group has not been vali- sodes of acute cellular rejection do not seem
dated in this setting, but essentially it is diag- to be a risk factor. A recent study did show
nosed taking into account the same an association between acute cellular
parameters. Of note, the performance and/or rejection and rAIH, but this was in the con-
reliability of this score may be suboptimal in text of a steroid‐free regimen of immuno-
the post‐transplant setting due to the effects suppression [17].
of immunosuppression, the need to exclude Moderate‐to‐severe inflammatory activity
other potential etiologies of graft dysfunction, in the explant [14], HLA‐DR3 or HLA‐DR4,
and the persistence of pre‐LT antibodies; and high IgG before transplant [14] have all
therefore, its use is not recommended. In the been associated with a higher risk of rAIH,
absence of validated criteria, rAIH is based whereas patients transplanted for acute liver
on the following characteristics. failure may have a lower risk of rAIH.
Treatment of AIH Recurrence After LT Prognostic Impact of AIH Recurrence

After LT
A timely diagnosis is important in order to
start treatment and modify the natural history Evidence to determine if rAIH affects LT out-
of rAIH. There are no controlled trials in rAIH, comes is discordant, but the balance seems
but in most cases when there is only mild to favor a negative impact on patient and
activity on liver biopsy an increase in the dose graft survival. Some studies show no
or recommencement of corticosteroids and/or difference in survival between patients with
optimization of azathioprine with monitoring and without rAIH [14], but others do show
of metabolite levels is sufficient [18]. In some that 5–18% of patients will develop fibrosis
non‐responders the addition of mycopheno- progression and 6.2–50% may experience
late has been useful. In cases where ciclosporin graft loss due to recurrence and require
is the backbone of immunosuppression, retransplantation.
switching to tacrolimus should be considered.
In difficult‐to‐treat cases where these mea-
sures failed, the use of sirolimus has shown Conclusions
acceptable results. Ultimately some patients
will progress to end‐stage liver disease and LT represents a life‐saving treatment for
require retransplantation. Of note, rAIH in the those patients with AILDs that progress to
retransplant is fairly common. end‐stage liver disease. In the case of PBC,
In the real world, most patients are kept on the rate of LT requirement has decreased
low‐dose prednisone indefinitely to lower the across the years, and this number will prob-
risk of rAIH. Balancing risks and benefits of ably decrease even more with the new
using long‐term steroids has to be done on an therapeutic options that are available (OCA
individual basis. This is a controversial topic, and bezafibrate). In the case of AIH and PSC,
and there is no good level of evidence to the rate of LT has remained constant, and
support this, but cohorts of patients kept on probably will remain the same until we have
steroids show lower recurrence rates when new and effective treatments, particularly for
compared with historic cohorts [15] PSC. Recurrence of AILD is seen in around
(Figure 13.1). Interestingly, a small case series 20–30% of LT and represent a unique oppor-
of AIH cases that underwent splenectomy at tunity to study the pathogenesis of these
the time of LT showed no evidence of recur- diseases in depth. As LT patient and graft
rence on follow‐up, suggesting a protective survival improves, it is possible that recur-
role of this strategy. Unfortunately, two patients rent AILD will become more prevalent and
died because of complications related to sple- more important in terms of adversely com-
nectomy, raising questions about the safety of promising graft function, and therefore more
this potential prophylactic measure. Finally, research is needed in this area in terms of
the role of protocol biopsies to diagnose silent potential preventive strategies and effective
rAIH in a timely fashion needs to be addressed. treatments.
References
1 Adam, R., Karam, V., Delvart, V. et al. 8 Hildebrand, T., Pannicke, N., Dechene, A.
(2012). Evolution of indications and et al. (2016). Biliary strictures and
results of liver transplantation in Europe. recurrence after liver transplantation for
A report from the European Liver primary sclerosing cholangitis: a
Transplant Registry (ELTR). J. Hepatol. retrospective multicenter analysis. Liver
57 (3): 675–688. Transpl. 22 (1): 42–52.
2 Montano‐Loza, A.J., Bhanji, R.A., 9 Trivedi, P.J., Reece, J., Laing, R.W. et al.
Wasilenko, S., and Mason, A.L. (2017). (2018). The impact of ileal pouch–anal
Systematic review: recurrent autoimmune anastomosis on graft survival following
liver diseases after liver transplantation. liver transplantation for primary sclerosing
Aliment. Pharmacol. Ther. 45 (4): 485–500. cholangitis. Aliment. Pharmacol. Ther. 48
3 Singal, A.K., Guturu, P., Hmoud, B. et al. (3): 322–332.
(2013). Evolving frequency and outcomes of 10 Ravikumar, R., Tsochatzis, E., Jose, S. et al.
liver transplantation based on etiology of liver (2015). Risk factors for recurrent primary
disease. Transplantation 95 (5): 755–760. sclerosing cholangitis after liver transplant
4 Webb, G.J., Rana, A., Hodson, J. et al. (2018). ation. J. Hepatol. 63 (5): 1139–1146.
Twenty‐year comparative analysis of 11 Joshi, D., Bjarnason, I., Belgaumkar, A.
patients with autoimmune liver diseases on et al. (2013). The impact of inflammatory
transplant waitlists. Clin. Gastroenterol. bowel disease post‐liver transplantation for
Hepatol. 16 (2): 278–287.e7. primary sclerosing cholangitis. Liver Int. 33
5 Montano‐Loza, A.J., Wasilenko, S., Bintner, (1): 53–61.
J., and Mason, A.L. (2010). Cyclosporine A 12 Lindstrom, L., Jorgensen, K.K., Boberg,
protects against primary biliary cirrhosis K.M. et al. (2018). Risk factors and
recurrence after liver transplantation. Am. J. prognosis for recurrent primary sclerosing
Transplant. 10 (4): 852–858. cholangitis after liver transplantation: a
6 Bosch, A., Dumortier, J., Maucort‐Boulch, Nordic multicentre study. Scand. J.
D. et al. (2015). Preventive administration of Gastroenterol. 53: 297–304.
UDCA after liver transplantation for 13 Trivedi, P.J., Scalera, I., Slaney, E. et al.
primary biliary cirrhosis is associated with a (2017). Clinical outcomes of donation after
lower risk of disease recurrence. J. Hepatol. circulatory death liver transplantation in
63 (6): 1449–1458. primary sclerosing cholangitis. J. Hepatol.
7 Montano‐Loza, A.J., Hansen, B.E., 67 (5): 957–965.
Corpechot, C. et al. (2019). Factors 14 Montano‐Loza, A.J., Mason, A.L., Ma, M.
associated with recurrence of primary et al. (2009). Risk factors for recurrence of
biliary cholangitis after liver transplantation autoimmune hepatitis after liver
and effects on graft and patient survival. transplantation. Liver Transpl. 15 (10):
Gastroenterology 156: 96–107.e1. 1254–1261.
15 Krishnamoorthy, T.L., Miezynska‐Kurtycz, J., 17 Satapathy, S.K., Jones, O.D., Vanatta, J.M.
Hodson, J. et al. (2016). Longterm et al. (2017). Outcomes of liver transplant
corticosteroid use after liver transplantation recipients with autoimmune liver disease
for autoimmune hepatitis is safe and using long‐term dual immunosuppression
associated with a lower incidence of recurrent regimen without corticosteroid.
disease. Liver Transpl. 22 (1): 34–41. Transplant. Direct 3 (7): e178.
6 Schramm, C., Bubenheim, M., Adam, R.
1 8 Stirnimann, G., Ebadi, M., Czaja, A.J., and
1
et al. (2010). Primary liver transplantation Montano‐Loza, A.J. (2019). Recurrent and
for autoimmune hepatitis: a comparative de novo autoimmune hepatitis. Liver
analysis of the European Liver Transplant Transpl. 25: 152–166.
Registry. Liver Transpl. 16 (4): 461–469.
247
14
Recurrent Autoimmune Liver Disease and its Scientific

Significance
Atsushi Tanaka1, Patrick S.C. Leung1, and M. Eric Gershwin2
1
Department of Medicine, Teikyo University School of Medicine, Tokyo, Japan
2
Division of Rheumatology, Allergy and Clinical Immunology, University of California School of Medicine, Davis, CA, USA
Abstract
The three most common autoimmune liver diseases (ALDs) are autoimmune hepatitis (AIH), pri
mary biliary cholangitis (PBC), and primary sclerosing cholangitis (PSC). This chapter discusses
the incidence, presentation, diagnosis, risk factors, and impact on outcomes of recurrence of PBC,
PSC and AIH after liver transplantation (LT). There is an urgent need to prevent recurrence of
ALD in the graft, especially in PSC and PBC, where recurrence is frequently linked to graft failure
and the necessity of retransplantation. While ALD patients are largely diagnosed when the disease
has progressed, those at the very early stages of disease onset can be followed longitudinally for an
extended period with protocol biopsy and blood sampling before and after LT to access disease
recurrence and perhaps identify biomarkers. The enigma of recurrence is an important immuno
logic issue, not only in ALD but also in other solid organs.
Keywords autoimmune hepatitis; blood sampling; graft failure; immunologic issue; liver
transplantation; primary biliary cholangitis; primary sclerosing cholangitis; protocol biopsy
Key Points
●● Recurrent autoimmune liver disease can ●● Understanding those factors that affect
be a significant cause of graft loss. recurrence have important implications
●● Risk factors for, and treatment of, recur for understanding liver autoimmunity;
rence is not well defined. these include the role of innate immunity,
●● Prospective studies are needed to better target organ expression of host antigens,
define risk factors, biomarkers and bystander mechanisms, and genetic simi
treatment. larities between donor and host.
Introduction natural course. The three most common

ALDs are autoimmune hepatitis (AIH), pri
Autoimmune liver disease (ALD) represents mary biliary cholangitis (PBC), and primary
about 5% of all liver diseases. ALDs are sclerosing cholangitis (PSC). Although
asymptotic chronic diseases with a long knowledge on the natural history of these
ALDs has been progressing, with a better resentation, diagnosis, risk factors, and
p
understanding of etiopathology [1–5] and impact on outcomes of recurrence of PBC,
long‐term outcomes with medical treatment PSC and AIH after LT.
remarkably improved [6], liver transplanta
tion (LT) remains the only and most effective
treatment option for patients with end‐stage Recurrence of PBC
ALDs. According to the registry data from
the USA, Europe and Japan, the 10‐year PBC is a chronic cholestatic liver disease,
survival rates after LT are similar (71–79%) mainly affecting small‐ to medium‐sized
in PBC, whereas those in PSC varied from intrahepatic bile ducts. Middle‐aged women
83% in the USA to 70% in Europe and 57% in are at the highest risk for developing PBC.
Japan [7–9] (Table 14.1). With regard to AIH, The diagnosis of PBC is made by meeting
the registries contain no available data except two of the following three criteria: (i) chronic
for the European registry, which shows elevation of cholestatic liver enzymes, i.e.
5‐ and 10‐year survival rates of 76% and 67%, alkaline phosphatase (ALP) and gamma‐
respectively [7] (Table 14.1). glutamyltransferase (GGT); (ii) detectable
Despite the high survival rate of ALD anti‐mitochondrial antibodies (AMAs); and
patients with LT, disease recurrence also (iii) characteristic histologic findings of
occurs and negatively affects graft as well as chronic non‐suppurative destructive cholan
overall survival. In particular, recurrence of gitis (CNSDC) and granulomas [13]. PBC is a
PSC develops in at least 25% of patients and slowly progressive disease, which can result
is definitely associated with poor outcome in cirrhosis and liver failure without appro
[10]. Although recurrence of PBC was priate treatment. Although ursodeoxycholic
believed to have little impact on graft failure acid (UDCA) improves LT‐free survival and
and prognosis [11], a recent international is recommended as first‐line treatment for
study with a large cohort demonstrated that PBC by clinical practice guidelines or
graft and patient survival are significantly guidance [13–15], about one‐third of patients
impaired by recurrence of PBC [12]. In this are refractory to UDCA and could progress
chapter, we discuss the incidence, to end‐stage liver disease requiring LT. In
Table 14.1 Patient and graft survival at 5 and 10 years after LT in PBC, PSC and AIHa.
Patient survival Graft survival
Region N 5 years 10 years 5 years 10 years
PBC Europe 4515 80 71 75 66

USA 3052 84 79 78 72
Japanb 710 79 74 NA NA
PSC Europe 3582 78 70 69 57
USA 3854 87 83 81 78
Japanb 232 73 57 NA NA
AIH Europe 1892 76 67 69 59
b
Japan 104 79 75 NA NA
a
Registry data from Europe [7], USA [8] and Japan [9].
b
Living‐related LT in all.
AIH, autoimmune hepatitis; LT, liver transplantation; PBC, primary biliary cholangitis; PSC, primary sclerosing
cholangitis.
Chapter 14 Recurrent Autoimmune Liver Disease and its Scientific Significance 249
2016, obeticholic acid (OCA) was officially Incidence and Diagnosis of

approved as a second‐line drug but long‐ Recurrent PBC
term efficacy has not been demonstrated On the other hand, recurrence of PBC after
[16]. Nevertheless, a recent study employing LT is not uncommon. The incidence of recur
the European Liver Transplant Registry rent PBC is reported to differ widely, bet
(ELTR) demonstrated a significant decrease ween 11% and 42% (Table 14.2) [12, 19–33].
in LT due to PBC over the last 30 years, after This difference is due to a combination of
the introduction of UDCA in clinical settings various factors, such as variation in sample
[17]. The proportion of LT performed for size, follow‐up period and, in particular, dif
PBC decreased from 20% of all LT cases in ficulty in the diagnosis of recurrent PBC.
1986 to 4% in 2015 (p <0.001). The absolute Indeed, diagnosis of recurrent PBC is
number of transplants was highest in 1994 extremely challenging. AMA, a serologic
(n = 279) which decreased to an average of hallmark of PBC, remains readily detectable
200 in the last decade. This decrease is strik after LT and is therefore not a reliable test for
ingly in contrast to the substantial increase in recurrent PBC. In addition, cholestatic liver
prevalence of PBC at the same time [18]. enzymes can be elevated due to multiple
Overall, the long‐term outcome after LT for other reasons. Therefore, liver histology
PBC is remarkable (Table 14.1) [7–9]. findings are vital for the diagnosis of disease
Table 14.2 Incidence and time to recurrence of PBC after LT.
Center sites Time period Year N Incidence Time to recurrence (years)a
Birmingham, UK [28] 1982–1999 2001 400 68 (17%) 3.0 (0.3–11.7)

Rochester, USA [33] 1985–1997 2003 100 17 (17%) 3.7 (0.3–7.9)
Dallas, USA [32] 1985–1999 2003 156 17 (10.9%) 4.1 (1.1–7.1)
Boston, USA [26] 1983–2001 2003 43 8 (18.6%) 6.7 (0.1–14)
Birmingham, UK [31] 1982–2002 2004 485 114 (23%) NAb
St. Louis, USA [23] 1985–1997 2005 48 17 (35.4%) 3.3 ± 2.8
Berlin, Germany [25] 1989–2003 2006 100 14 (14%) 5.1 (3.0–10.2)
Rochester, USA [21] 1985–2005 2007 154 52 (34%) 3.5 (0.3–18.1)
New York, USA [24] 1989–1999 2009 44 7 (15.9%) 2.8 (1.1–8.8)
London, UK [29] 1988–2008 2010 103 36 (35%) 3.7 (0.8–16.7)
Alberta, Canada [30] 1989–2008 2010 108 28 (26%) 5.8 (0.5–15.6)
Cambridge, UK [20] NA 2013 248 105 (42.3%) 5.1 (1.4–8.6)
Alberta, Canada [19] 1989–2010 2013 103 26 (25%) 15.3 ± 1.0
France and Switzerland [37] 1988–2010 2015 123 48 (53%) 6.4 ± 5.0
Japanc [22] 1994–2010 2016 444 65 (14.6%) 5.1
Toronto, Canadad [60] 2000–2015 2016 69 9 (13%) NA
Japan [27] 1994–2010 2017 388e 58 (14.9%) NA
North America and Europe [12] 1983–2016 2019 785 240 (31%) 4.4 [3.4–5.1]
a
Time to recurrence shown as median (range), median [interquartile range], or mean ± SD.
b
NA, not available.
c
All living donor liver transplantation (LDLT).
d
30 deceased donor liver transplantation, 39 LDLT.
e
Female patients only.
recurrence in PBC. Recurrent PBC is studies from Japan demonstrated that the
diagnosed with PBC‐specific histology, such increased frequency of recurrent PBC is
as florid duct lesions and lymphoplasmacytic associated with initial treatment with ciclo
infiltrates. However, other causes of bile duct sporin after LT [22, 27]. However, these two
injury may also mimic the typical findings of studies were performed by the same study
PBC, including ischemia–reperfusion injury, group, and subjects in these two studies
acute cellular rejection, humoral/chronic partially overlapped. Among 516 patients
rejection, drug‐induced liver injury, vascular who underwent LT for end‐stage PBC from
complications, and graft‐versus‐host disease. 1994 to 2010, living donor liver transplant
In addition, recurrence of PBC may develop (LDLT) and deceased donor liver transplant
few or no symptoms and elevation of (DDLT) of both genders were included in the
cholestatic enzymes may be absent. There first study [22]. However, only LDLT and
fore, protocol liver biopsy rather than event‐ only females were included, and patients who
driven biopsy is required for detection of did not survive for 1 year were excluded, in
recurrent PBC. It may have a huge impact on the second study [27]. The first study also
the difference in incidence of recurrent PBC demonstrated that the switch from tacroli
whether protocol biopsy is scheduled or mus to ciclosporin treatment decreased the
event‐driven biopsy is performed. frequency of recurrence. This finding sug
gests that the timing of selection or change of
calcineurin inhibitor and the way immuno
Risk Factors of Recurrent PBC
suppressive agents are used in clinical prac
Although several studies have reported risk tice may partly account for the difference in
factors associated with recurrence of PBC, the association with recurrence between the
most of them have consistently demonstrated East and the West. On the other hand, it is
that use of tacrolimus is definitely associated well known that genetics plays a crucial role
with an increased risk of recurrence [20, 21, in susceptibility of PBC [36], and this
28, 30–32]. For example, a recent study on difference in role of calcineurin inhibitors
785 PBC patients from North America and suggests that genetics may also contribute to
Europe who underwent LT from February recurrence of PBC after LT, as indicated by
1983 to June 2016 indicated that tacrolimus another previous study suggesting the
was linked to recurrence of PBC. While use of influence of the IL12A locus [20].
ciclosporin was protective, the 5‐year proba
bility of PBC recurrence was reported to be
Impact of Recurrent PBC on Long‐term
28% and 11% in patients receiving tacrolimus
Outcomes
and ciclosporin, respectively (p <0.001) [12].
The reasons for this phenomenon still remain Nevertheless, it is believed that recurrence
unclear. Since infection with bacteria or of PBC does not have a significant impact
viruses has been suggested to be associated on long‐term outcomes, such as overall
with tolerance breakdown against autoanti survival. For instance, a retrospective study
gens [34], the difference in PBC recurrence in the UK on 400 PBC patients who received
rates with these two immunosuppressive LT between 1983 and 1999 demonstrated
drugs is likely due to the more potent immu confirmation of recurrent PBC in 68 (17%)
nosuppressive effect of tacrolimus, which subjects but no effect on patient and graft
might facilitate alloreactivity against graft survival during, on average, 56 months of
liver through infection [35]. follow‐up [28]. Recently, a French–Swiss
On the other hand, the role of tacrolimus collaborative study was conducted on 123
as an agent that causes increasing rates of patients with PBC who underwent LT. Their
PBC recurrence does not seem to be the case data showed that recurrence of PBC was
in other ethnicities. Recently, two cohort observed in 48 subjects (53%) with an
average of 11.7 years of follow‐up; neither features differentiate PSC from PBC [10, 38].
recurrence nor preventive UDCA had a Unlike PBC, PSC is more common in males
significant impact on survival [37]. The two than in females; adolescents and young adults
Japanese studies described previously also are at the highest risk for developing PSC.
suggested that there is no impact of recur Disease‐specific biomarkers such as autoanti
rent PBC on survival [22, 27]. Surprisingly, bodies are not available in PSC (compare
another recent study on 785 PBC patients AMA in PBC). Although onion‐skin periduc
from 13 centers in North America and tal fibrosis is known to be a characteristic his
Europe who received LT with a median tologic finding in PSC, this feature is not
follow‐up of 6.9 years (interquartile range frequently observed in liver specimens, mak
6.1–7.9) reported the opposite and unex ing liver biopsy less valuable for diagnosis of
pected result that disease recurrence was PSC. As a result, the diagnosis of PSC is ulti
found in 240 patients (31%) and, impor mately dependent on cholangiography by
tantly, both graft and patient survival were endoscopic retrograde cholangiopancrea
significantly impaired in those with recur tography (ERCP) or magnetic resonance
rent PBC (p = 0.004 and 0.001, respectively) cholangiopancreatography (MRCP), which
[12]. Although the manner of biopsy (pro sometimes can be challenging. Indeed, the
tocol or event‐driven) depended on the natural course of PSC varies greatly depend
center, the median time for recurrence of ing on cases, a number of prognostic indices
PBC was similar in this study, which is by far have been proposed [10, 38], and no con
the largest cohort with long‐term follow‐up sensus on surrogate end points (ALP, bili
on PBC recurrence after LT. It is necessary rubin, histology, non‐invasive fibrosis
to remember that there might be a number markers, etc.) has been uniformly recom
of variables among centers, such as immu mended [39]. Additionally, cholangiocarci
nosuppressant protocols, diagnosis of noma and inflammatory bowel disease (IBD)
recurrent PBC, and the use of UDCA as are frequently observed as concomitant dis
prevention and/or treatment for PBC, and eases in patients with PSC, making definition
these could affect the conclusion. of surrogate end points more complicated.
Nevertheless, it is imperative to determine Regarding treatment, UDCA is probably used
whether recurrent PBC really has a signi as the first‐line drug in most cases with PSC
ficant impact on patient and graft survival. and indeed the efficacy of UDCA has been
Also, since UDCA treatment seems to be extensively studied, but no single drug
effective in improving markers of chole including UDCA has been proven to prolong
stasis, it will not effectively reduce the fre LT‐free survival of patients. It may be simply
quency and risk of recurrence of PBC after because of the ineffectiveness of UDCA, but
LT [37]; therefore, development of effective may possibly be due to lack of properly
strategies to halt recurrence of PBC is an designed clinical studies.
urgent need. In clinical settings when no medical
treatment is effective in patients with end‐
stage PSC, LT becomes the single treatment
Recurrence of PSC option. According to the registry data, the
long‐term outcomes of LT for patients with
PSC is a chronic cholestatic liver disease, PSC in Europe and the USA are comparable
mainly affecting large‐sized intrahepatic and to those in PBC; the respective 5‐ and 10‐
extrahepatic bile ducts. Although PBC and year patient survival rates are 78% and 70%
PSC share several clinical characteristics as in Europe, and 87% and 83% in the USA
cholestatic liver diseases, including elevation [7, 8]. The 5‐year survival in Europe has
of ALP and GGT and possibility of progres improved to 82% in recent years (1999–
sion to biliary cirrhosis, a number of distinct 2009) [7]. However, there is a striking
difference in the survival rates of Japanese failure thereafter. The percentage of LDLT
LT patients, with 5‐ and 10‐year patient was 100% in the Japanese registry [9], and
survival rates of 73% and 57%, respectively 87% (39/45) in a single‐center report in
[9]. Although graft survival rate was not Japan [40].
reported in this registry data, a recent
single‐center study in Japan reported that
Incidence and Diagnosis of Recurrent PSC
the 5‐year graft survival rate was 55.4% [40],
substantially lower than that in Europe Recurrence of PSC after LT is more fre
(69%) and the USA (81%). This marked quently observed compared to PBC. As
difference is partly explained by the fre shown in Table 14.3, PSC recurs in 8.6–44.7%
quent recurrence of PSC in LDLT and graft of patients after LT within a median of almost
Table 14.3 Incidence and risk factors of recurrence of PSC after LT.
Los Angeles, USA [41] 1984–1996 1997 127 11 (8.6%) NAb

Dallas, USA [42] 1985–1995 1998 100 18 (18%) 1.8 ± 0.6
Rochester, USA [43] 1985–1996 1999 120 24 (20%) 0.7 (0.3–3.5)
Gothenburg, Sweden [44] 1985–1998 2001 61 5 (8.2%) NA
San Francisco, USA [45] 1988–1997 2002 49 7 (14%) NA
Boston, USA [46] 1983–2000 2003 42 6 (14%) NA
Denver, USA [47] 1988–2000 2003 71 15 (21%) 4.4 (1.0–9.2)
Oslo, Norway [48] 1984–2003 2005 49 9 (18%) NA
c
Japan [49] 1993–2003 2007 44 11 (25%) NA
Seattle, USA [50] 1990–2003 2008 69 7 (10%) 5.7 (2.0–11.2)
Denver, USA [51] 1988–2006 2008 130 22 (16.9%) NA
London, UK [52] 1989–2004 2008 53 7 (13.2%) 5.0 (0.3–10.0)
Birmingham, UK [53] 1986–2006 2009 230 54 (23.5%) 4.6 (0.5–12.9)
New York, USA [54] 1995–2007 2010 58** 11 (19%) NA
Alberta, Canada [55] 1989–2006 2010 59 15 (25%) 3.4 (1.6–5.5)
Japand [56] 1996–2008 2011 96 26 (27%) 2.4 (0.7–6.6)
UK [57] 1990–2010 2015 565 81 (14%) NA
Germany [58] 1990–2006 2016 305 62 (20.3%) 4.6 (0.5–14.3)
North Americae [59] 1998–2013 2016 307 34 (11%) NA
Toronto, Canadaf [60] 2000–2015 2016 138 32 (23%) NA
Kyoto, Japan [40] 1996–2015 2017 40 16 (40%) 2.5 (0.8–6.3)
Nordic countries [61] 1984–2007 2018 440 85 (19%) NA
Prague, Czech Republic [62] 1994–2015 2018 47 21 (44.7%) 5.3 (1.0–15.0)
a
Time to recurrence shown as median (range) or mean ± SD.
b
NA, not available.
c
44 deceased donor liver transplantation (DDLT), 14 living donor liver transplantation (LDLT).
d
All LDLT.
e
65 DDLT, 242 LDLT.
f
70 DDLT, 68 LDLT.
1–5 years. It should be noted that lower for several reasons: sample size, extent and
recurrence rates were reported in the earlier depth of the collected dataset, difference in
studies with relatively shorter follow‐up patient populations, and ethnicities [65].
times [41, 44]. In contrast, the highest recur In Japan, where LDLT is dominant, the
rence rate, 44.7% (21/47), was recently recurrence rate of PSC after LT is relatively
reported from a single center, where the higher compared with other regions
follow‐up period of the patients after LT was (Table 14.3). Egawa et al. [56] identified first‐
a median of 10.2 (5.0–20.8) years, the longest degree relative donor as a significant risk
among all cohorts [62]. factor for recurrence. However, this is not the
The diagnosis of PSC is challenging. As case in North America. In a large‐scale North
described previously, there are no PSC‐ American cohort of patients with PSC
specific biomarkers that are available for (n = 307), it was observed that PSC recur
diagnosis and cholangiography by ERCP or rence rate was not significantly different for
MRCP is the most important tool for pre‐ LDLT versus DDLT recipients (p = 0.36). In
transplant diagnosis. After LT, imaging this study, the risk factors identified were
studies of the biliary tree may reflect a high MELD score, biliary complication, chol
variety of physiologic and pathologic condi angiocarcinoma, and high donor age in both
tions that mimic PSC, including organ pres LDLT and DDLT, while first‐degree relative
ervation, ischemia–reperfusion injury, donor was not significantly associated with
hepatic artery stenosis or thrombosis, recurrence [59]. Another recent cohort study
chronic ductopenia rejection, bile duct anas from Canada on 36 PSC patients who
tomosis, ABO incompatibility, and drug‐ received a graft from a first‐degree living‐
induced liver injury. For diagnosis of related donor, 32 PSC patients who received
recurrent PSC, the criteria proposed by a graft from a living but distinct/unrelated
Graziadei et al. [63] should be strictly used. donor, and 70 PSC patients who received a
Accordingly, (i) the diagnosis of PSC must graft from a deceased non‐related donor
be present before LT; (ii) cholangiographic indicated that first‐degree living‐related
findings (biliary stricturing, beading, and donor did not confer a significant increase in
irregularity lasting more than 90 days) and/ recurrent risk of PSC after LT [60].
or histologic features (fibrous cholangitis The impact of IBD on recurrent PSC is
and/or fibro‐obliterative lesions with or another important issue. Active IBD after
without ductopenia, biliary fibrosis, or bil LT is also associated with increasing recur
iary cirrhosis) are required for diagnosis of rence rate of PSC [40, 53, 57, 58, 66]. A
recurrent PSC; and (iii) other conditions recent study in the Czech Republic reported
mimicking PSC need to be excluded [63]. that de novo colitis was present in 25.5%
(12/47) of PSC patients between 17 and
87 months (median 45 months) post LT and
Risk Factors of Recurrent PSC
a significant number of them (11/12) were
A number of risk factors for recurrent PSC also confirmed with recurrent PSC at a
have been considered, including recipient’s age, median of 72 months post LT. [62]. These
gender, history of cholangiocarcinoma, Model observations surely underline the impor
for End‐stage Liver Disease (MELD) score, and tance of the gut–liver axis in PSC patho
status of IBD; and donor and recipient’s age, physiology, and also point to the
gender/cytomegalovirus mismatch and first‐ recommendation of preventable colectomy
degree relative donors [64, 65]. Acute cellular prior to LT [53, 61]. However, another
rejection (steroid‐resistant) is reported to be recent large‐scale cohort study did not show
another risk for recurrent PSC. It is strongly a preventable effect of colectomy for recur
noted, however, that there is great inconsis rence of PSC [57]. Further prospective
tency in the results of these studies, probably studies are warranted to determine whether
pre‐LT colectomy is significantly associated AIH, patients in childhood or adolescence

with a decreased risk of recurrent PSC. are not uncommon. Elevation of transaminases,
detectable autoantibodies such as anti‐nuclear
antibody (ANA) or anti‐smooth muscle anti
Impact of Recurrent PSC on Long‐term
body (ASMA), elevation of serum IgG levels,
Outcomes
and interface hepatitis or plasma cell infiltra
Unlike PBC, it is well established that recur tion in liver histology are characteristic fea
rent PSC has a huge impact on mortality and tures of AIH [69]. However, there is no
morbidity, and often requires retransplanta identified biomarker specific for the diagnosis
tion. In a UK study with a large cohort con of AIH and therefore criteria consisting of sev
sisting of 565 patients, the 5‐ and 10‐year eral items are used for diagnosis [70, 71]. AIH
graft survival rate was 88% and 82% in responds very well to immunosuppressive
patients without recurrent PSC, and 84% and drugs, and corticosteroids and/or azathioprine
56% in those with recurrent PSC, respec are the first‐line treatment [2, 72]. While
tively. The presence of recurrent PSC overall survival of patients with AIH is
increased the risk of both graft failure [hazard comparable to those in the general population
ratio (HR) 8.15, 95% confidence interval when treatment response is favorable, the out
(CI) 5.59–11.89] and of graft failure or death come is poor in patients with more than two
(HR 4.71, 95% CI 3.39–6.56) [57]. Another relapses even with corticosteroid treatment
large‐scale cohort involving 335 patients during clinical course [73].
with IBD who underwent LT for PSC in Patients with end‐stage liver disease due
Germany also indicated significantly poorer to AIH require LT. In addition, AIH pre
graft and recipient overall survival, with senting as acute hepatitis could lead to acute
91.4% versus 79% (p = 0.02) in recipient or acute‐on‐chronic liver failure, also neces
survival and 85.1% versus 61.3% (p <0.001) in sitating an emergent LT. Although data
graft survival [58]. As a result, a substantial regarding LT in patients with AIH com
portion of PSC patients after LT are awaiting pared to PBC or PSC are scarce, the overall
their second graft. survival rate seems to be excellent in AIH;
Currently, there are no medical treatments the 5‐ and 10‐year recipient survival rates
for PSC patients to prevent the development of are 76–79% and 67–75%, which is better
recurrent PSC after LT. Although UDCA is fre than for most other indications for LT [7–9]
quently used after LT, its efficacy in preventing (Table 14.1).
disease recurrence is unclear. Clearly, the
development of new drugs for preventing
Incidence and Diagnosis of Recurrent AIH
recurrence of PSC is an unmet need. In this
regard, it is notable that a retrospective study of Immunosuppressive drugs that are very
induction with rituximab for ABO‐incompat effective for treatment of AIH are inevitably
ible LDLT with 12 subjects demonstrated no administered after LT. Nevertheless, AIH can
recurrence of PSC, with excellent graft function develop in the graft after LT. It is very chal
in five recipients for more than 7 years [67]. lenging to determine the incidence of recur
rent AIH, but it is reported to be in the range
of 7–42% [74–84] (Table 14.4). The inconsis
Recurrence of AIH tency among studies is more prominent in
AIH compared to PBC and PSC, most likely
AIH is a chronic liver disease of unknown due to differences in diagnostic criteria, his
etiology in which autoimmune‐mediated tologic analysis (protocol or event‐driven
reactivities against hepatocytes presumably biopsy), small sample size in each study (no
play a crucial role [68]. While middle‐aged study with more than 100 patients enrolled),
women have the highest risk for developing and follow‐up time [64, 65]. The rate of
Table 14.4 Incidence and risk factors of recurrence of AIH after LT.
Spain [82] 1988–1996 1998 27 9 (33%) 2.6 ± 1.5

Birmingham, UK [79] NAb 1999 47 13 (28%) 2.4 (0.5–5.3)
Paris, France [83] 1985–1992 1999 15 3 (20%) 1.6 (1–2.5)
New York, USA [84] 1988–1995 2000 24 6 (25%) 1.3 ± 0.2
Boston, USA [74] 1983–1998 2000 12 5 (42%) NA
Rochester, USA [77] 1985–1998 2001 41 7 (17%) 4.6 ± 1
Dallas, USA [80] 1984–1998 2002 55 11 (20%) NA
Paris, France [76] 1985–1992 2003 17 7 (41%) 2.5 ± 1.7
Colorado, USA [75] 1988–2006 2008 66 23 (34.8%) 4.3
Alberta, Canada [81] NA 2009 46 11 (24%) 4 ± 1.3
Birmingham, UK [78] 1999–2014 2016 69 5 (7%) 3.8 (1.5–7.3)
a
Time to recurrence shown as median (range) or mean ± SD.
b
NA, not available.
recurrence increases as follow‐up time factor for recurrence in one study [86] but
increases after LT [75, 81, 82]. not in others [79–81, 84]. Withdrawal of cor
In the diagnosis of AIH, the revised AIH ticosteroids has been associated with higher
criteria [70] and the simplified score [71] risk of recurrent AIH [79, 82]. A recent study
have been proposed by the International AIH from the UK demonstrated that the 5‐ and
Group. The former is established for identi 10‐year recurrence rate after LT was 6% and
fying “pure” AIH for clinical studies but 11%, respectively, in a cohort consisting of 69
seems to be too complicated for daily patients with AIH, of which 87% were
practice, while the latter consists of only five receiving long‐term maintenance treatment
criteria and is more suitable in clinical set with corticosteroids after LT [78]. Compared
tings. However, both are not validated for the with the recurrence rate of 27% in their
diagnosis of AIH developed in the graft. previous report in 1999, when patients did
Meanwhile, it is reasonable to assume that not receive long‐term corticosteroid therapy
the criteria for a diagnosis of recurrent AIH [79], the authors concluded that long‐term
are the same as for a diagnosis of AIH before corticosteroid use in combination with
LT: elevation of transaminases, detection of immunosuppressive agents was associated
autoantibodies (ANA, ASMA), elevation of with a lower frequency of recurrence.
IgG, and histologic findings including inter
face hepatitis or perivenular lymphoplasma Impact of Recurrent AIH on Long‐
cytic infiltrates. A comprehensive analysis of term Outcomes
the clinical phenotypes is necessary in order In general, progression to cirrhosis and graft
to diagnose disease recurrence post LT [85]. failure requiring retransplantation is
uncommon even when AIH recurs in the graft
[64]. However, the mechanisms causing recur
Risk Factors of Recurrent AIH
rent AIH after LT are unclear. Furthermore,
A number of factors are reported to be asso there are also substantial differences between
ciated with recurrence of AIH, including adult and pediatric de novo AIH, which sub
severity of pre‐transplant AIH [74, 81]. HLA stantiates the need for more precise diagnostic
locus mismatching is identified as a risk guidelines in this area [87, 88]. When
r ecurrence develops in the graft, the strength Furthermore, these prospective cohorts
of immunosuppression should be reinforced would be valuable resources for further under
with readministration or increased dosing of standing the etiologic mechanisms of disease.
corticosteroids or addition of other immuno While ALD patients are largely diagnosed
suppressive agents. when the disease has progressed, those at the
very early stages of disease onset can be fol
lowed longitudinally for an extended period
Concluding Remarks with protocol (or even event‐driven) biopsy
and blood sampling before and after LT to
The shortage of donor livers is a major issue access disease recurrence and perhaps iden
for patients awaiting LT worldwide [89]. tify biomarkers. Again, the establishment of
Thus, there is an urgent need to prevent well‐designed and organized prospective reg
recurrence of ALD in the graft, especially in istries with clinical information as well as
PSC and PBC, where recurrence is frequently clinical samples are critical not only to
linked to graft failure and the necessity of improve the outcomes of post‐transplant
retransplantation. Furthermore, risk factors patients, but also to unveil the etiologic myths
associated with an increased frequency of of ALD. The enigma of recurrence is an
ALD recurrence are not clearly defined and important immunologic issue, not only in
robust approaches for the prevention of ALD ALD but also in other solid organs. Factors to
recurrence are not available. Large‐scale and be considered include innate immunity, target
multicenter studies with prospective cohorts organ expression of host antigens, bystander
of patients after LT are needed to draw con mechanisms, and genetic similarities between
clusive results, under the auspices of interna donor and host. Further studies are required
tional collaborative consortia. and the mechanisms may vary by disease.
References
1 Terziroli Beretta‐Piccoli, B., Stirnimann, 9 Japanese Liver Transplantation Society

G., Cerny, A. et al. (2018). (2015). Liver transplantation in Japan.
Geoepidemiology of primary biliary Registry by the Japanese Liver
cholangitis: lessons from Switzerland. Clin. Transplantation Society. [In Japanese].
Rev. Allergy Immunol. 54: 295–306. Ishoku 52: 134–147.
2 Wang, Z., Sheng, L., Yang, Y. et al. (2017). 10 Dyson, J.K., Beuers, U., Jones, D.E.J. et al.
The management of autoimmune hepatitis (2018). Primary sclerosing cholangitis.
patients with decompensated cirrhosis: Lancet 391: 2547–2559.
real‐world experience and a comprehensive 11 Levy, C. (2018). Primary biliary cholangitis
review. Clin. Rev. Allergy Immunol. 52: guidance update: implications for liver
424–435. transplantation. Liver Transpl. 24:
3 Gerussi, A., Cristoferi, L., Carbone, M. 1508–1511.
et al. (2018). The immunobiology of female 12 Montano‐Loza, A.J., Hansen, B.E.,
predominance in primary biliary Corpechot, C. et al. (2019). Factors
cholangitis. J. Autoimmun. 95: 124–132. associated with recurrence of primary
4 Tabibian, J.H. and Bowlus, C.L. (2017). biliary cholangitis after liver transplantation
Primary sclerosing cholangitis: a review and effects on graft and patient survival.
and update. Liver Res. 1: 221–230. Gastroenterology 156: 96–107.e1.
5 Takakura, W.R., Tabibian, J.H., and Bowlus, 13 Lindor, K.D., Bowlus, C.L., Boyer, J. et al.
C.L. (2017). The evolution of natural (2019). Primary biliary cholangitis: 2018
history of primary sclerosing cholangitis. practice guidance from the American
Curr. Opin. Gastroenterol. 33: 71–77. Association for the Study of Liver Diseases.
6 Cristoferi, L., Nardi, A., Ronca, V. et al. Hepatology 69: 394–419.
(2018). Prognostic models in primary biliary 14 (2014). Guidelines for the management of
cholangitis. J. Autoimmun. 95: 171–178. primary biliary cirrhosis: The Intractable
7 Adam, R., Karam, V., Delvart, V. et al. Hepatobiliary Disease Study Group
(2012). Evolution of indications and results supported by the Ministry of Health,
of liver transplantation in Europe. A report Labour and Welfare of Japan. Hepatol. Res.
from the European Liver Transplant 44 (Suppl S1): 71–90.
Registry (ELTR). J. Hepatol. 57: 675–688. 15 European Association for the Study of the
8 Singal, A.K., Guturu, P., Hmoud, B. et al. Liver (2017). EASL clinical practice
(2013). Evolving frequency and outcomes guidelines: the diagnosis and management
of liver transplantation based on etiology of of patients with primary biliary cholangitis.
liver disease. Transplantation 95: 755–760. J. Hepatol. 67: 145–172.
16 Nevens, F., Andreone, P., Mazzella, G. liver transplantation in 100 patients. Clin.
et al. (2016). A placebo‐controlled trial of Transpl. 20: 211–220.
obeticholic acid in primary biliary 26 Khettry, U., Anand, N., Faul, P.N. et al.
cholangitis. N. Engl. J. Med. 375: 631–643. (2003). Liver transplantation for primary
17 Harms, M.H., Janssen, Q.P., Adam, R. et al. biliary cirrhosis: a long‐term pathologic
(2019). Trends in liver transplantation for study. Liver Transpl. 9: 87–96.
primary biliary cholangitis in Europe over 27 Kogiso, T., Egawa, H., Teramukai, S. et al.
the past three decades. Aliment. (2017). Risk factors for recurrence of
Pharmacol. Ther. 49: 285–295. primary biliary cholangitis after liver
18 Boonstra, K., Beuers, U., and Ponsioen, transplantation in female patients: a
C.Y. (2012). Epidemiology of primary Japanese multicenter retrospective study.
sclerosing cholangitis and primary biliary Hepatol. Commun. 1: 394–405.
cirrhosis: a systematic review. J. Hepatol. 28 Liermann Garcia, R.F., Evangelista Garcia,
56: 1181–1188. C., McMaster, P., and Neuberger, J. (2001).
19 Bhanji, R.A., Mason, A.L., Girgis, S., and Transplantation for primary biliary cirrhosis:
Montano‐Loza, A.J. (2013). Liver retrospective analysis of 400 patients in a
transplantation for overlap syndromes of single center. Hepatology 33: 22–27.
autoimmune liver diseases. Liver Int. 33: 29 Manousou, P., Arvaniti, V., Tsochatzis, E.
210–219. et al. (2010). Primary biliary cirrhosis after
20 Carbone, M., Mells, G.F., Alexander, G.J. liver transplantation: influence of
et al. (2013). Calcineurin inhibitors and the immunosuppression and human leukocyte
IL12A locus influence risk of recurrent antigen locus disparity. Liver Transpl. 16:
primary biliary cirrhosis after liver 64–73.
transplantation. Am. J. Transplant. 13: 30 Montano‐Loza, A.J., Wasilenko, S., Bintner,
1110–1111. J., and Mason, A.L. (2010). Cyclosporine A
21 Charatcharoenwitthaya, P., Pimentel, S., protects against primary biliary cirrhosis
Talwalkar, J.A. et al. (2007). Long‐term recurrence after liver transplantation. Am.
survival and impact of ursodeoxycholic J. Transplant. 10: 852–858.
acid treatment for recurrent primary 31 Neuberger, J., Gunson, B., Hubscher, S., and
biliary cirrhosis after liver transplantation. Nightingale, P. (2004). Immunosuppression
Liver Transpl. 13: 1236–1245. affects the rate of recurrent primary biliary
22 Egawa, H., Sakisaka, S., Teramukai, S. et al. cirrhosis after liver transplantation. Liver
(2016). Long‐term outcomes of living‐ Transpl. 10: 488–491.
donor liver transplantation for primary 32 Sanchez, E.Q., Levy, M.F., Goldstein, R.M.
biliary cirrhosis: a Japanese multicenter et al. (2003). The changing clinical
study. Am. J. Transplant. 16: 1248–1257. presentation of recurrent primary biliary
23 Guy, J.E., Qian, P., Lowell, J.A., and Peters, cirrhosis after liver transplantation.
M.G. (2005). Recurrent primary biliary Transplantation 76: 1583–1588.
cirrhosis: peritransplant factors and 33 Sylvestre, P.B., Batts, K.P., Burgart, L.J. et al.
ursodeoxycholic acid treatment post‐liver (2003). Recurrence of primary biliary
transplant. Liver Transpl. 11: 1252–1257. cirrhosis after liver transplantation:
24 Hytiroglou, P., Gutierrez, J.A., Freni, M. histologic estimate of incidence and natural
et al. (2009). Recurrence of primary biliary history. Liver Transpl. 9: 1086–1093.
cirrhosis and development of autoimmune 34 Tanaka, A., Leung, P.S.C., and Gershwin,
hepatitis after liver transplant: a blind M.E. (2019). Pathogen infections and
histologic study. Hepatol. Res. 39: 577–584. primary biliary cholangitis. Clin. Exp.
25 Jacob, D.A., Neumann, U.P., Bahra, M. et al. Immunol. 195: 25–34.
(2006). Long‐term follow‐up after 35 Montano‐Loza, A.J. and Mason, A.L.
recurrence of primary biliary cirrhosis after (2017). Recurrence of primary biliary
cholangitis after liver transplantation: a for primary sclerosing cholangitis.

Japanese perspective. Hepatol. Commun. 1: Transplant. Proc. 33: 2452–2453.
391–393. 45 Renz, J.F. and Ascher, N.L. (2002). Liver
36 Tanaka, A., Leung, P.S.C., and Gershwin, transplantation for nonviral, nonmalignant
M.E. (2018). The genetics and epigenetics diseases: problem of recurrence. World
of primary biliary cholangitis. Clin. Liver J. Surg. 26: 247–256.
Dis. 22: 443–455. 46 Khettry, U., Keaveny, A., Goldar‐Najafi, A.
37 Bosch, A., Dumortier, J., Maucort‐Boulch, et al. (2003). Liver transplantation for
D. et al. (2015). Preventive administration primary sclerosing cholangitis: a long‐term
of UDCA after liver transplantation for clinicopathologic study. Hum. Pathol. 34:
primary biliary cirrhosis is associated with a 1127–1136.
lower risk of disease recurrence. J. Hepatol. 47 Kugelmas, M., Spiegelman, P., Osgood, M.J.
63: 1449–1458. et al. (2003). Different immunosuppressive
38 Karlsen, T.H., Folseraas, T., Thorburn, D., regimens and recurrence of primary
and Vesterhus, M. (2017). Primary sclerosing cholangitis after liver
sclerosing cholangitis: a comprehensive transplantation. Liver Transpl. 9: 727–732.
review. J. Hepatol. 67: 1298–1323. 48 Brandsaeter, B., Schrumpf, E., Bentdal, O.
39 Ponsioen, C.Y., Chapman, R.W., et al. (2005). Recurrent primary sclerosing
Chazouilleres, O. et al. (2016). Surrogate cholangitis after liver transplantation: a
endpoints for clinical trials in primary magnetic resonance cholangiography study
sclerosing cholangitis: review and results with analyses of predictive factors. Liver
from an international PSC study group Transpl. 11: 1361–1369.
consensus process. Hepatology 63: 49 Yamagiwa, S. and Ichida, T. (2007).
1357–1367. Recurrence of primary biliary cirrhosis and
40 Ueda, Y., Kaido, T., Okajima, H. et al. primary sclerosing cholangitis after liver
(2017). Long‐term prognosis and transplantation in Japan. Hepatol. Res. 37
recurrence of primary sclerosing (Suppl 3): S449–S454.
cholangitis after liver transplantation: a 50 Alexander, J., Lord, J.D., Yeh, M.M. et al.
single‐center experience. Transplant. (2008). Risk factors for recurrence of
Direct 3: e334. primary sclerosing cholangitis after liver
41 Goss, J.A., Shackleton, C.R., Farmer, D.G. transplantation. Liver Transpl. 14:
et al. (1997). Orthotopic liver 245–251.
transplantation for primary sclerosing 51 Campsen, J., Zimmerman, M.A., Trotter,
cholangitis. A 12‐year single center J.F. et al. (2008). Clinically recurrent
experience. Ann. Surg. 225: 472–481; primary sclerosing cholangitis following
discussion 481–483. liver transplantation: a time course. Liver
42 Jeyarajah, D.R., Netto, G.J., Lee, S.P. et al. Transpl. 14: 181–185.
(1998). Recurrent primary sclerosing 52 Cholongitas, E., Shusang, V.,
cholangitis after orthotopic liver Papatheodoridis, G.V. et al. (2008). Risk
transplantation: is chronic rejection part of factors for recurrence of primary sclerosing
the disease process? Transplantation 66: cholangitis after liver transplantation. Liver
1300–1306. Transpl. 14: 138–143.
43 Graziadei, I.W., Wiesner, R.H., Marotta, P.J. 53 Alabraba, E., Nightingale, P., Gunson, B.
et al. (1999). Long‐term results of patients et al. (2009). A re‐evaluation of the risk
undergoing liver transplantation for factors for the recurrence of primary
primary sclerosing cholangitis. Hepatology sclerosing cholangitis in liver allografts.
30: 1121–1127. Liver Transpl. 15: 330–340.
44 Lidén, H., Norrby, J., Gabel, M. et al. 54 Kashyap, R., Safadjou, S., Chen, R. et al.
(2001). Outcome after liver transplantation (2010). Living donor and deceased donor
liver transplantation for autoimmune and 63 Graziadei, I.W., Wiesner, R.H., Batts, K.P.
cholestatic liver diseases: an analysis of the et al. (1999). Recurrence of primary
UNOS database. J. Gastrointest. Surg. 14: sclerosing cholangitis following liver
1362–1369. transplantation. Hepatology 29: 1050–1056.
55 Moncrief, K.J., Savu, A., Ma, M.M. et al. 64 Montano‐Loza, A.J., Bhanji, R.A.,
(2010). The natural history of inflammatory Wasilenko, S., and Mason, A.L. (2017).
bowel disease and primary sclerosing Systematic review: recurrent
cholangitis after liver transplantation: a autoimmune liver diseases after liver
single‐centre experience. Can. J. transplantation. Aliment. Pharmacol.
Gastroenterol. 24: 40–46. Ther. 45: 485–500.
56 Egawa, H., Ueda, Y., Ichida, T. et al. (2011). 65 Visseren, T. and Darwish Murad, S. (2017).
Risk factors for recurrence of primary Recurrence of primary sclerosing
sclerosing cholangitis after living donor cholangitis, primary biliary cholangitis and
liver transplantation in Japanese registry. auto‐immune hepatitis after liver
Am. J. Transplant. 11: 518–527. transplantation. Best Pract. Res. Clin.
57 Ravikumar, R., Tsochatzis, E., Jose, S. et al. Gastroenterol. 31: 187–198.
(2015). Risk factors for recurrent primary 66 Cholongitas, E. and Burroughs, A.K.
sclerosing cholangitis after liver (2012). Recurrence of autoimmune liver
transplantation. J. Hepatol. 63: 1139–1146. diseases after liver transplantation: clinical
58 Hildebrand, T., Pannicke, N., Dechene, A. aspects. Auto Immun. Highlights 3:
et al. (2016). Biliary strictures and 113–118.
recurrence after liver transplantation for 67 Yamada, Y., Hoshino, K., Fuchimoto, Y.
primary sclerosing cholangitis: a et al. (2018). Rituximab induction to
retrospective multicenter analysis. Liver prevent the recurrence of PSC after liver
Transpl. 22: 42–52. transplantation: the lessons learned from
59 Gordon, F.D., Goldberg, D.S., Goodrich, ABO‐incompatible living donor liver
N.P. et al. (2016). Recurrent primary transplantation. Transplant. Direct 4: e342.
sclerosing cholangitis in the Adult‐to‐Adult 68 Manns, M.P., Lohse, A.W., and Vergani, D.
Living Donor Liver Transplantation Cohort (2015). Autoimmune hepatitis: Update
Study: comparison of risk factors between 2015. J. Hepatol. 62: S100–S111.
living and deceased donor recipients. Liver 69 Floreani, A., Restrepo‐Jimenez, P., Secchi,
Transpl. 22: 1214–1222. M.F. et al. (2018). Etiopathogenesis of
60 Aravinthan, A.D., Doyle, A.C., Issachar, A. autoimmune hepatitis. J. Autoimmun. 95:
et al. (2016). First‐degree living‐related 133–143.
donor liver transplantation in autoimmune 70 Alvarez, F., Berg, P., Bianchi, F. et al. (1999).
liver diseases. Am. J. Transplant. 16: International Autoimmune Hepatitis
3512–3521. Group Report: review of criteria for
61 Lindstrom, L., Jorgensen, K.K., Boberg, diagnosis of autoimmune hepatitis.
K.M. et al. (2018). Risk factors and J. Hepatol. 31: 929–938.
prognosis for recurrent primary sclerosing 71 Hennes, E.M., Zeniya, M., Czaja, A.J. et al.
cholangitis after liver transplantation: a (2008). Simplified criteria for the diagnosis
Nordic multicentre study. Scand. J. of autoimmune hepatitis. Hepatology 48:
Gastroenterol. 53: 297–304. 169–176.
62 Bajer, L., Slavcev, A., Macinga, P. et al. 72 European Association for the Study of the
(2018). Risk of recurrence of primary Liver (2015). EASL clinical practice
sclerosing cholangitis after liver guidelines: autoimmune hepatitis.
transplantation is associated with de novo J. Hepatol. 63: 971–1004.
inflammatory bowel disease. World J. 73 Yoshizawa, K., Matsumoto, A., Ichijo, T.
Gastroenterol. 24: 4939–4949. et al. (2012). Long‐term outcome of
Japanese patients with type 1 autoimmune autoimmune hepatitis after liver

hepatitis. Hepatology 56: 668–676. transplantation. Liver Transpl. 15:
74 Ayata, G., Gordon, F.D., Lewis, W.D. et al. 1254–1261.
(2000). Liver transplantation for 82 Prados, E., Cuervas‐Mons, V., de la Mata,
autoimmune hepatitis: a long‐term M. et al. (1998). Outcome of autoimmune
pathologic study. Hepatology 32: 185–192. hepatitis after liver transplantation.
75 Campsen, J., Zimmerman, M.A., Trotter, Transplantation 66: 1645–1650.
J.F. et al. (2008). Liver transplantation for 83 Ratziu, V., Samuel, D., Sebagh, M. et al.
autoimmune hepatitis and the success of (1999). Long‐term follow‐up after liver
aggressive corticosteroid withdrawal. Liver transplantation for autoimmune hepatitis:
Transpl. 14: 1281–1286. evidence of recurrence of primary disease.
76 Duclos‐Vallee, J.C., Sebagh, M., Rifai, K. et al. J. Hepatol. 30: 131–141.
(2003). A 10 year follow up study of patients 84 Reich, D.J., Fiel, I., Guarrera, J.V. et al. (2000).
transplanted for autoimmune hepatitis: Liver transplantation for autoimmune
histological recurrence precedes clinical and hepatitis. Hepatology 32: 693–700.
biochemical recurrence. Gut 52: 893–897. 85 Wang, Q., Yang, F., Miao, Q. et al. (2016).
77 Gonzalez‐Koch, A., Czaja, A.J., Carpenter, The clinical phenotypes of autoimmune
H.A. et al. (2001). Recurrent autoimmune hepatitis: a comprehensive review.
hepatitis after orthotopic liver J. Autoimmun. 66: 98–107.
transplantation. Liver Transpl. 7: 302–310. 86 Balan, V., Ruppert, K., Demetris, A.J. et al.
78 Krishnamoorthy, T.L., Miezynska‐Kurtycz, (2008). Long‐term outcome of human
J., Hodson, J. et al. (2016). Longterm leukocyte antigen mismatching in liver
corticosteroid use after liver transplantation transplantation: results of the National
for autoimmune hepatitis is safe and Institute of Diabetes and Digestive and
associated with a lower incidence of Kidney Diseases Liver Transplantation
recurrent disease. Liver Transpl. 22: 34–41. Database. Hepatology 48: 878–888.
79 Milkiewicz, P., Hubscher, S.G., Skiba, G. 87 Kerkar, N. and Vergani, D. (2018). De novo
et al. (1999). Recurrence of autoimmune autoimmune hepatitis: is this different in
hepatitis after liver transplantation. adults compared to children? J. Autoimmun.
Transplantation 68: 253–256. 95: 26–33.
80 Molmenti, E.P., Netto, G.J., Murray, N.G. 88 Kerkar, N. and Yanni, G. (2016). “De novo”
et al. (2002). Incidence and recurrence of and “recurrent” autoimmune hepatitis after
autoimmune/alloimmune hepatitis in liver liver transplantation: a comprehensive
transplant recipients. Liver Transpl. 8: review. J. Autoimmun. 66: 17–24.
519–526. 89 Neuberger, J. (2016). An update on liver
81 Montano‐Loza, A.J., Mason, A.L., Ma, M. transplantation: a critical review.
et al. (2009). Risk factors for recurrence of J. Autoimmun. 66: 51–59.
263
Section V
Controversies in Autoimmune Liver Diseases

265
15
Making Sense of Overlap and Crossover Syndromes

Olivier Chazouillères1,2,3,4
1
AP‐HP, Hôpital Saint‐Antoine, Service d’Hépatologie, Paris, France
2
Reference Center for Inflammatory Biliary Diseases and Autoimmune Hepatitis
3
INSERM, UMR‐S 938, CDR Saint‐Antoine, Paris, France
4
Sorbonne Université, Paris, France
Abstract
In a landmark review, J. Woodward and J. Neuberger emphasized that true overlaps should be dif
ferentiated from simple crossover or outlier syndromes. Moderate to severe interface hepatitis is a
fundamental component and histology is vital in evaluating patients with overlap presentation. The
low prevalence of overlap syndrome (OS) has made it impracticable to perform randomized con
trolled trials. For primary biliary cholangitis (PBC)/autoimmune hepatitis (AIH) OS, the European
Association for the Study of the Liver has provided diagnostic criteria and, in most cases, it is pos
sible to define one primary disorder, usually PBC. Primary sclerosing cholangitis/AIH OS is
assumed to exist in a considerable part of mainly young patients with autoimmune liver disease and
longterm progression toward cirrhosis seems to occur in the majority of cases. Treatment of OS is
empiric and includes ursodeoxycholic acid for the cholestatic component and immunosuppressive
agents for the hepatitic component, either simultaneously or sequentially.
Keywords autoimmune hepatitis; crossover syndromes; diagnostic criteria; immunosuppressive

agents; overlap syndromes; primary biliary cholangitis; primary sclerosing cholangitis;
ursodeoxycholic acid
Key Points
●● Some patients present with features of ●● The low prevalence of OS (roughly 10% of
both primary biliary cholangitis (PBC) or PBC or PSC) has made it impracticable to
primary sclerosing cholangitis (PSC) and perform randomized controlled trials.
autoimmune hepatitis (AIH) either simul ●● It remains unclear whether this syndrome
taneously or consecutively. forms a distinct entity or is a variant of
●● The term “overlap syndrome” (OS) is used PBC, PSC or AIH.
to describe these settings but lack of ●● Moderate to severe interface hepatitis is a
universal agreement on what precisely fundamental component and histology is
constitutes an OS has generated consider vital in evaluating patients with overlap
able confusion. presentation. Use of the International
266 Section V Controversies in Autoimmune Liver Diseases
Autoimmune Hepatitis Group criteria for cases have cholangiographic abnormal

the diagnosis of OS is not recommended. ities suggestive of PSC (autoimmune
●● For PBC/AIH OS, the European Association sclerosing cholangitis).
for the Study of the Liver has provided diag ●● Treatment of OS is empiric and includes
nostic criteria and, in most cases, it is pos ursodeoxycholic acid for the cholestatic
sible to define one primary disorder component (depending on local policy for
(“dominant” disease), usually PBC. PSC) and immunosuppressive agents for
●● Patients with OS seem to have a more the hepatitic component, either simulta
severe disease compared to conventional neously or sequentially.
PBC. ●● The dominant clinical feature should be
●● PSC/AIH OS is assumed to exist in a con treated first and therapy adjusted
siderable part of mainly young patients according to the response.
with autoimmune liver disease and long‐ ●● OS is not uncommon but should not be
term progression toward cirrhosis seems overdiagnosed in order not to unneces
to occur in the majority of cases. In chil sarily expose PBC or PSC patients to the
dren, the hepatitic feature can be very risk of steroid side effects. Therapy has to
dominant and up to 50% of pediatric AIH be individualized and not be static.
Introduction features between PBC and PSC have been

described only in a few case reports of vari
The three major autoimmune liver diseases able quality and do not represent a real issue.
are autoimmune hepatitis (AIH), primary In contrast, patients presenting with features
biliary cholangitis (PBC), and primary of PBC or PSC on the one hand and AIH on
sclerosing cholangitis (PSC). Within this the other, either simultaneously or consecu
spectrum of immune‐mediated disorders tively, have been repeatedly recognized. The
targeting hepatocytes or bile ducts, there are term “overlap syndrome” is often used to
two broad categories: those with predomi describe these variant forms. Unfortunately,
nant hepatocellular injury (AIH) and those lack of universal agreement on what precisely
with predominant cholestatic features (PBC constitutes an overlap syndrome has gener
and PSC) (Figure 15.1). These three diseases ated considerable confusion in the literature
are generally differentiated easily on the basis and the clinical phenotypes of patients with
of clinical, biochemical, serological, radio the same overlap syndrome designation
logical, and histological findings. Overlapping exhibit considerable heterogeneity [1]. As a
result, “overlap syndrome” is one of the most
abused descriptive terms currently used in
hepatology. Broadly similar pathogenic
PSC
themes of injury have been postulated for
AIH AIH, PBC and PSC and include environ
mental triggers, genetic predisposition, and
PBC failure of immune tolerance mechanisms,
whereby liver disease represents the result of
a cell‐ and antibody‐mediated immunologic
Figure 15.1 Overlap syndromes of the classical attack against liver‐specific targets [1].
autoimmune liver diseases. PBC/PSC overlap The pathogenesis of overlap syndrome is
syndrome is an extremely rare (and even
controversial) condition. AIH, autoimmune hepatitis; highly debated and it remains unclear whether
PBC, primary biliary cholangitis; PSC, primary two distinct diseases coexist in one patient,
sclerosing cholangitis. whether these forms are a single entity, or
Chapter 15 Making Sense of Overlap and Crossover Syndromes 267
whether they represent a variant form of diagnostic test (the possible exception being
either disease (PBC, PSC or AIH). The latter PBC). Their diagnosis is based on the
seems to be the most appropriate since a pre presence and relative absence of various
dominant phenotype can be identified in clinical, biochemical, serologic, and histo
most cases. For example, in PBC/AIH overlap, logic markers, with some being less
it has been proposed that overlap represents a categorical and objective than others [1]. As
“hepatitic” form of PBC in genetically suscep a result, there is intrinsic scope for individ
tible individuals (HLA‐B8, ‐DR3 or ‐DR4 uals to present with overlapping features of
positive) [2]. This would fit with the hypo more than one of these conditions although,
thesis that immune‐mediated disease can in most cases, it is possible to define one pri
develop (“secondary” AIH) in any susceptible mary disorder (“dominant” disease). In a
host if, for some reason, the local milieu landmark review, Woodward and Neuberger
becomes proinflammatory. In this regard, the [5] emphasized that “true overlaps” should
name “overlap,” which strongly suggests the be differentiated from simple “crossover” or
presence of two distinct diseases, could “outlier” syndromes (one clear diagnosis
be a misnomer. As a result, according while having one feature associated with
to the European Association for the Study of another). Overlapping presentations include
the Liver (EASL) AIH and PBC guidelines, the following.
the preferred terminology to describe these
●● Biochemical overlap: aspartate aminotrans
conditions is now variant forms, primarily
ferase (AST) or alanine aminotransferase
variant forms of the cholestatic autoimmune
(ALT) more than five times upper limit of
liver diseases PBC or PSC with autoimmune
normal (ULN) in patients with PSC or PBC;
features [3,4]. In contrast, recent British and
or alkaline phosphatase (ALP) more than
US PBC guidelines still use the term “overlap”.
three times ULN in patients with AIH.
Because of the absence of well‐validated
●● Serologic overlap: positive anti‐smooth
diagnostic criteria and publication bias, the
muscle antibody (ASMA) in anti‐mitoch
prevalence of overlap syndromes is difficult
ondrial antibody (AMA)‐positive PBC; or
to ascertain and diagnosis remains a
positive AMA in AIH.
challenge. It should be kept in mind that
●● Histologic overlap: interface hepatitis on
overlap syndrome should not be overdiag
liver biopsy with biliary lesions indicative of
nosed in order not to expose PBC or PSC
PBC or PSC; cholangiographic abnormal
patients unnecessarily to the risk of steroid
ities associated with clinical features of AIH.
side effects. On the other hand, the tragic
●● Varying combinations of the above.
consequences of a missed opportunity of
instituting immunosuppressive therapy in However, these overlap features have var
overlap patients have occasionally been ious significance, the weaker being probably
reported. The low prevalence of overlap immunoserology. Indeed, autoantibody pro
syndromes has made it impracticable to
file should never be used in isolation but
perform randomized controlled trials. As a rather interpreted in conjunction with
consequence, treatment of overlap syndrome biochemical, radiologic, and histologic fea
is largely empiric. tures. For example, patients with histologic
AIH and AMA positivity generally behave
like typical AIH and the same holds true for
General Considerations AMA‐negative but ANA‐ and/or ASMA‐
positive PBC (sometimes described as auto
The main features of the three classical phe immune cholangitis) when compared with
notypes of autoimmune liver disease are typical PBC. Laboratory features lack sensi
indicated in Table 15.1. A key point is that no tivity considering that cholestasis in itself can
autoimmune liver disease has an absolute cause raised ALT levels in the absence of
Table 15.1 Features of autoimmune liver diseases.
AIH PBC PSC
Gender Female > male (4 : 1) Female > male (9 : 1) Male > female (2 : 1)
Coexisting IBD 3–10% (PSC should Not characteristic Up to 80%a
be excluded)
ANA 70–80% 30–50% (some 30–70%
specific)
ASMA 70–80% May be present: 0–80%
<10%
AMA 5–10% 95% Coincidental
p‐ANCA Up to 90% 0–5% 25–95%
Immunoglobulins IgG elevated IgM elevated in IgG elevated (two‐thirds)
most and IgM (45%) elevated
Cholangiography Usually normal Normal Multifocal stricturing
(not in small‐duct PSC)
Interface hepatitis Characteristic Variably present Variably present
Biliary changes 10% Inflammatory duct Onion‐skin periductal
lesion fibrosis (<30%)
Response to Yes Mild Minimal
immunosuppression
a
Most characteristic features are indicated in bold.
AIH, autoimmune hepatitis; AMA, anti‐mitochondrial antibody; ANA, anti‐nuclear antibody; ASMA, anti‐smooth
muscle antibody; IBD, inflammatory bowel disease; pANCA, perinuclear anti‐neutrophil cytoplasmic antibody; PBC,
primary biliary cholangitis; PSC, primary sclerosing cholangitis.
inflammation and that cirrhosis can lead to reproducible ways to grade interface
high IgG levels in the absence of histologic hepatitis are lacking, although the non‐
hepatitis. In contrast, liver biopsy is the most specific hepatitis activity index (HAI score)
reliable way of diagnosing overlap. is frequently used [3]. A score of more than
4 out of 18 points is observed primarily in
patients with variant syndromes, mainly
Liver Biopsy
due to additional lobular inflammatory
Liver biopsy is considered a prerequisite for activities. Interface hepatitis is not disease
the diagnosis of AIH [3]. Interface hepatitis specific and patients with drug‐related, viral
(hepatitis at the portal–parenchymal inter or immune‐mediated disease may show
face) with dense plasma cell‐rich lympho similar features. Lymphocytic piecemeal
plasmacytic infiltrates is the typical hallmark necrosis (as opposed to biliary piecemeal
of AIH. However, there is no morphologic necrosis that is the consequence of major
feature that is pathognomonic of AIH and a cholestasis) can be seen in about 25% of
degree of interface hepatitis can be observed patients with PBC or PSC and, conversely,
across all liver autoimmune diseases. In some biliary injury is observed in 10–20% of
AIH, the majority of patients will have a patients with AIH [6]. It has been hypothe
significant degree of interface hepatitis, sized that these lesions may represent some
contrasting with the usual low intensity of collateral damage mediated via bystander
interface hepatitis in PBC or PSC mechanisms in the context of a marked
(Figure 15.2) [1]. Unfortunately, robust and inflammation targeted primarily against
Figure 15.2 Distribution and degree of Proportion of

interface hepatitis across the spectrum patients AIH
of autoimmune liver diseases. AIH, PBC
autoimmune hepatitis; PBC, primary High
biliary cholangitis; PSC, primary PSC
sclerosing cholangitis. Source: Trivedi
and Hirschfield [1]. Reproduced with
permission of John Wiley and Sons.
Low
Mild Moderate Severe
Interface hepatitis
either hepatocytes or biliary epithelial cells together with associated groups of general
[1]. A clinicopathologic study of PBC with inflammatory risk loci). Information about
interface hepatitis compared to AIH has genetic susceptibility in overlap syndromes is
suggested that, despite similar immunophe limited. The presence of the characteristic
notypes of infiltrating cells, the precise haplotypes of AIH (HLA‐DR3/DR4 and
mechanisms of hepatocellular injuries may HLA‐B8) has been reported in AIH/PBC
not be identical. Nevertheless, interface overlap patients and AIH patients (17/20 vs.
hepatitis is a fundamental component and 18/20, respectively), but these haplotypes
histology is vital in evaluating patients with were seen significantly less often in pure PBC
overlap presentation. It is generally assumed patients (4/20, p <0.05) [2].
that the degree of interface hepatitis can be In clinical practice, overlap syndromes
considered a measure of AIH‐like disease should be considered when a patient with
activity irrespective of the coexistence of autoimmune liver disease deviates from the
underlying cholestatic liver disease [6]. normal clinical phenotype and expected
Finally, interpretation of good‐quality liver response to therapy but it is important not to
biopsy is key and a specialist review of liver overdiagnose nor to frequently use non‐stan
biopsies has a major added value [7]. dard therapy with potential side effects. In
this regard, overlap syndromes should be
diagnosed conservatively and be differenti
AIH Scores ated from simple “crossover” or “outlier” syn
The International Autoimmune Hepatitis dromes that do not require treatment
Group (IAIHG) criteria (either revised or tailoring [5] (see following sections).
simplified) are an attractive tool for making
the diagnosis of overlap syndromes in patients
with an existing diagnosis of PBC or PSC. PBC/AIH Overlap Syndrome
However, although widely used in this setting,
the performance of these criteria appears low With no codified diagnostic approach,
and their use is not recommended for the reported prevalence figures are variable
diagnosis of overlap syndromes [6]. (2.1–19% in PBC patients) but it is generally
assumed that PBC/AIH overlap prevalence is
approximately 8–10% in adult patients with
Genetics
PBC or AIH [6]. The Paris criteria (Box 15.1)
The genetic architecture of PBC and PSC is are currently the most commonly used tool
similar to most autoimmune diseases, for diagnosing PBC/AIH overlap (see also
including AIH (i.e. a strong HLA association Figure 15.3 and Table 15.2).
Box 15.1 Paris criteria for diagnosing PBC/AIH overlap syndrome

The presence of at least two of the three accepted key criteria of each disease is required for
diagnosis
PBC
●● ALP ≥2 ULN or GGT ≥5 ULN
●● Presence of AMA or PBC‐specific ANA (i.e. anti‐gp210 or anti‐sp100)
●● Liver biopsy specimen showing florid bile duct lesions
AIH
●● ALT ≥5 ULN
●● Serum IgG levels ≥2 ULN or presence of ASMA
●● Liver biopsy showing moderate or severe periportal or periseptal lymphocytic piecemeal
necrosis) [8]
AIH, autoimmune hepatitis; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AMA, anti‐mitochondrial
antibody; ANA, anti‐nuclear anti‐body; ASMA, anti‐smooth muscle antibody; GGT, gamma‐glutamyltransferase;
PBC, primary biliary cholangitis; ULN, upper limit of normal.
Table 15.2 Proposed criteria for diagnosis of overlap

syndrome.
The presence of at least two of the three accepted

key criteria is required for diagnosis of each
disease
PBC
ALP ≥2 ULN and/or GGT ≥5 ULN
AMA ≥1 : 40 or PBC‐specific ANA
Florid bile duct lesions (liver biopsy)
PSC (causes of secondary sclerosing cholangitis
excluded)
ALP ≥2 ULN and/or GGT ≥5 ULN
Typical cholangiographic abnormalities
Figure 15.3 Histologic features of PBC/AIH overlap
Periductal fibrosis (liver biopsy) NB: some cases of
syndrome in a 38‐year‐old woman with chronic
overlap with small‐duct PSC
cholestasis, positive anti‐mitochondrial and anti‐
nuclear antibodies, gamma‐globulins of 15 g/l, and AIH
transaminase levels five times upper limit of normal. ALT ≥5 ULN
Diffuse lymphocytic interface hepatitis showing liver
IgG levels ≥2 ULNa or ASMA ≥1 : 80
cell plates thickened due to regeneration (hematoxylin
& eosin stain, original magnification ×100). Source: Moderate or severe periportal or periseptal
courtesy of Professor Dominique Wendum. lymphocytic piecemeal necrosis (liver biopsy)
(mandatory)
Most authors agree that these criteria pro a

20 g/l tends to be the usual proposed cutoff.
vide a diagnostic template that can be consis AIH, autoimmune hepatitis; ALP, alkaline phosphatase;
tently applied for a diagnosis of “true overlap” AMA, anti‐mitochondrial antibody; ANA, anti‐nuclear
anti‐body; ASMA, anti‐smooth muscle antibody; GGT,
and it has been shown that the Paris criteria gamma‐glutamyltransferase; PBC, primary biliary
can identify patients with a clinical diagnosis cholangitis; PSC, primary sclerosing cholangitis; ULN,
of AIH/PBC variant with high sensitivity upper limit of normal.
(92%) and specificity (97%), keeping in mind UDCA, progression of fibrosis, and liver‐
that it is difficult to assess diagnostic related mortality. In a large clinical PBC trial,
performance in the absence of a gold standard. the patients who were identified as having
The EASL guidelines endorsed these diag potential AIH overlap retrospectively had a
nostic criteria but specified that histologic evi clinical outcome similar to “pure” PBC patients
dence of moderate to severe lymphocytic but unfortunately histologic fibrosis course
piecemeal necrosis (interface hepatitis) was was not assessed and no firm conclusions can
mandatory and stated that overlap syndrome be drawn from this study. Although UDCA
should always be considered once PBC has monotherapy may induce biochemical
been diagnosed and in case of poor response response in some patients with PBC/AIH
to ursodeoxycholic acid (UDCA) because of overlap, most patients may require a
potential therapeutic implications [9]. combination of UDCA and immunosuppres
Nevertheless, there are still several areas of sive therapy to obtain a complete biochemical
uncertainty, including the cutoffs for IgG/ response. In our experience of overlap patients
gamma‐globulin and transaminase levels for with histologic follow‐up, the overall fibrosis
indicating liver biopsy and the grade of progression in non‐cirrhotic patients occurred
hepatitis activity for indicating immunosup significantly more frequently under UDCA
pression [4]. Indeed, the recent EASL guide monotherapy (13–15 mg/kg daily) than under
lines on AIH recommend treatment for combined therapy with predniso(lo)ne
patients with AIH at lower cutoffs for trans (0.5 mg/kg daily as starting dose then tapered)
aminase or IgG levels and a histologic HAI and azathioprine (50–100 mg/day) [10].
score as low as 4 [3]. Despite the lack of controlled trials, the 2009
Autoantibodies against soluble liver EASL guidelines, based on the results of small
antigen (SLA/LP) and double‐stranded DNA series, have recommended adding steroids
have been associated with presence of AIH in (eventually budesonide at an initial dose of
patients with PBC and these autoantibodies 9 mg/day) either at the time of diagnosis of
should be tested in the work‐up of PBC overlap syndrome or in case of inadequate
patients with suspected overlap. Additionally, biochemical response after 3 months of
overlap of AMA‐negative PBC with AIH has UDCA [9]. The more recent results of a large
also been reported [8]. retrospective multicenter study (88 patients
The simultaneous presence of features of defined according to Paris criteria) have
both diseases is the usual presentation. In underlined the predictive role of the degree of
most cases, the “dominant” disease pheno interface hepatitis: as first‐line therapy,
type is PBC [6]. Less commonly, the onset of 30 patients received UDCA alone and 58
AIH and PBC is temporally dissociated, usu patients a combination of UDCA and immu
ally with PBC presenting first with a variable nosuppression (prednisone with or without
interval of 6 months to 13 years before the azathioprine). In patients with moderate inter
onset of AIH. The development of AIH could face hepatitis, UDCA alone and combination
not be predicted from baseline characteris therapy had similar efficacy (80%) in terms of
tics and initial response to UDCA. biochemical response, whereas in patients
Alternatively, PBC may (rarely) occur after with severe hepatitis, efficacy of UDCA alone
the diagnosis of AIH. was much lower (14 vs. 71%, respectively).
Patients with PBC/AIH overlap syndrome Second‐line immunosuppressive agents
seem to have a more severe disease compared (ciclosporin, tacrolimus and mycophenolate)
to conventional PBC as illustrated by a higher led to biochemical remission in half of the
frequency of extensive fibrosis at presentation patients who were non‐responders to initial
(despite a younger age in some reports) and immunosuppression [11]. These findings
most series (but not all) support a worse strongly support the use of a combination of
outcome in terms of biochemical response to UDCA and immunosuppression as first‐line
therapy in overlap patients with severe inter features ranged from 7 to 14% [6]. On the
face hepatitis. The aim of the treatment is other hand, cholangiographic abnormalities
complete remission defined by normalization suggesting PSC are found in AIH patients at
of transaminases and IgG levels and minimal various prevalence depending on the age of
inflammation on histology. The optimal dura the patients evaluated: 2–10% in adults (41%
tion of immunosuppression in overlap syn if ulcerative colitis is present) and up to 50%
dromes is unclear. Interestingly, it has been in children [12]. The diagnosis of large‐duct
suggested that, in responders, doses of immu PSC should always be established on typical
nosuppressants in the long term could be cholangiographic findings, bearing in mind
lower and rate of successful withdrawal higher that an intrahepatic biliary tree which simu
than in classical AIH [10,11]. lates a sclerosing pattern can be observed in
Recently, obeticholic acid (OCA) has been any liver disease with extensive fibrosis.
approved as a second‐line therapy for PBC Some cases of small‐duct PSC (normal chol
patients with an inadequate response to angiogram)/AIH overlap syndrome have also
UDCA monotherapy. Impressive results of been reported but it can be argued that
fibrates have also been reported in these approximately 10% of patients with typical
patients. It is important to differentiate AIH, with or without ulcerative colitis, may
patients with “classical” PBC and non‐ have histologic features of bile duct injury. In
response to UDCA from those with overlap children, the hepatitic feature can be very
syndrome who are also non‐responsive to dominant and up to 50% of pediatric AIH
UDCA. Whether the pleiotropic effects of (clinical and/or evidence of liver disease
fibrates or farnesoid X receptor agonists like associated with circulating autoantibodies)
OCA have sufficient immunosuppressive have cholangiographic abnormalities sugges
capacities and could be beneficial for overlap tive of PSC including some (25%) without
syndromes is currently unknown. any histologic features of bile duct injury or
In UDCA‐treated PBC patients devel biochemical cholestasis [12]. Inflammatory
oping AIH (“sequential” overlap syndrome), bowel disease (IBD) was present in 44% of
use of immunosuppressive treatment is these children compared to 20% of those
mandatory [9]. with AIH alone. Thus, the term “autoim
mune sclerosing cholangitis” (AISC) was
introduced by Mieli‐Vergani’s group [12].
PSC/AIH Overlap Syndrome Evolution from AIH to AISC has been docu
mented, supporting the view they could be
PSC/AIH overlap syndrome has been part of the same pathogenic process. It has
described in both children and adults and is been proposed that at least some adult PSC
assumed to exist in a considerable number of cases may represent an advanced, at times
mainly young patients with autoimmune “burnt‐out”, stage of AISC but whether
liver disease. The diagnosis of overlap syn childhood AISC and adult PSC belong to the
drome is made in a patient with overt cholan same disease spectrum remains to be
giographic or histologic features of PSC, established. These findings also suggest the
alongside robust histologic features of AIH need for an investigation of the biliary tree at
concurrently or historically [1,6] (Table 15.2). least with magnetic resonance cholangio
Unfortunately, as opposed to PBC/AIH pancreatography (MRCP) in all children with
overlap syndrome, no precise and strict diag a diagnosis of AIH. At present, this variant
nostic criteria have been proposed. As a seems unique to children, as a prospective
result, reported prevalence figures vary study in adults with AIH was negative and
greatly. When the revised IAIHG criteria thus, in the absence of cholestatic indices,
were applied to large series of PSC patients, MRCP screening does not seem justified in
the prevalence of PSC/AIH overlapping adult‐onset AIH. However, particularly in
young adults with AIH and cholestatic fea is lower than in AIH (100%) [12]. In the series
tures or IBD, and in AIH patients with with the most homogeneous regimen
remaining cholestasis despite adequate (UDCA, prednisolone and azathioprine)
immunosuppression, MRCP for the detec including seven young adults with a mean
tion of possible underlying or coexistent PSC follow‐up of 8 years, the Mayo Risk Score did
is recommended [13]. not increase and transplant‐free survival was
In adults, AIH and PSC may be sequential much better (100%) than that of 34 classical
in their occurrence, typically with AIH pre PSC patients (43%) with the same follow‐up
senting first, as illustrated by a case series of and treated with UDCA. However, in the
AIH patients becoming cholestatic and resis long term (>10 years), progression toward
tant to immunosuppressive therapy. AIH is cirrhosis seems to occur in the majority of
more rarely diagnosed in patients with an patients.
original diagnosis of PSC. Taken together, all these data support the
Similar to PBC/AIH overlap, there are no use of UDCA in combination with an immu
double‐blind, randomized controlled trials in nosuppressive regimen in most patients with
PSC/AIH overlap syndromes. It should be PSC/AIH overlap syndrome despite the lack
kept in mind that although immunosuppres of adequate studies. However, the key point
sants benefit the hepatitic component of is that, even more than in PBC/AIH overlap
AIH, no survival benefit has been demon syndrome, treatments in PSC/AIH overlap
strated with UDCA in PSC. In addition, should be individualized based on bioch
unlike in PBC and AIH, biochemical improve emical, serologic, cholangiographic, and his
ment in PSC does not necessarily translate tologic findings. In patients with severe
into better clinical outcome. Various results interface hepatitis, use of immunosuppres
of therapy (usually prednisolone and azathio sants is mandatory. In other cases (moderate
prine with or without UDCA) have been interface hepatitis), my policy is, at present,
reported in patients with PSC/AIH overlap similar to that in PBC/AIH overlap and to
[6]. It is difficult to draw any firm conclusions start with UDCA monotherapy and add
because of the small number of patients, the immunosuppressants only in case of inade
usually retrospective nature of the studies, quate biochemical response after 3 months
and the heterogeneity of the regimens. The of UDCA.
combination of UDCA and immunosuppres
sive therapy may improve liver biochemistry
and this approach has been advocated by Liver Transplantation
EASL guidelines [9], whereas the American
Association for the Study of Liver Diseases Studies documenting outcomes of patients
(AASLD) guidelines recommend the use of who underwent liver transplantation for
corticosteroids and other immunosuppres overlap syndromes are limited. In the largest
sive agents and the IAIHG position is to con series, patients with overlap (n = 12) had a
sider immunosuppressive treatment with or higher probability of recurrent autoimmune
without UDCA [6]. Unsurprisingly, patients liver disease when compared to those with
with PSC/AIH overlap have a poorer out single autoimmune liver disease (at 5 years,
come when compared to those with (treated) 53% vs. 17%, respectively) together with a
AIH alone. However, their survival is appar shorter median time to recurrence but
ently similar or even better than in classical without lower patient survival [14].
PSC. In the pediatric AISC form treated with Interestingly, two patients had recurrence as
immunosuppressants, liver biopsies may overlap syndrome (one PBC/AIH and one
show improvement in inflammation but PSC/AIH) and experienced a lower graft
cholangiographic appearances may progress survival than those having recurrence as a
and transplant‐free survival at 10 years (65%) single autoimmune liver disease.
Conclusions Treatment decisions should be tailored to

the individual and not be static. In most
Liver overlap syndromes do exist but are cases, it is possible to define one primary
rare. Whatever the name used (for example, (dominant) disorder. As a rule, the domi
variant PBC or PSC with AIH features or nant clinical feature should be treated first
variant autoimmune hepatitis with PBC or and therapy should be individualized and
PSC features), recognition of autoim adjusted according to the response. In diffi
mune overlap syndromes is of interest not cult cases, referral to a specialist center
only from a classification standpoint but with a high volume of caseload with auto
also, and more importantly, because of immune liver diseases is recommended.
therapeutic implications. Overlap syn International efforts for collection of a large
dromes should be diagnosed conservatively database and discovery of more specific molec
by using as strict criteria as possible. ular signatures with the ability to identify sub
Appraisal has to be performed longitudi groups within the spectrum of autoimmune
nally rather than at a single point in time. liver disease should be encouraged.
References
1 Trivedi, P.J. and Hirschfield, G.M. (2012). 8 Chazouillères, O., Wendum, D., Serfaty, L.
Review article: overlap syndromes and et al. (1998). Primary biliary cirrhosis–
autoimmune liver disease. Aliment. autoimmune hepatitis overlap syndrome:
Pharmacol. Ther. 36: 517–533. clinical features and response to therapy.
2 Lohse, A.W., zum Buschenfelde, K.H., Franz, Hepatology 28: 296–301.
B. et al. (1999). Characterization of the 9 European Association for the Study of the
overlap syndrome of primary biliary Liver (2009). EASL clinical practice
cirrhosis (PBC) and autoimmune hepatitis: guidelines: management of cholestatic liver
evidence for it being a hepatitic form of PBC diseases. J. Hepatol. 51: 237–267.
in genetically susceptible individuals. 10 Chazouillères, O., Wendum, D., Serfaty, L.
Hepatology 29: 1078–1084. et al. (2006). Long term outcome and
3 European Association for the Study of the response to therapy of primary biliary
Liver (2015). EASL clinical practice cirrhosis–autoimmune hepatitis overlap
guidelines: autoimmune hepatitis. J. Hepatol. syndrome. J. Hepatol. 44: 400–406.
63: 971–1004. 11 Ozaslan, E., Efe, C., Heurgue‐Berlot, A.
4 European Association for the Study of the et al. (2014). Factors associated with
liver (2017). EASL clinical practice response to therapy and outcome of
guidelines: the diagnosis and management patients with primary biliary cirrhosis with
of patients with primary biliary cholangitis. features of autoimmune hepatitis. Clin.
J. Hepatol. 67: 145–172. Gastroenterol. Hepatol. 12: 863–869.
5 Woodward, J. and Neuberger, J. (2001). 12 Gregorio, G.V., Portmann, B., Karani, J.
Autoimmune overlap syndromes. et al. (2001). Autoimmune hepatitis/
Hepatology 33: 994–1002. sclerosing cholangitis overlap syndrome in
6 Boberg, K.M., Chapman, R.W., Hirschfield, childhood: a 16‐year prospective study.
G.M. et al. (2011). Overlap syndromes: the Hepatology 33: 544–553.
International Autoimmune Hepatitis 13 Lindor, K.D., Kowdley, K.V., and Harrison,
Group (IAIHG) position statement on a M.E. (2015). ACG clinical guideline:
controversial issue. J. Hepatol. 54: primary sclerosing cholangitis. Am. J.
374–385. Gastroenterol. 110: 646–659. quiz 660.
7 Paterson, A.L., Allison, M.E., Brais, R., and 14 Bhanji, R.A., Mason, A.L., Girgis, S., and
Davies, S.E. (2016). Any value in a Montano‐Loza, A.J. (2013). Liver
specialist review of liver biopsies? transplantation for overlap syndromes of
Conclusions of a 4‐year review. autoimmune liver diseases. Liver Int. 33:
Histopathology 69: 315–321. 210–219.
277
16
The Role of Extrahepatic Autoimmunity in Autoimmune

Liver Disease
Ewa Wunsch1 and Piotr Milkiewicz1,2
1
Translational Medicine Group, Pomeranian Medical University, Szczecin, Poland
2
Liver and Internal Medicine Unit, Medical University of Warsaw, Warsaw, Poland
Abstract
This chapter presents the knowledge on the role of extrahepatic autoimmunity in autoimmune
liver disease, even though systematic studies on large cohorts of patients are often lacking and
some experience is based only on case series reports. Awareness and early investigation of extrahe-
patic autoimmune diseases (EHAIDs) in patients with autoimmune liver disease is important.
Targeted testing should be considered according to the liver disease; it may be difficult to differen-
tiate between those EHAIDs that can impact the liver and drug-induced liver injury from primary
autoimmune liver disease. In primary sclerosing cholangitis (PSC), ulcerative colitis is a dominating
EHAID. The strong co-occurrence of PSC with inflammatory bowel disease (IBD), as well as
several other unique features of PSC-related IBD, led to the idea that it may be a distinct entity from
IBD, unrelated to PSC in terms of pathogenesis, phenotype and prognosis.
Keywords drug-induced liver injury; extrahepatic autoimmune diseases; inflammatory bowel
disease; primary biliary cholangitis; primary sclerosing cholangitis; ulcerative colitis
Key Points
●● Awareness and early investigation of ●● In autoimmune hepatitis, autoimmune thy-
extrahepatic autoimmune diseases (EHAIDs) roid disease is the most frequent EHAID.
in patients with autoimmune liver disease ●● In primary sclerosing cholangitis (PSC),
is important. ulcerative colitis (UC) is a dominating
●● Targeted testing should be considered EHAID. The clinical presentation of UC in
according to the liver disease; it may be association with PSC differs in many ways
difficult to differentiate between those from UC seen in patients without PSC.
EHAIDs that can impact the liver (such ●● In primary biliary cholangitis, autoimmune
as systemic lupus erythematosus) and thyroid disease, celiac disease, Sjögren syn-
drug‐induced liver injury from primary drome, and systemic sclerosis occur in a
autoimmune liver disease. significant proportion of patients.
Introduction administered medications are more common

causes. Accurate diagnosis has to be
Autoimmune diseases encompass a wide established according to respective criteria;
range of disorders that result from a however, the differentiation is sometimes
combination of genetic and environmental challenging, especially since the use of immu-
factors affecting various mechanisms of the nosuppressive therapy for rheumatic disor-
innate and acquired immune system. ders may distort the clinical and laboratory
Deprivation of immunologic self‐tolerance findings. Therefore, liver biopsy is often cru-
results in pathologic immune response to cial in selected cases.
organ‐specific or systemic self antigens and In this chapter, we present the current
thereby triggers various clinical manifesta- knowledge on the role of extrahepatic auto-
tions. However, divergent autoimmune con- immunity in AILD, even though systematic
ditions share some common genetic studies on large cohorts of patients are often
background and pathologic pathways. This lacking and some experience is based only on
results in the occurrence of more than one case series reports.
autoimmune condition in a single patient,
either as overlap syndromes or multiple
individual autoimmune diseases [1,2]. Epidemiology
Moreover, the same or diverse autoimmune
diseases may affect several family members, Concurrent EHAIDs exhibit the full range of
a phenomenon called autoimmunity clus- autoimmune conditions, from localized dis-
tering. Autoimmune liver diseases (AILDs) orders such as Sjögren syndrome (SS) to
are not the exception – they often coexist generalized autoimmune diseases such as
with other extrahepatic autoimmune dis- systemic lupus erythematosus (SLE) [3,4].
eases (EHAIDs). The high incidence of auto- Reported frequency is variable depending
immune comorbidity in patients with on underling AILD, geographic context, and
autoimmune hepatitis (AIH) was the reason the number of patients included in the
to incorporate EHAID in the revised scoring analysis (Table 16.1). In patients with

system of the International Autoimmune primary biliary cholangitis (PBC), SS and
Hepatitis Group. autoimmune thyroid disease are leading
On the other hand, in autoimmune concomitant autoimmune disorders.
systemic disorders such as connective tissue Autoimmune thyroid disease is the most
diseases (CTDs), concomitant liver abnor- common EHAID associated with AIH,
malities are not infrequent findings. They whereas inflammatory bowel disease (IBD),
may be associated with underlying AILD; mainly ulcerative colitis (UC), is a typical
however, non‐specific liver involvement in comorbidity observed in two‐thirds of
the course of CTDs or hepatotoxicity of the patients with primary sclerosing cholangitis
Table 16.1 The most frequent EHAIDs in patients with AILDs.
Inflammatory
Autoimmune Sjögren Systemic Systemic lupus Rheumatoid Celiac bowel
thyroid disease syndrome sclerosis erythematosus arthritis disease disease
AIH 10–15% 1.4–2.3% 1.2–3.5% Case series 1.8–2.3% 1.1–3.5% 2–8%

PBC 5.6–24% 3.5–27% 1.4–12.3% 1.0–3.7% 1.8–5.6% 5–10% 2–8%
PSC 11% Single Single cases Single cases 3.4% 1.6% 70%
cases
Chapter 16 Role of Extrahepatic Autoimmunity in Autoimmune Liver Disease 279
(PSC). Patients with both PBC and AIH with Autoimmune thyroid disease per se does
concomitant EHAIDs show significant not seem to significantly influence the natural
female predominance, whereas PSC‐related history of AILDs. Nevertheless, prior exacer-
IBD is diagnosed more commonly in males. bated hypothyroidism may cause a delay in
EHAIDs are more commonly found in AIH diagnosis of chronic liver diseases such as PBC
patients with first‐degree family history of and PSC as both conditions share common
EHAIDs. Concomitant autoimmunity may symptoms including fatigue, metabolic bone
predate the onset of AILDs or occur during disease, hyperlipidemia, and abnormalities in
the course and even years after the diagnosis cholestatic parameters, particularly alkaline
of AILDs. phosphatase (ALP) and bilirubin. On the other
hand, failure to recognize the onset of con-
comitant thyroid dysfunction in patients with
Autoimmune Thyroid Disease established AILD may lead to a misdiagnosis
of liver disease decompensation or lack of
The prevalence of thyroid dysfunction response to therapy. This can potentially lead
among patients with AILD is significantly to improper decisions about liver treatment
higher than in the general population, and or even timing for liver transplantation.
mainly manifests with hypothyroidism. The Therefore, the recognized association between
reported prevalence of abnormal thyroid AILDs and thyroid diseases should warrant
function has been most extensively studied the routine screening of patients with AILDs
in PBC and ranges between 13 and 22%. The for concomitant thyroid dysfunction to make
data among patients with PSC are scarce. A an accurate diagnosis according to their
study by Silveira et al. [5] showed 11% respective diagnostic criteria.
prevalence of thyroid dysfunction in this

group of patients; among them, 6% had
hypothyroidism and 5% hyperthyroidism.
Sjögren Syndrome
Autoimmunity plays a crucial role; other
causes of thyroid disease in AILDs are less SS is a progressive autoimmune exocrinopa-
common. As already mentioned, autoim- thy characterized by a lymphocytic infiltra-
mune thyroid diseases are the most frequent tion of the exocrine glands, leading to ocular
EHAID in patients with AIH, and the sec- dryness (xerophthalmia) and oral dryness
ond most frequent in patients with PBC. In (xerostomia). The disorder can be classified
the spectrum of thyroid autoimmunity, as primary and secondary SS. Primary SS
Hashimoto thyroiditis dominates, with represents a category on its own without

Graves disease being less frequent. The any other underlying condition, whereas
coincidence of autoimmune thyroid dis- secondary SS occurs as a concomitant
eases and AILD may be at least partially condition with another autoimmune disorder.
explained by a common genetic background. Sicca symptoms (i.e. eye and mouth dryness)
Our experience suggests the presence of and difficulty tasting and swallowing are
several single nucleotide polymorphisms reported by 47–73% of patients with PBC.
(located in genes PTPN22, IL2RA, MMEL1, Further investigation, such as Schirmer’s test
CTLA4, and RNASET2) that are associated and/or salivary gland biopsy, leads to the
with shared susceptibility to autoimmune diagnosis of secondary SS in 3.5–27% of them.
thyroid disease and PBC. The possible In fact, this is the most frequent extrahepatic
immunologic mechanism may lead to cross‐ condition seen in PBC [3,6]. The coincidence
reactivity of anti‐thyroid autoantibodies in with other AILDs is much less frequent, with
the presence of autoreactive T cells or sim- the prevalence in AIH being 1.4–2.3%, and
ilar epithelial antigens in both the liver and almost no cases observed in PSC. In
the thyroid. comparison, in the general population the
prevalence of SS ranges from 0.2 to 3%, sim- resulting in fibrosis that can involve the skin
ilar to other autoimmune diseases. The high and other organs but usually not the liver.
coincidence between SS and PBC may be Two major disease subcategories can be dis-
explained by the fact that both diseases share cerned, depending on the extent of skin
some pathogenic mechanisms, with apoptosis involvement: a limited cutaneous form
being likely the important element, and both (lcSSc; formerly CREST syndrome), associ-
are associated with chronic autoimmune epi- ated with limited skin thickening and little
theliitis, leading to destruction of both bile organ involvement; and diffuse cutaneous
ducts and certain exocrine glands [3]. The SSc (dcSSc), involving the skin and multiple
idea is further supported by other similarities internal organs. lcSSc dominates in patients
between both diseases such as female pre- with PBC, as up to 93% of patients with con-
dominance and age of onset in the fifth decade comitant SSc have this subtype. The autoim-
of life. Interestingly, when both diseases mune background of this association remains
coexist, the clinical course is usually mild, as to be elucidated. A positive anti‐centromere
symptoms of SS warrant therapy in only 11% antibody (ACA), a serologic hallmark of SSc,
of cases. Coexistent SS/AIH has been mainly can be observed in 9–30% of patients with
described in a cohort of Asiatic patients, PBC and in 22–25% of all patients with SSc,
among which AIH/PBC was common. The mostly in the limited cutaneous subtype.
clinical significance of the diseases’ Interestingly, patients with PBC and con-
association is not fully understood; however, comitant SSc show higher ACA positivity
anti‐Ro52 positivity has been considered a than in PBC or SSc alone. On the other hand,
negative prognostic factor in these patients. up to one‐fourth of patients with SSc are
On the other hand, among other organs, positive for anti‐mitochondrial antibody
the liver is a common non‐exocrine target in (AMA), although there is no cross‐reactivity
primary SS. Hepatomegaly is reported in between mitochondrial and centromere
2–20% of patients, whereas biochemical liver antigens. Genetic studies have detected asso-
abnormalities, cholestatic rather than hepato- ciations with only a few genes (STAT4 and
cellular, are reported in 49% of patients. These IRF5), which overlap between SSc and PBC
alterations are more common in patients with and which can explain the relative rarity of
positive antibodies to extractable nuclear SSc/PBC coexistence. Another gene,
antigens or evidence of systemic inflamma- PTPN22, that has been associated with SSc,
tion, as shown by an elevated erythrocyte but not with PBC, may infer susceptibility to
sedimentation rate. The abnormalities may the development of SSc in PBC patients.
be either persistent or intermittent but usu- Prognosis in PBC/SSc is not established.
ally have little clinical significance. In a Several studies have shown a significantly
minority of cases, further investigation led to better outcome, slower liver disease progres-
the diagnosis of AILDs or chronic viral sion, and lower rates of liver transplantation
hepatitis C. in PBC/SSc than in PBC alone. Conversely,
an increased mortality due to the coexistence
of SSc in patients with PBC has also been
Systemic Sclerosis reported, mainly due to non‐liver disease
complications. Another study has shown that
Systemic sclerosis (SSc) is an extrahepatic the prevalence of esophageal varices is signif-
manifestation of AILDs with a prevalence icantly higher in non‐cirrhotic patients with
ranging between 1.4 and 12.3% in PBC, and PBC/lcSSc compared with non‐cirrhotic
between 1.2 and 3.5% in AIH, but is seen only patients with PBC alone. In line with these
incidentally in patients with PSC [4]. It is a findings, ACA positivity was highlighted as
systemic CTD characterized by immune‐ a risk factor for an increased rate of
mediated extracellular matrix deposition portal hypertension. Summarizing these
bservations, early surveillance of gastro-

o use of immunosuppressive therapy for SLE
esophageal varices and management of may interfere with the clinical and laboratory
systemic SSc complications, such as findings of underlying AILD, leading to diag-
pulmonary hypertension or heart disorder, nostic delay or occulting the progression of
may lead to improved outcome in patients liver disease. Underlying AILDs, mainly AIH,
with AILDs associated with SSc. may be confirmed in 10–15% of patients
who have liver enzyme abnormalities. Drug‐
induced liver injury (DILI) and SLE‐associ-
Systemic Lupus Erythematosus ated hepatitis (lupus hepatitis) are the most
frequent causes of liver dysfunction in SLE
SLE is another disease that has been associ- patients. Lupus hepatitis is frequently con-
ated with AILDs [7]. This is a systemic auto- fused with lupoid hepatitis, a historical name
immune disease characterized by highly for AIH. These are separate clinical entities
variable multiorgan involvement and the with different clinical courses and prognoses.
presence of autoantibodies such as anti‐ They are compared in Table 16.2. The
nuclear antibodies (ANAs) and, more specifi- discrimination is often difficult, because both
cally, anti‐double‐stranded DNA antibodies conditions may overlap with clinical, labora-
(anti‐dsDNA) and anti‐Smith (anti‐Sm). The tory and systemic presentation. Both lupus
clinical presentation is very diverse and hepatitis and AIH show some similarities,
unpredictable, ranging from a mild disease such as polyarthralgia, ANA positivity,
characterized by arthritis to a severe life‐ hypergammaglobulinemia, hypertransami-
threatening course affecting multiple organs. nasemia, and response to glucocorticosteroid
The majority of SLE cases occur in women of therapy. Lupus hepatitis, however, is a dis-
childbearing age, and therefore the likelihood tinct manifestation of SLE occurring in 3–9%
of coexistence of PBC and SLE is relatively of patients, and frequently associated with
low as PBC affects predominantly middle‐ exacerbation of the underlying disease. In
aged women. Reported frequencies of PBC contrast, no obvious correlation between SLE
concomitant with SLE are less than 3.7%. The activity and the incidence of AILDs has been
largest study of 1032 patients with PBC identified. In cases of lupus hepatitis, eleva-
showed that 2.61% of these patients had con- tion of alanine aminotransferase is usually
comitant SLE, a significantly higher frequency mild and glucocorticosteroid therapy brings
than the 0.48% in the general population. The the values to within a normal range. Serology
coincidence with other AILDs is reported may be helpful in differential diagnosis,
mainly as case reports; systemic studies on wherein anti‐smooth muscle and anti‐liver
AIH and PSC are lacking. kidney microsomal antibodies are found in
Currently, it is not clear whether concomi- AIH, whereas low complement concentration
tant SLE and AILDs share a common immu- and ribosomal P antibodies are observed in
nologic and/or genetic background. Recent SLE. However, liver biopsy is sometimes
genome‐wide association studies have iden- necessary to establish the final diagnosis.
tified risk loci, such as IRF5‐TNPO3, which The prevalence of DILI is relatively high
may be associated with genetic susceptibility (approximately 30%) in SLE patients. The
to both SLE and PBC. causes often include therapeutic drugs for
Patients with SLE have a 25–50% risk of SLE, such as non‐steroidal anti‐inflammatory
developing abnormal liver function tests in drugs (NSAIDs), methotrexate, azathio-
their lifetime. Differential diagnosis of the prine, and other non‐glucocorticosteroids.
liver involvement is crucial for further Moreover, some studies report non‐alcoholic
treatment and prognosis; however, it is often steatohepatitis after exposure to large doses
challenging and requires careful clinical of glucocorticosteroids used for the treatment
investigation including liver histology. The of severe flare‐ups of SLE. The differential
Table 16.2 Differences between SLE‐associated hepatitis and autoimmune hepatitis.
SLE‐associated hepatitis Autoimmune hepatitis
ANA >99% >80% in type 1 AIH

Anti‐SMA 30% 60–80% in type 1 AIH
Anti‐LKM Negative Positive in type 2 AIH
Anti‐ribosomal P Positive Negative
Anti‐dsDNA 80–90% Rarely positive in AIH only. May
be positive in AIH concomitant
with SLE
Hypergammaglobulinemia Common Common, IgG elevated
Complement concentration Low in active disease Usually normal
Liver histology Lobular infiltrates with Interface hepatitis, periportal
paucity of lymphocytes. Mild piecemeal necrosis, lobular
chronic inflammation activity and rosette formation of
hepatocytes
Progression to liver cirrhosis No Yes
Response to steroids Favorable Favorable
AIH, autoimmune hepatitis; ANA, anti‐nuclear antibody; anti‐dsDNA, anti‐double‐stranded DNA; anti‐LKM,
anti‐liver kidney microsomal; anti‐SMA, anti‐smooth muscle antibody; SLE, systemic lupus erythematosus.
Source: adapted from Kaw et al. [9].
diagnosis is often difficult and requires liver Rheumatoid Arthritis

biopsy. The histologies of AIH and DILI have
certain similarities, including interface The literature provides inconsistent data on
hepatitis, inflammatory cell infiltration in the the prevalence of rheumatoid arthritis (RA)
portal area, and centrilobular 3 zone necrosis. in patients with autoimmune liver disorders.
However, in comparison with AIH, in DILI One of the factors that has made it difficult to
eosinophil infiltration is more common, but assess the true prevalence is the fact that
no obvious liver fibrosis is usually described. both clinical entities overlap with several
In terms of the clinical and prognostic clinical and serologic findings. Arthralgia,
significance of SLE associated with AILDs, it osteoporosis and chronic fatigue affect many
has been suggested that patients with both patients with various autoimmune disorders,
PBC and SLE may manifest less extensive including RA and AILDs. Moreover, a sero-
liver damage, reflected by lower gamma‐ logic marker of RA, rheumatoid factor (RF),
glutamyltransferase and IgM levels. Con has been reported to be present in a
versely, liver disease is not a major prognostic significant proportion of patients with PBC;
factor in SLE. Large multicenter studies conversely, AMA positivity is found in up to
have shown that liver dysfunction does not 18% of patients with RA. Finally, the majority
influence morbidity or mortality in SLE of published studies use different classification
patients. One of the possible explanations criteria for RA instead of the standardized
for this observation may be that end‐stage American College of Rheumatology (ACR)/
liver disease with concomitant SLE is generally European League Against Rheumatism
rare. A review of Japanese autopsy registry (EULAR) criteria. In effect, several studies
data for 1468 patients with SLE has shown have pointed toward an association between
that liver cirrhosis was present in only 1.1% PBC and RA, ranging between 1.8 and 5.6%.
of liver specimens [7]. On the other hand, some studies have
i ndicated that the prevalence of RA in PBC is disease (lower hemoglobin levels but higher
comparable to that observed in controls. The serum levels of IgG and ALP, and a higher
frequency of concomitant AIH and PSC ratio of RF positivity). However, this issue has
seems to be even lower. Larger studies are not been studied in detail.
needed to determine the real coexistence of
one disease with the other.
Although liver damage is not seen as a Celiac Disease
significant extra‐articular feature of RA,
18–50% of patients with RA have abnormal Celiac disease (CD) is the most common
liver biochemistry, associated with mostly autoimmune enteropathy. As well as clinical
mild and non‐specific changes in liver his- manifestations of a typical malabsorption
tology. In the vast majority of cases, the syndrome, CD may exhibit several extraintes-
underlying cause of these abnormalities is tinal symptoms [8]. The most common
something other than concomitant AILD, hepatic presentation, seen in 40% of patients,
most commonly associated with hepatotox- is isolated hypertransaminasemia with non‐
icity of the administered medications, mainly specific changes in liver histology (celiac
NSAIDs, methotrexate and sulfasalazine. hepatitis). A gluten‐free diet leads to the
DILI is most often manifested by asymptom- normalization of liver biochemistry and his-
atic elevation of aminotransferases. Severe tologic alterations in most cases. On the other
hepatic injury is rather rare; however, pro- hand CD often coexists with a spectrum of
gressive liver damage including liver cir- immunologically mediated liver diseases. The
rhosis has been associated with methotrexate prevalence of AIH in patients with CD is esti-
treatment. The risk of methotrexate‐induced mated at 1.6%, whereas typical CD autoanti-
liver injury is higher in patients with under- bodies are observed in 3.5% of patients with
lying liver disease, in particular those whose AIH in comparison to 0.35% in the general
treatment was prolonged over 2 years due to population. The relationship between CD and
an association with a high drug dosage. PBC is also well established. Epidemiologic
Genetic and genome‐wide association studies indicate that anti‐endomysium anti-
studies have found some common genetic bodies are positive in 10% of patients with
background of PBC and RA, although the PBC. Finally, in patients with PSC, subsequent
vast majority of implicated genes in both dis- studies also showed an increased risk of CD,
eases do not overlap. Overlapping genes with prevalence of 1.6% and, conversely, a
include HLA‐DQB1, STAT4, IRF5, MMEL1, fourfold to eightfold higher risk of PSC deve
CTLA4, and possibly CXCR5. Individuals lopment in patients with CD in comparison
carrying risk alleles may be more susceptible to the general population.
to developing concomitant RA and PBC. On The genetic and immunologic mechanisms
the other hand, patients with a genetic profile behind the AILD/CD association have been
contributing to RA or PBC, in whom genes extensively studied. The genetic link between
do not overlap, probably will not develop CD and AIH has been described by common
concomitant autoimmune disease. This is combinations of genes coding for class II
suggested by epidemiologic data that show a human leukocyte antigen (HLA) molecules
small proportion of patients with concomi- on chromosome 6. The main genetic marker
tant RA and AILD. of CD, HLA‐DQ2.5, has a strong linkage dis-
The data analyzing the influence of con- equilibrium with HLA‐DR4, HLA‐DR52, and
comitant RA in patients with AILD are HLA‐DR3. The last one is also a major HLA
scarce. Some observations suggest that the risk factor for AILDs. Moreover, several other
coexistence of RA may be a negative prog- gene polymorphisms outside the HLA com-
nostic factor for PBC, as PBC/RA patients plex were found in both CD and AILD,
show laboratory indicators of more severe suggesting shared dysregulated immune

responses contributing to the development of the upper limit of normal), laboratory signs
the two diseases. Moreover, the breakdown of of cholestasis, or physical and ultrasono-
gut–liver axis equilibrium has been suggested graphic signs suggesting a liver disorder. In
as a trigger in the immunopathologic pathway such cases, a thorough hepatologic examina-
toward immune‐mediated liver injury in CD. tion is necessary and liver biopsy should be
It is hypothesized that CD‐associated considered in selected cases. An early diag-
increased intestinal permeability allows the nosis with the appropriate treatment can
immunologically active molecules generated improve the outcome of these patients.
from the cross‐linking between tissue trans-
glutaminase and food or bacterial antigens to
reach the liver through the portal circulation. Inflammatory Bowel Disease
Molecular mimicry between bacterial and
hepatic antigens may further induce immune‐ The close association between PSC and IBD
mediated hepatic damage. Aberrant intestinal is well established [10]. The reported preva-
T‐lymphocyte homing to the liver may lence of IBD in PSC patients is usually esti-
further contribute in this process. mated at 70%, but the frequency ranges from
The clinical impact of CD on the progres- 50 to 99% across different studies, depending
sion of AILD remains uncertain. Studies with on the diagnostic criteria for IBD and geo-
a small number of pediatric patients with graphic differences. European and American
AIH/CD suggest that a gluten‐free diet facil- studies report a higher prevalence of IBD in
itates the achievement of treatment‐free PSC patients in comparison to populations
sustained remission of AIH and, conversely, a from Asia. Conversely, in patients with an
gluten challenge may promote AIH relapses established diagnosis of IBD, PSC is found to
in these children. be much less frequent, occurring in 2–8% of
The association between CD and liver dis- patients with UC and 3% of patients with
orders, both celiac hepatitis and concomi- Crohn’s disease. Intriguingly, IBD is not
tant AILD, deserves specific clinical closely associated with other AILDs, as it
attention. Clinicians need to consider CD in occurs in only about 2–8% of patients with
the differential of abnormal liver biochem- both AIH and PBC.
istry. Moreover, routine liver biochemistry is The strong co‐occurrence of PSC with
recommended in all patients with newly IBD, as well as several other unique features
diagnosed CD. Asymptomatic mild hyper- of PSC‐related IBD, led to the idea that it may
transaminasemia, together with a normal be a distinct entity from IBD, unrelated to
physical and ultrasound examination, PSC in terms of pathogenesis, phenotype and
strongly suggests liver manifestation of CD prognosis. In fact, patients with PSC‐related
rather than a coexistent AILD. Because of UC often demonstrate a right‐to‐left gra-
the non‐specific nature of the findings and dient of colonic inflammation, rectal sparing,
invasiveness of the procedure, in such cases and backwash ileitis. This pattern of inflam-
there is usually no need for liver biopsy. The mation is significantly different from individ-
normalization of liver biochemistry after uals with UC alone. Moreover PSC/IBD
6–12 months of gluten elimination confirms patients typically have an increased inci-
the diagnosis. An alternative etiology for dence of extensive colitis and pancolitis, irre-
liver injury should be explored in those spective of IBD subtype (i.e. UC or Crohn’s
patients with persistent hypertransaminase- disease), although reported symptoms are
mia despite gluten exclusion. Moreover, a usually mild (Figure 16.1). Even in patients
clinical profile not consistent with celiac with Crohn’s disease, most common is
hepatitis also requires an extended examina- colonic involvement rather than ileocolic
tion. These particularly include severe and isolated ileum disease localization.
hypertransaminasemia (more than five times Finally, the inflammatory activity in the
(a) (b)
Pancolitis More severe

Mild clinical clinical
symptoms symptoms
Left-sided
Right-sided colon cancer
colon cancer
Backwash Backwash
ileitis ileitis 7%
Right-to-left 51%
gradient Left-to-right
gradient
Rectal
Proctitis
sparing
95%
30%
Figure 16.1 Phenotypic features of (a) PSC‐related ulcerative colitis compared with (b) ulcerative colitis
without concomitant PSC.
intestine is usually low, which corresponds proctocolectomy and, conversely, IBD can
with milder symptoms, lower frequency occur many years after the initial diagnosis of
of strictures and penetrating disease. PSC or even after liver transplantation.
Intriguingly, PSC patients with concomitant In terms of prognosis, the association bet-
CD have a less progressive liver disease, ween PSC and IBD appears to increase the
favorable transplant‐free survival, and lower risk of developing colon cancer compared to
rate of hepatobiliary malignancy, and hence a the patients with IBD only. Colon cancer
better outcome than patients with PSC/UC associated with PSC and IBD shows distinct
or patients with PSC only. features, including right‐sided localization
Because the coexistence of PSC and IBD (up to 76%) and worse prognosis due to more
influences the IBD phenotype, some distinct advanced stage and younger age at diagnosis.
features of PSC may also be noted. Compared Moreover, the prolonged duration of IBD
to PSC patients alone, patients with PSC/IBD may also be associated with an increased risk
demonstrate more common combined intra- of cholangiocarcinoma in PSC patients,
hepatic and extrahepatic biliary involvement. irrespective of colectomy. It is unclear

Moreover, the diagnosis of biliary disease whether the observed elevated cancer risk
tends to be made more often in males and at may be the effect of the severity of intestinal
a younger age, when compared with PSC‐ inflammation.
only patients. The demographic data with The high coincidence of PSC and IBD sug-
respect to clinical onset of IBD related to gests a common underlying pathogenesis,
PSC are inconsistent. Some studies report most likely influenced by genetic predisposi-
that the mean age for diagnosis of IBD is tion, immune‐mediated mechanisms, and
higher in PSC/IBD compared with IBD alone. possibly altered gut microbiota. In fact, both
Conversely, a recent study indicates that diseases share common antibodies. In
PSC/UC patients have the clinical onset of particular, perinuclear anti‐neutrophil cyto-
UC at a significantly earlier age compared plasmic antibody (pANCA) has been
with patients with UC only. Usually, IBD pre- reported in 65–95% of PSC patients, mainly
dates the diagnosis of PSC; however, PSC in the context of concomitant UC. Recently,
may also be recognized many years after IgG to the serine protease proteinase‐3
(PR3), being a cytoplasmic ANCA target, symptoms. As the majority of PSC/IBD

was detected in up to 44% of patients with patients have clinically mild disease activity,
PSC. Formerly, PR3‐ANCA was detected in histologic sampling should be performed even
patients with IBD, specifically those with when normal endoscopic morphology of the
UC. Furthermore, antibodies against the colon is seen.
pancreatic major granule glycoprotein 2
(GP2), primarily reported in CD, were
recently found in a significant proportion of Conclusions
patients with PSC.
Currently, there are two immunopatho- Given the increased predisposition for extra-
logic hypotheses to explain the PSC/IBD hepatic autoimmune manifestations in
phenomenon: gut lymphocyte homing and the patients with AILDs, awareness and early
leaky gut. The “leaky gut” hypothesis is based investigation is crucial. As some of these
on the idea of autoimmune‐mediated bile EHAIDs negatively impact prognosis, proper
duct injury triggered by toxic or infectious diagnosis and close follow‐up may lead to
agents that gain access to the biliary tree improved outcome. Targeted testing should
through the inflamed and permeable colon. take into account the most frequent concom-
Dysbiosis seen in patients with IBD may con- itant autoimmunity seen in particular AILDs,
tribute to this process. Furthermore, findings including a thyroid panel and specific anti-
such as the development of PSC after colec- bodies (Table 16.3). Upon initial diagnosis,
tomy for IBD, or the onset of IBD after liver thorough specific investigations have to be
transplantation due to PSC, have led to the performed. In patients with PSC, regular
idea of gut lymphocyte homing in the patho- colonoscopies with histologic sampling
genesis of the PSC/IBD phenotype. Activated should be performed even in asymptomatic
lymphocytes from the gut may enter the patients. The management of patients with
enterohepatic circulation and persist as autoimmunity clustering should be under-
memory cells, triggering and maintaining bil- taken by a team of appropriate specialists.
iary inflammation. This idea is supported by
the fact that some chemokines and adhesion
Table 16.3 Initial serologic investigation
molecules are shared by the liver and intestine for suspicion of EHAIDs in AILDs.
and could contribute to lymphocyte binding
at both sites.
Suspicion of extrahepatic
Despite common immunopathologic path- manifestation Initial investigation
ways, genetic predisposition seems not to be
the major factor linking PSC and IBD. Large Thyroid disease TSH, T3, T4
genome‐wide association studies have identi- Sjögren syndrome Anti‐Ro/SSA,
fied almost 200 independent loci associated anti‐La/SSB
with IBD and 16 loci related to PSC; however, Systemic sclerosis ACA
only a few of them showed a robust association. Systemic lupus Anti‐dsDNA,
Given current knowledge, there are several erythematosus anti‐Sm
recommendations for the diagnosis of patients Rheumatoid arthritis Anti‐CCP
with PSC/IBD. The presence of persistent
Celiac disease IgA anti‐tTG
cholestasis obliges the investigation of
concurrent PSC, optimally with magnetic res- ACA, anti‐centromere antibody; anti‐dsDNA, anti‐
onance cholangiopancreatography. Similarly, double‐stranded DNA; anti‐Sm, anti‐Smith; anti‐CCP,
anti‐citrullinated peptide; IgA anti‐tTG,
in patients diagnosed with PSC, lifelong reg-
immunoglobulin A anti‐tissue transglutaminase; SSA,
ular colonoscopies with biopsies to exclude Sjögren syndrome type A antigen; SSB, Sjögren
concomitant IBD or any malignancy should syndrome type B antigen; T4, thyroxine; T3,
be performed, irrespective of gastrointestinal triiodothyronine; TSH, thyroid‐stimulating hormone.
Guidelines for Clinicians ●● In PBC, autoimmune thyroid disease, SS

and systemic sclerosis occur in a significant
●● Patients with AILDs should be routinely proportion of patients.
screened for concomitant EHAIDs, partic- ●● The diagnosis of concomitant EHAIDs
ularly those which occur most commonly should be made according to their respec-
in a given AILD. tive diagnostic criteria.
●● In AIH, thyroid disease of autoimmune ●● In selected cases liver biopsy is crucial in
background remains the most common distinguishing AILDs, CTDs-associated
EHAID. hepatic involvement or drug‐induced liver
●● Ulcerative colitis is a dominating EHAID injuries.
in patients with PSC; however, its clinical
presentation differs in many ways from UC
seen in patients without PSC.
References
1 Guo, L., Zhou, L., Zhang, N. et al. (2017). 6 Gershwin, M.E., Selmi, C., Worman, H.J.
Extrahepatic autoimmune diseases in et al. (2005). Risk factors and comorbidities
patients with autoimmune liver in primary biliary cirrhosis: a controlled
diseases: a phenomenon neglected by interview‐based study of 1032 patients.
gastroenterologists. Gastroenterol. Res. Hepatology 42: 1194–1202.
Pract. 2017: 2376231. 7 Matsumoto, T., Yoshimine, T., Shimouchi,
2 Teufel, A., Weinmann, A., Kahaly, G.J. K. et al. (1992). The liver in systemic lupus
et al. (2010). Concurrent autoimmune erythematosus: pathologic analysis of 52
diseases in patients with autoimmune cases and review of Japanese autopsy
hepatitis. J. Clin. Gastroenterol. 44: registry data. Hum. Pathol. 23: 1151–1158.
208–213. 8 Marciano, F., Savoia, M., and Vajro, P.
3 Selmi, C., Meroni, P.L., and Gershwin, (2016). Celiac disease‐related hepatic
M.E. (2012). Primary biliary cirrhosis and injury: insights into associated conditions
Sjogren’s syndrome: autoimmune and underlying pathomechanisms. Dig.
epitheliitis. J. Autoimmun. 39: 34–42. Liver Dis. 48: 112–119.
4 Rigamonti, C., Shand, L.M., Feudjo, M. 9 Kaw, R., Gota, C., Bennet, A. et al. (2006).
et al. (2006). Clinical features and Lupus‐related hepatitis: complication of
prognosis of primary biliary cirrhosis lupus or autoimmune association? Case
associated with systemic sclerosis. Gut 55: report and review of the literature. Dig. Dis.
388–394. Sci. 51 (4): 813–818.
5 Silveira, M.G., Mendes, F.D., Diehl, 10 Weismüller, T.J., Trivedi, P.J., Bergquist, A.
N.N. et al. (2009). Thyroid dysfunction et al. (2017). Patient age, sex, and
in primary biliary cirrhosis, primary inflammatory bowel disease phenotype
sclerosing cholangitis and non‐ associated with course of primary
alcoholic fatty liver disease. Liver Int. 29: sclerosing cholangitis. Gastroenterology
1094–1100. 152: 1975–1984.e8.
289
17
Symptoms, Chronic Disease, and Patient Management

David Jones
Newcastle University and Newcastle upon Tyne Hospitals, Newcastle upon Tyne, UK
Abstract
The chronic and incurable nature of autoimmune liver disease (AILD) needs to be recognized in
approaching the management of patients. This chapter explores the inherent complexity in terms
of patient impact and how this influences approaches to management, with a particular focus on
symptoms and their impact on quality of life. Symptoms can be stage-specific or non-stage-specific.
The commonest symptom in all AILD types is fatigue. Pruritus in cholestatic AILD is the symptom
that is easiest to treat, with sequential intervention approaches that are well described in treatment
guidelines. In AILD, the complex of symptoms related to fatigue has the most challenging effects
on quality of life. The critical issue is that the characteristic symptoms of AILD can each lead to
behavior patterns that can isolate people. There is a real need for the AILD community to explore
new approaches to structured care delivery utilizing digital care pathway approaches.
Keywords autoimmune liver disease; intervention approaches; non-stage-specific symptoms;
patients management; stage-specific symptoms
Key Points
●● Symptoms are an important aspect of important to prevent the former, but may
autoimmune liver disease (AILD) and in have little or no impact on the latter.
some patients are the major problem. They ●● The commonest symptom in all AILD
should always be considered when devel- types is fatigue. This is typically very dif-
oping management plans for individual ficult to treat. There is no single effec-
patients. tive therapy, although a structured
●● Symptoms can be stage‐specific (i.e. asso- supportive approach can improve
ciated with the development of cirrhosis) quality of life.
or non‐stage‐specific (i.e. present at any ●● Pruritus in cholestatic AILD is the
stage of the disease process). Effective symptom that is easiest to treat, with
treatment of the underlying disease is sequential intervention approaches that
are well described in treatment guide- ●● Social isolation can be an important

lines. There is evidence of undertreatment consequence of the symptoms of AILD.
in normal clinical practice, with failure by Patients should be warned about the
some clinicians to implement optimal importance of this and the need to develop
management. strategies to avoid it
Background Symptoms and quality of life impairment

are complex in AILD, broadly dividing into
The autoimmune liver diseases (AILDs) are those impacts which are directly related to
all chronic in nature and none can be “cured” the disease process (abdominal discomfort
using conventional treatment approaches, all arising from ascites, septic features arising
of which focus on disease control rather than from bacterial cholangitis in primary
reversal. Even transplantation represents a sclerosing cholangitis, etc.) and those which
shift in disease state rather than a cure, and is are indirectly linked (fatigue in particular).
accompanied by the increasingly recognized This dichotomy is important because it helps
risk of disease recurrence. The chronic and to explain the limitations of disease progres-
incurable nature of disease needs to be rec- sion‐controlling treatment in the domain of
ognized in approaching the management of symptoms and quality‐of‐life impairment.
patients. The search for optimal treatment Effective disease‐controlling treatment can
can lead to an excessive focus on specific limit the development of the former (e.g.
agents and approaches which, unless they patients who do not develop cirrhosis typi-
bridge the divide to fundamentally disease‐ cally do not experience ascites and its conse-
modifying paradigms, will only ever be part quences, patients with effectively managed
of the approach to management. The aim of extrahepatic strictures have a much reduced
this chapter is to explore the inherent com- risk of cholangitis) but typically has no
plexity of AILD in terms of patient impact impact on the latter.
and how this influences approaches to To add to the complexity of quality‐of‐life
management, with a particular focus on impairment in AILD, the potential impacts
symptoms and their impact on quality of life. of the treatments we use should be consid-
ered. Although ursodeoxycholic acid
(UDCA), the mainstay of treatment in pri-
Goals of Treatment mary biliary cholangitis (PBC), and with less
justification primary sclerosing cholangitis
The goals of treatment in AILD are, as is the (PSC), is well tolerated, there can be issues
case with any disease state, to prevent short- with second‐line therapy in PBC, with
ening of lifespan and to ensure that life is as decompensation risk in Childs C cirrhotic
close to normal in terms of its quality as pos- patients in particular with both obeticholic
sible. In terms of lifespan the goal is to prevent acid (OCA; licensed second‐line therapy)
or slow the progression to cirrhosis and its com- and fibrates (unlicensed but widely used)
plications, and to effectively manage those com- and pruritus risk with OCA [1]. In autoim-
plications to reduce morbidity and risk to life. mune hepatitis (AIH) it is becoming clear
Although much store is placed in transplanta- that treatment side effects are significant:
tion as an effective treatment for AILD, it is both corticosteroids and calcineurin inhibi-
limited in terms of its availability, is associated tors are associated with significant side
with intra-operative and postoperative risks, and effects that can limit health quality and
is not as effective at returning quality of life to which may contribute, through impacts on
normal as is sometimes thought to be the case. treatment adherence, to the poorer o utcomes
Chapter 17 Symptoms, Chronic Disease, and Patient Management 291
seen in this disease in practice than might be and language used is understandable to
anticipated given the apparent effectiveness patients and captures the specific impacts of
of the drugs [2]. the relevant disease state. The optimal
approach is a patient‐derived measure vali-
dated within the target population. In the
ymptoms, Quality of Life,
S area of AILD there is a high‐quality patient‐
and Health Utility: Key Concepts derived and validated measure, the PBC‐40
[4], which is widely used in research studies
The study of the impact of disease phenotype and clinical trials (a short‐form version for
on the patient’s quality‐of‐life experiences use in the clinic setting for symptom screen-
invokes the concepts of symptoms them- ing is under development). An equivalent
selves, quality of life assessment, and health measure is under development in PSC. As
utility. In simple terms the former can be yet there has been no work on measures in
regarded as the individual disease phe- AIH that we are aware of. All subjective mea-
nomena and the latter two measures of the sures can be prone to ceiling effects, where a
downstream impact of symptoms on the lives patient’s concept as to the worst symptom
of patients [3]. There is no automatic con- severity they can imagine is dictated by their
nection between symptom presence and experience to date; in simple terms, if a
downstream quality‐of‐life impairment. An patient scores a symptom as 10/10 because
example is pruritus in PBC where frequency they cannot envisage it getting worse, then if
of pruritus report, for example in the setting it does actually get worse it cannot be quanti-
of adverse event reporting, exceeds the take‐ fied using the measure.
up rate for pruritus therapy. Although this Objective symptom measures are highly
reflects, to some degree, poor clinical prac- attractive to the research community and to
tice, a significant element is patients regulators because of the degree to which
describing a symptom which is not severe they seemingly directly ascribe a value to a
enough in their eyes to warrant treatment as disease state. Caution must be applied, how-
it does not impact on their quality of life. It is ever, as the physical measure can be affected
critical to understand this concept to avoid by patient response to the symptom (e.g.
an over‐focus on low‐impact symptoms, the patients with pruritus who manage to control
treatment of which will not actually lead to their scratching, patients with fatigue who
patient improvement. make themselves keep going and thus show
reasonable physical activity levels).
A particular challenge in symptom
Symptoms
assessment relates to those states which can
All patients and clinicians are familiar with be part of normal life (e.g. tiredness or
the concept of symptoms (defined as fatigue) and which may be exaggerated by a
“physical or mental features which are disease state. In this setting, patients are
regarded as indicating a condition of disease, often left making a value judgment as to what
particularly such features that are apparent component of their experience results from
to the patient”). Symptom severity can be the disease and what is part of normal life. In
assessed using subjective and objective contrast, those symptom states which would
approaches. Subjective assessments rely on not be part of normal experience, exempli-
the patient’s stated view about symptom fied by pruritus, can present less of a
severity using simple scales (visual analog challenge. An important corollary of the
scales or numerical rating scales giving a challenge around those symptom states
score out of 10 to symptom severity) and which are and are not part of a continuum or
more complex assessment questionnaires. In normality is that the former can sometimes
all instances it is essential that the approach be seen as less “real” or disease‐related than
the latter. This issue is exemplified by the achieved in the decision‐making process
different perceptions of “reality” and the regarding drug reimbursement. Both PBC
importance of fatigue and pruritus among and AIH have recently undergone health
the PBC clinical community. utility assessments in the UK (PSC remains
to be studied using this approach). Both AIH
and PBC can be associated with normal
Health‐related Quality of Life
health utility (>0.95, 1.0 being deemed
The concept of quality‐of‐life measurement unachievable), suggesting that this should be
can be even more challenging than symptom a reasonable goal for management. Both are,
assessment, including, as it does, elements that typically, associated with significantly
go well beyond the impact of any disease or impaired utility. In the case of PBC the dom-
clinical condition. For this reason, the more inant effect is through non‐stage‐specific
focused concept of health‐related quality of life symptoms (in particular fatigue and
(HRQoL) is used, a standard definition for cognitive symptoms), although where pre-
which is “the measurable impact of a person’s sent end‐stage disease features such as
perception of his or her health and the effect ascites are associated with a substantial
that produces on satisfaction with life and well‐ additive effect. UDCA use, interestingly, is
being.” The more complex, abstract nature of associated with significantly higher levels of
this means that simple numerical scales are not health utility irrespective of UDCA response
helpful (their use would rely on all users status and of any impact on individual symp-
ascribing the same meaning to the term “quality toms. In AIH, although cirrhosis and its
of life”). Well‐validated, generic quality‐of‐life complications again have a marked effect,
measures, such as the SF‐36 and EQ5D, have the largest driver for poor utility is the use of
been used in the study of AILD, and have sug- corticosteroid and calcineurin inhibitor
gested levels of HRQoL impairment similar to therapy (independently of disease activity).
those seen in other chronic disease settings. As
well as being cumbersome to use in practice,
the scoring is not intuitive making judgment as ymptoms and Their Management
S
to the nature of quality‐of‐life impairment dif- in AILD
ficult to reach without detailed analysis.
However, they are the bedrock of quality‐of‐life The symptoms seen in AILD can be divided
assessment as required by regulatory bodies. into stage associated and non‐stage
associated.
Health Utility
Impact of Disease‐modifying Therapy
Health utility represents an analysis of
on Symptoms in AILD
HRQoL measures in which the scores are
converted to a linear scale ranging from 0 Although clearly optimal disease treatment
(equivalent to death) to 1 (perfect health) in AILD is necessary for the best overall out-
[2]. Health utilities make the study of comes, the extent to which it directly
HRQoL less abstract and, critically, allow us improves symptoms is limited; this is
to understand the effectiveness of interven- something which should be considered when
tion in the context of health economic eval- discussing expectations of treatment out-
uation (cost‐effectiveness). One year at a comes with patients. There is no evidence
health utility of 1 represents one quality‐ that UDCA improves either fatigue or pru-
adjusted life‐year (QALY). Reimbursement ritus in PBC [5,6], whilst OCA does not
bodies such as the National Institute for improve fatigue and can worsen pruritus [1].
Health and Care Excellence (NICE) in the Recent data from the UK suggests that
UK will define acceptable costs per QALY although ongoing disease activity is associated
with worse quality of life, so is the use of with the additional specific impact of pru-
long‐term steroid or calcineurin inhibitor ritus in the two cholestatic conditions.
therapy; indeed long‐term steroid therapy is
the major factor associated with impairment Pruritus
of quality of life in this disease [2]. UK‐wide data suggest that up to two‐thirds
of PBC patients will experience pruritus dur-
ing the course of their disease, with around
Stage‐associated Symptoms
half of these experiencing pruritus at any one
In all liver disease states yet studied, symptom time [7]. It is particularly prominent in the
impact is greater and HRQoL lower in cir- typically younger patients with aggressively
rhotic than in pre‐cirrhotic states [3]. The ductopenic disease. Mechanistic studies have
complications of advanced AIH closely pointed to the role of lysophosphatidic acid
mirror those seen in other hepatocellular (LPA) as the direct pruritogen, generated by
diseases, with encephalopathy (characteristic the actions of the enzyme autotaxin (ATX),
cognitive dysfunction and associated fatigue) which is present at elevated levels in patients
and ascites (abdominal discomfort and with cholestatic itch and which has a com-
impaired mobility) to the fore. In the chole- plex functional inter-relationship with bile
static conditions PBC and PSC, symptoms acids. Despite the apparent importance of
from advanced disease are less prevalent. ATX and LPA in the pathogenesis of itch,
This largely reflects the rapid, often understanding of the pathway has not, as yet,
exponential progression seen in cholestatic led to specific therapy. Cholestatic itch can
disease from compensated cirrhosis to death. also occur in PSC but its impact, perhaps
If PBC or PSC patients are seen with decom- paradoxically, appears lower than in PBC.
pensated cirrhosis, then the key question is In PSC, onset or worsening of itch should
whether they are suitable for transplantation always trigger reassessment of the bile duct
(the only effective intervention in this state) given the potential for stricturing as a driver
rather than how to ameliorate individual of cholestasis. The possibility of gallstone
symptoms. disease should also be considered in PBC
Because of the detrimental impact of cir- where there is a specific disease association.
rhosis on quality of life as well as length of Itch assessment in clinical practice can be
life, effective treatment to slow or prevent performed using a numerical rating scale
disease progression is essential in the (rating the impact from 0 to 10). The presence
management of HRQoL in AILD even if, as of pruritus is not always an indication for
already outlined, it may well not directly treatment as many patients with milder itch
improve other symptoms. are not troubled by it. Significant discomfort,
In PSC uniquely, occult cholangitis can be social impacts resulting from embarrass-
a driver of systemic symptoms, particularly ment around scratching in public, or
in older patients, and should always be con- significant night‐time itch (which can be a
sidered and should always be screened for. major contributor to poor sleep and fatigue)
are all indicators of a need for treatment.
The management of cholestatic itch is
Stage‐independent Symptoms
described in detail in national and interna-
The bulk of the HRQoL impairment in AILD tional guidelines ([8] (Table 17.1). There is no
comes not from the features of advanced dis- evidence to suggest that UDCA specifically
ease but from systemic symptoms which can improves pruritus and OCA can worsen pru-
be seen throughout the disease course. ritus or induce it in previously unaffected
Common themes around fatigue, cognitive patients. There are reports that bezafibrate
dysfunction, and emotional and social may improve pruritus, although published
dysfunction emerge across all the diseases, placebo‐controlled trial data with pruritus as
Table 17.1 Stepwise approach to the management of pruritus in PSC or PBC.
Intervention Notes
Step 1 Imaging to exclude obstruction Especially important in PSC because of structuring risk
to bile duct
Step 2 Bile acid sequestrant Space away from UDCA
(colestyramine 4 g up t.d.s.) Can be poorly tolerated because of taste. Consider
chilling or adding fruit juice
Step 3 Rifampicin 150 mg daily Check LFT after 4 weeks
Stop if significant ALT rise
Increase to maximum 600 mg daily if needed
Step 4 Naltrexone 50 mg daily Can be associated with “opiate withdrawal‐like” response
Cannot use with opiate analgesia (obviously)
Can be a problem in chronic pain states
Step 5 Sertraline 50 mg daily Can be increased if ineffective
Trial evidence positive but experience in practice
suggests limited benefit
Step 6 Experimental drug treatment Opportunity will depend on local activity
Bezafibrate has some promise
Step 7 Physical intervention Plasmapheresis, MARS or nasobiliary drainage
All are effective in the short term but costly, invasive and
with risk
Should only be considered as short‐term “salvage”
procedures
Step 8 Transplantation Instant itch control
Challenge to justify if liver function good
This represents the approach to systematic management of itch as recommended in relevant guidelines. It should
obviously be followed in addition to providing optimal underlying disease management. Steps 3 and 4 can be used
interchangeably based on local preference and experience.
ALT, alanine aminotransferase; LFT, liver function tests; MARS, molecular adsorbent recirculating systems; UDCA,
ursodeoxycholic acid.
an end point are currently lacking. This is key and this mirrors clinical experience of
given the scale of placebo effects in symptom‐ limited if any benefit.
modifying trials in PBC. All second‐line treatments have limitations
Bile acid sequestrants are licensed therapy and it is important to give an adequate trial of
and remain the first line of therapy despite bile acid sequestrants before moving to them.
limitations. Colestyramine has the best sup- Rifampicin, which probably acts through
porting evidence (although this is still very activation of the pregnane X receptor, has the
limited) but is compromised in practice by best supporting evidence (placebo‐controlled
its poor tolerability. Chilling the mixed resin, trials with positive meta‐analysis). However,
or adding fruit juice, can increase accept- tolerability can again be an issue, with rare
ability though many patients still struggle. It cases of significant liver dysfunction and
is also important to space colestyramine commoner gastrointestinal disturbance.
away from UDCA to prevent sequestration Patients should be started at low dose (150 mg
of UDCA in the bowel. Colesevelam, a tablet daily) with liver function monitoring at 2–4
formulation of bile acid sequestrant, offers weeks. If significant elevation of alanine ami-
theoretical advantages. However, a placebo‐ notransferase (ALT) is seen, then the agent
controlled trial failed to show benefit should be discontinued. If well tolerated and
effective, it should be continued at the plasmapheresis, molecular adsorbent recir-

starting dose. If the effect is incomplete, it culating systems, and nasobiliary drainage
can be cautiously increased to a maximum have short‐term benefits but are costly and
dose of 600 mg daily. Patients should be invasive, and in the case of nasobiliary drain-
warned that it can turn urine, tears and sweat age are associated with significant morbidity.
orange! These drawbacks mean that their use should
Oral opiate antagonists (e.g. naltrexone) be restricted to salvage intervention in pat-
have supportive evidence and are viable ents with extreme itch resistant to other
third‐line or alternative second‐line agents. modalities and undertaken only in special-
The role played by opiates in the pruritus sig- ized centers.
naling pathway is currently unclear. Proposed
models include release of endogenous opi- Fatigue and Cognitive Symptoms
ates by stressed biliary epithelial cells and a In AILD, the complex of symptoms related to
role in peripheral synaptic regulation. There fatigue have the most challenging effects on
is certainly evidence to suggest that there is quality of life [9]. Patients can variably expe-
elevated endogenous opiate tone in chole- rience symptoms relating to peripheral
static patients as opiate antagonism with fatigue (often characterized as loss of energy,
agents such as naltrexone can be associated loss of power, or as the “batteries running
with a significant opiate withdrawal reaction down” [10]) and to central fatigue, often
(“cold turkey”). This can be disturbing for associated with cognitive issues (loss of
patients and represents a major limitation for concentration and short‐term memory [11])
the class of drugs (along with increased pain and loss of motivation (crossing over into the
from intercurrent pain‐associated disease spectrum of depression). These features can
processes). It can be mitigated to a degree by occur in any of the AILDs (although more
starting at very low oral doses (limited by the frequently seen in PBC than AIH and PSC) at
degree to which the tablets can be divided) or any stage of the disease process. Clearly, in
even escalating intravenous naloxone. patients with advanced disease the impacts
For patients unable to tolerate, or under‐ of overt and minimal hepatic encephalopathy
responsive to, bile acid sequestrants, rifam- can complicate the clinical picture and
picin and opiate antagonists, the treatment should always be considered. It is incorrect,
landscape is complex. Sertraline has placebo‐ however, to assume that the presence of
controlled trial data to suggest benefit but cognitive symptoms in particular in chole-
this is not widely used because clinical expe- static patients automatically implies the
rience has been largely disappointing. In presence of advanced disease. Linear studies
contrast, gabapentin (intuitive because of the in PBC patients following their clinical phe-
overlap functionally between itch and notype through transplantation show a
peripheral pain nerve fibers) has trial evi- significant improvement but a failure to
dence suggesting lack of benefit but is widely return to normal. The interpretation of this
used in specialist clinics because of experi- observation is that the end‐stage associated,
ence of improvement and good tolerability. encephalopathy‐linked phenomena improve
Trials are ongoing, particularly of bile acid but the non‐stage‐associated changes do not.
uptake inhibitors active in the terminal ileum The epidemiology of fatigue and related
with evidence of benefit. features in the three conditions depends, to
For patients with intractable itch resistant an extent, on the assessment modality
to drug therapy, transplantation represents a used. Using well‐validated generic fatigue
highly effective intervention. However, it is assessment tools the prevalence of fatigue
essential that patients have had all medical appears to be higher, and the impact greater,
options fully explored before exploring this in PBC than in AIH and PSC, although
option. Physical interventions, such as patients with each of the conditions can be
impacted. Even in PBC, where around 50% of improve with exercise therapy; experience
patients experience fatigue, there are very suggests that this can be a challenging area
clearly unaffected patients. Fatigue severity, to obtain patient input.
with the exception of the added element seen ●● The central component is strongly associ-
in end‐stage disease, does not relate to dis- ated with autonomic dysfunction, sleep
ease activity or stage in PBC or PSC; active disturbance, and some features of depres-
inflammatory disease in AIH can be associ- sion; however, fatigue in AILD is not “just
ated with worse fatigue, although patients in depression,” a contention which can harm
biochemical remission can also be impacted, the relationship between patient and clini-
especially if treated with long‐term steroids. cian. There is clear evidence of organic
In each disease state severity of fatigue can CNS abnormality (in PBC in particular,
range from mild, raising questions as to the which is the best characterized of the dis-
extent to which this is disease‐related at all, to eases), with structural and functional
a debilitating symptom with profound impact imaging changes and neurophysiologic
on functional status. It is therefore important abnormality. The functional imaging
to quantify fatigue if it is to be managed effec- changes strongly suggest abnormality in
tively. As is the case with HRQoL in general, connectivity of the hippocampus with
the abstract nature of the symptom means higher brain areas. These findings raise the
that an overarching self‐assessment is possibility that there is a chronic organic
unhelpful and validated measures such as the process impacting on the brains of PBC
PBC‐40 fatigue domain or the Fatigue Impact patients. Recent studies have suggested
Scale should be used. that imaging change is present within
At present there is no consensus as to the months of the diagnosis of disease. As the
mechanism of fatigue in AILD and no specific importance and consequences of this pro-
therapy. The literature supports the follow- cess come to be understood, so it may
ing concepts. impact on disease treatment paradigms in
PBC, with a move toward better treatment
●● There is minimal if any link between earlier in the disease process.
fatigue severity and conventional markers ●● Critically, fatigue is both inherently com-
of disease severity (with the exception of plex (with many potential contributors to
the fatigue worsening seen in advanced the phenotype within and between
disease). patients) and can be sequential in nature:
●● There are both peripheral and central what starts as a central process could
components, with significant variation in become peripheral through the impact of
the relative contribution between patients. deconditioning. This implies that treatment
Central fatigue can overlap with cognitive of the central process, were it to become
symptomatology and includes loss of moti- available, might lose its effectiveness later
vation to undertake activities. Peripheral in the disease process when the secondary
fatigue is characterized by a sensation of deconditioning process has come to
loss of peripheral muscle power with diffi- predominate.
culty sustaining repeat exercise and even
muscle pain. There is currently no specific therapy for
●● The peripheral component is associated fatigue in AILD. Development and evalua-
with a significant bioenergetics abnor- tion of therapies is also challenging given the
mality in muscle, with evidence of over‐ complex phenotype, variation in the contrib-
reliance on anaerobic pathways and uting factors to the phenotype, the challenge
resulting lactic acid build‐up. This compo- in measuring it and, as in the case of pruritus,
nent may represent deconditioning significant placebo effect. However, a sup-
secondary to central fatigue and may portive management approach can help to
Table 17.2 Stepwise approach to the management of fatigue in PBC.
Intervention Notes
Step 1 (and Empathize with the Enter into a positive engagement with them. Do
throughout) patient not feel sorry for them but try to understand the
impact the symptom has on their lives. Reassure
them you are there to help them
Step 2 Screen for intercurrent Associated autoimmune conditions (e.g.
or associated conditions hypothyroid disease or autoimmune anemia)
also causing fatigue and Age‐associated disease such as heart failure or type
treat appropriately 2 diabetes mellitus
Step 3 Screen for associated These are features seen in sever fatigue and which
processes worsening worsen the patient experience. Addressing these
fatigue or its impact and will not reverse fatigue but will assist its control
treat appropriately Depression: Typically with SSRI
Autonomic dysfunction: Classically vasomotor
dysfunction (postural dizziness). Look for
inappropriate antihypertensives. Increasing fluid
intake can be helpful. If in doubt seek specialized
autonomic service input
Sleep disturbance: Typically daytime somnolence
that can be severe enough to impact on work.
Consider the possibility of obstructive sleep apnea.
Consider seeking sleep clinic input
Step 4 Help patient with coping It is critical that patients “own” the problem and
strategies the solution
Advise around energy conservation
Advise around scheduling (fatigue is typically
worse later in the day
Advise around maintaining social structures (social
isolation can worsen quality of life)
Advise around the potential value of exercise in
minimizing the peripheral manifestations of
fatigue
Step 5 Experimental drug Opportunity will depend on local activity
treatment
This represents a suggested approach to systematic management of fatigue in PBC. This goes beyond current guideline
recommendations and should therefore be regarded as experimental. There are fewer data around the impact and
pathogenesis of fatigue in PSC and AIH, although the basic approach is still applicable. The approach should again be
followed in addition to providing optimal underlying disease management. Consideration in AIH should be given to
moving away from long‐term steroid maintenance therapy given its association with worsened quality of life (although
clearly change in disease management should not be at the expense of effective disease control). Other than the stage‐
specific fatigue and cognitive symptoms of advanced liver disease, fatigue does not improve with liver transplantation
and this should not be considered as a potential treatment approach without other indications.
ameliorate the symptom (Table 17.2). The ●● Treating fatigue‐associated comorbidity:

key elements are as follows. all the AILDs are associated with other
conditions, frequently autoimmune, which
●● Optimal disease management: progression
are themselves associated with treatable
to cirrhosis is universally associated with
fatigue (examples include hypothyroidism
worsening in fatigue and associated quality
and anemias of autoimmune origin). There
of life. Preventing this is key.
are also age‐relevant comorbid conditions positive in their approach. Expressing views
that are fatigue associated (e.g. type 2 suggesting that fatigue is not real, or is a
diabetes). Potential confounding condi- lesser problem than cirrhosis (or indeed pru-
tions should be sought and treated ritus) can lead to a loss of relationship bet-
effectively. ween patient and clinician. Those clinicians
●● Treating associated processes: fatigue is who say they “never see fatigue” might use-
associated with sleep disturbance, fully consider the possibility that patients
autonomic dysfunction and depression. have learned to not mention it because of the
Clearly, cholestatic disease can also be response doing so in the past has elicited.
characterized by pruritus which can dis- One of the single most desirable advances
turb sleep and thus contribute to fatigue. in AILD would be a direct therapeutic agent
All such features should be explored and able to break the cycle of fatigue. At present
treated. Sleep clinic or formal autonomic there is no such licensed agent. The most
assessment (through the “falls and syn- interesting potential therapy is S‐adenosyl-
cope” services which are now widespread) methionine, for which older data suggest a
may be appropriate. One simple step is for beneficial effect in fatigue and for which
the liver clinic to ensure that there are no there are intriguing emerging data to suggest
direct contributing factors to both poor specific deficiency in fatigue‐associated
sleep (reflux, late evening caffeine‐ AILD states. Further trials with mechanistic
containing drinks or pruritus) and as well as symptom evaluation to mitigate the
autonomic dysfunction (excessive antihy- issues with placebo effects are warranted.
pertensive therapy).
●● Coping strategies: ultimately the symptom
Social Isolation Symptoms
is currently untreatable and it is critical that
There are emerging data from PBC to sug-
patients develop strategies that allow them
gest that although individual symptoms
to mitigate the effects. In essence this
clearly impact on patients, there are modi-
revolves around patients “owning the
fiers that limit their overall impact on per-
problem” and, supported by clinicians and
ceived quality of life. The most important of
others, developing coping strategies.
these is social isolation [12]. In simple terms,
Critical elements include pacing them-
highly fatigued PBC patients typically retain
selves (regarding energy as a valuable
good perceived quality of life provided they
resource to be used to its best advantage)
do not have symptoms of social isolation.
and daytime organization (fatigue in AILD
The critical issue is that the characteristic
typically worsens during the course of the
symptoms of AILD can each lead to behavior
day meaning that important activities in the
patterns that can isolate people. Pruritus can
morning should be scheduled for the morn-
result in people avoiding social situations
ing; night shift work should be avoided if at
because of embarrassment about scratching.
all possible). Exercise is beneficial (although
In the case of fatigue, people can look to con-
patients frequently lack the confidence to
serve energy by stopping working or avoiding
do it) and will help mitigate the effects of
social activities (more so if they take place in
deconditioning. As outlined in the follow-
the evening when fatigue might be at its
ing section, social isolation can be a
worse). The net result, however, can be
significant issue and its avoidance through
fatigue benefit at the expense of worsened
coping strategies is key. Patient groups such
overall life quality (“robbing Peter to pay
as the PBC Foundation and LIVErNORTH
Paul”). It is critical, as part of the approach to
in the UK can play a major role.
developing coping strategies, that patients
Behind all of these approaches is a need for are made aware of this risk and that the
clinicians to be empathetic to patients and importance of maintaining social networks is
emphasized. Patient groups may play a key The emerging complexity reflects increasing
role, as can online groups. opportunity to treat patients better. However,
each of the AILDs is rare with only limited
numbers of specialist centers. This raises
ffective Care Delivery
E important questions as to how cutting edge
in AILD therapy reaches all patients in an equitable
fashion. Currently, care delivery largely uses
Ultimately, the AILDs remain non‐curable traditional models. There is a real need for
chronic diseases. We are increasingly able to the AILD community to explore new
control the progression of disease and to approaches to structured care delivery uti-
limit the impact of symptoms. The lizing digital care pathway approaches.
approaches needed to do that are, however, Without a revolution in care delivery to
increasingly complex with multiple treatment mirror the revolution (in PBC at present and
domains to be considered and an increasing AIH in the near future) in treatment options,
number of treatment options in each domain. we are destined to fail to reap the benefits.
References
1 Nevens, F., Andreone, P., Mazzella, G. et al. 7 Hegade, V.S., Mells, G.F., Fisher, H. et al.
(2016). Obeticholic acid in primary biliary (2019). Pruritus in primary biliary cholangitis
cholangitis/cirrhosis: a phase 3 trial. N. Engl. is common and under‐treated: results from
J. Med. 375: 631–643. 2194 patients in the UK‐PBC cohort. Clin.
2 Wong, L.L., Fisher, H.F., Stocken, D.D. et al. Gastroenterol. Hepatol. 17: 1379–1387.e3.
(2018). The impact of autoimmune hepatitis 8 European Association for the Study of the
and its treatment on health utility. Liver (2017). EASL clinical practice
Hepatology 68: 1487–1497. guidelines: the diagnosis and management
3 Orr, J.G., Homer, T., Ternent, L. et al. (2014). of patients with primary biliary cholangitis.
Health related quality of life in people with J. Hepatol. 67: 145–172.
advanced chronic liver disease. J. Hepatol. 9 Goldblatt, J., Taylor, P.J.S., Lipman, T. et al.
61: 1158–1165. (2002). The true impact of fatigue in
4 Jacoby, A., Rannard, A., Buck, D. et al. primary biliary cirrhosis: a population
(2005). Development, validation and study. Gastroenterology 122: 1235–1241.
evaluation of the PBC‐40: a disease specific 10 Hollingsworth, K.G., Newton, J.L.,
health related quality of life measure for Robinson, L. et al. (2010). Loss of capacity to
primary biliary cirrhosis. Gut 54: recover from acidosis in repeat exercise is
1622–1629. strongly associated with fatigue in primary
5 Carbone, M., Mells, G., Pells, G. et al. biliary cirrhosis. J. Hepatol. 53: 155–161.
(2013). Sex and age are determinants of 11 Newton, J.L., Hollingsworth, K.G., Taylor, R.
theclinical phenotype of primary biliary et al. (2008). Cognitive impairment in primary
cirrhosis and response to ursodeoxycholic biliary cirrhosis: symptom impact and
acid. Gastroenterology 144: 560–569. potential aetiology. Hepatology 48: 541–549.
6 Mells, G., Pells, G., Newton, J.L. et al. 12 Dyson, J.K., Wilkinson, N., Jopson, L. et al.
(2013). The impact of primary biliary (2016). The inter‐relationship of symptom
cirrhosis on perceived quality of life: the severity and quality of life in 2055 patients
UK‐PBC national study. Hepatology 58: with primary biliary cholangitis. Aliment.
273–283. Pharmacol. Ther. 44: 1039–1050.
301
Index
Note: Page references in italics refer to Figures; those in bold refer to Tables
1,25‐dihydroxyvitamin‐D3 adaptive immunity anti‐asialoglycoprotein

(D3) 34 adaptive immune functions receptor antibody
2‐nonynoic acid 97 of the liver 25–8, (ASGPR) 79
2‐octynoic acid 97 26–7 antibiotics 11, 35, 80, 146
2‐oxo‐acid dehydrogenase role of liver as an adaptive in primary sclerosing
multienzyme immune organ 28 cholangitis
complex 78 adenosine triphosphate 150–1, 152
2‐oxoglutarate dehydrogenase (ATP)‐dependent antibody‐dependent cell‐
complex (OGDC) multidrug resistance mediated
78–9, 94, 126 protein (MRP)‐2 7 cytotoxicity
3‐hydroxy‐3‐methylglutaryl‐ adiponectin 223 (ADCC) 23
CoA (HMG‐ alanine aminotransferase 8, anti‐centromere antibodies
CoA) 8 111, 135, 191, 265, (ACA) 69
6‐bromohexonate 97 281, 292 anti‐ENA antibodies 69
6‐mercaptopurine (6‐MP) 112 albumin 8 antigen‐presenting cells
alcoholic liver disease (APCs) 23, 31, 37
a 79, 109 anti‐gp210 70
abdominal ultrasound alendronate 226, 228 anti‐Jo‐1 69
in autoimmune alkaline phosphatase (ALP) anti‐inflammatory treatment
pancreatitis 171 18, 19 in primary sclerosing
in autoimmune sclerosing age‐related differences cholangitis 152
cholangitis 111 in 110 anti‐La 69
in primary biliary in primary biliary anti‐liver cytosol type 1
cholangitis 128 cholangitis 126, antibody (anti‐LC1)
acinus 9 236, 246 67, 69, 75, 76, 106,
acute liver failure 8 in primary sclerosing 107, 109
AMA in 79 cholangitis anti‐liver‐kidney microsomal
ANA in 73 144, 221 antibody 74–5, 106
pediatric‐onset autoimmune α1‐antitrypsin 8 anti‐LKM1 67, 69, 74,
hepatitis and 109, American Lectures in Living 75, 78
110 Chemistry 2 anti‐LKM2 74
see also seronegative acute annexin 11 164 anti‐LKM3 74, 75
liver failure anti‐actin antibodies anti‐LP 76, 269
adalimumab 152, 163 73–4 anti‐LSP antibody 79
302 Index
anti‐mitochondrial antibodies apoptosis 14, 15, 36 HLA locus associations

(AMAs) 66, 78–9, Ascaridia galli 96 51
84–5, 90, 93, aspartate aminotransferase immune serology in 69
95–7, 100 8, 18 pathophysiology
in autoimmune hepatitis aspartate aminotransferase/ 106–7
110 alanine aminotrans- pediatric‐onset
liver transplantation ferase ratio 18 presentation 109
and 135 ATP‐binding cassette (ABC) prognosis 114–15
in primary biliary transporters 11 treatment 113
cholangitis AT‐rich interactive domain‐ first‐line therapy 113
124, 125, 126, containing protein 3A second‐line options
197, 246 (ARID3A, 113–14
anti‐MND 69–70 19p13.3) 58 withdrawal of
anti‐neutrophil cytoplasmic atypical severe combined therapy 113
antibody immunodeficiency Type 1 66, 106, 108
(ANCA) 65, 66, 86 Type 2 66, 74, 106, 108
77–8 autoantigens 34–9 see also autoimmune
anti‐nuclear antibody (ANA) loss of tolerance to hepatitis, recurrent
65, 68–71, 252 35–7 autoimmune hepatitis,
patterns role of gut microbiome 35 recurrent 252–4
multiple nuclear dots autoimmune diseases impact on long‐term
(MND) 69 factors for generation of 22 outcomes 253–4
rim‐like/membranous perpetuation of 37–9 incidence and diagnosis of
69, 70 autoimmune hepatitis 252–3, 253
anti‐nuclear factor see (AIH) 22, 48, after liver transplantation
antinuclear 65, 66, 105–15, 182 239–41, 245
antibody (ANA) adult‐onset prognostic
anti‐retroviral therapy 99 presentation 108–9 impact 241
anti‐ribonucleotide 69 prognosis 114 risk factors 240
anti‐Ro52 77 treatment 111–13 treatment 241
anti‐Scl‐70 69 second‐line risk factors 253, 253
anti‐Sjogren syndrome options 112–13 autoimmune hepatitis
antigen A 69 withdrawal of primary biliary
anti‐Sjogren syndrome antigen therapy 112 cholangitis overlap
B 69 candidate genes for 60 syndrome 101
anti‐Sm 69 definition 106–7 autoimmune liver diseases
anti‐smooth muscle antibody de novo 205–7, 207 (AILDs), recurrent
(ASMA) diagnosis 109–11 after liver transplantation
66, 73–4, 252 autoantibodies 233–41
anti‐soluble liver antigen 109–10 scientific significance
antibody autoimmune sclerosing 245–55
(anti‐SLA) cholangitis 111 see also autoimmune
68, 76–7, 269 histology 110 hepatitis; primary
anti‐sp100 70 immunofluores- biliary cholangitis;
‘anti‐synthetase cence 109 primary sclerosis
antibodies’ 69 environmental factors cholangitis
anti‐tumor necrosis factor 84, 99–101 autoimmune liver serology
(TNF) agents 112 epidemiology 107–8 testing, indications
aortitis 166 fertility in 200 for 79–81
Index 303
autoimmune pancreatitis BET inhibitors 40 bone mineral density in

(AIP) 162–3, 164 betaretroviruses 99 primary sclerosing
type 1 166 bezafibrate 218, 291 cholangitis 153
type 2 (idiopathic bile Bradford Hill criteria 96
duct‐centric functions 12 branched‐chain 2‐oxo acid
pancreatitis) 166 formation, secretion and dehydrogenase
autoimmune polyendocrine enterohepatic complex (BCOADC)
syndrome type 1 circulation 12–13 79, 126
(APS‐1) 29 bile acid 8 bromodomain and extratermi-
autoimmune bile acid sequestrants 292 nal (BET) family 40
polyendocrinopathy/ bile acid therapy 151–2 BTB domain and CNC
candidiasis/ bile salt export pump (BSEP or homolog 2
ectodermal ABCB11) 11 (BACH2, 6q15)
dystrophy biliary apoptosis 92 gene 59
(APECED) 75 biliary atresia 142 budesonide 111–12, 218, 222
autoimmune regulator biliary bicarbonate ion secretion, in primary biliary
(AIRE) transcription protective role 16 cholangitis 135
factor 29, 75 biliary cholangitis 2 bystander activation 86
autoimmune sclerosing biliary epithelial cells (BECs) 6
cholangitis (ASC) biliary piecemeal necrosis 266 c
66, 111, 270 biliary tree stem/progenitor cANCA 77
juvenile, treatment 156 cells (BTSCs) 15 carbohydrate metabolism 5,
autoimmune thyroid bilirubin diglucuronide 7 7–8
disease 84, 277 bilirubin metabolism and caspase recruitment domain
autoimmunity, definition 1 transport 7 family member 10
autoimmunity clustering 276 biliverdin 7 (CARD10) 60, 107
autoimmunosuppressive biologics, induction of cathelicidin 17
therapy 162 DILI 80 C–C chemokine receptor 6
autoreactive cells 36–7 bisphosphonates 225–7, 226 (CCR6) 56
autosomal recessive neonatal bone health 217–28 C–C motif chemokine 20
sclerosing assessment of bone (CCL20, 2q36.3)
cholangitis 142 disease 224 gene 56, 59
autotaxin (ATX) 291 diagnosis 219 CCDC88B gene 57, 59
azathioprine 111, 112, 113, 218 osteoporosis and fractures, CCDC113 gene 58
prevalence 220–2 CCL20 gene 56, 59
b autoimmune hepatitis CD8 T cells 25, 28
B regulatory cells (Bregs) 221–2 CD28 gene 56
31, 34 primary biliary cholangi- CD40 37
BACH2 gene 59 tis 220–1, 220 CD69 32
bacterial cholangitis 146 primary sclerosing CD80 (CD80, 3q13.33)
bacterial infection 34–5 cholangitis s 221 gene 56
in primary biliary pathogenesis 222–4, 223 CD226 antigen (CD226,
cholangitis prevention and treatment of 18q22.2) gene 60
88, 93–6 bone loss 225–7 celiac disease 55, 108, 281–2
Barr bodies 34 general measures 225 CENP‐B 70
B‐cell receptor (BCR) pharmacologic agents central B‐cell tolerance 29
editing 29 225–7, 226 central T‐cell tolerance 29
BCL2 like 11 (BCL2L11, 2q13) risk factors for bone central tolerance 29
gene 59 loss 224 ceruloplasmin 8
304 Index
chenodeoxycholic acid (CDCA) colon cancer 283 cytopenias, autoimmune 89

11, 13, 133 colorectal cancer 141, 145, cytotoxic T lymphocytes
Child–Pugh scores 208 147, 148, 149–50, 153 (CTLs) 25, 28
cholangiocarcinoma connective tissue diseases cytotoxic T lymphocyte antigen
in autoimmune sclerosing (CTDs) 69, 276 (CTLA)‐4 32
cholangitis 111 corticosteroids 8, 218 cytotoxic T‐lymphocyte
in primary sclerosing CREST syndrome 278 associated protein 4
cholangitis Crigler–Najjar syndrome 7 (CTLA4, 2q33.2)
84, 141, 144, 146, Crohn’s disease 55, 61, 149, gene 56, 59
147–8, 153, 249 282–3
cholangiocytes and immunity in juvenile sclerosing d
16–17 cholangitis 154 damage‐associated
cholangiography 111 in primary sclerosing molecular patterns
cholehepatic shunt pathway 13 cholangitis 143, 149 (DAMPs)
cholecystitis 146 crossover syndromes 263, 17, 23, 25, 34
cholelithiasis 146 265, 267 de novo autoimmune
cholic acid (CA) 13 cryptic antigens see hepatitis 66, 76,
chronic hypergammaglobulin- neoantigens 205–7, 207
emic hepatitis 68 cryptogenic chronic liver deceased donor liver
chronic non‐suppurative disease 190–5 transplant (DDLT)
destructive cholangitis diagnostic algorithm 196 in primary biliary
(CNSDC) 246 features suggesting cholangitis
ciclosporin 112 autoimmune recurrence 248
in autoimmune hepatitis pathogenesis in primary sclerosing
210, 241 190–2, 192 cholangitis
primary biliary cholangitis non‐alcoholic fatty liver recurrence 251
recurrence 236, 248 disease 192 dedicator of cytokinesis 1
cigarette smoking, primary seronegative autoimmune (DOCK1) 57
biliary cholangitis hepatitis 195–7 defensin 17
and 90 seronegative primary DENN domain‐containing
cirrhosis, in pregnancy biliary cholangitis protein 1B
207–8 197 (DENND1B, 1q31.3)
citrullination 36 Wilson disease 192–3 gene 56
Clichy criteria 188, 189 potential causes denosumab 227
clonal anergy 29, 30 190–1, 191 deoxycholic acid
cluster of differentiation 28 cryptogenic cirrhosis 181–2 11, 13
(CD28, 2q33.2) causes 191 diabetes, type 1 37, 55, 108
gene 56 C‐type lectin domain diffuse cutaneous systemic
coiled‐coil domain containing containing 16A sclerosis
88B (CCDC88B, (CLEC16A, 16p13.13) (dcSSc) 278
11q13.1) gene gene 58, 60 digoxin 9
57, 59 CTLA4 56, 59 DLEU1 58
coiled‐coil domain‐containing C‐X‐C chemokine receptor DNA repair protein RAD51
protein 113 type 5 (CXCR5, homolog 2
(CCDC113, 16q21) 11q23.3) gene 57 (RAD51B, 14q24.1)
gene 58 cystic fibrosis 17, 142 gene 58
colesevelam 292 cytochrome P4502D6 dominant stricture
colestyramine 8, 292 (CYP2D6) 75 in autoimmune sclerosing
collagen 25 cytomegalovirus 99 cholangitis 111
Index 305
in primary sclerosing esophagogastroduodenoscopy gallbladder polyps 146

cholangitis 146–7 (EGD) 208 gallstones 142, 148, 149, 166,
donation after circulatory estrogen levels 34 171, 291
death (DCD) 239 evolutionary survival γδT cells 25
drug metabolism 11–13 advantage 22 gamma‐glutamyltranspepti-
bile acid synthesis and experimental autoimmune dase (GGT) 18, 19
metabolism 13 allergic encephalo- in autoimmune sclerosing
enterohepatic bile acid myelitis models 85 cholangitis 156
circulation 13, 14 ‘exposome’ 32 in primary biliary
drug‐induced hepatitis 73 expression quantitative trait cholangitis 126,
drug‐induced liver injury loci (eQTLs) 55 125–6, 144, 145,
(DILI) 73, 80, 281 extractable nuclear antigens 246, 249
ductular reaction 16 (ENAs) 69 genetics 31–3
‘dysbiosis’ 34 extrahepatic autoimmune HLA risk alleles 32
diseases (EHAIDs) non‐HLA gene
e 276–7, 276 associations 32–3
Ehrlich, Paul 1, 2, 22 epidemiology 276–7, 276 genome‐wide association
ELISA assays 68–70, 68, 101, see also under names studies
126, 134, 197 (GWAS) 32, 35, 48,
ELMO1 57 f 49, 85, 106
emtricitabine 99 fanesoid X receptor (FXR) 10, of primary biliary
enalapril 9 11, 12–13, 133, 151 cholangitis 49
engulfment and cell motility fatigue in primary biliary of primary sclerosing
protein 1 (ELMO1, cholangitis 126 cholangitis 49
7p14.1) gene 57 Fatigue Impact Scale 293 of type 1 autoimmune
enhancer DNA 44 fertility in autoimmune hepatitis 50
enhancer RNA (eRNA) 33 hepatitis 200 ‘genome‐wide level of
enhancers 37–8 fexofenadine 9 significance’ 48
environmental factors 83‐ fibric acid derivatives in geo‐epidemiology of primary
link with autoimmune primary biliary biliary
hepatitis 84, cholangitis 134–5 cholangitis 92–3
99–101 fibrinogen 8 germinal centers 39
link with primary biliary FokI 34 glucagon‐like peptide
cholangitis forbidden clone concept 2 (GLP)‐1 13
84, 89–93, 91 Forkhead box P1 (FOXP1, glutathione‐S‐transferase 7
environmental triggers 85–6 3p13) gene 59 glycine‐conjugated cholate 11
enzyme‐linked immunosorbent formiminotransferase cyclode- glycogen 7–8
assay (ELISA) see aminase 76, 107 goals of treatment 288–9
ELISA FOXP1 59 guidelines for clinicians 285
epigenetic enhancer FOXP3 30, 31 gp210 70, 99, 197
regulation 40 fractures, prevalence 220–2 GPR35 59
epigenetics 33, 37–8 granulomas 246
epithelial cell‐induced g Graves disease 84, 277
transformation of G protein‐coupled receptor 35 gut lymphocyte homing 284
iTreg to Th17 (GPR35, 2q37.3)
Cells 39 gene 59 h
epitope spreading 38, 86 gabapentin 136, 293 hapten–carrier protein
Epstein–Barr virus 101 galectin‐3 164 complexes 36
Escherichia coli 93, 94, 96 gallbladder cancer 148, 153 haptens 36
306 Index
haptoglobin 8 hepatocellular carcinoma i

Hashimoto thyroiditis 277 (HCC) 13, 101, ibandronate 226–7
HDAC7 60 125, 128, 130, 234, iChip association study 48
health utility 290 237, 239 IgA secretion 17
health‐related quality of life in primary biliary IGF‐1 222
(HRQoL) 290, 294 cholangitis IgG4‐associated
Helicobacter pylori 165 138, 153 cholangitis(IAC)
heme 7 hepatocyte plasma 144
hemolytic anemia 1 membrane 19 IgG4‐related disease
hepatic harbor hepatic stem/ hepatocytes 6, 9 (IgG4‐RD)
progenitor cells cell death 14–15 2, 142, 161–74
(HpSCs) 15 liver regeneration 15–16, 15 alternative diagnoses
hepatic injury, biochemical herpesvirus‐ 95 170–1
markers and high mobility group box 1 blood tests 169–70
patterns of (HMGB1) 17 clinical characteristics of
17–18 Hill’s criteria 85 IgG4‐related
cholestasis 18–19 histone 77 hepatobiliary
hepatocellular necrosis 18 histone deacetylase 7 disease 166
hepatic lymph 25 (HDAC7, 12q13) disease outside the
hepatic metabolism 7–9 gene 60 hepatobiliary
bilirubin metabolism and HIV infection 89, 101 system 166
transport 7 HLA‐DRB1 gene 32 hepatobiliary
carbohydrate HMG‐CoA reductase 8 disease 166
metabolism 7–8 HMG1 77 diagnostic criteria
lipid metabolism 8 HMG2 77 166–8, 167
protein metabolism 8 Hodgkin disease 19 disease nomenclature
metabolic zonation 8–9 horror autotoxicus 22 163
hepatic sinusoidal endothelial human betaretrovirus etiology and pathogenesis
cells (HSECs) 6 (HBRV) 163
hepatic stellate cells (HSCs; 97, 98–9, 101 antigens that may drive
Ito cells; lipocytes) human herpesvirus 6 an IgG4‐RD
6, 7, 23 infection 101 Response 164–5
hepatic transport systems human leukocyte antigen development and
9–11 (HLA) complex characteristics
apical (canalicular) 50–3 of antibody response
transporters 9, 11 hyaluronan 17 163
basolateral (sinusoidal) hypercholesterolemia 8 IgG4 antibodies and
transporters hyperferritinemia 192 pathogenesis 165
9–11, 10 hypergammaglobu- HISORt criteria 167
hepatitis A virus 18, 101 linemia 76 histologic diagnosis 168–9,
hepatitis activity index hyperglycemia 8 168, 169
(HAI score) 266 hyperlipidemia in primary history 162–3
hepatitis B virus 18, 182 biliary cholangitis incidence and prevalence
NTCP in 11 125 of 165
hepatitis C virus (HCV) 28, hyperparathyroidism, malignancy and 171–2
75, 101, 109, 182 secondary management and treatment
hepatitis D virus 182 223 172–3
NTCP in 11 hyperpigmentation 125 monitoring and follow‐up
hepatitis E virus 101, 182 hyperthyroidism 84 173
Index 307
radiologic characteristics insulin resistance 8 IRF8 58

171 interface hepatitis 266–7, itch, cholestatic 291–2
IgG4‐related disease 267, 269 iTregs 31
responder interferon (IFN)‐γ 28, 107
index 173 interferon regulatory factor 5 j
IKZF3 58 (IRF5, 7q32.1) 57 Janus kinases (JAKs) 32
IL2 59 interferon regulatory juvenile sclerosing cholangitis
IL2RA 59 factor 8 (IRF8, 152, 153–7
IL7R 57 16q24.1) 58 autoimmune sclerosing
IL12A 57 interleukin‐2 (IL2, 4q27) cholangitis 154
IL12B 57 gene 59 clinical features 156
IL12RB2 56 interleukin‐2 receptor subunit diagnosis 154
IL16 58 alpha (IL2RA, epidemiology 153
IL21 57 CD25, 10p15.1) prognosis 157
IL21R 58 gene 59 treatment 156
immune complex disease 89 interleukin‐7 receptor subunit k
immune deviation 37 alpha (IL7R, 5p13.2) Kaiser‐Fleischer rings 125
immune homeostasis mainte- gene 57 kernicterus 7
nance 30, 107 interleukin‐12 receptor kinetochore 70
immune tolerance 29 subunit beta 2 King’s College criteria 188,
immunoblot (Western (IL12RB2, 1p31.3) 189
blot) 68 gene 56 Koch’s postulates 84, 85, 96,
immunodysregulation 30 interleukin‐12 subunit alpha 98
immunodysregulation (IL12A, 3q25.33) Kupffer cells 6, 16
polyendocrinopathy gene 57
enteropathy interleukin‐12 subunit beta l
X‐linked (IPEX) (IL12B) gene 57 lactoferrin 78
syndrome 30, 86 interleukin‐16 (IL16, 15q25.1) laminin 164
immunosuppression 151 gene 58 Langerhans cell histiocytosis
indirect immunofluorescence interleukin‐21 (IL21, 4q27) 142
(IIF) 66–7, 68–70, gene 57 leaky gut 283
73–6 interleukin‐21 receptor lecithin‐cholesterol
inducible tolerance 29 (IL21R, 16p12.1) acyltransferase 8
inflammatory bowel disease gene 58 leukemia‐associated protein
(IBD) 77, 84, 108, interleukins (IL) 1 (DLEU1, 13q14.2)
276, 277, 282–4 IL‐1β 25 gene 58
juvenile early onset 86 IL‐6 25 leukocyte proteinase 3, 77
in primary sclerosing IL‐10 25 LFA3 56
cholangitis 142, deficiency 86 limited cutaneous systemic
143, 221, 249 IL‐12 25, 28, 38, 85, 86 sclerosis (lcSSc) 278
in primary sclerosing IL‐12Rβ1 33 lipid metabolism 5, 8
cholangitis after IL‐17 28 lipoic acid 97
transplanta- IL‐18 25, 28 lipopolysaccharide (LPS) 93
tion 238 IL‐20 38–9 lipoteichoic acid 93
infliximab 113, 114 IL‐23 33 lithocholic acid 13
innate immune organ, liver IL23R gene 33 liver cell types and
as 25 international normalized ratio organization 6–7
innate immunity 23–6, 24 (INR) 18 liver function tests (LFTs)
inositol 1,4,5‐trisphosphate 56 IRF5 57 17–18
308 Index
liver sinusoidal endothelial m mucosal invariant T (MAIT)

cells (LSECs) 25, magnetic resonance cells 28
26, 28 cholangiography multidrug resistance
‘liver specific protein’ 111 MDR1 (ABCB1) 11
(LSP) 79 major histocompatibility MDR3 (ABCB4) 11
liver synthesis 5 complex (MHC) 85 multidrug resistance
liver transplantation 2, 90, MAPT 58 protein (MRP) 9
114–15, 233–41, maternal–fetal immune MRP1 9
235, 237 homeostasis 2 MRP2 7, 11
autoimmune membrane metalloendopepti- MRP3 9
hepatitis 239–41 dase like 1 (MMEL1, MRP4 9
overlap syndrome 271 1p36.32) gene multiple sclerosis (MS),
in primary biliary 56, 59 lncRNAs and 33
cholangitis 135 menstruation, autoimmune Mycobacterium gordonae
primary sclerosing hepatitis and 200 heat shock
cholangitis 237–9 metabolic zonation 5, 8–9 protein 65
recurrent primary biliary methotrexate 9, 173, 281, 283 (Hsp65) 95
cholangitis 234–7, methylation quantitative trait Mycobacterium tuberculosis
236 loci (mQTLs) 55 95, 96
liver X receptors 17 mevalonate 8 mycophenolate mofetil
living‐donor liver transplant microbiota (MMF) 112,
(LDLT) modulators 151–2 113–14
in primary biliary microRNA 33, 38
cholangitis microsomal heme n
recurrence 248 oxygenase 7 naltrexone 293
in primary sclerosing microtubule‐associated natural antibodies 25
cholangitis protein tau (MAPT, natural tolerance 29
recurrence 251 17q21.31) necrosis 14
long non‐coding RNAs gene 58 neoantigens (cryptic antigens)
(lncRNAs) 33 Mikulicz syndrome 166 15, 36, 86
lopinavir 99 ‘missing heritability’ 53 neonatal systemic LE
LTBR gene 58 MIT3 79 syndrome 77
lung cancer 19 mitochondrial oxo‐acid neutrophil extracellular
lupoid hepatitis (‘chronic dehydrogenase 99 traps (NETs)
hypergammaglobulin- mitogen‐activated protein 35
emic hepatitis’) 68 kinase (MAPK)/ NFKB1 57, 59
lupus erythematosus 68, 77 activator protein 1 NK cells 23
see also systematic lupus (AP‐1) 17 non‐HLA associations
erythematosus mixed connective tissue 53–61
lymphocyte function‐ disease 69 non‐alcoholic fatty liver
associated antigen 3 MMEL1 56, 59 disease(NAFLD)
(LFA3, CD58, Model for End‐Stage Liver 80, 182, 192
1p13.1) Disease (MELD) non‐alcoholic steatohepatitis
gene 56 130, 135, 189, 191, (NASH)
lymphocytic piecemeal 208, 211 126, 192, 281
necrosis 266 molecular mimicry non‐receptor tyrosine‐
lymphopenia 89 86, 93 protein kinase
lysophosphatidic acid mouse mammary tumor virus (TYK2, 19p13.2)
(LPA) 291 (MMTV) 89, 99 gene 58
Index 309
Novosphingobium outlier syndromes 262, peroxisome proliferator‐

aromaticivorans 95 265, 267 activated receptors
nuclear anti‐neutrophil overlap syndrome 110, (PPARs) 17, 134
antibody 263–72, 263 phospholipase C‐like protein 2
(NANA) 77 AIH Scores 267 (PLCL2, 3p24.3)
nuclear factor kappa B subunit criteria for diagnosis 268 gene 56
1 (NFKB1, 4q24) features of 266 phospholipid flippase 11
57, 59 genetics 267 Pirquet, Clemens von 1
nuclear receptor family 17 liver biopsy 266 pit cells 6, 7
nucleotide binding PLCL2 gene 56
oligomerization p pleiotropy analysis 55
domain (NOD)‐like pamidronate disodium 227 polyendocrinopathy 30
receptor P3 pancreatic cancer in primary polymyositis 69
(NLRP3) 23, 25 sclerosing portal fibroblasts 16
cholangitis 148 POU domain class
o pancreatitis 2‐associating
obeticholic acid (OCA) 218, autoimmune pancreatitis factor 1 57
247, 288 (AIP) 162–3, pravastatin 9
in primary biliary 164, 166 prednis(ol) 111
cholangitis 133–4 idiopathic duct‐centric 166 pregnancy
O‐phosphoseryl‐tRNA (Sec) post‐ERC 152 autoimmune hepatitis
selenium parenchymal cells 6 in 199–208
transferase 77 pathogen‐associated in cirrhosis 207–8
optioid peptides 9 molecular patterns de novo autoimmune
oral tolerance prevention of (PAMPs) 17, 22–3, hepatitis 205–7,
autoimmunity 25, 34 207
39–40 pattern recognition receptors fertility in autoimmune
organic anion‐transporting (PRRs) 23 hepatitis 205–7,
polypeptides pediatric‐onset autoimmune 207
(OATPs) 9 hepatitis gut microbiota
OATP1B1 10 presentation 109 manipulation
OATP1B3 10 prognosis 114–15 in 39
OATP2B1 9 treatment 113 liver related outcomes
osteocalcin 224 first‐line therapy 113 in 204–7, 206
osteomalacia 218, 219, 228 second‐line medication safety
osteopenia 151, 219, options 113–14 in 208–11
222, 224 withdrawal of outcomes in autoimmune
osteoporosis 218–19 therapy 113 hepatitis 200–4,
in autoimmune perbiliary glands (PBGs) 15 201–2
hepatitis 221–2 perinuclear anti‐neutrophil pregnane X receptor
assessment 224 cytoplasmic (PXR) 11, 12, 136
diagnosis 219 antibody (pANCA) prevention of autoimmunity
prevalence 220–2 77, 283 39–40
in primary biliary perinuclear anti‐neutrophil primary biliary cholangitis
cholangitis 137–8, nuclear antibody (PBC) 8, 22, 48, 66,
220–1, 220 (pANNA) 109, 123–40
in primary sclerosing 77, 78 ASPR in 79
cholangitis 221 peripheral tolerance 30–1 asymptomatic patients 126
osteoprotegerin(OPG) 223 perisinusoidal cells 6 candidate genes for 56–9
310 Index
primary biliary cholangitis prognostic impact of 237 PRKCB 58

(PBC) (cont’d) risk factors 235–6, 236 PRKD2 60
cancer and 90 treatment 236–7 procollagen type 1 N‐terminal
clinical presentation primary biliary cholangitis / propeptide
126–7 AIH overlap (P1NP) 224
complications of liver syndrome 267–70 progesterone 34
disease 137–8 Paris criteria 267, 268, programmed cell death 1
definition 124 268 (PD‐1) receptors 28
diagnosis 125–8, 225 primary sclerosing cholangitis prolactin 34
autoantibodies 126 (PSC) 2, 8, 22, 48, promyelocytic leukemia
biochemical tests 125–6 66, 77, 84, 141–57 protein 69–70
differential diagnosis candidate genes for 54, 55, proteasome assembly chap-
128 59–60 erone 1 (PSMG1,
imaging 128 definition 143 21q22.2) gene 60
liver biopsy 126–8 diagnosis 142, 144–5, protein kinase C beta type
environmental factors 84, 144–5 (PRKCB, 16p12.2)
89–99, 91 epidemiology 143 gene 58
bacterial infection 88, malignancy 147–8, 152–3 protein kinase D2 (PRKD2,
93–6 presentation 145–7 19q13.32) gene 60
geo‐epidemiology 92–3 prognosis 148–50 protein metabolism 5, 8
viruses 97–9 ALP and bilirubin 149 pruritis 291–3
xenobiotics and 96–7 natural history management 291–3, 292
epidemiology 124 models 149 in primary biliary cholangitis
etiopathogenesis 125–6 non‐invasive evaluation 126, 289
history 124 of fibrosis 149–50 PSMG1 60
natural history 128–9 risk stratification 148–9 psoriasis 2, 33, 55
risk loci and candidate recurrence 249–52, 250 pyridoxal 5‐phosphate
genes for 52–3, 54 after liver transplanta- deficiency 18
risk stratification 130–2, tion 237–9, pyruvate dehydrogenase
131 249–50 complex (PDC‐E2)
symptom management risk 48 15, 36, 79, 85 90, 87,
135–7, 137 surveillance for 92, 97, 126
symptomatic patients 126 malignancy 152–3 binding protein 90
treatment 132–7, 225 treatment 150–2
UDCA and 129–30, endoscopic 152 r
133–4 medical 79, 150–1 RAD51B 58
primary biliary cholangitis, variant phenotypes 145 raloxifene 227
recurrent 245, elevated IgG4 145 Raynaud syndrome 84, 90
246–9 features of autoimmune reactive ductular cells
incidence and diagnosis hepatitis 145 (RDCs) 16
247–8, 247 small‐duct primary receptor activatory of nuclear
long‐term outcomes sclerosing kapp‐B ligand
248–9 cholangitis 145 (RANKL) 223, 227
risk factors 248–9 see also juvenile sclerosing receptor‐related orphan
primary biliary cholangitis, cholangitis; prima receptor (ROR)γ 31
recurrent, after liver primary sclerosing cholangitis receptor tyrosine kinases
transplantation /AIH overlap (RTKs) 32
234–7 syndrome 144, recombination activating gene
diagnosis 234–5 270–1 proteins
Index 311
RAG‐1 29, 37 emergency liver sp140 70

RAG‐2 proteins 29, 37 transplantation SPIB 59
Regan isoenzyme 19 191–2, 191 STAT4 gene 56
regulatory dendritic cells potential causes statins 80
(DCregs) 31 183–5, 183 steatohepatitis 192
renal cell carcinoma 19 risk stratification 188 STK4 59
retinoic acid receptor 17 King’s College criteria super enhancers (SEs) 32, 33,
retinoid X receptor (RXR) 17 188, 189 37–8
retroperitoneal fibrosis 166 Clichy criteria 188, 189 superantigen 35, 89
rheumatoid arthritis (RA) 1, Model for End‐stage symptoms 289–90
2, 55, 89, 282–3 Liver Disease management 290–7
IgG4 response to monoclonal (MELD) 189, 191 effective care delivery
antibodies 163 see also cryptogenic chronic in 297
lncRNAs and 33 liver disease fatigue and cognitive
RIC8 guanine nucleotide seronegative autoimmune 293–6, 295
exchange factor B hepatitis 195–7 impact of disease‐
(RIC8B, 12q23.3) seronegative primary biliary modifying therapy
gene 60 cholangitis 197 on 290–1
rifampicin 292 sertraline 293 social isolation 296–7
risk factors 31–4 sex hormones 33–4 stage‐associated
rituximab 112, 113, 114, SGSM1 60 symptoms 291
165, 174 SH2B adapter protein 3 stage‐independent
(SH2B3, 12q24.12) symptoms
s gene 32, 58, 60 291–6, 295
salt inducible kinase 2 (SIK2, shock liver 18 synaptogyrin‐1 (SYNGR1,
11q23.1) gene sicca syndrome 126 22q13.1) gene 59
59–60 sickle cell anemia 142 systemic lupus erythematosus
sarcoidosis 55, 89 signal transducer and activator (SLE) 89, 90, 276,
scarlet fever 1 (STAT)‐4 32 281–82
scleroderma 84 signal transducer and activator lncRNAs and 33
sclerosing cholangitis see of transcription 4 associated hepatitis (lupus
IgG4‐related disease (STAT4, 2q32.2) hepatitis) 281, 282
self‐antigens, tolerance gene 56 systemic sclerosis 278–9
to 28–31 SIK2 59–60
senescence‐associated sinusoidal cells 6 t
secretory phenotype sirolimus 114 T follicular helper (Tfh) cell
(SASP) 17 Sjögren syndrome (SS) 84, 37, 39
serine/threonine kinase 4 89, 90, 276, 277–8 tacrolimus 112, 113
(STK4, small and medium‐vessel primary biliary cholangitis
MST1,3p21.31) vasculitis 77 recurrence and 248
gene 59 small G protein signaling taurine‐conjugated
seronegative acute liver modulator 1 cholate 11
failure 182–3, 197 (SGSM1, 22q11.23) T‐cell mediated immune
autoimmune pathogen- gene 60 regulation 30
esis 185, 184 small‐duct disease 142 T‐cell receptor revision
clinical features 188 sodium taurocholate cotrans- 29, 37
management 189–92 porting polypeptide T‐cell ubiquitin ligand
N‐acetylcysteine 190 (NTCP) 9, 11–12 (TULA) family 60
corticosteroids 190–1 sp‐100 69 tenofovir 99
312 Index
teriparatide 228 tumor necrosis factor in primary biliary cholangitis

tertiary lymphoid structures α‐induced protein 2 recurrence
(TLSs) 39 (TNFAIP2, 14q32.32) 246, 249
testosterone 34 gene 58 in primary sclerosing
Th1 28, 31, 37 tumor necrosis factor α‐ cholangitis 249
Th17 cells 28, 31, 33, 39 induced protein 3 overlap syndrome 269–70
thiopurine S‐methyltransferase (TNFAIP3, ustekinumab 85
(TPMT) 6q23.3) urinary tract infections in
deficiency 113 gene 57 primary biliary
thyroid disease 90 tumor necrosis factor ligand cholangitis
ticrynafen (tielinic acid) 75 superfamily member patients 93
tissue memory T cells 38 15 (TNFSF15,
tissue‐resident memory T 9q32) gene 57 v
(TRM) cells 38 tumor necrosis factor receptor varices in primary biliary
TNF apoptosis‐inducing superfamily member cholangitis 138
ligand (TRAIL) 28 1A (TNFRSF1A, vasculitides 89
TNFAIP2 58 12p13.31) viral infections 34–5
TNFAIP3 57 gene 57 in primary biliary
TNFAIP3 32 tumor necrosis factor receptor cholangitis 97–9
TNFRSF1A 57 superfamily member vitamin D deficiency 34, 39
TNFSF15 57 3 (LTBR, 12p13.31) vitamin D receptor
Toll‐like receptors (TLRs) 17 gene 58 (VDR) 17, 34
TLR4 17 TYK2 58 vitamin D3, in prevention of
toxic waste and primary biliary autoimmunity 39
cholangitis 92–3 u vitamin K deficiency 223
transcription factor Spi‐B ubiquitin associated and SH3
w
(SPIB, 19q13.33) domain containing A
waterborne risk factor of
gene 59 (UBASH3A,
primary biliary
transcription factor 21q22.2) 60
cholangitis 92
enhancers 33 ulcerative colitis 55, 276,
Western blot 68, 99
transferrin 8 282, 283
Wilson disease 192–3
transjugular intrahepatic post‐transplantation 238
Wiskott–Aldrich
portosystemic shunt in primary sclerosing
syndrome 86
(TIPS) 208 cholangitis 143
Witebsky’s postulates 84, 85
Tregs 28, 30, 31 uridine diphosphate glucuro-
inducible Tregs (iTregs) 30 nosyltransferase x
natural Tregs (nTregs) 30 (UGT) 7, 36 X monosomy 34
trigger mechanisms of uridine 5′‐diphosphate xanthelasma 8, 125
autoimmunity glucuronosyltrans- xanthoma 8, 125
86–9 ferase 1 (UGT‐1) xenobiotics and primary
tubulointerstitial family 75 biliary cholangitis
nephritis 166 urobilin 7 96–7
tumor necrosis factor (TNF)‐α urobilinogens 7
25, 32 ursodeoxycholic acid (UDCA) z
tumor necrosis factorα‐ 218, 222, 288 zinc finger protein Aiolos
induced protein in primary biliary (IKZF3, 17q21.1)
3‐interacting protein cholangitis gene 58
1 (TNIP1) 32 132–3, 150, 234 zoledronic acid 227

Liver

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Liver

Uploaded by

Copyright:

Available Formats

Autoimmune Liver Disease

Autoimmune Liver Disease

Management and Clinical Practice

Limit of Liability/Disclaimer of Warranty

Library of Congress Cataloging‐in‐Publication Data

Set in 10/12pt Warnock by SPi Global, Pondicherry, India

Introduction: The Paradigm and Paradox of Liver Autoimmunity 1

Section I Scientific Basis of Clinical Autoimmune Liver Diseases 3

1 Introduction to the Physiology, Immunology and Pathology

2 Concepts of Autoimmunity Relevant to Autoimmune Liver Diseases 21

Cytokines Promoting Chronic Inflammation and Autoimmunity 38

3 Genetics and Risk of Autoimmune Liver Diseases 47

4 Autoantibodies and Understanding of Autoimmune Liver Diseases 65

5 Environmental Exposure and Risk in Autoimmune Liver Diseases 83

6 Autoimmune Hepatitis 105

Second‐line Options 115

7 Primary Biliary Cholangitis 123

8 Primary Sclerosing Cholangitis 143

9 IgG4‐Related Liver and Biliary Disease 163

Utility of Blood Tests, Including Serum IgG4 Levels

Section III Specific Clinical Challenges 181

10 Managing Acute and Chronic Seronegative Liver Disease 183

11 Managing Pregnant Women with Autoimmune Liver Disease 201

12 Bone Health in Patients with Autoimmune Liver Diseases 219

13 Recurrent Autoimmune Liver Disease and Its Impact on Clinical Practice 235

14 Recurrent Autoimmune Liver Disease and its Scientific Significance 247

Risk Factors of Recurrent PBC 250

Section V Controversies in Autoimmune Liver Diseases 263

15 Making Sense of Overlap and Crossover Syndromes 265

16 The Role of Extrahepatic Autoimmunity in Autoimmune Liver Disease 277

17 Symptoms, Chronic Disease, and Patient Management 289

­Symptoms, Quality of Life, and Health Utility: Key Concepts 291

David Jones MD, PhD, OBE Carlos Moctezuma‐Velázquez MD

Atsushi Tanaka MD Mette Vesterhus MD, PhD

AASLD American Association for the BCOADC branched‐chain 2‐oxo acid

DCreg regulatory dendritic cell HLA human leukocyte antigen

MELD Model for End‐stage Liver OR odds ratio

SNP single nucleotide polymorphism TORC transducer of regulated CREB

From an evolutionary perspective, the (TCR). This magnitude of diversity is a result

Scientific Basis of Clinical Autoimmune Liver Diseases

Introduction to the Physiology, Immunology and Pathology

­Liver Cell Types and Organization transporters (e.g. ileal BA ­ transporter), and

Bilirubin Metabolism and Transport Carbohydrate Metabolism

of the liver parenchyma, known as meta‑ ­Hepatic Transport Systems

Hepatocellular transporters are subject (FXR)‐mediated ­ signaling, thereby pre‑

c­onjugation with glutathione, methyl or heavy metals, as well as exogenous drugs,

Our understanding of cholestatic liver d­ iseases

Canals of Hering Peribiliary glands

Hepatic stem/progenitor cells Biliary tree stem/progenitor cells

• Location: Canals of Hering • Location: Peribiliary glands – large intrahepatic

aggression, or be the initiators of an involved in the immune response of BECs.

­ isease. LFTs that more accurately reflect liver

Concepts of Autoimmunity Relevant to Autoimmune

Knowledge of the concepts of autoimmunity can aid gastroenterologists and hepatologists in

Keywords adaptive immunity; autoantigens; autoimmune liver diseases; autoimmunity prevention;

­Introduction cholangitis (PSC) are classified as autoimmune

Table 2.1 Comparison of classic autoimmunity with AIH, PBC or PSC.

Classic autoimmunity AIH PBC PSC

Disease‐specific autoantigens Yesa Yes Yesb

Table 2.2 Comparison of selected features of innate and adaptive immunity.

Innate immunity Adaptive immunity

IL-6 Th17 IL-17, IL-22, TNFα Cytokine-Mediated

(TGFβ) IL-1β TGFβ

Symptoms, Quality of Life, and Health Utility: Key Concepts 291

Liver Cell Types and Organization transporters (e.g. ileal BA transporter), and

of the liver parenchyma, known as meta‑ Hepatic Transport Systems

Hepatocellular transporters are subject (FXR)‐mediated signaling, thereby pre‑

conjugation with glutathione, methyl or heavy metals, as well as exogenous drugs,

Our understanding of cholestatic liver d iseases

isease. LFTs that more accurately reflect liver

Introduction cholangitis (PSC) are classified as autoimmune

Peripheral Tolerance FoxP3 Tregs. FoxP3 Tregs secrete immuno

romote adaptive immune responses to

Introduction that they result from the complex interaction

PBC (European) R B39:06‐DRB108:01‐DQA104:01‐DQB104:02

PBC (Japanese) R DRB108:03‐DQB106:01

Type 1 AIH R A1‐B8‐DRB301:01‐DRB103:01‐DQAl05:01‐DQBl02:01

Type 2 AIH R DRB103:01‐DQB102

Introduction immunosuppressive treatment is usually

Introduction antibodies are detected in the majority of

been clearly documented in both infectious lymphopenia‐induced autoimmunity is found

including Epstein–Barr virus, cytomegalo PBC‐specific phenotype. It is also notable

Autoimmune hepatitis (AIH) is a progressive inflammatory hepatopathy characterized by hyper

Introduction by hypergammaglobulinemia, specific