You are on page 1of 261

Advances in Experimental Medicine and Biology 1050

Advances in Microbiology, Infectious Diseases and Public Health

Paola Mastrantonio
Maja Rupnik Editors

Updates on
difficile in Europe
Advances in Microbiology, Infectious Diseases
and Public Health Volume 8
Advances in Experimental Medicine
and Biology
Volume 1050

Series Editors
Irun R. Cohen, The Weizmann Institute of Science, Rehovot, Israel
N.S. Abel Lajtha, Kline Institute for Psychiatric Rrch, Orangeburg, NY, USA
John D. Lambris, University of Pennsylvania, Philadelphia, PA, USA
Rodolfo Paoletti, University of Milan, Milan, Italy
Nima Rezaei, Tehran University of Medical Sciences, Tehran, Iran

Advances in Microbiology, Infectious Diseases

and Public Health

Subseries Editor
Gianfranco Donelli, Microbial Biofilm Laboratory, Fondazione Santa Lucia
IRCCS, Rome, Italy

Subseries Editorial Board

Murat Akova (Turkey), Massimo Andreoni (Italy), Beate Averhoff
(Germany), Joana Azeredo (Portugal), Fernando Baquero (Spain), George
Belibasakis (Switzerland), Emilio Bouza (Spain), Maria Rosaria
Capobianchi (Italy), Tom Coenye (Belgium), Anne Collignon (France),
Rita Colwell (USA), Mahmoud Ghannoum (USA), Donato Greco (Italy),
Jeffrey B. Kaplan (USA), Vera Katalinic-Jankovic (Croatia), Karen Krogfelt
(Denmark), Maria Paola Landini (Italy), Paola Mastrantonio (Italy), Teresita
Mazzei (Italy), Elefterios Mylonakis (USA), Jiro Nakayama (Japan), Luisa
Peixe (Portugal), Steven Percival (UK), Mario Poljak (Slovenia), Edoardo
Pozio (Italy), Issam Raad (USA), Evangelista Sagnelli (Italy), Stefania
Stefani (Italy), Paul Stoodley (USA), Jordi Vila (Spain)
This book series focuses on current progress in the broad field of medical
microbiology, and covers both basic and applied topics related to the study of
microbes, their interactions with human and animals, and emerging issues
relevant for public health. Original research and review articles present and
discuss multidisciplinary findings and developments on various aspects of
microbiology, infectious diseases, and their diagnosis, treatment and
The book series publishes review and original research contributions,
short reports as well as guest edited thematic book volumes. All contributions
will be published online first and collected in book volumes. There are no
publication costs.
Advances in Microbiology, Infectious Diseases and Public Health is a
subseries of Advances in Experimental Medicine and Biology, which has
been publishing significant contributions in the field for over 30 years and is
indexed in Medline, Scopus, EMBASE, BIOSIS, Biological Abstracts, CSA,
Biological Sciences and Living Resources (ASFA-1), and Biological
Sciences. 2016 Impact Factor: 1.881

More information about this series at

Paola Mastrantonio • Maja Rupnik

Updates on Clostridium
difficile in Europe
Advances in Microbiology, Infectious
Diseases and Public Health Volume 8
Paola Mastrantonio Maja Rupnik
Former Research Director in the Department for Microbiological
Department of Infectious Diseases Research
National Institute of Health (ISS) National Laboratory of Health,
Rome, Italy Environment and Food (NLZOH)
Maribor, Slovenia

ISSN 0065-2598 ISSN 2214-8019 (electronic)

Advances in Experimental Medicine and Biology
ISSN 2365-2675 ISSN 2365-2683 (electronic)
Advances in Microbiology, Infectious Diseases and Public Health
ISBN 978-3-319-72798-1 ISBN 978-3-319-72799-8 (eBook)

Library of Congress Control Number: 2016935504

# Springer International Publishing AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, express or implied, with respect to the material
contained herein or for any errors or omissions that may have been made. The publisher
remains neutral with regard to jurisdictional claims in published maps and institutional

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

The chapters of this book were planned to cover the most important issues to
be addressed in the study of infections due to Clostridium difficile, a micro-
organism still feared not only as the cause of nosocomial diarrhea related to
protracted antibiotic administration but more and more frequently of
diarrheal diseases unrelated to the hospital environment, including those
affecting animals. In the last decades, a growing number of clinicians,
microbiologists, and epidemiologists have investigated this topic, as
evidenced by the large amount of scientific publications still in an upward
trend over the years. In particular, this book has been focused on the clinical
and experimental activities carried out in Europe for a better knowledge of
this pathogen and its molecular characteristics, associated pathologies, and
possible transmission routes, as well as to build up preventive and diagnostic
strategies and efficacious therapeutic approaches for the treatment of
C. difficile infection (CDI).
Thanks also to the foundation of the European Study Group on C. difficile
(ESGCD) in 2000, in the framework of the European Society of Clinical
Microbiology and Infectious Diseases (ESCMID), European clinicians and
researchers, together with experts from all over the world, were able to
consolidate the already existing, positive collaboration that led in recent
years to the establishment of a European network for the epidemiological
surveillance, the molecular characterization, and the evaluation of the antibi-
otic resistance profile of the clinical isolates, with obvious advantages for the
continuous updating of the treatment strategies of CDI. To emphasize the
positive role of this study group in the fight against C. difficile infection, an
invited chapter written by both the current and the past president of ESGCD
has been included at the end of this book.
We are grateful to all the authors for their significant contributions to the
book. In our view and intention, they ideally represent also the work of many
other European experts in this field who did not get involved on this occasion
for obvious limits of space.

Rome, Italy Paola Mastrantonio

Maribor, Slovenia Maja Rupnik


Economic Burden of Clostridium difficile Infection

in European Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Elena Reigadas Ramı́rez and Emilio Santiago Bouza
The Need for European Surveillance of CDI . . . . . . . . . . . . . . . . . 13
Camilla Wiuff, A-Lan Banks, Fidelma Fitzpatrick,
and Laura Cottom
Diagnostic Guidance for C. difficile Infections . . . . . . . . . . . . . . . . 27
Monique J.T. Crobach, Amoe Baktash, Nikolas Duszenko,
and Ed J. Kuijper
Ribotypes and New Virulent Strains Across Europe . . . . . . . . . . . 45
Jeanne Couturier, Kerrie Davies, Cécile Gateau,
and Frédéric Barbut
Comparative Genomics of Clostridium difficile . . . . . . . . . . . . . . . 59
Sandra Janezic, Julian R. Garneau, and Marc Monot
Cellular Uptake and Mode-of-Action of Clostridium
difficile Toxins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Panagiotis Papatheodorou, Holger Barth, Nigel Minton,
and Klaus Aktories
Clostridium difficile Biofilm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Claudia Vuotto, Gianfranco Donelli, Anthony Buckley,
and Caroline Chilton
European Practice for CDI Treatment . . . . . . . . . . . . . . . . . . . . . 117
Fidelma Fitzpatrick, Mairead Skally, Melissa Brady, Karen Burns,
Christopher Rooney, and Mark H. Wilcox
Antibiotic Resistances of Clostridium difficile . . . . . . . . . . . . . . . . 137
Patrizia Spigaglia, Paola Mastrantonio, and Fabrizio Barbanti
Probiotics for Prevention and Treatment of Clostridium
difficile Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Lorena Valdés-Varela, Miguel Gueimonde,
and Patricia Ruas-Madiedo

viii Contents

Faecal Microbiota Transplantation as Emerging Treatment

in European Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
Marcello Maida, James Mcilroy, Gianluca Ianiro,
and Giovanni Cammarota
Immunization Strategies Against Clostridium difficile . . . . . . . . . . 197
Jean-François Bruxelle, Séverine Péchiné, and Anne Collignon
Non-human C. difficile Reservoirs and Sources: Animals, Food,
Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Cristina Rodriguez Diaz, Christian Seyboldt, and Maja Rupnik
The ESCMID Study Group for Clostridium difficile:
History, Role and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
John E. Coia and Ed J. Kuijper

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Economic Burden of Clostridium difficile
Infection in European Countries

Elena Reigadas Ramı́rez and Emilio Santiago Bouza

Abstract Keywords
Clostridium difficile infection (CDI) remains C. difficile infection · Economic costs ·
a considerable challenge to health care Economic burden · Length of stay · Europe
systems worldwide. Although CDI represents
a significant burden on healthcare systems in
Europe, few studies have attempted to esti-
mate the consumption of resources associated
with CDI in Europe. The reported extra costs
attributable to CDI vary widely according to 1 Introduction
the definitions, design, and methodologies
used, making comparisons difficult to per- In this chapter, the economic burden of
form. In this chapter, the economic burden of healthcare facility-associated Clostridium diffi-
healthcare facility–associated CDI in Europe cile infection (CDI) in Europe will be assessed,
will be assessed, as will other less explored as will other less explored areas such as recurrent
areas such as the economic burden of recur- CDI (R-CDI), community-acquired CDI, pediat-
rent CDI, community-acquired CDI, pediatric ric CDI, and CDI in outbreaks.
CDI, and CDI in outbreaks. Despite advances in the diagnosis and treat-
ment of CDI and prevention efforts to reduce the
incidence of CDI, the disease remains a signifi-
cant challenge to health care systems worldwide

E.S. Bouza (*)

Department of Clinical Microbiology and Infectious
E. Reigadas Ramı́rez (*) Diseases, Hospital General Universitario Gregorio
Department of Clinical Microbiology and Infectious Marañón, Madrid, Spain
Diseases, Hospital General Universitario Gregorio Medicine Department, School of Medicine,
Marañón, Madrid, Spain Universidad Complutense de Madrid (UCM),
Medicine Department, School of Medicine, Madrid, Spain
Universidad Complutense de Madrid (UCM), Instituto de Investigación Sanitaria Gregorio Marañón,
Madrid, Spain Madrid, Spain
Instituto de Investigación Sanitaria Gregorio Marañón, CIBER de Enfermedades Respiratorias (CIBERES CB06/
Madrid, Spain 06/0058), Madrid, Spain
e-mail: e-mail:

# Springer International Publishing AG 2018 1

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
2 E. Reigadas Ramı́rez and E.S. Bouza

(Dubberke and Olsen 2012; Bouza 2012). From separate the costs of resources due to CDI from
an economic point of view, CDI increases patient those generated by the underlying disease.
healthcare costs as a result of extended length of Therefore, comparisons need to be made with
hospital stay (LOS), re-admission, laboratory caution and limited to results obtained in a simi-
tests, and medication (Wiegand et al. 2012; lar manner.
Gabriel and Beriot-Mathiot 2014; Nanwa et al. A clearer understanding of the healthcare and
2015). C. difficile infection is costly, not only to economic burden of CDI is of value to hospital
third-party payers and hospitals, but also to soci- administrators, infection prevention teams, and
ety as a whole (McGlone et al. 2012). persons involved in antimicrobial stewardship
Most of the existing literature is from the programs, who can use this key information to
United States, where an in silico economic determine the appropriate degree of investment
model, reported in 2012, suggested that the in infection control measures and in other prior-
annual US economic burden of CDI would be ity areas.
$496 million from a hospital perspective, $547 Future studies should follow standard meth-
million from a third-party payer perspective, and odology, include other indirect cost perspectives
$796 million from a societal perspective such as societal and patient perspectives, and
(McGlone et al. 2012). Regrettably, few examine poorly explored populations, such as
published studies have attempted to estimate the individuals with community-acquired CDI.
consumption of resources associated with CDI in
Europe (Wiegand et al. 2012) and it has been
estimated that the annual cost of CDI in Europe 2 Economic Burden of Hospital-
is €3 billion per year (Jones et al. 2013); conse- Acquired CDI in European
quently, approaches that can reduce Countries
CDI-associated resource use and costs are of
interest. A wide range of CDI costs in Europe have been
Although antibiotics are a key component of reported, ranging from €5798–€11,202/episode
therapy for CDI, they currently represent a mini- (Wiegand et al. 2012). Data are only available
mal cost in the overall budget for CDI manage- from six European countries (Ireland, England,
ment, and the main extra associated cost reported Wales, Germany, Spain, and Italy). Table 1
in most studies is the extended LOS attributable summarizes CDI costs by study and country.
to CDI (Wiegand et al. 2012; Asensio et al. 2013,
2015; Wilcox et al. 1996; Hubner et al. 2015).
CDI-related costs are also likely to increase as
the population ages. In a systematic European 2.1 Primary Episodes
meta-analysis on clinical and economic burden,
the authors reported that the incremental cost of The economic burden of primary episodes in
CDI may have increased by £1857–£4266 Europe is reviewed below by region, as defined
(27–93%) over a 12-year period (Wiegand et al. by EuroVoc. The most abundant literature comes
2012). In a review by Kuijper et al., the potential from Western Europe, followed by Southern
cost of CDI was estimated to be €3 billion/year Europe.
and is expected to almost double over the next
four decades, assuming a European Union popu- 2.1.1 Western Europe
lation of 457 million inhabitants (Kuijper et al.
2006). A recent study conducted in a tertiary referral
The reported extra costs attributable to CDI hospital in Ireland during August 2015 showed
vary widely according to the definitions, design, that the total incremental cost of CDI was
and methodologies used (Ghantoji et al. 2010; €75,680, with a mean cost of €5820 per patient
Wiegand et al. 2012). Most studies do not (Ryan et al. 2017).
Economic Burden of Clostridium difficile Infection in European Countries 3

Table 1 Summary of European Clostridium difficile infection costs by study and country
Author Country CDI cases Study
(year) or region examined Study population period Cost
Ryan et al. Ireland N ¼ 13 Healthcare CDI August €5820/CDI
(2017) 2015
Al-Eidan Ireland N ¼ 87 Healthcare CDI 1994–1995 ₤2860/CDI
et al. (2000)
Wilcox England N ¼ 50 Healthcare CDI 1994–1995 ₤4107/CDI
et al. (1996)
Wilcox United N ¼ 128 Healthcare CDI, 2012–2014 ₤6294/CDI
et al. (2017) Kingdom recurrences ₤7539/recurrent CDI
Vonberg Germany N ¼ 116 Healthcare CDI 2006 €7147/CDI
et al. (2008)
Hubner Germany N ¼ 43 Healthcare CDI 2010 €5262.96/CDI
et al. (2015)
Grube et al. Germany N ¼ 2767 Healthcare CDI, 2011 €4132/CDI as primary diagnosis
(2015) recurrences €19,381/CDI as secondary diagnosis
€20,755/recurrent CDI
Le Monnier France N ¼ 1097 Healthcare CDI, 2011 €9575/CDI (€6056 CDI as primary
et al. (2015) recurrences diagnosis/€11,251 CDI as secondary
€9625/recurrent CDI
Asensio Spain N ¼ 7601 Healthcare CDI, 2012 €3901/CDI
et al. (2013) recurrences €4875/first recurrent CDI
€5916/second recurrent CDI
Asensio Spain N ¼ 232 Healthcare CDI, 2011–2013 €4265/CDI case (Spain)
et al. (2015) and Italy (Spain) recurrences, (adults)
N ¼ 145 children 2006–2012 €14,936/adult CDI case (Italy)
(Italy) (pediatrics) €17,714/recurrent CDI case (Italy)
€3545/pediatric CDI case (Italy)

Another study conducted in Ireland (Le Monnier et al. 2015). The costs of CDI
established the mean cost per treated case of incurred by public insurance and by the hospital
CDI in terms of bed occupancy, laboratory itself (euros) were based on full unit cost per
requests, and treatment to be £4577 (2010 GBP) diagnosis-related group in hospitals at 2010
(Al-Eidan et al. 2000). values. The annual incidence of CDI based on
It has been estimated that the cost for CDI is laboratory reporting was estimated at 3.74 cases
€5000–€15,000 per case in England (Kuijper per 10,000 patient-days. In cases where CDI was
et al. 2006). The earliest data on economic bur- the primary diagnosis, the mean cost per stay was
den from England were communicated by €6056 (median €4410) and the cumulative cost
Wilcox et al. (1996), who performed a study in for the whole set of stays observed in 2011 for the
geriatric wards. Cases and controls were matched 12 hospitals was €823,656. In patients where
for age, sex, and distribution of the main CDI was considered a secondary diagnosis, the
diagnoses. The total identifiable increased cost mean extra cost adjusted for age, sex, and
of CDI was £6986 in 2010 GBP. diagnosis-related groups in cases without CDI
A retrospective multicenter study analyzed a was €11,251 (median: €8822) per stay
sample of 12 large, public, acute-care hospitals in (Le Monnier et al. 2015). The extrapolated
France representing 5.82% of the cumulative annual nationwide cost of CDI in 2011 in France
annual number of patient-days spent in public was €163.1 million.
acute-care hospitals in France in 2011
4 E. Reigadas Ramı́rez and E.S. Bouza

A single-center retrospective analysis of data per episode of CDI was €3901 for initial or
from patients with nosocomial CDI carried out primary CDI episodes.
over a 1-year period at a teaching hospital in More recently, another study assessed the
Germany showed an additional cost of €5262/ impact of CDI on hospital resources and costs
case (Hubner et al. 2015). in both Spain and Italy (Asensio et al. 2015).
Another single-center German study showed Each patient was matched with two randomly
that costs for CDI patients were significantly selected uninfected controls in the same institu-
higher than for their matched controls (median: tion. Data were collected for 232 adult infected
€7147) (Vonberg et al. 2008). A large multicen- patients and 426 matched non-infected patients
ter study conducted in 37 German hospitals in Spain (n ¼ 106) and Italy (n ¼ 126).
based on data from the German DRG system CDI-associated costs were due to excess hospi-
analyzed 2767 CDI cases grouped according to talization. The difference in LOS between the
whether CDI was a primary or secondary diag- two countries resulted in a significant variation
nosis (Grube et al. 2015). For comparison, in costs.
non-CDI cases from the same hospitals during Hospitalization costs attributable to CDI in
the same year were matched using propensity Spain were €4265 per patient for all patients,
score matching. €2882/patient for patients aged 65 years, and
Patients from the primary diagnosis group €4885 for those aged >65 years (Asensio et al.
(n ¼ 817) showed a mean cost per case of 2015).
€4132 (€536 more than controls), while the For Italy, the total cost attributable to CDI was
secondary diagnosis group (n ¼ 1840) had €14,023 per patient for all patients. The cost was
costs of €19,381 (€13,082 for controls) (Grube €15,668 for those aged 65 years and €13,862
et al. 2015). The authors extrapolated their data for those aged >65 years, with the difference in
and declared that CDI generates a yearly cost cost being due to differences in LOS (21 vs.
burden of €464 million for the German 19 days, respectively). The authors estimated a
healthcare system. cost of CDI in Italy of €32,371 per 10,000
patient-days (Asensio et al. 2015).
2.1.2 Southern Europe A recent multicenter Italian cost analysis
study has been performed in hospitalized patients
Evidence regarding the impact of CDI on from the hospital’s perspective (Poli et al. 2015).
healthcare resources in southern Europe is gen- This study showed that the mean total incremen-
erally scarce. In the case of Spain, few studies tal cost for a patient with CDI was €3270
have assessed the economic burden of CDI. An per case.
economic model analysis performed in 2012 by
Asensio et al.(Asensio et al. 2013) assessed the
cost of CDI in adult patients (18 years) treated 2.2 Recurrent Episodes
with metronidazole or vancomycin from the per-
spective of the Spanish National Health System One of the first studies to assess the cost of
Service. The resources used in clinical practice recurrent CDI in an European country was a
were obtained through a Delphi panel of Spanish Spanish study in which the cost of the initial
clinicians with expertise in CDI. Unit costs CDI episode was estimated to be €3901, the
(€2012) were obtained from Spanish sources. cost of the first recurrence was €4875, and that
This study estimated that 7601 episodes of of the second recurrence was €5916 (Asensio
CDI occur annually in Spain (incidence of 17.1 et al. 2013).
episodes/year/10,000 hospital discharges) with In an Italian multicenter study including
an estimated annual cost to the Spanish National recurrences (Asensio et al. 2015), the cost attrib-
Health System Service of €32,157,093. The cost utable to recurrent CDI was €17,714 per patient,
Economic Burden of Clostridium difficile Infection in European Countries 5

while for patients with a single episode of CDI, excess LOS attributable to CDI are limited. Not
the cost was €14,936. In this study, a total of many studies assess the attributable LOS,
34 adult patients (12.5%) and 2 pediatric patients reporting only total LOS. It was recently
(10.5%) experienced a first recurrence of CDI. suggested that, compared with newer statistical
Three of the 34 adult patients and 1 of the 2 pedi- models, models that were previously used to
atric patients had an additional recurrence. determine the LOS attributable to CDI
A French multicenter study estimated the overestimated the additional LOS (Mitchell and
median extra cost per stay with CDI to be Gardner 2012). Therefore, future studies must
€7514, i.e., approximately €9.5 million in 2011 take this into account. Table 2 shows the
for the 12 facilities included. The fraction of that European studies reporting LOS attributable to
total cost attributable to recurrences was 12.5% CDI; the mean incremental LOS attributable to
(Le Monnier et al. 2015). Recurrences occurring CDI ranged from 4.2 to 20 days (Ryan et al.
in acute-care settings were present in 12.0% of 2017).
hospital stays with CDI. In addition, 9.3% As for recurrent CDI, the mean incremental
(11/118) of recurrences were coded as the pri- LOS in Europe is 9.1–26 days (Asensio et al.
mary diagnosis and led to readmission of the 2013, 2015). Although data may vary, most stud-
patients, which resulted in prolonged LOS and ies agree that recurrent CDI presents longer LOS
additional medical costs. than primary episodes. In a recent study
Data from 37 German hospitals revealed high conducted in England, Wilcox et al. observed a
costs for recurrent CDI of €20,755 vs. €13,101 median LOS of 21 days for recurrent CDI in
for matched controls from the same hospitals contrast to 15.5 days for first episodes (Wilcox
during the same year (Grube et al. 2015). et al. 2017).
Wilcox et al. recently analyzed the impact of Few studies have assessed differences in extra
recurrent CDI in terms of hospital resource use costs between mild to moderate CDI cases and
and health-related quality of life associated severe CDI cases. A study conducted by van
with hospitalizations for recurrent CDI in six Kleef et al. in a large English teaching hospital
UK acute-care hospitals (Wilcox et al. 2017). showed that severe cases had an average excess
The median cost per patient during a 28-day LOS which was twice that of the nonsevere cases
post-index period was £7539 for recurrent CDI (11.6 days [95% CI, 3.6–19.6] vs. approximately
and £6294 for first CDI episodes (Wilcox et al. 5 days [95% CI: 1.1–9.5]) (van Kleef et al. 2014).

2.4 Distribution of Costs

2.3 Length of Stay
The expense associated with CDI stems mainly
In their review, Wiegand et al. (2012) estimated from extended LOS. Various studies in Europe
the average LOS in Europe to be 15 days. When place the additional cost of LOS at 43.2–95.6%
examined by country, they found that of the total extra costs of the CDI episode (Ryan
Switzerland had the lowest LOS (12 days), et al. 2017; Asensio et al. 2013, 2015; Wilcox
followed by Belgium, France, Ireland (17 days), et al. 1996; Poli et al. 2015). Figure 1 represents
and Spain (18 days), while the highest LOS were the distribution of CDI costs of the above men-
observed for The Netherlands (21 days), tioned studies.
Germany (27 days), and the UK (37 days) In contrast, cost for CDI antibiotics account
(Wiegand et al. 2012). for a low percentage of the total cost, ranging
Even though LOS values are more reproduc- from 0.43% to 13.3% (Ryan et al. 2017; Asensio
ible between studies than costs, data on the et al. 2013, 2015; Wilcox et al. 1996; Poli et al.
6 E. Reigadas Ramı́rez and E.S. Bouza

Table 2 Length of stay (LOS) attributable to Clostridium difficile by study and country
CDI cases Study
Author (year) Country examined Study population period LOS attributable to CDI (days)
Eckmann Netherlands N ¼ 270 Healthcare CDI 2008–2009 Mean 12.58
et al. (2013)
Ryan et al. Ireland N ¼ 13 Healthcare CDI August 2015 Mean 4.2
Al-Eidan Ireland N ¼ 87 Healthcare CDI 1994–1995 Mean 13
et al. (2000)
Eckmann England N ¼ 10,602 Healthcare CDI 2007–2009 Mean 16.09
et al. (2013)
van Kleef England N ¼ 157 Healthcare CDI 2012 Mean 7.2 (all CDI)
et al. (2014) Mean 11.6 (severe CDI)
Mean 5.3 (non severe CDI)
Vonberg Germany N ¼ 116 Healthcare CDI 2006 Median 7
et al. (2008)
Eckmann Germany N ¼ 109,526 Healthcare CDI 2008–2010 Mean 15.47
et al. (2013)
Hubner et al. Germany N ¼ 43 Healthcare CDI 2010 Mean 11.4
Le Monnier France N ¼ 1097 Healthcare CDI, 2011 Mean 8.9
et al. (2015) recurrences
Eckmann Spain N ¼ 830 Healthcare CDI 2008–2010 Mean 13.56
et al. (2013)
Asensio et al. Spain N ¼ 7601 Healthcare CDI, 2012 Mean 7.4 (CDI)
(2013) recurrences Mean 9.1 (first recurrent CDI)
Mean 10.8 (second recurrent
Asensio et al. Spain and N ¼ 232 Healthcare CDI, 2011–2013 Median 6.4 (Madrid)
(2015) Italy (Spain) recurrences, children (adults)
N ¼ 145 2006–2012 Median 20.0 (Barcelona)
(Italy) (pediatrics) Median 20.0 (Rome)
Median 26.0 for first recurrent
CDI case (Spain and Italy)
Median 5.0 for pediatric case

2015). Figure 2 illustrates the distribution of CDI-specific drugs was higher in R-CDI patients
costs in patients with CDI antibiotics as percent (£376/patient) than first-case CDI (£46/patient)
of total cost per country. Most of these studies (Wilcox et al. 2017).
only include vancomycin and metronidazole as
treatment for CDI, probably because they were
conducted before fidaxomicin was licensed in
3 Economic Burden
those countries. Only one recent study conducted
of Community-Acquired CDI
in Ireland included fidaxomicin as treatment for
Community-acquired CDI is a growing problem,
Regarding distribution of costs for R-CDI, a
and additional data are needed to accurately
recent study conducted in England observed that
quantify the contribution of this subpopulation
the cost of hospital admissions and emergency
to the overall burden of CDI. Few studies provide
department visits accounted for more than 85%,
insight on this understudied patient group (Kuntz
similar to first-case CDI. The median cost for
et al. 2012; Sammons et al. 2013; Nanwa et al.
Additional Antimicrobial
Radiology Laboratory
2.60% Ireland Personnel
England CDI
4.20% 1.14%
Higienic Antimicrobial
measures CDI
2.70% 13.30%

Length of Stay
43.20% Length of

Higienic Surgery or
measures Germany CDI
Additional Italy other
4.83% Personnel treatments
0.43% 1.10%
Higienic 0.10%
measures CDI
8.00% 1.10%

Length of Stay diagnosis
Isolation 48.56% 2.50%
45.85% Length of Stay

Surgery or
other Spain Antimicrobial
treatments CDI
2.80% 0.50%

Length of Stay

Fig. 1 Distribution of costs of Clostridium difficile infection

Fig. 2 Costs of antibiotics

for CDI treatment as
percentage of the total costs
attributable to Clostridium
difficile infection
8 E. Reigadas Ramı́rez and E.S. Bouza

2017), and none have been performed in for hospital-onset CDI and 5.55 days and
European patients. In addition, across studies, $18,900 for community-onset CDI. Although
the case definition of community-acquired CDI mortality rates did not differ between those with
may differ depending on the time between a community-onset CDI and matched unexposed
previous hospital admission and whether the subjects, community-onset CDI patients had sig-
case of CDI was an incident case (Kutty et al. nificantly longer LOS and total hospital costs
2010; Freeman et al. 2010). (Sammons et al. 2013).
The most recent and largest study is a Kuntz et al. performed a population-based
population-based matched cohort study examin- study in which they identified 3067 CDIs and
ing the mortality and costs of patients with classified CDI by whether it was identified in
community-onset CDI identified based on emer- the outpatient or inpatient healthcare setting
gency department visits and hospital admissions (Kuntz et al. 2012). A total of 1712 (56%) were
in Ontario, Canada (Nanwa et al. 2017). In this identified in the outpatient setting. These patients
study, Nanwa et al. studied 7950 subjects with tended to be younger, with fewer comorbid
community-onset CDI and found that up to conditions than patients with CDI identified in
1 year after the index date, the disease was the inpatient setting. Eleven percent of patients
associated with 1.9- to 5.1-fold higher mean with outpatient-identified CDI were hospitalized
costs (CDN$10,700 in 2014) than in uninfected with a CDI-related diagnosis code during the
subjects. The largest cost components were follow-up period. These hospitalizations
hospitalizations and physician visits. occurred, on average, 27 days after outpatient
Differences in mortality and costs remained identification of CDI and lasted an average of
3 years after index date, probably owing to 10 days.
recurrences of CDI, treatment failure, need for As expected, the impact of CDI on healthcare
colectomy, and increased susceptibility to other utilization and cost was most notable in the
conditions resulting from CDI. Mean attributable setting in which the patient’s infection had been
costs were higher among those aged >65 years, identified. Outpatient care costs were higher
those infected in 2008 (year of an outbreak in among persons with CDI identified in the outpa-
Ontario caused by a particularly virulent CDI tient setting, with drugs representing the greatest
strain (Pillai et al. 2010)), and those who died percentage of these costs in both groups. Simi-
within 1 year after the index date. However, this larly, patients with inpatient-identified CDI had
study had a major limitation, namely, the authors higher inpatient costs than patients with
were not able to identify subjects whose only outpatient-identified CDI ($10,708.40 vs
contact with the healthcare system was a visit to $837.40). Total costs for community-onset CDI
their family physician. Consequently, mortality were $1697 vs $11,315 of hospital-onset CDI
and the economic burden of community-onset (in US $2009 per patient) (Kuntz et al. 2012).
CDI per subject were likely overestimated,
since all of the cases included were probably
more severe. 4 Pediatric Population
Sammons et al. examined a cohort of children
and performed a subanalysis on community- Data on the burden of CDI in children are even
onset and hospital-onset CDI (Sammons et al. more limited and most of the literature on this
2013). They found that patients with topic comes from studies performed in the
community-onset CDI comprised 54% of cases United States. In Italy, Asensio et al. (Asensio
(2414 cases). Patients with hospital-onset CDI et al. 2015) reported separate data on the eco-
had significantly higher mortality rates and lon- nomic burden of CDI in children, although they
ger LOS than those with community-onset CDI, found that the number of patients included was
and mean differences in LOS and total low (n ¼ 19). Most cases of CDI in children were
standardized costs were 21.60 days and $93,600 community-acquired as opposed to nosocomial.
Economic Burden of Clostridium difficile Infection in European Countries 9

Disease characteristics were generally compara- $12,801 (P < 0.001), accounting for 8295 days
ble to those of adults, although the incidence of spent in the hospital and $18.4 million (2012
ulceration and bowel wall thickening was higher USD) in annual expenditure (Kulaylat et al.
than in adults. The authors found that the median 2017).
LOS attributable to CDI was lower than in adults
(5 vs. 19 days in Rome), as was the frequency of
isolation and admission to the ICU, probably 5 Economic Costs of CDI
because most cases were community-acquired. Outbreaks
Therefore, although daily costs of care are higher
for children than adults, the overall burden of Few data have been published on the costs
CDI in the pediatric population in Italy is lower derived from outbreaks. One of the few studies
than in adults. The total cost attributable to CDI to assess this situation was that conducted in
in pediatric patients in Naples was €3545 per Ireland by Ryan et al. (2017). The authors col-
patient (Asensio et al. 2015). lected data on LOS, diagnosis, diagnosis-related
The only data on the economic burden of group codes at discharge, time in isolation
pediatric CDI in larger populations are from because of CDI, additional measures because of
American studies. In their multicenter cohort CDI (medications, consultations, investigations,
study, Sammons et al. found that CDI was and procedures), unit costs (laboratory testing,
associated with worse outcomes among personal protective equipment, single room
hospitalized children who were otherwise similar accommodation, and cleaning/decontamination),
in the main demographic and clinical and personnel time.
characteristics, although the difference was This study covered only a 1-month period
most pronounced in children with hospital onset (August 2015), during which they observed that
disease. The presence of CDI was associated the CDI outbreak resulted in additional costs of
with >6-fold higher mortality rates among €46,967. The outbreak resulted in 58 bed days
those with healthcare-onset CDI and resulted in lost due to bed closures on the outbreak ward,
significantly longer LOS and increased total hos- with an estimated value of €34,585. Five out-
pital costs, corresponding to a mean difference in break control meetings were held, each with a
total standardized costs of $48,500 between mean duration of 47 min and supported by 15 h
matched exposed and unexposed patients of administrative input. All meetings involved a
(Sammons et al. 2013). consultant microbiologist, a senior laboratory
In another study performed in acute care scientist, a senior antimicrobial specialist phar-
hospitals in the Michigan Health and Hospital macist, an assistant director of nursing, multiple
Association, children younger than 5 years of clinical nursing managers, and a number of other
age had mean charges of $148,525, compared staff members. The mean personnel cost per
with $56,796 for discharges of patients who meeting was €546, and the aggregate cost was
were aged 65 years, probably because of longer €2728. The cost of outbreak-related cleaning/
LOS: children younger than 5 years of age were decontamination during August was €9654
hospitalized for a mean of more than 25 days per (Ryan et al. 2017).
discharge vs 14.2 days for the remaining age For the patients involved in the CDI outbreak,
groups reported (VerLee et al. 2012). excluding the value of the 58 bed days lost
A large propensity score–matching analysis in (€34,585), costs were 30% higher (€7589 per
313,664 patients aged 1–18 years was performed patient) than those not involved in the outbreak
to evaluate the influence of CDI on mortality, during the same period (Ryan et al. 2017).
LOS, and costs in hospitalized surgical pediatric Van Beurden et al. assessed the costs of an
patients. The authors observed that after propen- outbreak of C. difficile ribotype 027 at the VU
sity score matching, the mean excess LOS and University medical center, a 750-bed tertiary
costs attributable to CDI were 5.8 days and care center in The Netherlands, from May 2013
10 E. Reigadas Ramı́rez and E.S. Bouza

to May 2014 (van Beurden et al. 2017). Several prolonged LOS is a major cost. The economic
control measures were implemented, such as and healthcare impact of loss of revenue is very
reinforcement of infection control, the introduc- difficult to determine, and closed beds prevent
tion of hydrogen peroxide as disinfectant, extra inpatient accommodation, with the resultant
cleaning, optimization of CDI diagnosis, optimi- morbidity and mortality (Singer et al. 2011). In
zation of CDI treatment, and antibiotic steward- addition, increased bed usage by medical
ship. Twelve meetings of the outbreak specialties is associated with cancelled elective
management team (consisting of five medical surgeries (Robb et al. 2004; Nasr et al. 2004).
specialists, one infection prevention specialist, Outbreak control generates extra work, which
one care manager, and two co-workers from often relies on staff already overburdened with
facility management) were held during the administrative tasks from patient care activities.
study period. Several beds had to be closed to Extra cleaning measures and multidisciplinary
ensure that every patient with suspected CDI was infection control teams are key elements for out-
placed in contact isolation in a single room. After break control (Barbut et al. 2011, 2015).
the implementation of these control measures, Healthcare facilities should be able to assess the
the incidence of CDI decreased to around 1.5 economic impact of an outbreak, and knowing
cases per 10,000 patient days in early 2014. the costs of additional measures will make it
Missed revenue due to prolonged LOS among possible to establish a cost-efficient program for
CDI patients, costs of the outbreak meetings, outbreak control, with adequate resource
extra surveillance, contact isolation material allocation.
(compared with the same period 1 year earlier It is obvious to state that accounting of LOS,
and 1 year later), and additional microbiological cost of antimicrobial agents and other expenses
diagnostics (compared with the same period of the healthcare system are unable to quantify
1 year earlier) were calculated directly from the cost of pain, human suffering, and death.
available data for the entire outbreak. Overall
costs for additional cleaning, contact isolation,
and missed revenue due to closed beds were
extrapolated from the costs incurred during the
previous 3 months of the outbreak. Attributable Al-Eidan FA, McElnay JC, Scott MG, Kearney MP
costs per item (in 2014 euros) were assessed over (2000) Clostridium difficile-associated diarrhoea in
a 365-day period. hospitalised patients. J Clin Pharm Ther 25(2):101–
The total identifiable costs of this C. difficile 109
Asensio A, Bouza E, Grau S, Rubio-Rodriguez D, Rubio-
outbreak were €1,222,376. Most costs (36%) Terres C (2013) Cost of Clostridium difficile
stemmed from the loss of revenue resulting associated diarrhea in Spain. Rev Esp Salud Publica
from decreased hospital capacity because of the 87(1):25–33.
increased LOS of CDI patients and the closure of 000100004
Asensio A, Di Bella S, Lo Vecchio A, Grau S, Hart WM,
multiple beds to ensure contact isolation of a Isidoro B, Scotto R, Petrosillo N, Watt M, Nazir J
single CDI patient. Twenty-five percent of the (2015) The impact of Clostridium difficile infection
costs were from extra surveillance and the work on resource use and costs in hospitals in Spain and
of the department of infection control, 24% were Italy: a matched cohort study. Int J Infect Dis 36:31–
for extra cleaning of the affected wards, 6% for Barbut F (2015) How to eradicate Clostridium difficile
extra microbiological diagnostic procedures, 3% from the environment. J Hosp Infect 89(4):287–295.
for the outbreak meetings, and 3% for the use of
extra gloves and aprons. Extra antibiotic treat- Barbut F, Jones G, Eckert C (2011) Epidemiology and
control of Clostridium difficile infections in healthcare
ment of CDI patients counted for 2% of the total settings: an update. Curr Opin Infect Dis 24(4):370–
costs (van Beurden et al. 2017). 376. 834748e5
As can be seen in both studies, the cost of one Bouza E (2012) Consequences of Clostridium difficile
missed hospital admission due to closed beds or infection: understanding the healthcare burden. Clin
Economic Burden of Clostridium difficile Infection in European Countries 11

Microbiol Infect 18(Suppl 6):5–12. Kutty PK, Woods CW, Sena AC, Benoit SR, Naggie S,
1111/1469-0691.12064 Frederick J, Evans S, Engel J, McDonald LC (2010)
Dubberke ER, Olsen MA (2012) Burden of Clostridium Risk factors for and estimated incidence of commu-
difficile on the healthcare system. Clin Infect Dis 55 nity-associated Clostridium difficile infection, North
(Suppl 2):S88–S92. Carolina, USA. Emerg Infect Dis 16(2):197–204.
Eckmann C, Wasserman M, Latif F, Roberts G, Beriot-
Mathiot A (2013) Increased hospital length of stay Le Monnier A, Duburcq A, Zahar JR, Corvec S, Guillard
attributable to Clostridium difficile infection in T, Cattoir V, Woerther PL, Fihman V, Lalande V,
patients with four co-morbidities: an analysis of hos- Jacquier H, Mizrahi A, Farfour E, Morand P, Marcade
pital episode statistics in four European countries. Eur G, Coulomb S, Torreton E, Fagnani F, Barbut F (2015)
J Health Econ 14(5):835–846 Hospital cost of Clostridium difficile infection includ-
Freeman J, Bauer MP, Baines SD, Corver J, Fawley WN, ing the contribution of recurrences in French acute-
Goorhuis B, Kuijper EJ, Wilcox MH (2010) The care hospitals. J Hosp Infect 91(2):117–122. https://
changing epidemiology of Clostridium difficile
infections. Clin Microbiol Rev 23(3):529–549. McGlone SM, Bailey RR, Zimmer SM, Popovich MJ, Tian Y, Ufberg P, Muder RR, Lee BY (2012) The
Gabriel L, Beriot-Mathiot A (2014) Hospitalization stay economic burden of Clostridium difficile. Clin
and costs attributable to Clostridium difficile infection: Microbiol Infect 18(3):282–289.
a critical review. J Hosp Infect 88(1):12–21. https:// 1111/j.1469-0691.2011.03571.x Mitchell BG, Gardner A (2012) Prolongation of length of
Ghantoji SS, Sail K, Lairson DR, DuPont HL, Garey KW stay and Clostridium difficile infection: a review of the
(2010) Economic healthcare costs of Clostridium dif- methods used to examine length of stay due to
ficile infection: a systematic review. J Hosp Infect healthcare associated infections. Antimicrob Resist
74(4):309–318. Infect Control 1(1):14.
10.016 2994-1-14
Grube RF, Heinlein W, Scheffer H, Rathmayer M, Schepp Nanwa N, Kendzerska T, Krahn M, Kwong JC, Daneman
W, Lohse AW, Stallmach A, Wilke MH, Lerch MM N, Witteman W, Mittmann N, Cadarette SM, Rosella
(2015) Economic burden of Clostridium difficile L, Sander B (2015) The economic impact of Clostrid-
enterocolitis in German hospitals based on routine ium difficile infection: a systematic review. Am J
DRG data. Z Gastroenterol 53(5):391–397. https:// Gastroenterol 110(4):511–519. 1038/ajg.2015.48
Hubner C, Hubner NO, Muhr M, Claus F, Leesch H, Nanwa N, Sander B, Krahn M, Daneman N, Lu H, Austin
Kramer A, Flessa S (2015) Cost analysis of PC, Govindarajan A, Rosella LC, Cadarette SM,
hospitalized Clostridium difficile-associated diarrhea Kwong JC (2017) A population-based matched cohort
(CDAD). GMS Hyg Infect Control 10:Doc13. https:// study examining the mortality and costs of patients with community-onset Clostridium difficile infection
Jones AM, Kuijper EJ, Wilcox MH (2013) Clostridium identified using emergency department visits and hos-
difficile: a European perspective. J Inf Secur pital admissions. PLoS One 12(3):e0172410. https://
019 Nasr A, Reichardt K, Fitzgerald K, Arumugusamy M,
Kuijper EJ, Coignard B, Tull P (2006) Emergence of Keeling P, Walsh TN (2004) Impact of emergency
Clostridium difficile-associated disease in North admissions on elective surgical workload. Ir J Med
America and Europe. Clin Microbiol Infect 12(Suppl Sci 173(3):133–135
6):2–18. Pillai DR, Longtin J, Low DE (2010) Surveillance data on
01580.x outbreaks of Clostridium difficile infection in Ontario,
Kulaylat AN, Rocourt DV, Podany AB, Engbrecht BW, Canada, in 2008–2009. Clin Infect Dis 50(12):1685–
Twilley M, Santos MC, Cilley RE, Hollenbeak CS, 1686. author reply
Dillon PW (2017) Costs of Clostridium difficile infec- 1686
tion in pediatric operations: a propensity score- Poli A, Di Matteo S, Bruno GM, Fornai E, Valentino MC,
matching analysis. Surgery 161(5):1376–1386. Colombo GL (2015) Economic burden of Clostridium difficile in five hospitals of the Florence health care
Kuntz JL, Johnson ES, Raebel MA, Petrik AF, Yang X, system in Italy. Risk Manag Healthc Policy 8:207–
Thorp ML, Spindel SJ, Neil N, Smith DH (2012) 213.
Epidemiology and healthcare costs of incident Clos- Robb WB, O’Sullivan MJ, Brannigan AE, Bouchier-
tridium difficile infections identified in the outpatient Hayes DJ (2004) Are elective surgical operations can-
healthcare setting. Infect Control Hosp Epidemiol celled due to increasing medical admissions? Ir J Med
33(10):1031–1038. Sci 173(3):129–132
12 E. Reigadas Ramı́rez and E.S. Bouza

Ryan P, Skally M, Duffy F, Farrelly M, Gaughan L, Flood difficile infection: a multi-state modelling approach. J
P, McFadden E, Fitzpatrick F (2017) Evaluation of Hosp Infect 88:213–217
fixed and variable hospital costs due to Clostridium VerLee KE, Finks JL, Wilkins MJ, Wells EV (2012)
difficile infection: institutional incentives and Michigan Clostridium difficile hospital discharges:
directions for future research. J Hosp Infect 95 frequency, mortality, and charges, 2002–2008. Public
(4):415–420. Health Rep 127(1):62–71
016 Vonberg RP, Reichardt C, Behnke M, Schwab F, Zindler
Sammons JS, Localio R, Xiao R, Coffin SE, Zaoutis T S, Gastmeier P (2008a) Costs of nosocomial Clostrid-
(2013) Clostridium difficile infection is associated ium difficile-associated diarrhoea. J Hosp Infect 70
with increased risk of death and prolonged hospitali- (1):15–20.
zation in children. Clin Infect Dis 57(1):1–8. https:// Wiegand PN, Nathwani D, Wilcox MH, Stephens J, Shelbaya A, Haider S (2012) Clinical and economic
Singer AJ, Thode HC Jr, Viccellio P, Pines JM (2011) The burden of Clostridium difficile infection in Europe: a
association between length of emergency department systematic review of healthcare-facility-acquired
boarding and mortality. Acad Emerg Med 18 infection. J Hosp Infect 81(1):1–14.
(12):1324–1329. 10.1016/j.jhin.2012.02.004
2011.01236.x Wilcox MH, Cunniffe JG, Trundle C, Redpath C (1996)
Van Beurden YH, Bomers MK, van der Werff SD, Pompe Financial burden of hospital-acquired Clostridium dif-
EA, Spiering S, Vandenbroucke-Grauls CM, Mulder ficile infection. J Hosp Infect 34(1):23–30
CJ (2017) Cost analysis of an outbreak of Clostridium Wilcox MH, Ahir H, Coia JE, Dodgson A, Hopkins S,
difficile infection ribotype 027 in a Dutch tertiary care Llewelyn MJ, Settle C, McLain-Smith S, Marcella
centre. J Hosp Infect 95(4):421–425. SW (2017) Impact of recurrent Clostridium difficile
10.1016/j.jhin.2016.12.019 infection: hospitalization and patient quality of life. J
Van Kleef E, Green N, Goldenberg SD, Robotham JV, Antimicrob Chemother 72:2647–2656.
Cookson B, Jit M, Edmunds WJ, Deeny SR (2014) 10.1093/jac/dkx174
Excess length of stay and mortality due to Clostridium
The Need for European Surveillance of CDI

Camilla Wiuff, A-Lan Banks, Fidelma Fitzpatrick,

and Laura Cottom

For surveillance systems to be useful, they must adapt to the changing environment in
which they operate and accommodate emerging public health requirements that were not
conceived previously.
Joseph S. Lombardo and David L. Buckeridge

Abstract have prevented the true burden of disease from

Since the turn of the millennium, the epidemiol- being appreciated.
ogy of Clostridium difficile infection (CDI) has Acceptance that a multi-country surveillance
continued to challenge. Over the last decade programme and optimised diagnostic strategies
there has been a growing awareness that are required not only to detect and control CDI in
improvements to surveillance are needed. The Europe, but for a better understanding of the
increasing rate of CDI and emergence of epidemiology, has built the foundations for a
ribotype 027 precipitated the implementation of more robust, unified surveillance. The concerted
mandatory national surveillance of CDI in the efforts of the European Centre for Disease Pre-
UK. Changes in clinical presentation, severity of vention and Control (ECDC) CDI networks, has
disease, descriptions of new risk factors and the lead to the development of an over-arching long-
occurrence of outbreaks all emphasised the term CDI surveillance strategy for 2014–2020.
importance of early diagnosis and surveillance. Fulfilment of the ECDC priorities and targets
However a lack of consensus on case will no doubt be challenging and will require
definitions, clinical guidelines and optimal labo- significant investment however the hope is that
ratory diagnostics across Europe has lead to the both a national and Europe-wide picture of CDI
underestimation of CDI and impeded compari- will finally be realised.
son between countries. These inconsistencies
F. Fitzpatrick
Department of Clinical Microbiology, The Royal College
of Surgeons in Ireland, Dublin, Ireland
Department of Clinical Microbiology, Beaumont
Hospital, Dublin, Ireland
C. Wiuff (*) · A.-L. Banks
Strategic Lead Microbiology, NHS National Services L. Cottom
Scotland, Health Protection Scotland, HAI & IC Section, Department of Microbiology, Glasgow Royal Infirmary,
Glasgow, UK Glasgow, UK
e-mail:; e-mail:

# Springer International Publishing AG 2018 13

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
14 C. Wiuff et al.

Keywords that two separate lineages of 027, FQR1 and

Surveillance · Epidemiology · FQ2, had emerged in North America within a
Standardisation · Capacity building · short period of time, after acquiring the same
Collaborative effort fluoroquinolone resistance mutation, of which
one spread throughout the United States (US),
South Korea and Switzerland and the other
spread more widely across continents throughout
1 Epidemiology of CDI in Europe Europe and Australia (He et al. 2013). Isolates
obtained prior to the emergence of these two
Since 1978, Clostridium difficile has been lineages were not associated with any hospital
recognised as a leading infectious cause of outbreaks, suggesting that they represented
antimicrobial-associated diarrhoea with symptoms pre-epidemic lineages of ribotype 027. These
ranging from mild or moderate diarrhoea to findings highlighted the important role of selec-
pseudomembranous colitis (PMC). Until the end tive pressure from flouroquinolone use in the
of the millennium interest in this pathogen was evolution and spread of these two lineages in
primarily in relation to health care and impact on healthcare settings and highlighted the intercon-
morbidity and mortality in the elderly. However, nectedness of the global healthcare systems due
since 2000 there has been an explosion in reports on to human travel.
C. difficile infection (CDI) as a consequence of In the 2011/2012 European Centre for Disease
large increases in CDI cases (and incidence), sig- Prevention and Control (ECDC) acute hospital
nificant changes in the clinical presentation of CDI point prevalence survey of hospital-acquired
including more severe disease, occurrence of infection (HAI) C. difficile was the most fre-
outbreaks and descriptions of new risk factors quently reported pathogen associated with
(Freeman et al. 2010). The changes in the epidemi- healthcare associated gastrointestinal disease in
ology of CDI leading to several outbreaks in North European hospitals(accounting for 48% of all
America and Europe have mainly been attributed to gastrointestinal disease) (ECDC 2013a). Based
the emergence of a new hypervirulent strain PCR on this data it was estimated that 152,905 new
ribotype 027 (Kuijper et al. 2006), and to a lesser cases of CDI occur every year in Europe with an
extent, PCR ribotype 078 (Goorhuis et al. 2008). incidence of 30 cases per 100,000 population.
Ribotype 027 was associated with more severe CDI was associated with considerable short or
disease, higher mortality, increased risk of relapse long term disability, and 8382 deaths per year
and higher colectomy rates (Kuijper et al. 2006; (Cassini et al. 2016). In addition, CDI occurred
Ricciardi et al. 2007; Warny et al. 2005). However, increasingly in persons in the community without
other ribotypes of C. difficile also caused outbreaks typical risk factors such as antimicrobial treat-
and contributed to the spread of this infection in ment and recent hospitalisation (Bauer et al.
Europe and worldwide (Bauer et al. 2011). 2009; Wilcox et al. 2008) (Fig. 1).
Outbreaks caused by PCR ribotype 027 were
first reported in Europe in England and the
Netherlands followed by a series of reports 2 Development of European CDI
from several other European countries (Kuijper Guidance
et al. 2007); and, by 2008 a total of 16 countries
had reported this ribotype, including outbreaks in In response to the changing epidemiology of CDI
nine countries (Kuijper et al. 2008). However, in Europe and North America, the European
the reasons for its emergence and rapid global Study Group on C. difficile (ESGCD) published
spread remained unexplained until the genomes a standardised approach to detect, monitor and
of a global collection of C. difficile 027 isolates control CDI (Kuijper et al. 2006). In the follow-
from hospital patients between 1985 and 2010 ing year, evidence-based guidance on infection
was sequenced. Phylogenetic analysis showed prevention and control measures to limit the
Fig. 1 Global transmission events of C. difficile PCR ribotype 027 (with permission from authors). Arrows indicate individual introductory transmission events of FQR1 and
FQR2 (He et al. 2013)
16 C. Wiuff et al.

spread of CDI was also developed. In this context 3 Approaches to Diagnosing

CDI surveillance was identified as one of the ten and Monitoring Cases of CDI
most important measures in preventing and
controlling CDI. Surveillance allows continuous In the 1990s, a large number of diagnostic tests
monitoring and identification of increases in inci- for C. difficile became commercially available,
dence and severity of disease at an early stage in including faecal culture on selective media,
order to implement changes in practice and mon- detection of GDH (glutamate dehydrogenase) a
itor the impact of infection prevention and con- non-specific antigen, direct detection of toxin A
trol efforts (Vonberg et al. 2008). and B from stool using enzyme immune assay or
Timely feedback of surveillance data and its cytotoxicity assay (Delmee 2001) but system-
interpretations not only to the clinical and infec- wide or national surveillance programmes
tion prevention and control teams but also to remained rare. In 1993, a French multicentre
senior management, governing boards and point prevalence study identified C. difficile
administrators via the established communica- from 11.5% of 3921 diarrhoeal stool cultures
tion system is considered essential to preventing sent to 11 microbiology laboratories (Barbut
and control CDI in hospitals (Commission. et al. 1996). Stool assays for toxin A and B
2006). However, when CDI emerged as a serious became quickly the main clinical test for
threat to public health and patient safety at the diagnosing CDI while stool cultures were used
start of this millennium, comparison of the bur- mainly for epidemiological investigations (Kelly
den of CDI between healthcare facilities and and LaMont 1998). However, the majority of the
countries was problematic for a number of available testing methods were associated with
reasons. This included suboptimal case ascertain- either low sensitivity or specificity, or both (see
ment and inconsistent patient sampling, inade- also Chap. 4), and some required culture
quate laboratory diagnosis and use of facilities. Moreover, at that time there was no
non-standardised denominators for calculation consensus across Europe in terms of diagnostic
of incidence rates. Standardisation of laboratory testing and surveillance due to the lack of
testing methodology and adherence to agreed guidance.
surveillance definitions is needed for accurate In 2002, the ESGCD carried out a survey of
monitoring of trends in hospitals and other 212 laboratories in eight countries (B, DK, F,
healthcare settings (and for comparison between NL, G, I, SP, GB) to obtain an overview of
hospitals in their country and between countries). diagnostic methods used and to estimate the
The suite of guidance documents on CDI average incidence of CDI across Europe (Barbut
diagnostics, infection prevention and control et al. 2003). A high proportion (87.7%) of
and treatment developed by ESGCD and laboratories performed C. difficile diagnosis on
supported by the European Society of Clinical a routine basis, although laboratories in smaller
Microbiology and Infectious Diseases hospitals often relied on sending samples to a
(ESCMID), has provided the evidence platform bigger centre. Laboratory methods used in the
for the development of European surveillance of surveyed hospitals included direct toxin detec-
CDI now undertaken and coordinated by ECDC tion (93%), culture (55%), glutamate dehydroge-
(ECDC 2017) (Fig. 2). nase (GDH) (5.9%) but testing strategies
(i.e. application of standalone tests
vs. combination of tests) varied considerably
between laboratories and between countries.
Moreover, criteria for investigation for CDI
The Need for European Surveillance of CDI 17

Fig. 2 Targets in the ECDC long-term surveillance strategy for 2014–2020 (ECDC 2013b)
18 C. Wiuff et al.

varied extensively with 58% of laboratories only surveillance protocols at the time. Fourteen
testing for CDI if specifically requested by a countries were found to have a total of 18 surveil-
physician while 40.7% of laboratories routinely lance systems in place (of which some had more
tested for CDI when specific criteria (determined than one data collection system) (Kola et al.
by the microbiologist) were met; most commonly 2016). The majority of the European surveillance
loose or watery stools (40.3%), stools from systems were continuous and prospective; and
patients with a history of antibiotic therapy 11/18 used mandatory reporting while seven
(45.5%) and stools from nosocomial diarrhoea used voluntary reporting. Key features of the
(57.1%). Within most countries the proportion surveillance systems varied widely with consid-
of laboratories that used criteria for investigating erable variation in case definitions data collec-
CDI also varied between countries (ranging from tion methods, reporting and availability of
13% to 67%); only in the United Kingdom (UK), reference typing. In total, 12/18 countries used
testing was routinely done according to specific the ECDC/ Centers for Disease Control and Pre-
criteria determined by the microbiologist vention (CDC) case definitions, nine used the
(in 95% of laboratories). Ability to type ECDC definitions for community associated
C. difficile was infrequent with only 10.7% of and healthcare associated CDI while the remain-
laboratories reporting experience with typing. der had different cut-off for healthcare associa-
The inconsistent approach to diagnosing and typ- tion (including 48 h, 72 h, >3 or 4 days
ing CDI, including variation in the criteria for after admission). Thirteen systems had a defini-
testing, laboratory methodology and strategy for tion for severe disease while 11 had a definition
testing and possible bias in the study for recurrence but both definitions varied
(by inclusion of only the most responsive between countries. Despite the increasingly
laboratories) raised concern of under- recognised role of CDI in community settings
ascertainment due to un-diagnosed and only two countries (Austria and Scotland)
mis-diagnosed cases and inaccurate estimates of engaged General Practitioners in their surveil-
the overall burden of disease and highlighted the lance systems. Descriptive enhanced patient
need for international guidance. data were only collected in 6/18 systems and
The ESCGD review of the emergence of CDI death within 30 days in five. Reference typing
in North America and Europe (Kuijper et al. was performed routinely in 13/14 countries using
2006) specified for the first time a case definition various different criteria for submission includ-
for CDI, (including healthcare and community ing the presence of severe CDI, outbreaks or a
association), provided advice on optimal diag- more systematic periodic collection of a repre-
nostic testing and recommended that each mem- sentative sample of cases. Finally, the reporting
ber state should develop systematic and of the CDI burden varied widely with the use of a
comprehensive surveillance systems in order to non-standardised denominators and stratification
detect, monitor and respond to changes in the by geographical region, healthcare facility or
epidemiology of CDI, and in particular PCR laboratory making comparisons over time and
ribotype 027 at both national and European between regions and facilities difficult (Kola
levels. Following 2006 national surveillance et al. 2016).
systems were developed or expanded in countries Although, the overall capability and capacity
across Europe. for monitoring CDI has increased tremendously
In 2011, the European C.difficile Infection across Europe between 2003 and 2011, as of
Surveillance Network (ECDIS-net) surveyed today there is still scope for improvement in
the national surveillance systems through a diagnostic and surveillance setups in the majority
web-based questionnaire and reviewed extant of European countries.
The Need for European Surveillance of CDI 19

4 Diagnostic Capability: A capacity to diagnose, sub-type, report, collect

Pre-requisite for Surveillance patient risk factor data and monitor CDI across
Europe and as a consequence the true burden of
The attention given to diagnostic procedures and CDI and distribution of ribotypes is unclear.
surveillance of CDI varied widely between
countries. In 2008, with the support of ECDC a
Europe-wide survey (involving 106 laboratories 5 Benefits of Mandatory
in 34 countries) assessed the epidemic prepared- Surveillance: Experiences from
ness and current CDI epidemiology aiming to United Kingdom
ultimately build capacity for diagnosis and sur-
veillance of CDI in each country (Bauer et al. Prior to the year 2000, data on C. difficile was
2011). The frequency of testing varied between collected on a voluntary basis. In the UK, a steady
countries from 3 to 141 CDI tests conducted per increase in laboratory reports was observed during
10,000 patient days; and a correlation between the 1990s (Department of Health and Health Pro-
testing rate and CDI incidence was identified tection Agency 2008; Health Protection Scotland
(resulting in North European countries reporting 2006). In England, this was suggested to reflect a
the highest incidence rates). When a subset of failure to implement guidelines published in 1994,
C. difficile isolates were typed centrally as well as the result of increased testing and
(in Leiden, NL), a higher than expected diversity awareness of CDI, and an increase in
of PCR ribotypes was observed with ribotypes community-associated CDI (Department of
001, 014/020, and 078 ribotype being the most Health and Health Protection Agency 2008).
prevalent and 027 being only the 6th most com- The increasing CDI rates and emergence of
mon type (4.8% of examined isolates). ribotype 027 precipitated the implementation of
Optimum laboratory diagnosis of CDI depends mandatory national surveillance of CDI by
on testing patients at the correct time using appro- England, Wales and Northern Ireland in 2004,
priate testing methodology and strategy. A point and by Scotland in 2006. Initially, the surveil-
prevalence study in multi-centre setting in Spain lance programmes included only those aged
evaluated 988 unformed stools (from 897 patients) 65 years and above, but have since expanded to
found 66% of CDI episodes were undiagnosed or include all ages except the very young (Pearson
misdiagnosed due to lack of clinical suspicion 2009; Health Protection Scotland 2010).
(48%) or due to using a non-sensitive test (19%) Between 2003 and 2007, several large hospital
(Alcala et al. 2012). In the Europe-wide point prev- outbreaks of CDI occurred, involving ribotype
alence study (EUCLID) 3800 unformed stools 027 (two in England and one in Scotland),
(from >450 hospitals in 20 countries) were tested which brought CDI to the public attention
using the recommended two-step diagnostic algo- (Healthcare Commission 2006; The Vale of
rithm. In total, 25% of samples had not been tested Leven Hospital Inquiry 2014). Among the many
due to lack of clinical suspicion and 23% of patients key findings and recommendations contained
had been misdiagnosed due to using an inadequate within the critical reports that followed was an
laboratory test. It was estimated that on a single day acknowledgement of a lack of appropriate sur-
82 patients with diarrhoea due to C. difficile in veillance mechanisms, both locally and nation-
hospitals across Europe were not diagnosed due to ally, that could have identified an outbreak, and
lack of suspicion (Davies et al. 2014). In addition, the need for formal communication channels to
only 32% of participating hospitals used the opti- be in place to allow information on CDI numbers
mum diagnostic method at the first measurement and severity to be quickly disseminated. These
(in 2011–2012) whereas this had improved at the major incidents were quickly followed by the
second measurement (in 2012–2013) when 48% setting of national targets within the UK to
used the optimum method. reduce CDI rates by 30% (Duerden 2011; Scot-
These two recent studies highlighted again tish Government 2012).
variation in awareness and capability and
20 C. Wiuff et al.

Around the same time as the UK was Protection Scotland 2012); both downward
implementing national surveillance schemes, trends that have continued to date.
the ECDC and the U.S. CDC produced In order to respond to the public health need
recommendations for surveillance of CDI and to provide more detailed epidemiological
(Kuijper et al. 2006; McDonald et al. 2007). information on circulating strains of C. difficile
The publication of these documents enabled a a network of reference laboratories was
standardised surveillance case definition to be established in England (the Clostridium difficile
developed as well as definitions for severe CDI, Ribotyping Network, CDRN) with collaborative
recurrence, outbreaks and origin of infection that links to a single reference laboratory in Scotland.
could be used as necessary within a surveillance Investigations and isolate typing criteria
programme. Shortly thereafter, evidence-based focussed on severe cases of CDI, clusters of
recommendations for infection prevention and cases and unexplained increases in incidence in
control of CDI were published (Vonberg et al. both countries. In the first 3 years after
2008), with strong recommendations for the establishing these laboratory services, the preva-
implementation of routine surveillance of CDI, lence of ribotype 027 decreased markedly in
including the setting of thresholds to identify England (from 55% to 21%). This change in
outbreaks, emphasis on the importance of early distribution of ribotypes in England coincided
diagnosis, and awareness of changes in incidence with a 61% reduction in reports of CDI cases
or severity of disease. The foundations were laid (from 36,095 in 2008–2009 to 21,698 in
for the development of a range of tools and 2010–2011) and a decrease in reports of
strategies to deal with the CDI epidemic (Depart- complications, including mortality (Wilcox
ment of Health and Health Protection Agency et al. 2012). Likewise, the three major epidemic
2008; Health Protection Scotland 2009). Contin- ribotypes 027, 001 and 106 were gradually
uous and prospective surveillance at national replaced with other less prevalent ribotypes
level in healthcare and community settings was while rates of CDI were reducing in Scotland
mandated by governments in England and (Wiuff et al. 2011, 2014). It has been argued
Scotland; and real-time ‘local surveillance’ that the timely provision of ribotype information
(by ward, unit or facility) to monitor the number to infection prevention and control teams has
of cases, disease severity, surgery and mortality enabled the targeting of interventions and
rates with a duty for the multidisciplinary clinical resources on high incidence settings and in par-
and infection prevention control team to investi- ticular those with a high prevalence of
gate the root cause of any anomalies or 027 (Wilcox et al. 2012). However, there might
‘exceedances’ identified at local level in order also have been an additive effect of heightened
rectify deficiencies in patient care and/or infec- awareness and an improved understanding of the
tion control (Department of Health and Health need for clinical vigilance and aggressive inter-
Protection Agency 2008; Health Protection vention due to CDI caused by virulent strains
Scotland 2014). The heightened focus on local such as 027.
surveillance was a result of recommendations The overall decrease in CDI can be attributed
emerged from investigations of previous hospital to a multi-disciplinary approach including
outbreaks (The Vale of Leven Hospital Inquiry evidence-based guidance for the treatment and
2014; Healthcare Commission 2006). management of CDI patients, restrictive antimi-
CDI incidence rates in the UK peaked during crobial stewardship policies, and, arguably due to
2007/2008, and have since been steadily declin- the government targets for reducing CDI
ing (McDonald et al. 2007; Vonberg et al. 2008; (Duerden 2011; Nathwani et al. 2011; Lawes
Scotland 2014). In the 4-year period from 2007 et al. 2017). The establishment of mandatory
to 2010, significant reductions in the incidence of surveillance systems across the UK driven by
CDI were observed in England (54%) and government policy was instrumental to the devel-
Scotland (72%) (Duerden 2011; Health opment of standardised, evidence based
The Need for European Surveillance of CDI 21

diagnostic testing and expansion of national ref- agencies from the Netherlands, Germany and
erence laboratory services. The success of the the UK in collaboration with ECDC (that
UK surveillance programmes has undoubtedly evolved from the earlier ECDIS study Group)
been due to the rapid and joined up development was established in 2010 to support capacity
of diagnostic and surveillance capability and building for surveillance at the European level.
capacity with coverage of all healthcare settings. The formation of this multination surveillance
Standardised national surveillance collaboration was a result of heightened aware-
programmes are crucial to enable the monitoring ness in clinical communities across Europe fol-
of trends within and between countries, as well as lowing the culmination of a decade of published
facilitating the monitoring of interventions for literature detailing the emerging challenges of
improving care and outcomes of CDI patients. CDI, with papers highlighting the changing epi-
Central to all of this has been the adoption within demiology of CDI, the increased reporting of
the UK national surveillance programmes of outbreaks and the identification of hypervirulent
standardised protocols for sampling, testing, typ- ribotypes such as 027.
ing of isolates, reporting and feeding back data in One of the formative documents, which
management structures. This has resulted in included collaboration with CDC, published,
more solid reporting and accountability agreed CDI case definitions and issued
structures that lead to rapid responses to recommendations regarding surveillance
increases in CDI. (Kuijper et al. 2006). Thereafter, the ECDIS
Study Group survey of European hospitals
(Bauer et al. 2011) identified considerable varia-
6 The Need for European tion in methodology used in national surveillance
Surveillance of CDI programs; the significant variation in hospital
procedures and the availability of typing which
Suboptimal laboratory diagnostics, a previous also limited comparison between countries. Per-
lack of consensus on optimal testing methodol- haps the most startling finding was a CDI inci-
ogy for CDI and availability of typing across dence of 70% higher than previously reported
Europe has lead to under-diagnosis and impeded studies from 2005 (Barbut et al. 2007).
comparison between countries. Underestimation Importantly, the European Study Group
of CDI has also resulted from a deficiency in (ESGCD) coordinated these developments in
uniformity of case definitions, clinical algorithms collaboration with an increasing number of
and recognition amongst clinicians of when to national surveillance and laboratory coordinators
suspect CDI. These inconsistencies have from the participating countries and created a
prevented the true burden of disease from being professional network that met and communicated
appreciated. frequently allowing extensive discussions of a
Acceptance that a multi-country surveillance wide range of aspects of CDI paving the way
program is required not only to detect and control for achieving consensus on the current evidence
CDI in Europe, but for a better understanding of base on CDI. A need for standardisation was
the epidemiology, has paved the way forward. further supported by a review of existing national
Deserved attention afforded to this infection and CDI surveillance schemes which showed large
the concerted efforts to optimise diagnostic variations in the surveillance definitions used,
strategies have built the foundations for a more especially with regards to inclusion criteria for
robust, unified surveillance. The hope now is that cases and tying, choice of denominator, and ori-
both a national and Europe-wide picture of CDI gin of infection of CDI (Kola et al. 2016). As a
can be finally realised. result of these Europe-wide collaborative efforts
With funding from the ECDC, the ECDIS-Net the first ECDIS-Net protocol, a precursor for a
a consortium of universities and government European protocol, was developed. A draft
22 C. Wiuff et al.

protocol for CDI surveillance based on the above (a) Target 2 – Machine-to-machine reporting to
recommendations was piloted in 2013, with the The European Surveillance System
results being published in 2016 and highlighted (TESSy) in use by a majority of Member
that the foundations for Europe-wide surveil- States.
lance were in place. Furthermore, consideration (b) Target 3 – Data processing is semi-
was given to resource-poor countries that lacked automated while retaining a high quality,
facilities to perform internationally standardised enabling ECDC routine surveillance
PCR for ribotyping. outputs to be timelier, more easily available,
These foundations referred to by van Dorp user-customisable and thus perceived to be
et al., are the cumulation of the body of work more useful by stakeholders.
supported by ECDC over the past decade in (c) Target 10 – European event-based surveil-
developing an evidence based consensus lance detects, assesses and monitors com-
approach which could be applied across a myriad municable disease threats to public health in
of healthcare and laboratory settings to allow a near-real time.
high-level surveillance system to come to fru- (d) Target 12: European surveillance data are
ition. In terms of progress, while a 2002 surveil- used to monitor and evaluate prevention
lance study of diagnostic methods and protocols, programmes against agreed indicators.
found that only 55% of laboratories were capable
of culturing C.difficile (Barbut et al. 2003), in All of these targets require significant infra-
2013, 95% performed CDI diagnostics (van structure investment/realignment to achieve, and
Dorp et al. 2016a). run the risk of resource-poor member states
EDCD developed an over-arching long-term being left behind. The EU carries a diverse mix
surveillance strategy for 2014–2020 (ECDC of countries all with individual priorities and
2013b), in which the ECDC ‘European Surveil- levels of expenditure. Most EU countries fall
lance of Clostridium difficile infections – surveil- within the World Bank definition of a high-
lance protocol 2.3’ plays a disease specific role income country. Three Countries: Bulgaria,
(Control ECDC 2017). The strategy has six Croatia and Romania fall into the category of
priorities: upper-middle income, and while being wealthy
when compared to low income nations, their
1. Consolidating surveillance, increasing its effi- relative health expenditure is typically 20% of
ciency and enhancing the outputs and their that of high income EU nations (World Bank
impact 2017). Even amongst high income EU members
2. Developing standards, improving data quality GDP per capita can vary by as much as 400%.
and sharing best practices in surveillance Priorities at an EU and national level will have to
3. Promoting use of surveillance data be decided to minimise the effect of expected
4. Strengthening capacity in surveillance budget cuts and regional variation on expendi-
5. Controlling expansion ture. Capital funding from central European
6. Monitoring the strategy organisations or the ECDC will not be available
to fund member states public health infrastruc-
The overall strategic approach has been fur- ture to address any imbalance. As such, nations
ther subdivided into 17 individual targets to help with robust and successful surveillance networks
guide the development of surveillance will need to lead a concerted effort in sharing
programmes. technical expertise and advice, to allow other
Perhaps the most challenging of these for nations to develop their own sustainable and
regional and national programmes looking to integrated surveillance systems which can be
contribute to the Europe-wide level program are: integrated into an EU wide programme.
The Need for European Surveillance of CDI 23

7 Conclusions Barbut F, Corthier G, Charpak Y et al (1996) Prevalence

and pathogenicity of Clostridium difficile in
hospitalized patients. A French multicenter study.
Significant reductions in CDI have been reported Arch Intern Med 156(13):1449–1454
in countries across Europe, however, incidence Barbut F, Delmee M, Brazier JS et al (2003) A European
rates vary widely between countries and their survey of diagnostic methods and testing protocols for
Clostridium difficile. Clin Microbiol Infect 9(10):989–
capacity for surveillance, diagnostics and epide- 996
miological typing is highly variable (Kola et al. Barbut F, Mastrantonio P, Delmee M et al (2007) Pro-
2016; van Dorp et al. 2016b; Bauer et al. 2011). spective study of Clostridium difficile infections in
Therefore, there is a need to strengthen the Europe with phenotypic and genotypic
characterisation of the isolates. Clin Microbiol Infect
capacity for surveillance of CDI within Europe; 13(11):1048–1057.
this is both feasible and manageable (Control 0691.2007.01824.x
2017; van Dorp et al. 2016a). Bauer MP, Veenendaal D, Verhoef L et al (2009) Clinical
A standardised approach to surveillance, and microbiological characteristics of community-
onset Clostridium difficile infection in The
diagnostics and typing would allow the estima- Netherlands. Clin Microbiol Infect 15(12):1087–
tion of the total burden of CDI in Europe (and its 1092.
member states) and the continuous monitoring of 02853.x
incidence, severity, outcome and risk factors for Bauer MP, Notermans DW, Van Benthem BH et al (2011)
Clostridium difficile infection in Europe: a hospital-
developing CDI would lead to improved man- based survey. Lancet 377(9759):63–73. https://doi.
agement and control of CDI. org/10.1016/s0140-6736(10)61266-4
In countries where large reductions in CDI Cassini A, Plachouras D, Eckmanns T et al (2016) Burden
incidence have been achieved comprehensive of six healthcare-associated infections on European
population health: estimating incidence-based disabil-
national surveillance programmes have been a ity-adjusted life years through a population
key driver in the standardisation of diagnostic prevalence-based modelling study. PLoS Med 13
approach, sampling and reporting practices and (10):e1002150.
in developing coordinated approaches and 1002150
Davies KA, Longshaw CM, Davis GL et al (2014) Under-
resources to infection prevention and manage- diagnosis of Clostridium difficile across Europe: the
ment of CDI by highlighting the evolving epi- European, multicentre, prospective, biannual, point-
demic of CDI. prevalence study of Clostridium difficile infection in
Additional benefits of laboratory based surveil- hospitalised patients with diarrhoea (EUCLID). Lan-
cet Infect Dis 14(12):1208–1219.
lance could be achieved from the introduction of 1016/S1473-3099(14)70991-0
genetic finger printing using multi-locus variable- Delmee M (2001) Laboratory diagnosis of Clostridium
number tandem repeat (MLVA typing) and whole difficile disease. Clin Microbiol Infect 7(8):411–416
genome sequencing to investigate clusters and Duerden BI (2011) Contribution of a government target to
controlling Clostridium difficile in the NHS in
cross-transmission routes (Fawley et al. 2011; England. Anaerobe 17(4):175–179.
Eyre et al. 2013; Eyre and Walker 2013) and iden- 10.1016/j.anaerobe.2010.12.004
tify regional and inter-continental spread of new Department of Health and Health Protection Agency
potentially epidemic and virulent lineages of (2008) Clostridium difficile infection: how to deal
with the problem
C. difficile (He et al. 2013). ECDC (2013a) European Centre for Disease Prevention
and Control. Point prevalence survey of healthcare-
associated infections and antimicrobial use in
European acute care hospitals. Stockholm
References ECDC (2013b) European Centre for Disease Prevention
and Control. Long-term surveillance strategy
Alcala L, Martin A, Marin M et al (2012) The undiag- 2014–2020. Stockholm
nosed cases of Clostridium difficile infection in a ECDC (2017) European centre for disease prevention and
whole nation: where is the problem? Clin Microbiol control. European surveillance of Clostridium difficile
Infect 18(7):E204–E213. infections. Surveillance protocol version 2.3.
1469-0691.2012.03883.x Stockholm
24 C. Wiuff et al.

Eyre DW, Walker AS (2013) Clostridium difficile Kola A, Wiuff C, Akerlund T et al (2016) Survey of
surveillance: harnessing new technologies to control Clostridium difficile infection surveillance systems in
transmission. Expert Rev Anti-Infect Ther 11 Europe, 2011. Euro Surveill 21(29).
(11):1193–1205. 2807/
2013.845987 Kuijper EJ, Coignard B, Tull P (2006) Emergence of
Eyre DW, Fawley WN, Best EL et al (2013) Comparison Clostridium difficile-associated disease in North
of multilocus variable-number tandem-repeat analysis America and Europe. Clin Microbiol Infect 12(6):2–
and whole-genome sequencing for investigation of 18.
Clostridium difficile transmission. J Clin Microbiol Kuijper EJ, Coignard B, Brazier JS et al (2007) Update of
51(12):4141–4149. Clostridium difficile-associated disease due to PCR
01095-13 ribotype 027 in Europe. Euro Surveill 12(6):E1–E2
Fawley WN, Wilcox MH, Clostridium difficile Ribotyping Kuijper EJ, Barbut F, Brazier JS et al (2008) Update of
Network for E et al (2011) An enhanced DNA finger- Clostridium difficile infection due to PCR ribotype
printing service to investigate potential Clostridium 027 in Europe, 2008. Euro Surveill 13(31):pii: 18942
difficile infection case clusters sharing the same PCR Lawes T, Lopez-Lozano JM, Nebot CA et al (2017) Effect
ribotype. J Clin Microbiol 49(12):4333–4337. https:// of a national 4C antibiotic stewardship intervention on the clinical and molecular epidemiology of Clostrid-
Freeman J, Bauer MP, Baines SD et al (2010) The chang- ium difficile infections in a region of Scotland: anon-
ing epidemiology of Clostridium difficile infections. linear time-series analysis. Lancet Infect Dis 17
Clin Microbiol Rev 23(3):529–549. (2):194–206.
1128/CMR.00082-09 30397-8
Goorhuis A, Bakker D, Corver J et al (2008) Emergence McDonald LC, Coignard B, Dubberke E et al (2007)
of Clostridium difficile infection due to a new hyper- Recommendations for surveillance of Clostridium dif-
virulent strain, polymerase chain reaction ribotype ficile-associated disease. Infect Control Hosp
078. Clin Infect Dis 47(9):1162–1170. https://doi. Epidemiol 28(2):140–145.
org/10.1086/592257 511798
He M, Miyajima F, Roberts P et al (2013) Emergence and Nathwani D, Sneddon J, Malcolm W et al (2011) Scottish
global spread of epidemic healthcare-associated Clos- Antimicrobial Prescribing Group (SAPG): develop-
tridium difficile. Nat Genet 45(1):109–113. https://doi. ment and impact of the Scottish National Antimicro-
org/10.1038/ng.2478 bial Stewardship Programme. Int J Antimicrob Agents
Healthcare Commission (2006) Investigation into 38(1):16–26.
outbreaks of Clostridium difficile at Stoke Mandeville 2011.02.005
Hospital. Trust, Buckinghamshire Hospitals NHS Pearson A (2009) Historical and changing epidemiology
Health Protection Scotland (2006) CDI background. of healthcare-associated infections. J Hosp Infect 73 (4):296–304.
resourceid¼1446 016
Health Protection Scotland (2009) Guidance on Preven- Ricciardi R, Rothenberger DA, Madoff RD et al (2007)
tion and Control of Clostridium difficile Infection Increasing prevalence and severity of Clostridium dif-
(CDI) in Healthcare Settings in Scotland. Health Pro- ficile colitis in hospitalized patients in the United
tection Network Scottish Guidance 6. Health Protec- States. Arch Surg 142(7):624–631.
tion Scotland. http://www.processcleaningsolutions. 1001/archsurg.142.7.624. discussion 631
com/pdf/clostridium-difficile-infection-guidelines% Scottish Government (2012) NHSScotland local delivery
20Scotland.pdf plan guidance 2013/14
Health Protection Scotland (2010) The annual surveil- The Vale of Leven Hospital Inquiry (2014) The Vale of
lance of healthcare associated infection report Leven Hospital inquiry report. The Rt Hon Lord
January – December 2009. http://www.documents. MacLean van Dorp SM, Kinross P, Gastmeier P et al (2016a)
haireport-2009.pdf Standardised surveillance of Clostridium difficile
Health Protection Scotland (2012) The annual surveil- infection in European acute care hospitals: a pilot
lance of healthcare associated infections report. study, 2013. Euro Surveill 21(29).
January–December 2011. http://www.documents. 2807/ van Dorp SM, Notermans DW, Alblas J et al (2016b)
haireport-2011.pdf Survey of diagnostic and typing capacity for Clostrid-
Health Protection Scotland (2014) Prevention and Control ium difficile infection in Europe, 2011 and 2014. Euro
of Clostridium difficile (CDI) in care settings in Surveill 21(29).
Scotland. 2.1 2016.21.29.30292
Kelly CP, LaMont JT (1998) Clostridium difficile infec- Vonberg RP, Kuijper EJ, Wilcox MH et al (2008) Infec-
tion. Annu Rev Med 49:375–390. tion control measures to limit the spread of
The Need for European Surveillance of CDI 25

Clostridium difficile. Clin Microbiol Infect 14(Suppl surveillance scheme in England. Clin Infect Dis 55
5):2–20. (8):1056–1063.
01992.x Wiuff C, Brown DJ, Mather H et al (2011) The epidemi-
Warny M, Pepin J, Fang A et al (2005) Toxin production ology of Clostridium difficile in Scotland. J Inf Secur
by an emerging strain of Clostridium difficile 62(4):271–279.
associated with outbreaks of severe disease in North 015
America and Europe. Lancet 366(9491):1079–1084. Wiuff C, Banks A, Brown DJ, Henderson D,Coia JE, Inkster T (2014) The changing epidemiology of Clos-
Wilcox MH, Mooney L, Bendall R et al (2008) A case- tridium difficile ribotypes in Scotland between 2009
control study of community-associated Clostridium and 2013. Paper presented at the European Congress
difficile infection. J Antimicrob Chemother 62 on Clinical Microbiology and Infectious Diseases,
(2):388–396. 2014, Barcelona
Wilcox MH, Shetty N, Fawley WN et al (2012) Changing World Bank (2017) GNI (nominal) per capita for year
epidemiology of Clostridium difficile infection follow- 2016, World Development Indicators database, Atlas
ing the introduction of a national ribotyping-based method, World Bank, revised 01 July 2017
Diagnostic Guidance for C. difficile

Monique J. T. Crobach, Amoe Baktash, Nikolas Duszenko,

and Ed J. Kuijper

Abstract multi-step algorithmic testing has now been

Diagnosis of Clostridium difficile infection advocated by international guidelines in order
(CDI) can be challenging. First of all, there to optimize diagnostic accuracy. Despite these
has been debate on which of the two reference recommendations, testing methods between
assays, cell cytotoxicity neutralization assay hospitals vary widely, which impacts CDI
(CCNA) or toxigenic culture (TC) should be incidence rates. CDI incidence rates are also
considered the gold standard for CDI detec- influenced by sample selection criteria, as
tion. Although the CCNA suffers most from several studies have shown that if not all
suboptimal storage conditions and subsequent unformed stool samples are tested for CDI,
toxin degradation, TC is reported to falsely many cases may be missed due to an absence
increase CDI detection rates as it cannot dif- of clinical suspicion. Since methods for
ferentiate CDI patients from patients asymp- diagnosing CDI remain imperfect, there has
tomatically colonised by toxigenic C. difficile. been a growing interest in alternative testing
Several rapid assays are available for CDI strategies like faecal biomarkers, immune
detection and fall into three broad categories: modulating interleukins, cytokines and imag-
(1) enzyme immunoassays for glutamate ing methods. At the moment, these alternative
dehydrogenase, (2) enzyme immunoassays methods might play an adjunctive role, but
for toxins A/B and (3) nucleic acid amplifica- they are not suitable to replace conventional
tion tests detecting toxin genes. All three CDI testing strategies.
categories have their own limitations, being
suboptimal specificity and/or sensitivity or the Keywords
inability to discern colonised patients from Clostridium difficile · Diagnostics · Testing ·
CDI patients. In light of these limitations, Algorithmic testing

E. J. Kuijper
M. J. T. Crobach (*) · A. Baktash · N. Duszenko ESCMID Study Group for C.difficile (ESGCD), Basel,
Department of Medical Microbiology, Centre for Switzerland
Infectious Diseases, Leiden University Medical Centre, Department of Medical Microbiology, Center of
Leiden, The Netherlands Infectious Diseases, Leiden University Medical Center,
e-mail:; A.Baktash.MM@lumc. Leiden, The Netherlands
nl; e-mail:

# Springer International Publishing AG 2018 27

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
28 M. J. T. Crobach et al.

1 Introduction cells or Hep-2 cells. At 24- and 48-h intervals,

these cultures are evaluated for the characteristic
Diagnosis of Clostridium difficile infection (CDI) rounding effect engendered by toxin B. Reversal
is challenging, as there is no optimal laboratory of this effect by toxin B antitoxin demonstrates
assay and even no universal reference test. Due the specific role of toxin B in inducing the cyto-
to imperfect assays, combinations of assays to pathic effects observed, and thus its presence
optimize their performance have been proposed. (Delmee 2001; Burnham and Carroll 2013). The
However, diverse testing strategies are applied reference test for detection of toxigenic
across laboratories. These diverse testing C. difficile is toxigenic culture (TC) (Planche
strategies may impact CDI incidence rates. In and Wilcox 2011; Burnham and Carroll 2013).
addition to the conventional testing methods, For TC stool samples are inoculated onto selec-
alternative methods are sometimes applied either tive media and incubated for at least 48 h (Hink
to diagnose CDI or as an aid to predict severity. et al. 2013). Colonies suspected of being
Here, we will describe the diverse testing C. difficile, by e.g. Gram staining, colony mor-
strategies with their advantages and limitations phology, odour or more sophisticated techniques,
and clinical relevance. are isolated. Their toxigenic potential is assessed
by testing for in vitro toxin production via the
aforementioned CCNA, by enzyme
2 Reference Tests immunoassays (EIA) for toxins A/B, or by test-
ing for toxin-producing genes via nucleic acid
The diagnosis of CDI relies on one of two amplification tests (NAAT) (Burnham and
approaches: demonstrate the presence of toxins Carroll 2013).
responsible for the clinical manifestations of During the last years, there has been debate on
CDI, or demonstrate the presence of C. difficile which of these two reference tests represents true
which is capable of producing toxins (so-called disease, as the CCNA detects in vivo toxins
toxigenic C. difficile) (Planche and Wilcox 2011) while TC detects in vitro toxin production
(Table 1). The reference test for detection of (Planche and Wilcox 2011). There is a growing
toxins in stools is the cell cytotoxicity body of evidence demonstrating that toxigenic
neutralisation assay (CCNA) (Planche and strains are often carried asymptomatically
Wilcox 2011; Burnham and Carroll 2013). For (Kyne et al. 2000; Loo et al. 2011). TC is not
CCNA, stool sample filtrate is inoculated onto an able to make a distinction between asymptomatic
in vitro cell monolayer, using cell lines such as carriage of toxigenic C. difficile strains and true
Vero cells, HeLa cells, human foreskin fibroblast infection. Studies have shown that patients with

Table 1 Available assays for CDI detection

Type of assay Target of detection Detected condition
Culture C. difficile C. difficile colonisation,
can be CDI
Glutamate dehydrogenase enzyme Glutamate dehydrogenase C. difficile colonisation, can
immunoassay (GDH EIA) be CDI
Toxins A/B enzyme immunoassay Toxins A and B CDI
(Tox A/B EIA)
Nucleic acid amplification test TcdB and/or TcdA genes, sometimes cdt and Toxigenic C. difficile
(NAAT) deletion in tcdC colonisation, can be CDI
Cell cytotoxicity neutralization assay Toxin B CDI
Toxigenic culture (TC) C. difficile and thereafter in vitro toxin Toxigenic C. difficile
production by Tox A/B EIA, NAAT or CCNA colonisation, can be CDI
CDI Clostridium difficile infection
Diagnostic Guidance for C. difficile Infections 29

positive CCNA or Tox A/B EIA have a worse sensitive (pooled sensitivity compared to
prognosis than patients who test only positive in CCNA and TC 94% and 96%, respectively)
TC, indicating that this latter category may actu- (Crobach et al. 2016). However, they cannot
ally be colonised patients instead of patients with make a distinction between the presence of toxi-
true CDI (Planche et al. 2013; Polage et al. genic or non-toxigenic strains and are thus less
2015). Although CCNA may therefore better specific to detect true disease. This was
reflect true CDI, it is this reference test that demonstrated by a specificity of only 90% in
suffers most from lack of standardization and comparison to CCNA (Crobach et al. 2016).
suboptimal storage or collection conditions, NAATs include PCR assays, helicase-
thereby possibly generating false-negative dependent amplification assays and loop-
results. Both reference tests are laborious and mediated isothermal amplification assays. Most
expensive and require trained personnel. There- of these assays target conserved regions within
fore, easy-to perform rapid assays have been the gene for toxin B (tcdB), although some target
developed. These include enzyme immunoassays a highly conserved sequence of the toxin A gene
for GDH, enzyme immunoassays for Toxins A/B (tcdA). Assays that detect the ribtoype 027/NAP1
and during the last decade, NAATs for toxin strain (and related ribotypes) are also available,
genes have become available. Given their ease these detect the genes for binary toxin (cdt) and
of use and rapid turnaround time, these rapid the deletion at nucleotide 117 on the regulatory
tests have become the mainstays of CDI diagno- gene tcdC.
sis in a clinical setting. NAATs are sensitive (sensitivity compared to
CCNA and TC 96% and 95%, respectively)
(Crobach et al. 2016). As they only detect toxi-
genic strains instead of all C. difficile, they are
3 Rapid Assays
more specific than GDH EIA (specificity com-
pared to CCNA and TC 94% and 98%, respec-
Reference methods are accurate, but the lengthy,
tively) (Crobach et al. 2016). However, NAATs
laborious nature of such testing precludes its
only detect the presence of toxin genes and hence
application in a clinical setting. Rapid tests are
the toxin producing capacity of C. difficile.
ideally suited for clinical use, but each suffers
Therefore, a major drawback of NAAT is that
from its own shortcomings. Tox A/B EIAs
in addition to CDI cases, it will also detect
directly detect free toxins in stools and are there-
asymptomatic carriers of toxigenic C. difficile.
fore believed to correlate to clinical symptoms
(Polage et al. 2015). They are cheap and easy to
use. However, sensitivity of Tox A/B EIAs is
suboptimal. Compared to CCNA, pooled sensi- 4 Recommended Testing
tivity of Tox A/B EIA was 83%. In comparison Algorithms
to toxigenic culture, pooled sensitivity of Tox
A/B EIA was as low as 57%. Pooled specificity Although it would be the easiest to use one of the
of Tox A/B EIAs was however reported to be as rapid assays for CDI detection in daily practice,
high as 99%, both compared to CCNA and TC this will falsely impact CDI detection rates. First
(Crobach et al. 2016). of all, GDH EIA and NAAT results do not
GDH EIAs are also easy to perform and directly correlate with clinical symptoms possi-
cheap. They detect glutamate dehydrogenase, bly leading to over diagnosis of CDI. Second, all
an enzyme that is produced by both toxigenic of these three tests, even the very specific Tox
and non-toxigenic strains. GDH EIAs are A/B EIAs, are not specific enough to be used as a
30 M. J. T. Crobach et al.

stand-alone test (Crobach et al. 2016). Namely, the first step, optionally followed by TC or
most of the samples submitted for CDI testing NAAT in case of ambiguous results, is men-
will not have the disease. Assuming a CDI prev- tioned as a suitable equivalent (Crobach et al.
alence rate of 5% among submitted samples, 2016) (Fig. 1).
positive predictive values of the most specific The gains in diagnostic accuracy achieved by
assays (Tox A/B EIA) range from 69% to 81%, such algorithmic testing are substantial. It was
indicating that 19–31% of samples with a posi- calculated that in a typical endemic setting of 5%
tive test result actually do not have the disease CDI prevalence among submitted samples, PPV
(Crobach et al. 2016). and NPV of the most accurate algorithm, NAAT
In light of these limitations of the rapid followed by Tox A/B EIA, are 98.5% and 98.9%,
assays, common guidelines for CDI diagnosis respectively. In comparison, PPV and NPV of
put forth by the European Society of Clinical standalone NAAT are 45.7% and 99.8%, respec-
Microbiology and Infectious Diseases tively; PPV and NPV of standalone Tox A/B EIA
(ESCMID) and the Society for Healthcare Epi- are 81.4% and 99.1%, respectively (Crobach
demiology of America/the Infectious Diseases et al. 2016).
Society of America (SHEA/IDSA) recommend Algorithmic testing does have its own draw-
the use of multi-step algorithmic testing to maxi- back: increased turnaround time. While patients
mize diagnostic accuracy (Crobach et al. 2016; with a negative result can quickly be ruled out for
Cohen et al. 2010). The premise of this strategy is CDI, actually establishing a CDI diagnosis
sequential testing that most efficiently uses requires two positive tests, inevitably requiring
molecular tests’ different strengths. First, stool more time, especially if CCNA is used as the
samples are screened by a sensitive test. second test as recommended by IDSA/SHEA
According to the ESCMID guidelines this could guidelines. This is a non-trivial drawback, as it
either be either be GDH EIA or NAAT, while has been shown that decreasing the time to diag-
SHEA/IDSA guidelines recommend the use of nosis positively affects patient outcomes (Barbut
GDH EIA as a first step (Cohen et al. 2010; et al. 2014). Numerous studies have found an
Crobach et al. 2016). The high sensitivity of association between low CT values and toxin
these tests provides them a high negative predic- presence or outcome (Chung and Lee 2017;
tive value (NPV) with which to be reasonably Jazmati et al. 2016; Reigadas et al. 2016; Dionne
confident that a negative test is in fact indicative et al. 2013; Leslie et al. 2012; Kaltsas et al.
of no CDI. In this manner, a large proportion of 2012). Efforts have been made to address the
diarrheal cases can be quickly ruled out for CDI. longer turnaround time of algorithms by examin-
If the first test is positive, reflex testing occurs by ing whether quantitation of NAAT results by
Tox A/B EIA (Crobach et al. 2016) or CCNA cycle threshold (CT), the point during a PCR
(Cohen et al. 2010), a test of high specificity with when product begins being fluorescently detect-
a correspondingly high positive predictive value able that serves as an indirect measure of the
(PPV) as it is now used in selected samples with a starting number of DNA copies in a sample, can
higher pre-test probability of CDI. Thus, a posi- be used by itself to establish a CDI diagnosis
tive result on this second test is likely indicative (Senchyna et al. 2017; Crobach et al. submitted).
of CDI. In the event of a positive first test and a Although studies indicate that NAAT CT values
negative second, the result is considered an can be used to predict the toxin status, the rela-
ambiguous one in need of resolution by clinical tionship between the two is not strong enough to
evaluation or further testing, e.g. via TC. In the negate the need for toxin testing by a second test
ESCMID guidelines, an alternative algorithm at his moment (Senchyna et al. 2017; Crobach
starting with both GDH and Tox A/B EIA in et al. submitted). For now, the increased
Diagnostic Guidance for C. difficile Infections 31

Step 1:
Highly sensitive test: NAAT or GDH EIA

Positive test result Negative test result

Step 2: No further testing required:

Highly specific test: CDI is unlikely to be
Toxin A/B EIA present

Positive test result Negative test result

Clinical Step 3 (optional):

CDI is likely to
evaluation: CDI Perform TC or NAAT (in
be present
or carriage of case first test was a GDH
(toxigenic) C. EIA)
difficile is

Step 1:
Highly sensitive test: GDH and Tox A/B

Both negative GDH positive, Tox Both positive

A/B negative

No further testing No further testing

Step 2 (optional):
required: CDI is required: CDI is likely to
unlikely to be present be present

Negative test result Positive test result

CDI is unlikely evaluation: CDI
to be present or carriage of
(toxigenic) C.
difficile is

Fig. 1 Algorithms for CDI testing as recommended by immunoassay (Figure reprinted from Crobach et al. CMI
ESCMID guidelines. (a) GDH or NAAT- Tox A/B algo- 2016;22:S63,,
rithm, (b) GDH and Tox A/B – NAAT/TC algorithm. available under a Creative Commons Attribution-
CDI Clostridium difficile infection, GDH glutamate dehy- NonCommercial-NoDerivates License (CC BY NC
drogenase, NAAT nucleic acid amplification test, TC toxi- ND),
genic culture, Tox A/B, toxin A/B; EIA enzyme legalcode)
32 M. J. T. Crobach et al.

turnaround time of algorithms must be accepted, variable-number of tandem-repeat analysis

as algorithms seem to represent the most accu- (MLVA), multilocus sequence typing (MLST)
rate, clinically implementable testing strategy for and whole-genome sequencing (WGS) are
CDI diagnosis. mainly of use for outbreak investigations
There are still controversies concerning the (Knetsch et al. 2013; van den Berg et al. 2007;
use of NAATs in CDI diagnosis. ESCMID Maiden et al. 1998). Furthermore, TC may be
guidelines recommend against their use as needed to resolve discrepant results of algorith-
stand-alone tests based on the limitations that mic testing where C. difficile is detected by GDH
we described above, but they do recommend EIA or NAAT but toxin is not. A positive TC
their use as a first step in an algorithm (Crobach result rules out a false positive GDH EIA/NAAT
et al. 2016). The older SHEA/IDSA guidelines result in these patients. In that case, clinical eval-
indicate that NAAT testing may ultimately uation is needed; these patients can either be CDI
address testing concerns although more data are patients with a false negative Tox A/B EIA result
needed before this methodology can be due to low toxin levels or degradation of toxins,
recommended for routine testing (Cohen et al. or C. difficile carriers.
2010). However, others prefer PCR testing over
Tox A/B EIA testing and over GDH-Tox A/B
algorithms, because of superior sensitivity 5 Selection of Stool Samples
(Surawicz et al. 2013). They do indicate that
NAAT should only be applied in patients with Testing for CDI should only be performed on
diarrhoea, to overcome the problem of false pos- unformed stools as the presence of clinical
itive results (Surawicz et al. 2013). symptoms is a prerequisite to diagnose CDI
Although TC is not an efficient method for (Crobach et al. 2016; Cohen et al. 2010;
screening large numbers of diarrheal samples for Surawicz et al. 2013). However, it can be diffi-
potential CDI, it nevertheless remains an impor- cult to assess which unformed stools should be
tant technique for laboratories to be able to carry tested. A large study in 482 hospitals across 20 -
out. Isolating C. difficile by TC serves several European countries showed that 23% of samples
post-diagnostic purposes. These include antimi- positive for CDI were not diagnosed by the local
crobial susceptibility testing and molecular typ- hospital because of an absence of clinical suspi-
ing of isolates. For molecular typing, pulsed-field cion (Davies et al. 2014). It was reported that
gel electrophoresis (PFGE) was considered the mostly younger patients and patients who are
standard method in North America, with the not hospitalized or have been hospitalized for
resulting banding patterns described as “North <3 days are inadvertently not tested for CDI
American pulse-field” (NAP) types (Killgore (Davies et al. 2014; Alcala et al. 2012). In gen-
et al. 2008; Kristjansson et al. 1994). In Europe eral practice, CDI is also often missed due to lack
PCR ribotyping is most commonly applied, with of suspicion, as was shown in a study among
the resulting patterns described as PCR ribotypes 12,714 unformed stool samples (Hensgens et al.
(Stubbs et al. 1999; Bidet et al. 1999). Recently, 2014). In this study, general practitioners
reference laboratories in Canada and the US have requested CDI testing in 7% of unformed stool
also applied PCR ribotyping, using a samples, thereby detecting only 40% of all CDI
standardized protocol for capillary- cases (Hensgens et al. 2014). In light of these
electrophoresis PCR ribotyping (Fawley et al. problems, testing of all submitted unformed stool
2015). While PFGE and PCR ribotyping are the samples is now endorsed by the ESCMID
methods of choice for surveillance purposes, guidelines (Crobach et al. 2016). This approach
additional typing methods like multilocus has been shown to increase the diagnostic yield
Diagnostic Guidance for C. difficile Infections 33

(Davies et al. 2014; Reigadas et al. 2015). positive in a repeat sample submitted within
Restricting CDI testing to liquid samples instead 7 days (Aichinger et al. 2008; van Prehn et al.
of all unformed samples seems to be too stringent 2015). These samples constitute around 9% of all
and may cause the diagnosis of CDI to be missed positive samples (Aichinger et al. 2008; van
(Berrington and Settle 2007). Prehn et al. 2015). Although the former studies
A special situation exists for patients with were performed in endemic situations, a study
ileus due to CDI. In this case, formed stools or performed during an outbreak situation
rectal swabs can be tested for CDI (McFarland demonstrated that there was a definite diagnostic
et al. 1987; Rogers et al. 2013). Although yield of retesting in such a situation; of all
perirectal swabs have also been proposed as suit- samples submitted for repeat Tox A/B EIA test-
able alternatives, their use may depend on the ing, 8.2% tested positive. These samples
presence of faecal staining on the swab (Rogers constituted 5% of all positive CDI samples
et al. 2013; Kundrapu et al. 2012). (Debast et al. 2008).
In young children, high C. difficile The utility of repeat NAAT testing has been
colonisation rates have been described (Enoch evaluated in several studies, too. The percentages
et al. 2011). Young children can also test positive of samples that were positive within 7 days after
for toxins, without clinical significant disease. a negative test range from 0.9% to 2.9% (Green
On the other hand, the incidence of CDI among et al. 2014; Luo and Banaei 2010; Khanna et al.
hospitalized children has been increasing 2012; Aichinger et al. 2008; van Prehn et al.
(Schutze and Willoughby 2013). CDI testing is 2015). The number of CDI cases detected by a
therefore burdensome in young children, and repeat test range from 1.7% to 4.5% (Aichinger
should always include clinical evaluation. Rou- et al. 2008; van Prehn et al. 2015). The chance of
tine testing for CDI in children <1 year should be turning positive was lower in the first 7 days after
avoided, according to guidelines launched by the a negative NAAT result than in the 7–14 days
American Academy of Pediatrics (Schutze and period after the negative test result (Luo and
Willoughby 2013). For children between 1 and Banaei 2010; Khanna et al. 2012). In one study,
3 years of age with diarrhoea, CDI testing can be a history of CDI seemed to increase the risk of a
considered, but testing for other causes, particu- positive repeat NAAT result within 7 days after a
larly viral infections, is recommended first first negative test (Green et al. 2014).
(Schutze and Willoughby 2013). For children The general consensus is that in a
above 3 years of age, normal testing procedures non-epidemic situation, the diagnostic yield of
can be followed (Schutze and Willoughby 2013; repeat testing by both Tox A/B EIA and NAAT
Crobach et al. 2016). is too low, and therefore, repeat testing should be
discouraged (Cohen et al. 2010; Crobach et al.
2016; Surawicz et al. 2013). If an algorithm is
6 Repeat Testing used instead of stand-alone NAAT or Tox A/B
EIA, the even higher predictive values make
Before the introduction of algorithms, lack of repeat testing redundant. However, in epidemic
confidence in Tox A/B EIAs led to the submis- situations, or in patients with very high clinical
sion of multiple stool samples during one suspicion, repeat testing may be of value
diarrheal episode. Several studies sought to (Crobach et al. 2016).
determine the yield of such repeat testing. Diag- Sometimes, repeat samples are taken after
nostic yield can either be expressed in the per- CDI treatment as a test of cure. However, after
centage of first test negative samples converting resolution of diarrhoea, patients can still test
to positive in a repeat test, or the percentage of positive for toxins (Wenisch et al. 1996). Fur-
positive samples that is detected by repeat test- thermore, patients can become asymptomatic
ing. After a first negative Tox A/B EIA result, it carriers after treatment for CDI: one small study
was reported that 0.9–2.5% of samples test showed that 1–4 weeks after treatment, 29/56
34 M. J. T. Crobach et al.

(56%) of patients were found to be asymptomatic These observations hold true when the same
carriers of C. difficile (Sethi et al. 2010). Testing samples are concomitantly tested with both
for cure is therefore not recommended in current stand-alone NAAT and an algorithm. In one
guidelines (Crobach et al. 2016; Cohen et al. study of 1321 stool samples, the CDI positivity
2010; Schutze and Willoughby 2013). rate by NAAT was 6.4%, while the CDI positiv-
ity rate by a GDH and Tox A/B EIA – CCNA
algorithm on the same samples was 4.2%. The
7 Consequences of Testing overall incidence rates were 8.9 and 5.8 per
Strategy on CDI Incidence/ 10,000 patient-days for stand-alone NAAT and
Reporting Rates the algorithm, respectively (Longtin et al. 2013).
When stand-alone NAAT was compared to
Despite the common recommendations of stand-alone Tox A/B EIA, higher CDI positivity
ESCMID and SHEA/IDSA advocating the use of rates and higher CDI incidence rates for NAAT
algorithmic testing in CDI diagnosis, testing compared to Tox A/B EIA were reported, too
methods between hospitals vary widely. A large (Grein et al. 2014). Even so, hospitals that switch
study across 60 European hospitals found that from non-molecular tests to stand-alone NAAT
only 64% of hospitals use a recommended testing testing are reported to experience an increase in
algorithm for CDI testing (Davies et al. 2016). Data their CDI incidence rates (Moehring et al. 2013).
from the US show, that in 2012, 51% of hospitals The implications of testing method-dependent
was still relying on stand-alone Tox A/B EIA CDI incidence rates are consequential. Besides
(CDC 2012). Standalone use of Tox A/B EIAs the obvious effect of interfering with attempts to
decreased in response to recognition that low accurately monitor CDI for surveillance
sensitivities were leading to CDI under-diagnosis, purposes, financially tangible effects also result.
and consequently commercially available NAATs For instance, UK hospitals can be assessed finan-
began emerging as viable replacements, particu- cial penalties for excessive numbers of hospital-
larly because the high sensitivity of NAATs acquired CDI cases (Davies et al. 2016). Simi-
would directly address the shortcoming of Tox larly, in the US, the Centres for Medicare and
A/B EIA. For example, in Canada, a cross- Medicaid Services (CMS) value-based purchas-
sectional study across Quebec showed that the ing program are affected by reported incidence
number of hospitals detecting toxigenic rates (Marra et al. 2017). In the latter’s case, an
C. difficile instead of C. difficile toxins increased attempt to normalize rates by factoring in testing
significantly between 2010 and 2014, and in 2014 method has been made, although the study
stand-alone NAAT was the most common applied demonstrated the inadequacy of such normaliza-
assay (21% of hospitals) (Bogaty et al. 2017). But tion and stressed the need for refinement.
recent work has suggested that NAATs may now In conclusion, CDI incidence is clearly
be causing CDI over-diagnosis, leading to an over- affected by testing method. Given the heteroge-
estimation of CDI incidence in hospitals using neity of such methods between institutions, and
NAAT rather than algorithmic testing. the importance of correctly ascertaining CDI
In the study across 60 European hospitals, a incidence, it is necessary to somehow normalize
2.5-fold higher CDI positivity rate was incidence rates in a way that takes into consider-
demonstrated when stand-alone or GDH/NAAT ation testing method.
were used instead of a recommended algorithm.
This was reflected in the subsequent incidence
rates; hospitals relying on NAAT or GDH/NAAT 8 Alternative Testing Strategies
reported a mean incidence rate of 5.2 per 10,000
patient-days, while hospitals relying on an algo- Methods for diagnosing CDI remain imperfect,
rithm reported a lower mean incidence rate of 2.0 which naturally has spurned an interest in alter-
per 10,000 patient-days, despite similar testing native testing strategies. Alternative testing
frequencies (Davies et al. 2016). strategies cannot only possibly aid in the
Diagnostic Guidance for C. difficile Infections 35

diagnosis of CDI, but might also be able to pre- marker of inflammation due to release into the
dict severity or prognosis of CDI. These testing gut lumen by neutrophils during infiltration and
methods include faecal biomarkers, immune can be measured in stool (Popiel et al. 2015).
modulating interleukins and cytokines and imag- However, infection cannot be differentiated
ing methods. Their role is discussed below. from inflammation by this marker, since both
give a rise in faecal calprotectin (FCP) levels
(Usacheva et al. 2016). The role of calprotectin
in evaluating disease severity has been well stud-
8.1 Calprotectin
ied in IBD (Vrabie and Kane 2014). Several
studies evaluated the role of FCP in CDI testing
Calprotectin, a calcium-and zinc-binding pro-
(Table 2). First, the usefulness of FCP testing to
tein, is found predominantly in the cytosol of
diagnose CDI was evaluated in several studies. In
neutrophils (Usacheva et al. 2016; Popiel et al.
all studies, median FCP levels were found to be
2015; Whitehead et al. 2014). In vitro studies
significantly higher in CDI patients than in
have shown that it has bacteriostatic and
diarrhoeal patients who tested negative for CDI
fungostatic properties (Peretz et al. 2016). It is a

Table 2 Overview of studies evaluating the role of FCP in patients with CDI
Study Type of study Detection of CDI Number of cases/controls Results
Kim et al. Retrospective NAAT for toxin 30 pts. with severe CDI (group CDI diagnosis
Ann Lab cohort study gene 1), 50 pts. with mild CDI Median levels of FCP were
Med (group 2) and 71 CDI neg significantly higher in group
(2016) healthy controls (group 3) 1 than in group 2 and group
3, 1391.5 μg/g
(170.0–2088.1 μg/g) vs
188.2 μg/g (41.4–188.2 μg/g)
and 35.0 μg/g (10.7–108.9 μg/
g) respectively
Optimal cut-off value for CDI
diagnosis 112.5 μg/g, ROC
curve AUC 0.821 sens 75%
and spec.79%
CDI severity
Median levels of FCP were
significantly higher in group
1 than in group 2, 1391.5 μg/g
(173.5–2075.9 μg/g) vs
188.2 μg/g (41.4–591.6 μg/g),
Optimal cut-off value for
differentiating mild from
severe CDI
729.8 μg/g, ROC curve, AUC
0.746, sens 70% and spec 80%
Peretz et al. Retrospective NAAT for toxin 29 pts. with CDI (7 CDI Overal mean levels of FCP
BMC cohort study gene and ribotype 027, 22 other 331.4 μg/g (21–932 μg/g)
Infect Dis identification ribotype) Mean levels of FCP were
(2016) 027 strains significantly higher in 027 pos
group than in 027 neg group,
331.4 μg/g (21–932 μg/g) vs
249 μg/g (155–498 μg/g),
A trend was found between
higher FCP levels and higher
Clostridium severity score
36 M. J. T. Crobach et al.

Table 2 (continued)
Study Type of study Detection of CDI Number of cases/controls Results
Popiel Prospective NAAT for toxin 44 CD-PCR pos vs Median levels of higher-range
et al. JCM exploratory gene 20 CD-PCR neg assay of FCP (assay range,
(2015) observational 100–1800 μg/g) were
study significantly higher in
CD-PCR+ than in CD-PCR-,
983 μg/g (351- > 1800 μg/g)
vs <100 μg/g (<100–194 μg/
g) and also in the lower range
assay of FCP (assay range,
30–300 μg/g) >300 μg/g
(>300- >300 μg/g) vs 77.5 μg/
g (30–238 μg/g)
Optimal cut-off value 135 μg/g
High range FCP ROC curve
AUC 0.82 sens. 88.6% and
spec. 75%
Whitehead Prospective Phase 1: toxin EIA 75 pts. toxin EIA pos (group Median levels of FCP were
et al. J Med cohort study (N ¼ 75) 1), 45 pts. GDH-EIA/NAAT significantly higher in group
Microbiol Phase 2: GDH EIA pos (group 2), 99 pts. negative 1 than in group 2, 336 μg/g
(2014) + NAAT for toxin for C. difficile (group 3), (208–536 μg/g) vs 249 μg/g
gene (N ¼ 45) group 3: 99 cases in CDI (155–498 μg/g), respectively.
Change of negative Both were significantly higher
departmental than in group 3, 106
C. difficile testing (46–176 μg/g)
methodology Optimal cut-off value 176 μg/g
during evaluation and 169 μg/g, ROC curve AUC
0.84 and 0.80, sens 81% and
73%, spec 77% and 77% for
group 1 and 2, respectively
Swale et al. Prospective NAAT for toxin 164 CDI cases vs 52 AAD CDI diagnosis
PLOS One cohort study gene and toxin EIA controls Median levels of FCP were
(2014) significantly higher in CDI
cases vs AAD, 684.8 μg/g
(203.7–1581.0 μg/g) vs
66.5 μg/g (23.1–145.7 μg/g),
Optimal cut-off value 148 μg/g
ROC curve AUC 0.86 4 sens
81,8% spec 76.9% PPV
91.5%, NPV 57.4%
Sub-group analyses:
8 severe CDI cases vs CDI severity
116 non-severe CDI cases Median levels of FCP were not
significantly higher in severe
CDI cases vs non-severe CDI
cases, 969.3 μg/g vs 512.7 μg/
g), respectively.
C. difficile isolates recovered Ribotype 027
from 149 CDI cases Median levels of FCP were not
72 cases with ribotype 027 vs significantly higher in ribotype
77 non-ribotype 027 027 cases vs non-ribotype
027 cases, 1011 μg/g vs
658 μg/g), respectively
Diagnostic Guidance for C. difficile Infections 37

Table 2 (continued)
Study Type of study Detection of CDI Number of cases/controls Results
Darkoh Prospective AAD stools: CDI-positive stools (N ¼ 50), FCP concentration in CDI pos
et al. Clin cohort study CCNA, NAAT for CDI- negative stools (N ¼ 50) stools, 18 μg/g (2.8–70.2 μg/g)
Vaccine toxin genes and hospitalized patients without was 3-fold higher than in
Immunol toxin EIA diarrhea (N ¼ 45) CDI-neg stools, 6.5 μg/g
(2014) Control stools: (2.0–31.0 μg/g) and 2-fold
NAAT for toxin higher than of hospitalized pts.
gene and toxin EIA without diarrhea, 8.7 μg/g
(1.8–33.2 μg/g)
FCP levels of 80% of the
CDI-pos stools and 30% of the
CDI-neg stools higher than
hospitalized pts. without
AAD antibiotic-associated diarrhea, CCNA cell cytotoxicity neutralization assay, CDI Clostridium difficile infection,
FCP fecal calprotectin, NAAT nucleic acid amplification test

and in non-diarrhoeal controls (Kim et al. 2017; 2014). The correlation between FCP levels and
Popiel et al. 2015; Darkoh et al. 2014; Swale CDI severity has also been evaluated in three
et al. 2014; Whitehead et al. 2014). Studies who small studies, but results were conflicting (Kim
calculated optimal FCP cut-off points for et al. 2017; Swale et al. 2014; Peretz et al. 2016).
distinguishing CDI from non-CDI samples A correlation between CDI due to ribotype
reported sensitivities ranging from 77% to 88% 027 and FCP levels was also evaluated in two
and specificities ranging from 75% to 79% studies (Peretz et al. 2016; Swale et al. 2014).
(Popiel et al. 2015; Kim et al. 2017; Swale Significantly higher FCP levels compared to
et al. 2014; Whitehead et al. 2014). However, in non-027 CDI were found in one small study
two of these studies the discriminative power of comprising 7 ribotype 027 cases and 22 -
FCP might have been attenuated as the group of non-ribotype 027 cases (Peretz et al. 2016), the
CDI patients might have included CD carriers same trend was shown in a bit larger study, but
due to testing for CDI by NAAT only (Popiel results were not significant (Swale et al. 2014).
et al. 2015; Kim et al. 2017). On the other hand, In conclusion, there is also insufficient evi-
the use of healthy controls instead of patients dence for the use of FCP levels to predict severity
suspected of CDI might have falsely increased or presence of ribotype 027.
the specificity in one study (Kim et al. 2017).
Overall, the suboptimal sensitivity and specific-
ity demonstrated in these observational studies, 8.2 Lactoferrin
of which several with limitations or small sample
sizes, does not provide enough evidence for the Lactoferrin is a glycoprotein and resides in
use of FCP to detect CDI. neutrophils. It is released upon neutrophil activa-
Interestingly, besides the expected suboptimal tion. The faecal lactoferrin (FL) levels can be
specificity of FCP, sensitivity is also moderate. measured in stool and correlate with the number
One study reported that in 20% of CDI patients, of infiltrated neutrophils. Multiple studies have
FCP levels were lower than in hospitalised proven that it can be an accurate marker of intes-
patients without diarrhoea (Darkoh et al. 2014). tinal inflammation and useful in diagnosis of
Another study reported that from 120 CDI inflammatory diarrhoea (Usacheva et al. 2016).
subjects only five had normal FCP levels The usefulness of FL to detect CDI was
(<50 μg/g) and speculated that these cases evaluated in a handful studies (Table 3). All stud-
might represent mild disease (Whitehead et al. ies report higher median FL levels in CDI samples
Table 3 Overview of studies evaluating the role of FL in patients with CDI
Study Study type of CDI Number of cases/controls Results
Darkoh et al. Prospective AAD 50 pts. with CDI-positive stools, FL concentration in CDI- pos
Clin Vaccine cohort study stools: 50 pts. with CDI-negative stools, stools, 31.4 μg/g (3.0–155.2 μg/
Immunol CCNA, 45 hospitalized pts. without g) was significantly different and
(2014) NAAT and diarrhea was 5-fold higher than in
toxin EIA CDI-neg stools, 6.3 μg/g
control (0.6–140.3 μg/g) and 6-fold
stools: higher than of hospitalized pts.
NAAT and without diarrhea, 5.6 μg/g
toxin EIA (0.5–35.0 μg/g)
FL levels of 88% of the CDI-pos
stools and 44% of the CDI-neg
stools higher than in hospitalized
pts. without diarrhea
Swale et al. Prospective toxin EIA 164 CDI cases vs 52 AAD Median levels of FL were
PLOS One cohort study controls significantly higher in CDI
(2014) cases 57.9 μg/ml
(11.4–177.5 μg/ml) vs AAD
2.7 μg/ml (0.7–7.8 μg/ml)
Optimal cut-off value 8.06 ng/ml
ROC curve AUC 0.859, Sens
81,7% Spec 76.9%, PPV 91.8%,
NPV 57.1%
Sub-group analysis:
8 severe CDI cases vs CDI severity
116 non-severe CDI cases Median levels of FL were
significantly higher in severe
CDI cases vs non-severe CDI
cases, 104.6 μg/ml vs 40.1 μg/
ml, respectively
C. difficile isolates recovered Ribotype 027
from 149 CDI cases Median levels of FL were not
72 cases with ribotype 027 vs significantly higher in ribotype
77 non-ribotype 027 027 cases vs non-ribotype
027 cases, 83.2 μg/ml vs
51.0 μg/ml, respectively
Archbald- Prospective Not 79 pts. with 41 severe CDI Overal mean concentration of FL
Pannone, J cohort study described vs. 38 pts. with non-severe CDI in the cohort was 388.8 μg/ml
Geriatr Paliat Mean levels of LF in severe CDI
Care (2014) pts. 580 μg/ml (sd 989.0 μg/ml)
were significantly higher than in
non-severe CDI pts. 181.7 μg/ml
(sd 244.2 μg/ml)
Boone et al. Prospective NAAT and N ¼ 210 FL concentration in group
Eur J Clin cohort study TC 129 TC+&CCNA + (group 1), 1 (90 μg/g) was significantly
Microbiol 62 TC+&CCNA- (group 2) and higher than in group 2 (24 μg/g)
Infect Dis 19 TC – &CCNA- (group 3) and group 3 (20 μg/g)
Boone et al. Prospective GDH N ¼ 98 96% of pts. with pos toxin stool
Eur J Clin cohort study Membrane (85 toxigenic strains, had elevated LF and 59% of pts
Microbiol based EIA 6 non-toxigenic, 6 neg for C. negative stool toxin had elevated
Infect Dis and toxin difficile, 1 mixed infection) levels
(2013) EIA 85 toxigenic (21 severe CDI, Mean levels of severe CDI
57 moderate, 7 mild) (961 μg/g, SE 303 μg/g) were
significantly higher than in
moderate CDI, (292 μg/g, SE
42 μg/g), and mild CDI (73 μg/g,
SE 52 μg/g)
38 pts. had a 027 infection (45%) There is a significant difference
in LF between pts. with 027 and
Diagnostic Guidance for C. difficile Infections 39

Table 3 (continued)
Study Study type of CDI Number of cases/controls Results
LaSala et al. J Retrospective GDH EIA, N ¼ 112 Median levels of LF were
Clin cohort study toxin EIA 43 GDH neg (group 1), 14 GDH significantly higher in group
Microbiol and NAAT pos/toxin neg/PCR neg (group 3, 80 μg/ml (3–124 μg/ml) than
(2013) 2), 25 GDH& toxin pos (group in group 1, 13 μg/ml (3–143 μg/
3), 30 GDH pos/toxin neg/PCR ml), group 2, 18 μg/ml (4–78 μg/
pos (group 4) ml) and group 4, 24 μg/ml
(4–160 μg/ml)
AAD antibiotic-associated diarrhea, CCNA cell cytotoxicity neutralization assay, CDI Clostridium difficile infection,
EIA enzyme immunoassay, FL fecal lactoferrin, GDH glutamate dehydrogenase, NAAT nucleid acid amplification test,
TC toxigenic culture

than in control samples (either diarrheal samples parts may be ascribed to variation in disease
without CDI or non-diarrheal samples) (Swale severity, while other parts are due to laboratory
et al. 2014; Darkoh et al. 2014; Boone et al. handling and the volume of diluent. Another
2014; LaSala et al. 2013). However, a substantial problem is that FL can be elevated due to
proportion of CDI-negative patients have elevated co-morbidities, such as ulcerative colitis and
FL levels, too (Boone et al. 2014; Darkoh et al. Crohn’s disease.
2014). This was also reflected in the suboptimal Some studies also report an association
specificity of 77% that was found when an opti- between elevated FL and CDI severity
mal cut-off point to distinguish CDI from (Archbald-Pannone 2014; Boone et al. 2013).
patients with non-CDI antibiotic associated diar- However these studies have had small sample
rhoea was determined (Swale et al. 2014). sizes, and more research is needed. To our
Whether FL could be used as a marker for knowledge, there are no studies that observed
severe CDI was also evaluated in some studies. that FL, on its own, is a predictor of severity or
Severe CDI was found to be associated with mortality. Therefore more research is needed to
higher median FL levels in two small studies understand the role of lactoferrin.
(Archbald-Pannone 2014; Boone et al. 2013). In
addition, higher FL levels were associated with a
higher white blood cell count and decreased 8.3 Faecal Leukocyte Test
serum albumin (Boone et al. 2013), but no asso-
ciation with mortality was demonstrated The faecal leukocyte test is performed on stool
(Archbald-Pannone 2014), possibly due to small specimens, which are smeared on slides and
cohorts. Furthermore, it was demonstrated that Wright stained. The test takes approximately
patients with CDI due to ribotype 027 and posi- 1 h and samples are positive when >1
tive stool toxin had significantly higher FL levels WBC/highfield are observed (Reddymasu et al.
and WBC counts than non-027 CDI patients 2006). However, in a study evaluating 263 stool
(Boone et al. 2013, 2014). In patients with CDI samples from patients suspected of CDI for the
due to ribotype 027, patients with positive stool diagnosis of CDI, the faecal leukocyte test
toxin and elevated FL had a higher mortality risk showed a sensitivity and specificity of 30% and
(Boone et al. 2014). 74.9% respectively compared to toxin EIA
To conclude, all of the studies report an asso- (Reddymasu et al. 2006). A larger study
ciation between elevated FL and CDI. However, (n ¼ 797 stool samples) observed a sensitivity
the reported specificity is insufficient for imple- and specificity of 14% and 90% respectively
mentation in the diagnosis of CDI. Furthermore, (Savola et al. 2001). Thus, faecal leukocyte test-
as the studies report different median FL levels, ing is not a good test for CDI and a poor predictor
this would reduce predictive accuracy. Some of the toxin assay result.
40 M. J. T. Crobach et al.

8.4 Interleukins and Chemokines diarrhoeal persistence and longer time to diar-
rhoea resolution. The levels were also higher in
IL 8 is a chemoattractant and recruits neutrophils patients with CDI in the prior 90 days than in
to sites of infection. Activated dendritic cells and patients with no history of CDI (El Feghaly et al.
macrophages produce IL 23. This interleukin is 2013).
involved in host defence against bacterial So, it seems that markers of inflammation play
infections and the development of chronic a role in CDI and may correlate to disease sever-
inflammation. Darkoh and colleagues tested ity. However more research is needed to confirm
CDI stools, diarrheal non-CDI stools and these associations.
non-diarrheal stools for interleukins both by a
cytokine assay and by a quantitative EIA
(Darkoh et al. 2014). Both IL-8 and IL-23 were 8.5 CT-Imaging
detected in more CDI-positive stools than
CDI-negative stools. The cytokine assay showed CT imaging can be useful in diagnosing fulmi-
that the relative amount of IL-8 was higher in the nant CDI and pseudomembranous colitis (PMC).
50 CDI-positive stools, compared to 50 CDI- Several features are suggestive of advanced PMC
negative stools. This in contrast with IL-23, such as colonic-wall thickening, pericolonic
were the relative amount was higher in the stranding, the accordion sign, the double-halo
CDI-negative stools. When the findings were sign and ascites (Bartlett and Gerding 2008;
confirmed by EIA, they found that CDI-positive Kirkpatrick and Greenberg 2001). The radiogra-
stools showed a significantly higher amount of IL phy is usually normal in absence of ileus or toxic
8 (mean 318.2 pg/ml) in stools compared to the megacolon.
CDI-negative stools (mean 84.7 pg/ml) and Kirkpatrick et al. evaluated whether diagnosis
hospitalized patients without diarrhoea (mean of C. difficile colitis could be made with CT
79.8 pg/ml). In contrast, IL 23 was significantly (Kirkpatrick and Greenberg 2001). They
higher in CDI-negative stools and hospitalized included 110 patients of which 54 had a positive
patients without diarrhea than in the CDI positive stool assay and 56 patients a negative stool assay.
stools, 946.7 pg/ml (185.5–2016 pg/ml), The sensitivity at their centre was 52%, the spec-
1617 pg/ml (489.0–6810 pg/ml) and 722 pg/ml ificity 93% and the positive and negative predic-
(110.0–7069 pg/ml), respectively. This study tive values were respectively 88% and 67%. CT
shows that IL-8 plays a role in CDI and that imaging is less sensitive when compared with
increased levels are associated with more severe NAAT or stool toxin tests but can be useful
forms of CDI. In contrast, IL-23 amounts during when there is a need for quick results (Bartlett
CDI may be inadequate to sustain sufficient cel- and Gerding 2008).
lular immunity. Therefore, lower concentrations
of IL-23 may show a lack of immunological
response in a proportion of CDI patients and 8.6 Endoscopy
may explain also recurrence (Darkoh et al. 2014).
CXCL-5 is a CXC chemokine and recruits and Nearly all cases of PMC are caused by CDI (Tang
activates neutrophils. El Feghaly and colleagues et al. 2016) though other causes are sometimes
studied the correlation between intestinal inflam- found, such as chemotherapy, toxin producing
mation and disease severity in hospitalized Staphylococus aureus and cytomegalovirus infec-
patients with symptomatic CDI (El Feghaly tion (Sundar and Chan 2003; Pressly et al. 2016).
et al. 2013). They found that faecal CXCL-5 PMC is not very common and not all CDI will
mRNA and IL-8 mRNA were associated with develop in PMC (Bartlett 2002). Therefore,
Diagnostic Guidance for C. difficile Infections 41

endoscopy is a relatively insensitive procedure. Fur- Dis 46(Suppl 1):S12–S18.

thermore, in one third of the patients PMC is missed 521863
Berrington A, Settle CD (2007) Which specimens should
by sigmoidoscopy because of involvement of the be tested for Clostridium difficile toxin? J Hosp Infect
right colon; making colonoscopy the preferred 65(3):280–282.
endoscopic procedure. Endoscopy is an invasive 12.011
procedure with perforation risks and is often expen- Bidet P, Barbut F, Lalande V, Burghoffer B, Petit JC
(1999) Development of a new PCR-ribotyping method
sive (Bartlett 2002; Bartlett and Gerding 2008). for Clostridium difficile based on ribosomal RNA gene
Endoscopy is therefore not recommended to sequencing. FEMS Microbiol Lett 175(2):261–266
diagnose CDI. Bogaty C, Levesque S, Garenc C, Frenette C, Bolduc D,
Galarneau LA, Lalancette C, Loo V, Tremblay C,
Trudeau M, Vachon J, Dionne M, Villeneuve J,
Longtin J, Longtin Y (2017) Trends in the use of
8.7 Histopathology laboratory tests for the diagnosis of Clostridium diffi-
cile infection and association with incidence rates in
Quebec, Canada, 2010–2014. Am J Infect Control 45
CDI is more likely when pseudomembranes are (9):964–968.
detected histologically. Pseudomembrane lesions 002
are microscopically visualized as “mushroom” Boone JH, DiPersio JR, Tan MJ, Salstrom SJ, Wickham
like and consist of pus, mucin and fibrin. Their KN, Carman RJ, Totty HR, Albert RE, Lyerly DM
(2013) Elevated lactoferrin is associated with moder-
reported sensitivity is 44% and the specificity is ate to severe Clostridium difficile disease, stool toxin,
89% (Wang et al. 2013). Biopsy is not needed for and 027 infection. Eur J Clin Microbiol Infect Dis 32
the diagnosis of CDI. However, histologic (12):1517–1523.
findings of pseudomembranes may suggest CDI 1905-x
Boone JH, Archbald-Pannone LR, Wickham KN, Carman
and should stimulate stool testing. RJ, Guerrant RL, Franck CT, Lyerly DM (2014)
Ribotype 027 Clostridium difficile infections with
measurable stool toxin have increased lactoferrin and
are associated with a higher mortality. Eur J Clin
References Microbiol Infect Dis 33(6):1045–1051. https://doi.
Aichinger E, Schleck CD, Harmsen WS, Nyre LM, Patel R Burnham CA, Carroll KC (2013) Diagnosis of Clostrid-
(2008) Nonutility of repeat laboratory testing for detec- ium difficile infection: an ongoing conundrum for
tion of Clostridium difficile by use of PCR or enzyme clinicians and for clinical laboratories. Clin Microbiol
immunoassay. J Clin Microbiol 46(11):3795–3797. Rev 26(3):604–630. 00016-13
Alcala L, Martin A, Marin M, Sanchez-Somolinos M, CDC (2012) MMWR Morb Wkly Rep 61:157–162
Catalan P, Pelaez T, Bouza E (2012) The undiagnosed Chung HS, Lee M (2017) Evaluation of the performance
cases of Clostridium difficile infection in a whole of C. DIFF QUIK CHEK COMPLETE and its useful-
nation: where is the problem? Clin Microbiol Infect ness in a hospital setting with a high prevalence of
18(7):E204–E213. Clostridium difficile infection. J Investig Med 65
0691.2012.03883.x (1):88–92.
Archbald-Pannone LR (2014) Quantitative fecal lactoferrin Cohen SH, Gerding DN, Johnson S, Kelly CP, Loo VG,
as a biomarker for severe Clostridium difficile infection McDonald LC, Pepin J, Wilcox MH (2010) Clinical
in hospitalized patients. J Geriatr Palliat Care 2(1):3. practice guidelines for Clostridium difficile infection in adults: 2010 update by the society for healthcare
Barbut F, Surgers L, Eckert C, Visseaux B, Cuingnet M, epidemiology of America (SHEA) and the infectious
Mesquita C, Pradier N, Thiriez A, Ait-Ammar N, diseases society of America (IDSA). Infect Control
Aifaoui A, Grandsire E, Lalande V (2014) Does a Hosp Epidemiol 31(5):431–455.
rapid diagnosis of Clostridium difficile infection 1086/651706
impact on quality of patient management? Clin Crobach MJ, Planche T, Eckert C, Barbut F, Terveer EM,
Microbiol Infect 20(2):136–144. Dekkers OM, Wilcox MH, Kuijper EJ (2016) Euro-
1111/1469-0691.12221 pean society of clinical microbiology and infectious
Bartlett JG (2002) Clinical practice. Antibiotic-associated diseases: update of the diagnostic guidance document
diarrhea. N Engl J Med 346(5):334–339. https://doi. for Clostridium difficile infection. Clin Microbiol
org/10.1056/NEJMcp011603 Infect 22(Suppl 4):S63–S81.
Bartlett JG, Gerding DN (2008) Clinical recognition and j.cmi.2016.03.010
diagnosis of Clostridium difficile infection. Clin Infect
42 M. J. T. Crobach et al.

Darkoh C, Turnwald BP, Koo HL, Garey KW, Jiang ZD, Microbiol 52(11):3853–3855.
Aitken SL, DuPont HL (2014) Colonic JCM.01659-14
immunopathogenesis of Clostridium difficile Grein JD, Ochner M, Hoang H, Jin A, Morgan MA,
infections. Clin Vaccine Immunol 21(4):509–517. Murthy AR (2014) Comparison of testing approaches for Clostridium difficile infection at a large commu-
Davies KA, Longshaw CM, Davis GL, Bouza E, Barbut nity hospital. Clin Microbiol Infect 20(1):65–69.
F, Barna Z, Delmee M, Fitzpatrick F, Ivanova K,
Kuijper E, Macovei IS, Mentula S, Mastrantonio P, Hensgens MP, Dekkers OM, Demeulemeester A, Buiting
von Muller L, Oleastro M, Petinaki E, Pituch H, Noren AG, Bloembergen P, van Benthem BH, Le Cessie S,
T, Novakova E, Nyc O, Rupnik M, Schmid D, Wilcox Kuijper EJ (2014) Diarrhoea in general practice: when
MH (2014) Underdiagnosis of Clostridium difficile should a Clostridium difficile infection be considered?
across Europe: the European, multicentre, prospec- Results of a nested case-control study. Clin Microbiol
tive, biannual, point-prevalence study of Clostridium Infect 20(12):O1067–O1074.
difficile infection in hospitalised patients with diar- 1469-0691.12758
rhoea (EUCLID). Lancet Infect Dis 14(12):1208– Hink T, Burnham CA, Dubberke ER (2013) A systematic
1219. evaluation of methods to optimize culture-based
Davies K, Davis G, Barbut F, Eckert C, Petrosillo N, recovery of Clostridium difficile from stool specimens.
Wilcox MH (2016a) Variability in testing policies Anaerobe 19:39–43.
and impact on reported Clostridium difficile infection obe.2012.12.001
rates: results from the pilot longitudinal European Jazmati N, Hellmich M, Licanin B, Plum G, Kaasch AJ
Clostridium difficile infection diagnosis surveillance (2016) PCR cycle threshold value predicts the course
study (LuCID). Eur J Clin Microbiol Infect Dis 35 of Clostridium difficile infection. Clin Microbiol
(12):1949–1956. Infect 22(2):e7–e8.
2746-1 2015.09.012
Debast SB, van Kregten E, Oskam KM, van den Berg T, Kaltsas A, Simon M, Unruh LH, Son C, Wroblewski D,
Van den Berg RJ, Kuijper EJ (2008) Effect on diag- Musser KA, Sepkowitz K, Babady NE, Kamboj M
nostic yield of repeated stool testing during outbreaks (2012) Clinical and laboratory characteristics of Clos-
of Clostridium difficile-associated disease. Clin tridium difficile infection in patients with discordant
Microbiol Infect 14(6):622–624. diagnostic test results. J Clin Microbiol 50(4):1303–
1111/j.1469-0691.2008.01999.x 1307.
Delmee M (2001) Laboratory diagnosis of Clostridium Khanna S, Pardi DS, Rosenblatt JE, Patel R, Kammer PP,
difficile disease. Clin Microbiol Infect 7(8):411–416 Baddour LM (2012) An evaluation of repeat stool testing
Dionne LL, Raymond F, Corbeil J, Longtin J, Gervais P, for Clostridium difficile infection by polymerase chain
Longtin Y (2013) Correlation between Clostridium reaction. J Clin Gastroenterol 46(10):846–849. https://
difficile bacterial load, commercial realtime PCR
cycle thresholds, and results of diagnostic tests based Killgore G, Thompson A, Johnson S, Brazier J, Kuijper E,
on enzyme immunoassay and cell culture cytotoxicity Pepin J, Frost EH, Savelkoul P, Nicholson B, van den
assay. J Clin Microbiol 51(11):3624–3630. https://doi. Berg RJ, Kato H, Sambol SP, Zukowski W, Woods C,
org/10.1128/JCM.01444-13 Limbago B, Gerding DN, MD LC (2008) Comparison
El Feghaly RE, Stauber JL, Deych E, Gonzalez C, Tarr PI, of seven techniques for typing international epidemic
Haslam DB (2013) Markers of intestinal inflamma- strains of Clostridium difficile: restriction endonucle-
tion, not bacterial burden, correlate with clinical ase analysis, pulsed-field gel electrophoresis, PCR-
outcomes in Clostridium difficile infection. Clin Infect ribotyping, multilocus sequence typing, multilocus
Dis 56(12):1713–1721. variable-number tandem-repeat analysis, amplified
cit147 fragment length polymorphism, and surface layer pro-
Enoch DA, Butler MJ, Pai S, Aliyu SH, Karas JA (2011) tein A gene sequence typing. J Clin Microbiol 46
Clostridium difficile in children: colonisation and dis- (2):431–437.
ease. J Inf Secur 63(2):105–113. Kim J, Kim H, Oh HJ, Kim HS, Hwang YJ, Yong D,
1016/j.jinf.2011.05.016 Jeong SH, Lee K (2017) Fecal calprotectin level
Fawley WN, Knetsch CW, MacCannell DR, Harmanus C, reflects the severity of Clostridium difficile Infection.
Du T, Mulvey MR, Paulick A, Anderson L, Kuijper EJ, Ann Lab Med 37(1):53–57.
Wilcox MH (2015) Development and validation of an alm.2017.37.1.53
internationally-standardized, high-resolution capillary Kirkpatrick ID, Greenberg HM (2001) Evaluating the CT
gel-based electrophoresis PCR-ribotyping protocol for diagnosis of Clostridium difficile colitis: should CT
Clostridium difficile. PLoS One 10(2):e0118150. guide therapy? AJR Am J Roentgenol 176(3):635– 639.
Green DA, Stotler B, Jackman D, Whittier S, Della-Latta Knetsch CW, Lawley TD, Hensgens MP, Corver J,
P (2014) Clinical characteristics of patients who test Wilcox MW, Kuijper EJ (2013) Current application
positive for Clostridium difficile by repeat PCR. J Clin and future perspectives of molecular typing methods
Diagnostic Guidance for C. difficile Infections 43

to study Clostridium difficile infections. Euro Surveill Epidemiol 38(1):109–111.

18(4):20381 ice.2016.227
Kristjansson M, Samore MH, Gerding DN, DeGirolami McFarland LV, Coyle MB, Kremer WH, Stamm WE
PC, Bettin KM, Karchmer AW, Arbeit RD (1994) (1987) Rectal swab cultures for Clostridium difficile
Comparison of restriction endonuclease analysis, surveillance studies. J Clin Microbiol 25(11):2241–
ribotyping, and pulsed-field gel electrophoresis for 2242
molecular differentiation of Clostridium difficile Moehring RW, Lofgren ET, Anderson DJ (2013) Impact
strains. J Clin Microbiol 32(8):1963–1969 of change to molecular testing for Clostridium difficile
Kundrapu S, Sunkesula VC, Jury LA, Sethi AK, Donskey infection on healthcare facilityassociated incidence
CJ (2012) Utility of perirectal swab specimens for rates. Infect Control Hosp Epidemiol 34(10):1055–
diagnosis of Clostridium difficile infection. Clin Infect 1061.
Dis 55(11):1527–1530. Peretz A, Tkhawkho L, Pastukh N, Brodsky D, Halevi
cis707 CN, Nitzan O (2016) Correlation between fecal
Kyne L, Warny M, Qamar A, Kelly CP (2000) Asymp- calprotectin levels, disease severity and the hypervir-
tomatic carriage of Clostridium difficile and serum ulent ribotype 027 strain in patients with Clostridium
levels of IgG antibody against toxin A. N Engl J difficile infection. BMC Infect Dis 16:309. https://doi.
Med 342(6):390–397. org/10.1186/s12879-016-1618-8
nejm200002103420604 Planche T, Wilcox M (2011) Reference assays for Clos-
LaSala PR, Ekhmimi T, Hill AK, Farooqi I, Perrotta PL tridium difficile infection: one or two gold standards? J
(2013) Quantitative fecal lactoferrin in toxin-positive Clin Pathol 64(1):1–5.
and toxin-negative Clostridium difficile specimens. J 2010.080135
Clin Microbiol 51(1):311–313. Planche TD, Davies KA, Coen PG, Finney JM, Monahan
1128/jcm.02735-12 IM, Morris KA, O’Connor L, Oakley SJ, Pope CF,
Leslie JL, Cohen SH, Solnick JV, Polage CR (2012) Role Wren MW, Shetty NP, Crook DW, Wilcox MH (2013)
of fecal Clostridium difficile load in discrepancies Differences in outcome according to Clostridium dif-
between toxin tests and PCR: is quantitation the next ficile testing method: a prospective multicentre diag-
step in C. difficile testing? Eur J Clin Microbiol Infect nostic validation study of C difficile infection. Lancet
Dis 31(12):3295–3299. Infect Dis 13(11):936–945.
s10096-012-1695-6 s1473-3099(13)70200-7
Longtin Y, Trottier S, Brochu G, Paquet-Bolduc B, Polage CR, Gyorke CE, Kennedy MA, Leslie JL, Chin
Garenc C, Loungnarath V, Beaulieu C, Goulet D, DL, Wang S, Nguyen HH, Huang B, Tang YW, Lee
Longtin J (2013) Impact of the type of diagnostic LW, Kim K, Taylor S, Romano PS, Panacek EA,
assay on Clostridium difficile infection and complica- Goodell PB, Solnick JV, Cohen SH (2015) Overdiag-
tion rates in a mandatory reporting program. Clin nosis of Clostridium difficile Infection in the molecu-
Infect Dis 56(1):67–73. lar test era. JAMA Intern Med 175(11):1792–1801.
Loo VG, Bourgault AM, Poirier L, Lamothe F, Michaud Popiel KY, Gheorghe R, Eastmond J, Miller MA (2015)
S, Turgeon N, Toye B, Beaudoin A, Frost EH, Gilca Usefulness of adjunctive fecal calprotectin and serum
R, Brassard P, Dendukuri N, Beliveau C, Oughton M, procalcitonin in individuals positive for Clostridium
Brukner I, Dascal A (2011) Host and pathogen factors difficile toxin gene by PCR assay. J Clin Microbiol 53
for Clostridium difficile infection and colonization. N (11):3667–3669.
Engl J Med 365(18):1693–1703. 15
1056/NEJMoa1012413 Pressly KB, Hill E, Shah KJ (2016) Pseudomembranous
Luo RF, Banaei N (2010) Is repeat PCR needed for colitis secondary to methicillin-resistant Staphylococ-
diagnosis of Clostridium difficile infection? J Clin cus aureus (MRSA). BMJ Case Rep 2016. https://doi.
Microbiol 48(10):3738–3741. org/10.1136/bcr-2016-215225
1128/jcm.00722-10 Reddymasu S, Sheth A, Banks DE (2006) Is fecal leuko-
Maiden MC, Bygraves JA, Feil E, Morelli G, Russell JE, cyte test a good predictor of Clostridium difficile
Urwin R, Zhang Q, Zhou J, Zurth K, Caugant DA, associated diarrhea? Ann Clin Microbiol Antimicrob
Feavers IM, Achtman M, Spratt BG (1998) Multilocus 5:9.
sequence typing: a portable approach to the identifica- Reigadas E, Alcala L, Marin M, Burillo A, Munoz P,
tion of clones within populations of pathogenic Bouza E (2015) Missed diagnosis of Clostridium diffi-
microorganisms. Proc Natl Acad Sci U S A 95 cile infection; a prospective evaluation of unselected
(6):3140–3145 stool samples. J Inf Secur 70(3):264–272. https://doi.
Marra AR, Edmond MB, Ford BA, Herwaldt LA, org/10.1016/j.jinf.2014.10.013
Algwizani AR, Diekema DJ (2017) Failure of risk- Reigadas E, Alcala L, Valerio M, Marin M, Martin A,
adjustment by test method for C. difficile laboratory- Bouza E (2016) Toxin B PCR cycle threshold as a
identified event reporting. Infect Control Hosp predictor of poor outcome of Clostridium difficile
44 M. J. T. Crobach et al.

infection: a derivation and validation cohort study. J R, Parry CM, Pirmohamed M (2014) Calprotectin
Antimicrob Chemother 71(5):1380–1385. https://doi. and lactoferrin faecal levels in patients with Clostrid-
org/10.1093/jac/dkv497 ium difficile infection (CDI): a prospective cohort
Rogers DS, Kundrapu S, Sunkesula VC, Donskey CJ study. PLoS One 9(8):e106118.
(2013) Comparison of perirectal versus rectal swabs 1371/journal.pone.0106118
for detection of asymptomatic carriers of toxigenic Tang DM, Urrunaga NH, von Rosenvinge EC (2016)
Clostridium difficile. J Clin Microbiol 51(10):3421– Pseudomembranous colitis: not always Clostridium
3422. difficile. Cleve Clin J Med 83(5):361–366. https://
Savola KL, Baron EJ, Tompkins LS, Passaro DJ (2001)
Fecal leukocyte stain has diagnostic value for Usacheva EA, Jin JP, Peterson LR (2016) Host response
outpatients but not inpatients. J Clin Microbiol 39 to Clostridium difficile infection: diagnostics and
(1):266–269. detection. J Glob Antimicrob Resist 7:93–101.
Schutze GE, Willoughby RE (2013) Clostridium difficile van den Berg RJ, Schaap I, Templeton KE, Klaassen CH,
infection in infants and children. Pediatrics 131 Kuijper EJ (2007) Typing and subtyping of Clostrid-
(1):196–200. ium difficile isolates by using multiple-locus variable-
Senchyna F, Gaur RL, Gombar S, Truong CY, Schroeder number tandem-repeat analysis. J Clin Microbiol 45
LF, Banaei N (2017) Clostridium difficile PCR cycle (3):1024–1028.
threshold predicts free toxin. J Clin Microbiol 55 van Prehn J, Vandenbroucke-Grauls CM, van Beurden
(9):2651–2660. YH, van Houdt R, Vainio S, Ang CW (2015) Diag-
Sethi AK, Al-Nassir WN, Nerandzic MM, Bobulsky GS, nostic yield of repeat sampling with immunoassay,
Donskey CJ (2010) Persistence of skin contamination real-time PCR, and toxigenic culture for the detection
and environmental shedding of Clostridium difficile of toxigenic Clostridium difficile in an epidemic and a
during and after treatment of C. difficile infection. non-epidemic setting. Eur J Clin Microbiol Infect Dis
Infect Control Hosp Epidemiol 31(1):21–27. https:// 34(12):2325–2330. 015-2484-9
Stubbs SL, Brazier JS, O’Neill GL, Duerden BI (1999) Vrabie R, Kane S (2014) Noninvasive markers of disease
PCR targeted to the 16S-23S rRNA gene intergenic activity in inflammatory bowel disease. Gastroenterol
spacer region of Clostridium difficile and construction Hepatol 10(9):576–584
of a library consisting of 116 different PCR ribotypes. Wang T, Matukas L, Streutker CJ (2013) Histologic
J Clin Microbiol 37(2):461–463 findings and clinical characteristics in acutely symp-
Sundar S, Chan SY (2003) Cholestatic jaundice and tomatic ulcerative colitis patients with superimposed
pseudomembranous colitis following combination Clostridium difficile infection. Am J Clin Pathol 140
therapy with doxorubicin and docetaxel. Anti-Cancer (6):831–837.
Drugs 14(4):327–329. Wenisch C, Parschalk B, Hasenhundl M, Hirschl AM, Graninger W (1996) Comparison of vancomycin,
Surawicz CM, Brandt LJ, Binion DG, Ananthakrishnan teicoplanin, metronidazole, and fusidic acid for the
AN, Curry SR, Gilligan PH, McFarland LV, Mellow treatment of Clostridium difficile-associated diarrhea.
M, Zuckerbraun BS (2013) Guidelines for diagnosis, Clin Infect Dis 22(5):813–818
treatment, and prevention of Clostridium difficile Whitehead SJ, Shipman KE, Cooper M, Ford C, Gama R
infections. Am J Gastroenterol 108(4):478–498. (2014) Is there any value in measuring faecal quiz 499 calprotectin in Clostridium difficile positive faecal
Swale A, Miyajima F, Roberts P, Hall A, Little M, samples? J Med Microbiol 63(Pt 4):590–593. https://
Beadsworth MB, Beeching NJ, Kolamunnage-Dona
Ribotypes and New Virulent Strains Across

Jeanne Couturier, Kerrie Davies, Cécile Gateau,

and Frédéric Barbut

Abstract describe the new virulent C. difficile strains

Clostridium difficile is a major bacterial cause circulating in Europe, as well as other poten-
of post-antibiotic diarrhoea. The epidemiol- tial emerging strains described elsewhere.
ogy of C. difficile infections (CDI) has dra- Standardized typing methods and surveillance
matically changed since the early 2000s, with programmes are mandatory for a better under-
an increasing incidence and severity across standing and monitoring of CDI in Europe.
Europe. This trend is partly due to the emer-
gence and rapid worldwide spread of the Keywords
hypervirulent and epidemic PCR ribotype C. difficile · PCR ribotypes · Emerging
027. Profiles of patients with CDI have also strains · European epidemiology · Binary
evolved, with description of community- toxin
acquired (CA) infections in patients with no
traditional risk factors for CDI. However,
recent epidemiological studies indicated that 1 Introduction
some European countries have successfully
controlled the dissemination of the 027 clone Clostridium difficile is the main bacterial cause
whereas other countries recently reported the of hospital-acquired diarrhoea; it is responsible
emergence of other virulent or unusual strains. for 15–25% of post-antibiotic diarrhoea and for
The aims of this review are to summarize the virtually all cases of pseudomembranous colitis
current European CDI epidemiology and to (Bartlett and Gerding 2008). C. difficile infection

J. Couturier (*) · F. Barbut

National Reference Laboratory for C. difficile, Hôpital
Saint-Antoine, Paris, France
Université Paris Descartes, Faculté de Pharmacie, Paris,
K. Davies
Healthcare Associated Infections Research Group, Leeds C. Gateau
Teaching Hospitals NHS Trust and University of Leeds, National Reference Laboratory for C. difficile, Hôpital
Leeds, UK Saint-Antoine, Paris, France
e-mail: e-mail:

# Springer International Publishing AG 2018 45

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
46 J. Couturier et al.

(CDI) epidemiology has dramatically changed in 2 C. difficile Typing Methods

Europe since the beginning of the 2000s. The
incidence has increased over the last 10 years 2.1 PCR Ribotyping
from 2.45 cases per 10,000 patient-days in 2005
(Barbut et al. 2007), to 4.1 in 2008 (Bauer et al. PCR ribotyping is the reference method for
2011) and 7.0 in 2012–2013 (Davies et al. 2014). C. difficile typing in Europe. It relies on the
Nevertheless, the incidence of CDI varies widely presence of several alleles of the rRNA operon
across European countries from 0.7 to 28.7 per in the C. difficile genome. The length polymor-
10,000 patient-bed days per hospital. This trend phism of the intergenic spacer region between
is likely to result from a combination of several 16S and 23S rRNA genes results in RT-specific
factors, including the level of awareness of CDI patterns after genomic amplification and migra-
among physicians, the type of methods/algorithm tion (Bidet et al. 1999). PCR ribotyping was first
for CDI diagnosis implemented in each country, developed using agarose gel electrophoresis, but
and the global spread of the PCR ribotype the capillary gel-based electrophoresis method
(RT) 027 clone. A European study showed that has now been widely adopted. The latter enables
there is still a substantial underdiagnosis of CDI better standardization and easier comparison
coupled with large variations in testing policies between laboratories and is recommended as
among European countries (Davies et al. 2014). the reference technique in Europe (Fawley et al.
In Europe, the hypervirulent epidemic RT 2015).
027 strain (or REA type BI/NAP1/toxinotype III) Most European countries use a common
was first reported in England in 2005 (Smith 2005) nomenclature, but some laboratories developed
and has since rapidly spread in other European their own local databases. An online database
countries. RT 027 is characterized by an 18 bp containing internationally recognised capillary
deletion and a deletion at position 117 in tcdC electrophoresis RT profiles is available
gene, resulting in the inactivation of the toxin (WEBRIBO,, Indra
repressor TcdC and higher amounts of toxin pro- et al. 2008). However, there is no standardized
duction (Warny et al. 2005), although the role of protocol since several primer sets were published
tcdC mutation in toxin overproduction is currently (Stubbs et al. 1999; Bidet et al. 1999), some of
debated (Murray et al. 2009; Cartman et al. 2012). them enabling direct PCR ribotyping from stool
Moreover, epidemic 027 strains also produce an samples (Janezic et al. 2011). Harmonization of
additional toxin, the binary toxin, and are resistant the PCR ribotyping method and nomenclature is
to erythromycin and moxifloxacin, which may have therefore essential and needs to be improved in
conferred a selective advantage. The same combi- Europe, in order to detect emergence of new
nation of genetic and phenotypic features can be unreferenced RT in a timely manner.
found in other rare RT, such as RT 176 (Krutova
et al. 2015; Drabek et al. 2015). RT 027-related CDI
are associated with a higher rate of complications
and recurrences (Sundram et al. 2009). The RT 2.2 Other Methods Used
027 has disseminated throughout Europe, with a for C. difficile Typing
clear shift in its regional repartition from United
Kingdom and Ireland in 2008 (Bauer et al. 2011) Toxins A and B, which are considered as the
to Eastern Europe in 2012–2013 (Davies et al. main virulence factors of C. difficile (Pruitt and
2016b). Some countries have successfully con- Lacy 2012), are encoded by tcdA and tcdB genes
trolled its spread and decreased its prevalence located within a locus of pathogenicity (PaLoc).
(Hensgens et al. 2009; Fawley et al. 2016), while The PaLoc also contains tcdR (positive regulator
other were recently hit by large outbreaks (Bouza of toxin expression), tcdE (holin required for
et al. 2017). In addition, other virulent or unusual toxin secretion), and tcdC (potential negative
PCR ribotypes are emerging. regulator). The genetic polymorphism of the
Ribotypes and New Virulent Strains Across Europe 47

PaLoc can be explored by toxinotyping, which is transferability of data and the diversity of poten-
a PCR-restriction based method (Rupnik et al. tial applications, such as comparative genome
1998). Toxinotypes are defined according to analysis and lineages analysis, this method is
differences in the PaLoc compared to the refer- increasingly being used for C. difficile typing
ence strain VPI 10463 (nonvariant toxinotype 0). and could spread widely in the coming years
To date, 34 toxinotypes have been described (Knetsch 2013). WGS has successfully and rap-
(Rupnik and Janezic 2016; idly identified transmission of healthcare-
si/mf/tox/profile.html). Toxinotyping and PCR associated infection and could become a valuable
ribotyping are well correlated since most of the tool in routine clinical practice (Eyre et al. 2012).
strains in a given RT have similar changes in the
PaLoc and thus belong to a single toxinotype.
The analysis of 123 strains showed that in a few 3 Global Distribution
cases, PCR ribotyping can be more discrimina- of C. difficile PCR Ribotypes
tory than toxinotyping, whereas RT include sev- in Europe
eral toxinotypes less frequently (Rupnik et al.
2001). To avoid ambiguities, a revised The European C. difficile infection study (Bauer
toxinotyping nomenclature was recently et al. 2011) and the EUCLID study (Davies et al.
published (Rupnik and Janezic 2016). 2014, 2016b) are two major European surveys
PFGE (Pulsed-field gel electrophoresis) is a aiming at describing the epidemiology of CDI
genotype-based typing method developed in the including prevalence, diagnosis and RT
1980s and mostly used in North America. There distribution.
is good concordance between results of PFGE and The first pan-European study on C. difficile
PCR ribotyping (Bidet et al. 2000). PFGE has a was performed in 2008 in 106 laboratories from
higher discriminatory power than PCR ribotyping 34 countries (Bauer et al. 2011). The incidence of
(Killgore et al. 2008) but the interpretation of CDI and the RT distribution varied greatly
genetic relatedness is comparable between both between hospitals, as well as the density testing
typing methods. However, some strains are for CDI. The authors could differentiate 65 RT
non-typeable with this method, and degradation of among 389 C. difficile isolates. One of the main
genomic DNA can hinder the analysis findings of this study was that RT 027 was not
(Kristjánsson et al. 1994). PFGE is also very predominant in 2008, representing only 5% of
labour-intensive and the lack of standardisation the isolates. The most common RT were
makes inter-laboratory data comparison difficult. 014/020 (16%), 001, (9%), and 078 (8%). Some
The discriminatory power of PCR ribotyping RT seemed to spread regionally, such as RT
is not sufficient to prove the nosocomial trans- 106 mostly described in UK and Ireland.
mission of a strain, particularly when a RT is The EUCLID study (European, multicentre,
endemic at a regional or national level. In that prospective, biannual, point-prevalence study of
case, another more discriminant typing method CDI in hospitalized patients with diarrhoea) was
has to be used, such as multilocus variable- conducted in 2012–2013 and included 482 hospitals
number tandem repeat (VNTR) analysis from 19 European countries (Davies et al. 2016b).
(MLVA). MLVA relies on the variability of the The objectives were to measure the underdiagnosis
VNTR at different loci. The genetic relatedness of CDI and to assess the diversity of RT repartition
of isolates is appreciated through the sum of in Europe. During two sampling days (one in win-
tandem repeat number differences (STRD) ter and one in summer), participating hospitals sent
(Marsh et al. 2006). every diarrhoeal stool sample, irrespective of the
Whole genome sequencing (WGS) can distin- request to test for C. difficile by the physician, to a
guish between strains at the single nucleotide national coordinating laboratory. The RT diversity
level, highly increasing the discriminatory was much higher than in the previous study, with
power over other typing schemes. Given the 125 RT identified among 1196 isolates.
48 J. Couturier et al.

Interestingly, the most common RT was (Alcalá et al. 2012). Among 42 toxigenic strains,
027 (19%), highlighting the rapid spread of this RT 014/020, 001 and 078/126 were the most preva-
strain at a global scale. An inverse correlation was lent (20.5%, 18.2% and 18.2% respectively). RT
noted between the rate of testing and prevalence of 027 was not found.
ribotype 027 across north, south, east, and west The characterization of 498 clinical isolates
quadrants of Europe, which suggests that increased from 20 hospitals in Portugal showed that RT
awareness of CDI and use of optimum testing 027 was predominant with 18.5% of all the
methods and policies can reduce the dissemination strains, and 19.6% of HA-associated CDI. RT
of epidemic strains (Davies et al. 2014). The com- 014 was the second most frequent overall
parison with the 2008 data indicated a shift in the (9.4%) and the most frequent among CA-CDI
frequency of RT 027 from UK and Ireland (12%). The prevalence of RT 126 and 078 was
(decreasing prevalence) to Eastern Europe low (3.8% and 2.8% respectively) (Santos et al.
countries (increasing prevalence). RT 001/072 2016). The authors described a great heterogene-
(11%) and 014/020 (10%) were the second and ity of the RT distribution through the country
third most prevalent RT, consistent with the 2008 with a higher diversity in the north, where RT
results, however the prevalence of RT 078 dropped 027 was not predominant.
from 8% in 2008 to 3% in 2012–2013. The distri- The geographic distribution of C. difficile
bution of causative RT was country-specific as genotypes in Germany was assessed using
shown in the Fig. 1 (Davies et al. 2016b). 393 isolates sent to the national advisory labora-
Besides these two large epidemiological stud- tory for diagnostic reason between 2011 and
ies, several other recent European studies 2013 (von Müller et al. 2015). The typing method
analysed RT distribution at a national level. The used was surface-layer protein A sequence typ-
results of these national studies are summarized ing, with strain assignment to RT for better com-
in the Table 1. parison with international data. RT 001 (35%)
A multicentre study characterized 3333 toxi- and 078 (8%) were prevalent nationwide; RT
genic strains isolated between 2010 and 2015 in 027 (26%) and 014/066 (9%) were detected in
110 Belgian hospitals (Neely et al. 2017). RT almost all regions.
027 (4.2%) and 078 (7.0%) were associated In France, a multicentre study conducted in
with a higher rate of complications (unadjusted 2009 in 78 healthcare facilities showed that the
data) and higher levels of in-vitro toxin produc- most prevalent RT were 014/020/077 (18.7%),
tion from cultured isolates. followed by 078/126 (12.1%) (Eckert et al.
A study compared epidemiological data for 2013). The prevalence of RT 027 strains remained
community-associated (CA)-CDI and healthcare- low (3.1%), and they were only isolated in North-
associated (HA)-CDI in 113 laboratories across ern France, where RT 027 emergence was first
England between 2011 and 2013 (Fawley et al. described in 2006 (Coignard et al. 2006; Birgand
2016). A total of 703 C. difficile toxin-positive et al. 2010). These results are consistent with the
faecal samples from CA-CDI cases were analysed more recent LuCID (Longitudinal European Clos-
and the results were compared to HA-CDI data tridium difficile Infection Diagnosis) surveillance
(n ¼ 10,754) obtained from the C. difficile study (Davies et al. 2016a), during which RT
Ribotyping Network. RT distribution was similar 014/020/077 and 078/126 were the most prevalent
in cases of CA- and HA-CDI, but RT 002 was more in France (21.9% and 9.5% respectively) (Eckert
likely to cause CA-CDI, while RT 027 was more et al. 2015).
often associated with HA-CDI. In conclusion, RT 014/020 and 001/072 are
In Spain, Alcalá et al. performed C. difficile endemic in almost all European countries while
cultures on 807 unformed stool specimens sent to there is a national or regional specificity for other
118 Spanish microbiology laboratories on a single RT. Moreover, the RT diversity is significantly
day, regardless of the prescription by the clinician increasing across Europe.
Ribotypes and New Virulent Strains Across Europe 49

Fig. 1 Geographical distribution of C. difficile PCR proportion of the most common ribotypes per country and
ribotypes, by participating European country, EUCLID the number in the centre of the charts is the number of
2012–2013 and 2013 (n ¼ 1196) (Reproduced with per- typed isolates in the country
mission from Davies et al. 2016b) Pie charts show the

The first cases of RT 176-associated CDI were

4 Emerging PCR Ribotypes
described in 2008 in Poland (Obuch-Woszczatyński
et al. 2014), in 2009 in the Czech Republic (Nyč
4.1 PCR Ribotype 176
et al. 2011) and in 2015 in Croatia (Rupnik et al.
2016). The first RT 176-related outbreak was
RT 176 strains are closely related to RT 027 (Stabler
recently described in France (Couturier et al.
et al. 2006). They belong to toxinotype III, produce
2017). Four strains isolated in two geographically
the binary toxin and bear a deletion at position
close hospitals, previously identified as RT 027 with
117 of the tcdC gene, leading to a potential RT
the agarose gel method, were reassigned as RT
027 misidentification with commonly used molecu-
176 by capillary gel-based electrophoresis. MLVA
lar assays such as Xpert® C. difficile (Cepheid).
analysis showed that those four strains formed a
Moreover, their similar banding pattern (only one
clonal complex (STRD 2), and were genetically
band difference) after gel electrophoresis can be
related to RT 027 strains (STRD 10).
confusing for RT attribution (Valiente et al. 2012).
50 J. Couturier et al.

Table 1 National epidemiological studies on Clostridium difficile PCR ribotypes repartition

N PCR ribotyping
Country strains method Most prevalent RT (%) References
Belgium 3333 Agarose gel 014 (11.6), 020 (8.5), 002 (7.6), 078 (7.0), Neely et al.
electrophoresis 027 (4.2), 005 (3.5), 106 (3.4) (2017)
United 11,457 Agarose gel 015 (10.2), 002 (9.1), 014 (9.1), 078 (8.0), Fawley et al.
Kingdom electrophoresis 005 (7.4) and 027 (6.4) (2016)
Spain 42 Agarose gel 014/020 (20.5), 001 (18.2), 078/126 (18.2) Alcalá et al.
electrophoresis (2012)
Portugal 498 Capillary 027 (18.5), 014 (9.4), 020 (5.6), 017 (5.2) Santos et al.
electrophoresis (2016)
Germany 393 slpAST with 001 (35), 027 (26), 014/066 (9), 078 (8) von Müller et al.
assignment to RT (2015)
France 224 Agarose gel 014/020/077 (18.7), 078/126 (12.1), 015 (8.5), Eckert et al.
electrophoresis 002 (8), 005 (4.9) (2013)
Czech 774 Capillary 176 (29), 001 (24) Krutova et al.
Republic electrophoresis (2016)
slpAST surface-layer protein A sequence typing

The results of the EUCLID study showed a HA-infections in the elderly, 078 strains are
regional specificity of RT 176, isolated mostly in more frequently associated with CA-infections
the Czech Republic where it accounted for 38% in a younger population. CA-CDI due to
of the strains (Davies et al. 2016b). In 2014, a 078 strains were also described in England
study among 18 Czech hospitals showed that (Fawley et al. 2016) (see “Clostridium difficile
29% of C. difficile isolates belonged to RT infection in the community” below). Finally, RT
176, and 24% to RT 001 (Krutova et al. 2016, 078 strains are frequently resistant to
Table 1). Further typing analysis by MLVA, fluoroquinolones and erythromycin, partly
indicated that both RT formed clonal complexes explaining this epidemiological success (Baldan
in several hospitals, suggesting a rapid spread of et al. 2015).
these clones at a national level.
These results suggest a rapid nosocomial
spread of RT 176 strains through Europe,
4.3 PCR Ribotype 126
stressing the need for a common data base for
PCR ribotyping.
RT 078 and 126 are highly related: they share
similar banding patterns in agarose gel electro-
phoresis method, and can only be differentiated
4.2 PCR Ribotype 078 with the capillary gel-based electrophoresis.
Consequently, they are often reported together
RT 078 strains can produce toxins A and B, as as RT 078/126. Like RT 078 strains, RT
well as the binary toxin and belong to toxinotype 126 strains belong to toxinotype V and are con-
V. They are characterized by a 39 bp deletion in sidered as “hypervirulent” (Knetsch et al. 2011).
tcdC. RT 078 was reported as predominant in They also produce the binary toxin and are
Greece in 2005 (Barbut et al. 2007), and was characterized by a 39 bp deletion in tcdC.
the third most common RT in the 2008 The prevalence of RT 126 strains in animals
European study (Bauer et al. 2011). A recent in Germany is high, suggesting the potential zoo-
study showed that RT 078 strains co-circulate notic spread of this RT (Schneeberg et al. 2013).
with the hypervirulent 027 strains in Southern MLVA analysis showed that most of those
France (Cassir et al. 2017). While 027 strains strains are genetically related to RT 078 strains
are mostly responsible for outbreaks of (STRD 10), and some of them belong to the
Ribotypes and New Virulent Strains Across Europe 51

same clonal complex (STRD 2). RT 126 strains Xpert® C. difficile assay, which detects binary
are also frequently resistant to antibiotics, includ- toxin genes, will possibly enable a better identi-
ing erythromycin, moxifloxacin and tetracyclin fication of toxinotype XI strains.
(Álvarez-Pérez et al. 2017).

4.5 PCR Ribotype 018

4.4 PCR Ribotype 033/Toxinotype XI
RT 018 has recently been reported as an
PCR ribotype 033 strains belong to toxinotype emerging RT responsible for outbreaks in Italy,
XI. They are characterized by the absence of where RT 126 was previously predominant
TcdA and TcdB expression and therefore cannot (Spigaglia et al. 2010). The EUCLID study
be detected by EIA (enzyme immunoassay) (Davies et al. 2016b) showed that prevalence of
methods for toxins. These strains were first RT 018 was high in Italy (22%), as opposed to
described in 2001 (Rupnik et al. 2001). In 2014, other European countries. In addition, Baldan
six symptomatic CDI cases due to toxinotype XI et al. characterized 312 C. difficile isolates from
strains were reported by the French National a large Italian teaching hospital between 2009
Reference Laboratory for C. difficile (Eckert and 2013, and observed that RT 018 was pre-
et al. 2014). In four cases, the patient was suc- dominant. After epidemiological investigation of
cessfully treated by oral metronidazole. These the outbreaks, RT 018 represented 42% of index
strains were characterized by PCR ribotyping, CDI cases and virtually all secondary cases (due
amplification of tcdA, tcdB, cdtA and cdtB to nosocomial transmission). The transmission
genes and toxinotyping. The six strains were index (number of secondary cases divided by
defined as RT 033 (or 033-like) and were nega- number of index cases) of RT 018 was signifi-
tive for TcdA and TcdB. The binary toxin genes cantly higher than that of RT 078 (0.640 and
were present and a 39 bp deletion was identified 0.0606, respectively) (Baldan et al. 2015).
in the tcdC gene. The six strains were Another study comparing RT 018, RT 126 and
characterized by major deletions of the 50 region RT 078 demonstrated that RT 018 strains pro-
of the PaLoc including tcdB, tcdE and tcdR; only duced higher levels of toxins, showed increased
a remnant part of tcdA (A2 and A3 fragments) adhesion to cells and became endemic in a short
and tcdC could be amplified. time (Barbanti and Spigaglia 2016). Moreover,
The pathogenicity of toxinotype XI strains RT 018 strains were all multidrug resistant (resis-
remains controversial. Studies on the role of the tance to erythromycin, clindamycin and
binary toxin as a virulence factor in animal moxifloxacin). Together, these results suggest
models gave contradictory results. In the rabbit that RT 018 strains have phenotypic traits con-
ileal loop model, an enterotoxic response was ferring an adaptive advantage and are able to
observed after inoculation of supernatants from spread widely. RT 018 strains were indeed
culture of ABCDT+ strains. However, despite reported in Southern Europe (Spain, Austria and
colonization, no symptoms occurred in Slovenia) and are associated with a higher rate of
clindamycin-treated hamsters challenged with complicated infections (Bauer et al. 2011).
these strains (Geric et al. 2006). Although the
prevalence of ABCDT+ strains in Europe
seems rather low (Barbut et al. 2007; Bauer 4.6 PCR Ribotype 017
et al. 2011), surveillance of this unusual strains
is required. Indeed, the atypical genomic organi- RT 017 strains belong to toxinotype VIII and are
zation of the PaLoc can lead to a false negative part of C. difficile clade 4; they lack toxin A
diagnosis, more particularly when methods rely- production and binary toxin genes (Cairns et al.
ing on the presence of toxin A and/or toxin B 2012). The clinical relevance and the prevalence
only are used. However, the increasing use of the of this clone has been unclear for many years,
52 J. Couturier et al.

since it was mainly found in asymptomatic spread: among 20 strains, 13 were isolated in
infants (Depitre et al. 1993; Kato et al. 1998). the United Kingdom and 5 in Ireland (Bauer
However, it has now been established that RT et al. 2011). In a Southern England healthcare
017 strains are predominant in Asian countries facility, 38% of C. difficile isolates (n ¼ 97) were
such as Korea, China and Japan (Collins et al. identified as RT 106, the second most prevalent
2013), and that they have spread worldwide. RT RT after 027 (45%) (Sundram et al. 2009).
017-related outbreaks have been reported in Almost all of these RT 106 strains were resistant
England (Cairns et al. 2015), The Netherlands to ciprofloxacin and erythromycin. Moreover, in
(Kuijper et al. 2001), Poland (Pituch et al. the Belgian multicentre study (Neely et al. 2017),
2001), and Ireland (Drudy et al. 2007). RT recurrences were more frequent with RT
017-related CA-CDI appear to be more likely to 106-related CDI.
affect younger patients (Fawley et al. 2016). Other data reported the emergence of RT
Severe RT 017-related CDI have been described 001 strains with reduced susceptibility to metro-
in Germany, although RT 027 was the most nidazole, raising concerns about the potential
prevalent strain in this study (Arvand et al. 2009). spread of these strains due to this selective
advantage (Baines et al. 2008). In Southern
Germany, the prevalence of RT 001 strains
exhibiting resistance to erythromycin, ciproflox-
4.7 Other Emerging PCR Ribotypes
acin and moxifloxacin is high in both in- and
out-patients (Borgmann et al. 2008; Arvand
RT 244 strains belong to the same hypervirulent
et al. 2009).
clade as RT 027 (clade 2) (Lim et al. 2014). They
Given their pathogenic and epidemic poten-
produce the binary toxin and bear a single nucle-
tial, the emergence of these RT should be closely
otide deletion at position 117 in tcdC. Severe
followed in European countries.
CA-CDI and outbreaks due to RT 244 strains
The genetic and epidemiological features of
were recently reported in Australia and
the emerging RT described above are
New Zealand, where it was previously uncom-
summarized in the Table 2.
mon (De Almeida et al. 2013; Eyre et al. 2015).
Eyre et al. showed that a strain isolated in a
patient recently returned from Australia to the
UK was phylogenetically related to their out- 4.8 Emerging Strains with a AþB-
break, highlighting the potential rapid spread of CDT- Unusual Profile
RT 244 via international travel.
The previously quoted French multicentre Recently, three clinical strains with an atypical
survey showed that among 224 toxigenic strains, PaLoc structure were described in France
19 (8.5%) belonged to RT 015 which was the (Monot et al. 2015), including the first variant
third most frequent RT (Eckert et al. 2013). strain producing only toxin A (A+BCDT).
Fawley et al. showed that RT 015 was also pre- WGS analysis of this strain showed that its
dominant in England (Fawley et al. 2016). PaLoc only contained tcdA and tcdR. None of
Although RT 015 accounted for only 2% of the the three strains belonged to any of the most
strains analysed in the EUCLID study, it seems frequent RT. Moreover, the authors described
that RT 015 strains can spread and become pre- variability in the sequence of the toxin genes,
dominant at a national scale. which may lead to potential false negative
RT 106 strains represented 5% of all toxigenic results with the most commonly used diagnostic
isolates in the 2008 hospital-based European methods (immunoenzymatic or molecular
study, but their distribution showed a regional assays).
Ribotypes and New Virulent Strains Across Europe 53

Table 2 Characteristics of currently circulating and emerging PCR ribotypes in Europe

Toxins Binary Deletion
RT Toxinotype A and B toxin in tcdC Main circulation area
027 III +/+ + 18 bp/ Europe, mostly Eastern Europe Davies et al. (2016b)
176 III +/+ + 18 bp/ Poland, Czech Republic Nyč et al. (2011), Obuch-
Δ117 Woszczatyński et al. (2014)
078 V +/+ + 39 bp/ Community-onset infections Eckert et al. (2011), Fawley et al.
A117T (2016)
126 V +/+ + 39 bp/ Eckert et al. (2011)
033 XIa/XIb / + 39 bp Low prevalence in Europe Eckert et al. (2014)
018 XIX +/+  ND Italy Spigaglia et al. (2010), Rupnik and Janezic (2016)
017 VIII /+  ND Asia Collins et al. (2013), Ireland Drudy et al. (2007), England
(Cairns et al. (2015), The Netherlands Kuijper et al. (2001),
Poland Pituch et al. (2001), Germany
244 IXb +/+ + ND/ Australia Lim et al. (2014), Rupnik and Janezic (2016)
015 NA +/+  18 bp France Eckert et al. (2013)
or ND
106 NA +/+  18 bp United Kingdom, Ireland Bauer et al. (2011)
or ND
001 XXIX +/+  ND Germany, multidrug resistant strains Borgmann et al. (2008),
Rupnik and Janezic (2016)
ND not deleted, NA not available

(Debast et al. 2009). In Canada, RT 078 epidemic

5 C. difficile Infection
strains (identified as pulsotype NAP7 by PFGE)
in the Community
were found in vegetables from grocery stores
(Metcalf et al. 2010). RT 078 has also been
The epidemiology of CA-CDI is poorly known,
described in the environment; it was the most fre-
since C. difficile testing is rarely requested in
quently isolated RT in wastewater treatment plants
stool samples from community patients. How-
in Switzerland (Romano et al. 2012). RT 078 was
ever, recent data suggest that the incidence of
the commonest (19.0%) in 42 CA-CDI cases in a
CA-CDI is rising (Chitnis et al. 2013). In addi-
prospective study conducted in Scotland, followed
tion, CDI were recently described among young
by RT 014/020 (16.7%), 015 (14.3%) and
patients from community settings without the
001 (11.9%) (Taori et al. 2014). However, in a US
traditional risk factors (antibiotic exposure,
study of 984 CA-CDI cases, NAP1/RT 027 was the
recent hospitalization, co-morbidities) (Wilcox
most frequent strain isolated (21.7%), while less
et al. 2008; Gupta and Khanna 2014).
than 7% of the isolates belonged to NAP7/RT
Fawley et al. showed that RT 002, 020 and
078 (Chitnis et al. 2013). In 2011, population- and
056 were largely responsible for CA-CDI, whereas
laboratory-based surveillance for CDI was
RT 027 was most associated with HA-CDI (Fawley
conducted in 10 US areas (Lessa et al. 2015). A
et al. 2016). RT 078 strains have been reported in
total of 1364 strains were characterized. The most
animals in the Netherlands (Goorhuis et al. 2008),
common strains were NAP1/RT 027 (18.8% of
and by using MLVA analysis, Debast et al. showed
CA-CDI and 30.7% of HA-CDI), NAP4/RT
that RT 078 strains found in animals and in humans
020 (11.4% and 10.3%) and NAP11/RT
were genetically highly related, suggesting a
106 (10.7% and 10.0%). Less than 4% of the strains
foodborne interspecies transmission of C. difficile
in both settings belonged to NAP7/RT 078. These
54 J. Couturier et al.

results show a large overlapping of the RT distribu- emerging type. Anaerobe 42:123–129. https://doi.
tion in HA- and CA-CDI, suggesting the existence org/10.1016/j.anaerobe.2016.10.003
Barbut F, Mastrantonio P, Delmee M, Brazier J,
of common reservoirs and multiple transmission Kuijper E, Poxton I (2007) Prospective study of Clos-
routes between community and hospital settings. tridium difficile infections in Europe with phenotypic
and genotypic characterisation of the isolates. Clin
Microbiol Infect 13:1048–1057
Bartlett JG, Gerding DN (2008) Clinical recognition and
6 Conclusion diagnosis of Clostridium difficile infection. Clin Infect
Dis 46:S12–S18.
Bauer MP, Notermans DW, van Benthem BH, Brazier JS,
In conclusion, there is a large diversity of RT Wilcox MH, Rupnik M, Monnet DL, van Dissel JT,
across Europe, although some specific RT are Kuijper EJ (2011) Clostridium difficile infection in
able to disseminate at a regional or national Europe: a hospital-based survey. Lancet 377:63–73.
level. A national and European clinical surveil-
Bidet P, Barbut F, Lalande V, Burghoffer B, Petit J-C
lance system, associated with microbiological (1999) Development of a new PCR-ribotyping method
characterization of strains, is essential in order for Clostridium difficile based on ribosomal RNA gene
to monitor the constantly changing epidemiology sequencing. FEMS Microbiol Lett 175:261–266.
of CDI. A common European data base of the
Bidet P, Lalande V, Salauze B, Burghoffer B, Avesani V,
circulating RT would be very helpful to detect Delmee M, Rossier A, Barbut F, Petit JC (2000) Com-
emergence of new virulent clones in a timely parison of PCR-ribotyping, arbitrarily primed PCR,
manner. and pulsed-field gel electrophoresis for typing Clos-
tridium difficile. J Clin Microbiol 38:2484–2487
Birgand G, Blanckaert K, Carbonne A, Coignard B,
Barbut F, Eckert C, Grandbastien B, Kadi Z,
Astagneau P (2010) Investigation of a large outbreak
References of Clostridium difficile PCR-ribotype 027 infections in
northern France, 2006–2007 and associated clusters in
Alcalá L, Martin A, Marin M, Sánchez-Somolinos M, 2008–2009. Eurosurveillance 15(25):8–13
Catalán P, Peláez T, Bouza E (2012) The undiagnosed Borgmann S, Kist M, Jakobiak T, Reil M, Scholz E, von
cases of Clostridium difficile infection in a whole Eichel-Streiber C, Gruber H, Brazier JS, Schulte B
nation: where is the problem? Clin Microbiol Infect (2008) Increased number of Clostridium difficile
18:E204–E213. infections and prevalence of Clostridium difficile
2012.03883.x PCR ribotype 001 in southern Germany.
Álvarez-Pérez S, Blanco JL, Harmanus C, Kuijper E, Eurosurveillance 13(49):11–15
Garcı́a ME (2017) Subtyping and antimicrobial sus- Bouza E, Alcalá L, Marı́n M, Valerio M, Reigadas E,
ceptibility of Clostridium difficile PCR ribotype Muñoz P, Vecchio MG-D, Egea V de (2017) An
078/126 isolates of human and animal origin. Vet outbreak of Clostridium difficile PCR ribotype
Microbiol 199:15–22. 027 in Spain: risk factors for recurrence and a novel
vetmic.2016.12.001 treatment strategy. Eur J Clin Microbiol Infect Dis
Arvand M, Hauri AM, Zaiss NH, Witte W, Bettge-Weller 36(10):1777–1786. doi:
G (2009) Clostridium difficile ribotypes 001, 017, and s10096-017-2991-y
027 are associated with lethal C. difficile infection in Cairns MD, Stabler RA, Shetty N, Wren BW (2012) The
Hesse, Germany. Eurosurveillance 14(45):27–30 continually evolving Clostridium difficile species.
Baines SD, O’Connor R, Freeman J, Fawley WN, Future Microbiol 7:945–957.
Harmanus C, Mastrantonio P, Kuijper EJ, Wilcox fmb.12.73
MH (2008) Emergence of reduced susceptibility to Cairns MD, Preston MD, Lawley TD, Clark TG, Stabler
metronidazole in Clostridium difficile. J Antimicrob RA, Wren BW (2015) Genomic epidemiology of a
Chemother 62:1046–1052. protracted hospital outbreak caused by a toxin
jac/dkn313 A-negative Clostridium difficile sublineage PCR
Baldan R, Trovato A, Bianchini V, Biancardi A, ribotype 017 strain in London, England. J Clin
Cichero P, Mazzotti M, Nizzero P, Moro M, Ossi C, Microbiol 53:3141–3147.
Scarpellini P, Cirillo DM (2015) Clostridium difficile JCM.00648-15
PCR ribotype 018, a successful epidemic genotype. J Cartman ST, Kelly ML, Heeg D, Heap JT, Minton NP
Clin Microbiol 53:2575–2580. (2012) Precise manipulation of the Clostridium diffi-
1128/JCM.00533-15 cile chromosome reveals a lack of association between
Barbanti F, Spigaglia P (2016) Characterization of Clos- the tcdC genotype and toxin production. Appl Environ
tridium difficile PCR-ribotype 018: a problematic
Ribotypes and New Virulent Strains Across Europe 55

Microbiol 78:4683–4690. De Almeida MN, Heffernan H, Dervan A, Bakker S,

AEM.00249-12 Freeman JT, Bhally H, Taylor SL, Riley TV, Roberts
Cassir N, Fahsi N, Durand G, Lagier J-C, Raoult D, SA (2013) Severe Clostridium difficile infection in
Fournier P-E (2017) Emergence of Clostridium diffi- New Zealand associated with an emerging strain,
cile tcdC variant 078 in Marseille, France. Eur J Clin PCR-ribotype 244. N Z Med J 126:9–14
Microbiol Infect Dis:1–4. Debast SB, LAMG VL, Goorhuis A, Harmanus C,
s10096-017-3022-8 Kuijper EJ, Bergwerff AA (2009) Clostridium difficile
Chitnis AS, Holzbauer SM, Belflower RM, Winston LG, PCR ribotype 078 toxinotype V found in diarrhoeal
Bamberg WM, Lyons C, Farley MM, Dumyati GK, pigs identical to isolates from affected humans. Envi-
Wilson LE, Beldavs ZG, Dunn JR, Gould LH, ron Microbiol 11:505–511.
MacCannell DR, Gerding DN, McDonald LC, Lessa 1462-2920.2008.01790.x
FC (2013) Epidemiology of community-associated Depitre C, Delmee M, Avesani V, Roels A, L’haridon R,
Clostridium difficile infection, 2009 through 2011. Popoff M, Corthier G (1993) Serogroup F strains of
JAMA Intern Med 173:1359–1367. Clostridium difficile produce toxin B but not toxin A. J
10.1001/jamainternmed.2013.7056 Med Microbiol 38:434–441.
Coignard B, Barbut F, Blanckaert K, Thiolet JM, Poujol I, 00222615-38-6-434
Carbonne A, Petit JC, Desenclos JC (2006) Emer- Drabek J, Nyc O, Krutova M, Stovicek J, Matejkova J,
gence of Clostridium difficile toxinotype III, Keil R (2015) Clinical features and characteristics of
PCR-ribotype 027-associated disease, France, 2006. Clostridium difficile PCR-ribotype 176 infection:
Euro Surveill Bull Eur Sur Mal Transm Eur Commun results from a 1-year university hospital internal
Dis Bull 11:E060914.1 ward study. Ann Clin Microbiol Antimicrob. https://
Collins DA, Hawkey PM, Riley TV (2013) Epidemiology
of Clostridium difficile infection in Asia. Antimicrob Drudy D, Harnedy N, Fanning S, Hannan M, Kyne L
Resist Infect Control 2:21. (2007) Emergence and control of fluoroquinolone-
2047-2994-2-21 resistant, toxin A–negative, toxin B–positive Clostrid-
Couturier J, Eckert C, Barbut F (2017) Spatio-temporal ium difficile. Infect Control Am Hosp Epidemiol
variability of the epidemic 027 Clostridium difficile 28:932–940.
strains in France based on MLVA typing. Anaerobe. Eckert C, Tessier C, Chassaing D, Barbut F (2011) Is deletion at 117 of the tcdC gene specific of
Davies KA, Longshaw CM, Davis GL, Bouza E, PCR-ribotype 027 strains?
Barbut F, Barna Z, Delmée M, Fitzpatrick F, Eckert C, Coignard B, Hebert M, Tarnaud C, Tessier C,
Ivanova K, Kuijper E, Macovei IS, Mentula S, Lemire A, Burghoffer B, Noel D, Barbut F (2013)
Mastrantonio P, von Müller L, Oleastro M, Clinical and microbiological features of Clostridium
Petinaki E, Pituch H, Norén T, Nováková E, Nyč O, difficile infections in France: the ICD-RAISIN 2009
Rupnik M, Schmid D, Wilcox MH (2014) Underdiag- national survey. Médecine Mal Infect 43:67–74.
nosis of Clostridium difficile across Europe: the
European, multicentre, prospective, biannual, point- Eckert C, Emirian A, Le Monnier A, Cathala L, De
prevalence study of Clostridium difficile infection in Montclos H, Goret J, Berger P, Petit A, De
hospitalised patients with diarrhoea (EUCLID). Lan- Chevigny A, Jean-Pierre H, Nebbad B, Camiade S,
cet Infect Dis 14:1208–1219. Meckenstock R, Lalande V, Marchandin H, Barbut F
S1473-3099(14)70991-0 (2014) Prevalence and pathogenicity of binary toxin–
Davies K, Davis G, Barbut F, Eckert C, Petrosillo N, positive Clostridium difficile strains that do not pro-
Wilcox MH (2016a) Variability in testing policies duce toxins A and B. New Microbes New Infect
and impact on reported Clostridium difficile infection 3:12–17.
rates: results from the pilot Longitudinal European Eckert C, Bildan, M-A, Quach, S, Youssouf, A, Barbut, F,
Clostridium difficile Infection Diagnosis surveillance C. difficile study group (2015) Caractérisation des
study (LuCID). Eur J Clin Microbiol Infect Dis souches de Clostridium difficile circulant en France
35:1949–1956. en 2014 et 2015: résultats d’une étude multicentrique
2746-1 (ref 350) – 35ème RICAI
Davies KA, Ashwin H, Longshaw CM, Burns DA, Davis Eyre DW, Golubchik T, Gordon NC, Bowden R, Piazza P,
GL, Wilcox MH, on behalf of the EUCLID study Batty EM, CLC I, Wilson DJ, Didelot X, O’Connor L,
group (2016b) Diversity of Clostridium difficile PCR Lay R, Buck D, Kearns AM, Shaw A, Paul J, Wilcox
ribotypes in Europe: results from the European, MH, Donnelly PJ, Peto TEA, Walker AS, Crook DW
multicentre, prospective, biannual, point-prevalence (2012) A pilot study of rapid benchtop sequencing of
study of Clostridium difficile infection in hospitalised Staphylococcus aureus and Clostridium difficile
patients with diarrhoea (EUCLID), 2012 and 2013. for outbreak detection and surveillance. BMJ Open
Euro Surveill 21(29) 2(3):e001124.
7917.ES.2016.21.29.30294 001124
56 J. Couturier et al.

Eyre DW, Tracey L, Elliott B, Slimings C, Huntington B-positive Clostridium difficile by PCR. J Clin
PG, Stuart RL, Korman TM, Kotsiou G, McCann R, Microbiol 36:2178–2182
Griffiths D, Fawley WN, Armstrong P, Dingle KE, Killgore G, Thompson A, Johnson S, Brazier J, Kuijper E,
Walker AS, Peto TE, Crook DW, Wilcox MH, Riley Pepin J, Frost EH, Savelkoul P, Nicholson B, van den
TV (2015) Emergence and spread of predominantly Berg RJ, Kato H, Sambol SP, Zukowski W, Woods C,
community-onset Clostridium difficile PCR ribotype Limbago B, Gerding DN, McDonald LC (2008) Com-
244 infection in Australia, 2010 to 2012. Euro Surveill parison of seven techniques for typing international
Bull Eur Sur Mal Transm Eur Commun Dis Bull epidemic strains of Clostridium difficile: restriction
20:21059 endonuclease analysis, pulsed-field gel electrophore-
Fawley WN, Knetsch CW, MacCannell DR, Harmanus C, sis, PCR-ribotyping, multilocus sequence typing,
Du T, Mulvey MR, Paulick A, Anderson L, Kuijper multilocus variable-number tandem-repeat analysis,
EJ, Wilcox MH (2015) Development and validation of amplified fragment length polymorphism, and surface
an internationally-standardized, high-resolution capil- layer protein A gene sequence typing. J Clin
lary gel-based electrophoresis PCR-ribotyping proto- Microbiol 46:431–437.
col for Clostridium difficile. PLoS One. https://doi. 01484-07
org/10.1371/journal.pone.0118150 Knetsch EC for DP and C (ECDC)-HCU-E editorial
Fawley WN, Davies KA, Morris T, Parnell P, Howe R, (2013) Current application and future perspectives of
Wilcox MH, Clostridium difficile Ribotyping Net- molecular typing methods to study Clostridium diffi-
work (CDRN) Working Group (2016) Enhanced sur- cile infections.
veillance of Clostridium difficile infection occurring ViewArticle.aspx?ArticleId¼20381. Accessed 2 Dec
outside hospital, England, 2011 to 2013. Euro 2014
Surveill Bull Eur Sur Mal Transm Eur Commun Dis Knetsch CW, Hensgens MPM, Harmanus C, van der Bijl
Bull. MW, Savelkoul PHM, Kuijper EJ, Corver J, van
29.30295 Leeuwen HC (2011) Genetic markers for Clostridium
Geric B, Carman RJ, Rupnik M, Genheimer CW, Sambol difficile lineages linked to hypervirulence. Microbiol-
SP, Lyerly DM, Gerding DN, Johnson S (2006) Binary ogy 157:3113–3123.
toxin–producing, large clostridial toxin–negative 051953-0
Clostridium difficile strains are enterotoxic but do Kristjánsson M, Samore MH, Gerding DN, De Girolami
not cause disease in hamsters. J Infect Dis PC, Bettin KM, Karchmer AW, Arbeit RD (1994)
193:1143–1150. Comparison of restriction endonuclease analysis,
Goorhuis A, Bakker D, Corver J, Debast SB, Harmanus C, ribotyping, and pulsed-field gel electrophoresis for
Notermans DW, Bergwerff AA, Dekker FW, Kuijper molecular differentiation of Clostridium difficile
EJ (2008) Emergence of Clostridium difficile infection strains. J Clin Microbiol 32:1963–1969
due to a new hypervirulent strain, polymerase chain Krutova M, Matejkova J, Tkadlec J, Nyc O (2015) Anti-
reaction ribotype 078. Clin Infect Dis Off Publ Infect biotic profiling of Clostridium difficile ribotype 176 –
Dis Soc Am 47:1162–1170. a multidrug resistant relative to C. difficile ribotype
592257 027. Anaerobe 36:88–90.
Gupta A, Khanna S (2014) Community-acquired Clos- anaerobe.2015.07.009
tridium difficile infection: an increasing public health Krutova M, Matejkova J, Kuijper EJ, Drevinek P, Nyc O,
threat. Infect Drug Resist 7:63–72. Czech Clostridium difficile study group (2016)
2147/IDR.S46780 Clostridium difficile PCR ribotypes 001 and 176 –
Hensgens MP, Goorhuis A, Notermans DW, van Benthem the common denominator of C. difficile infection epi-
BH, Kuijper EJ (2009) Decrease of hypervirulent demiology in the Czech Republic, 2014. Euro
Clostridium difficile PCR ribotype 027 in the Surveill.
Netherlands. Eurosurveillance 14(45):7–9 21.29.30296
Indra A, Huhulescu S, Schneeweis M, Hasenberger P, Kuijper E, Weerdt J, Kato H, Kato N, Dam A, Vorm E,
Kernbichler S, Fiedler A, Wewalka G, Allerberger F, Weel J, Rheenen C, Dankert J (2001) Nosocomial
Kuijper EJ (2008) Characterization of Clostridium outbreak of Clostridium difficile-associated diarrhoea
difficile isolates using capillary gel electrophoresis- due to a clindamycin-resistant enterotoxin A-negative
based PCR ribotyping. J Med Microbiol strain. Eur J Clin Microbiol Infect Dis 20:528–534.
Janezic S, Strumbelj I, Rupnik M (2011) Use of modified Lessa FC, Mu Y, Bamberg WM, Beldavs ZG, Dumyati
PCR ribotyping for direct detection of Clostridium GK, Dunn JR, Farley MM, Holzbauer SM, Meek JI,
difficile ribotypes in stool samples. J Clin Microbiol Phipps EC, Wilson LE, Winston LG, Cohen JA,
49:3024–3025. Limbago BM, Fridkin SK, Gerding DN, McDonald
Kato H, Kato N, Watanabe K, Iwai N, Nakamura H, LC (2015) Burden of Clostridium difficile infection in
Yamamoto T, Suzuki K, Kim S-M, Chong Y, Wasito the United States. N Engl J Med 372:825–834. https://
EB (1998) Identification of toxin A-negative, toxin
Ribotypes and New Virulent Strains Across Europe 57

Lim SK, Stuart RL, Mackin KE, Carter GP, Kotsanas D, Romano V, Pasquale V, Krovacek K, Mauri F, Demarta A,
Francis MJ, Easton M, Dimovski K, Elliott B, Riley Dumontet S (2012) Toxigenic Clostridium difficile PCR
TV, Hogg G, Paul E, Korman TM, Seemann T, Stinear ribotypes from wastewater treatment plants in Southern
TP, Lyras D, Jenkin GA (2014) Emergence of a Switzerland. Appl Environ Microbiol 78:6643–6646.
ribotype 244 strain of Clostridium difficile associated
with severe disease and related to the epidemic Rupnik M, Janezic S (2016) An update on Clostridium
ribotype 027 strain. Clin Infect Dis 58:1723–1730. difficile toxinotyping. J Clin Microbiol 54:13–18.
Marsh JW, O’Leary MM, Shutt KA, Pasculle AW, Rupnik M, Avesani V, Janc M, von Eichel-Streiber C,
Johnson S, Gerding DN, Muto CA, Harrison LH Delmée M (1998) A novel toxinotyping scheme
(2006) Multilocus variable-number tandem-repeat and correlation of toxinotypes with serogroups of
analysis for investigation of Clostridium difficile Clostridium difficile isolates. J Clin Microbiol
transmission in hospitals. J Clin Microbiol 36:2240–2247
44:2558–2566. Rupnik M, Brazier JS, Duerden BI, Grabnar M, Stubbs SL
Metcalf DS, Costa MC, Dew WMV, Weese JS (2010) (2001) Comparison of toxinotyping and PCR
Clostridium difficile in vegetables, Canada. Lett Appl ribotyping of Clostridium difficile strains and descrip-
Microbiol 51:600–602. tion of novel toxinotypes. Microbiol Read Engl
1472-765X.2010.02933.x 147:439–447.
Monot M, Eckert C, Lemire A, Hamiot A, Dubois T, 2-439
Tessier C, Dumoulard B, Hamel B, Petit A, Rupnik M, Tambic Andrasevic A, Trajkovska Dokic E,
Lalande V, Ma L, Bouchier C, Barbut F, Dupuy B Matas I, Jovanovic M, Pasic S, Kocuvan A, Janezic S
(2015) Clostridium difficile: new insights into the evo- (2016) Distribution of Clostridium difficile PCR
lution of the pathogenicity locus. Sci Rep. https://doi. ribotypes and high proportion of 027 and 176 in
org/10.1038/srep15023 some hospitals in four South Eastern European
Murray R, Boyd D, Levett PN, Mulvey MR, Alfa MJ countries. Anaerobe 42:142–144.
(2009) Truncation in the tcdC region of the Clostrid- 1016/j.anaerobe.2016.10.005
ium difficile pathLoc of clinical isolates does not pre- Santos A, Isidro J, Silva C, Boaventura L, Diogo J,
dict increased biological activity of toxin B or toxin Faustino A, Toscano C, Oleastro M (2016) Molecular
A. BMC Infect Dis 9:103. and epidemiologic study of Clostridium difficile
1471-2334-9-103 reveals unusual heterogeneity in clinical strains
Neely F, Lambert M-L, Van Broeck J, Delmée M (2017) circulating in different regions in Portugal. Clin
Clinical and laboratory features of the most common Microbiol Infect 22:695–700.
Clostridium difficile ribotypes isolated in Belgium. J 1016/j.cmi.2016.04.002
Hosp Infect 95:394–399. Schneeberg A, Neubauer H, Schmoock G, Baier S,
jhin.2016.12.011 Harlizius J, Nienhoff H, Brase K, Zimmermann S,
Nyč O, Pituch H, Matějková J, Obuch-Woszczatynski P, Seyboldt C (2013) Clostridium difficile genotypes in
Kuijper EJ (2011) Clostridium difficile PCR ribotype piglet populations in Germany. J Clin Microbiol
176 in the Czech Republic and Poland. Lancet 51:3796–3803.
377:1407. Smith A (2005) Outbreak of Clostridium difficile infec-
60575-8 tion in an English hospital linked to hypertoxin-
Obuch-Woszczatyński P, Lachowicz D, Schneider A, producing strains in Canada and the US. Euro Surveill
Mól A, Pawłowska J, Ożdżeńska-Milke E, Bull Eur Sur Mal Transm Eur Commun Dis Bull 10:
Pruszczyk P, Wultańska D, Młynarczyk G, E050630.2
Harmanus C, Kuijper EJ, van Belkum A, Pituch H Spigaglia P, Barbanti F, Dionisi AM, Mastrantonio P
(2014) Occurrence of Clostridium difficile (2010) Clostridium difficile isolates resistant to
PCR-ribotype 027 and it’s closely related fluoroquinolones in Italy: emergence of PCR ribotype
PCR-ribotype 176 in hospitals in Poland in 018. J Clin Microbiol 48:2892–2896.
2008–2010. Anaerobe 28:13–17. 10.1128/JCM.02482-09
1016/j.anaerobe.2014.04.007 Stabler RA, Gerding DN, Songer JG, Drudy D, Brazier
Pituch H, van den Braak N, van Leeuwen W, van JS, Trinh HT, Witney AA, Hinds J, Wren BW (2006)
Belkum A, Martirosian G, Obuch-Woszczatyński P, Comparative phylogenomics of Clostridium difficile
Łuczak M, Meisel-Mikołajczyk F (2001) Clonal dis- reveals clade specificity and microevolution of hyper-
semination of a toxin-A-negative/toxin-B-positive virulent strains. J Bacteriol 188:7297–7305. https://
Clostridium difficile strain from patients with
antibiotic-associated diarrhea in Poland. Clin Stubbs SL, Brazier JS, O’Neill GL, Duerden BI (1999)
Microbiol Infect 7:442–446. PCR targeted to the 16S-23S rRNA gene intergenic
j.1198-743x.2001.00312.x spacer region of Clostridium difficile and construction
Pruitt RN, Lacy DB (2012) Toward a structural under- of a library consisting of 116 different PCR ribotypes.
standing of Clostridium difficile toxins A and B. Front J Clin Microbiol 37:461–463
Cell Infect Microbiol 2:28. Sundram F, Guyot A, Carboo I, Green S, Lilaonitkul M,
fcimb.2012.00028 Scourfield A (2009) Clostridium difficile ribotypes
58 J. Couturier et al.

027 and 106: clinical outcomes and risk factors. J surveillance and German-wide genotyping of recent
Hosp Infect 72:111–118. isolates provided to the advisory laboratory for diagnostic
jhin.2009.02.020 reasons. Int J Med Microbiol 305:807–813. https://doi.
Taori SK, Wroe A, Hardie A, Gibb AP, Poxton IR (2014) org/10.1016/j.ijmm.2015.08.035
A prospective study of community-associated Clos- Warny M, Pepin J, Fang A, Killgore G, Thompson A,
tridium difficile infections: the role of antibiotics and Brazier J, Frost E, McDonald LC (2005) Toxin pro-
co-infections. J Infect 69:134–144. duction by an emerging strain of Clostridium difficile
1016/j.jinf.2014.04.002 associated with outbreaks of severe disease in North
Valiente E, Dawson LF, Cairns MD, Stabler RA, Wren America and Europe. Lancet 366:1079–1084. https://
BW (2012) Emergence of new PCR ribotypes from
the hypervirulent Clostridium difficile 027 lineage. J Wilcox MH, Mooney L, Bendall R, Settle CD, Fawley
Med Microbiol 61:49–56. WN (2008) A case-control study of community-
jmm.0.036194-0 associated Clostridium difficile infection. J
von Müller L, Mock M, Halfmann A, Stahlmann J, Simon A, Antimicrob Chemother 62:388–396.
Herrmann M (2015) Epidemiology of Clostridium diffi- 10.1093/jac/dkn163
cile in Germany based on a single center long-term
Comparative Genomics of Clostridium

Sandra Janezic, Julian R. Garneau, and Marc Monot

Abstract Keywords
Clostridium difficile, a gram-positive spore- Genomics · Evolution · Transmission ·
forming anaerobic bacterium, has rapidly Recurrence · CRISPR/Cas · Nontoxigenic
emerged as the leading cause of nosocomial strains · Epidemiology
diarrhoea in hospitals. The availability of
genome sequences in large numbers, mainly
due to the use of next-generation sequencing 1 Introduction
methods, have undoubtedly shown their
immense advantages in the determination of Clostridium difficile infection (CDI) is currently
the C. difficile population structure. The imple- the most frequently occurring nosocomial diar-
mentation of fine-scale comparative genomic rhoea in healthcare environments (Davies et al.
approaches have paved the way to global trans- 2016). This major pathogen synthesizes two
mission and recurrence studies, but also more toxins, encoded in a pathogenicity locus
targeted studies such as the PaLoc or the (PaLoc), that are generally recognised as its
CRISPR/Cas systems. In this chapter, we pro- main virulence factors. Over the last decade,
vide an overview of the recent and significant the incidence and severity of CDI have signifi-
findings on C. difficile using comparative geno- cantly increased mainly due to the emergence
mics studies with implication for the epidemiol- of new strain variants. Molecular typing
ogy, infection control and understanding of the methods were extensively used to understand
evolution of C. difficile. its epidemiology, genetic diversity and

S. Janezic
National Laboratory for Health, Environment and Food
(NLZOH), Maribor, Slovenia
M. Monot (*)
University of Maribor, Faculty of Medicine, Maribor, Département de Microbiologie, Institut Pasteur,
Slovenia Laboratoire Pathogenèse des bactéries anaérobies, Paris,
J. R. Garneau
Département de Microbiologie, Institut Pasteur, Département de Microbiologie et d’infectiologie, Faculté
Laboratoire Pathogenèse des bactéries anaérobies, Paris, de Médecine et des Sciences de la Santé, Université de
France Sherbrooke, Sherbrooke, QC, Canada
Département de Microbiologie et d’infectiologie, Faculté Université Paris Diderot, Sorbonne Paris Cité, Paris,
de Médecine et des Sciences de la Santé, Université de France
Sherbrooke, Sherbrooke, QC, Canada e-mail:

# Springer International Publishing AG 2018 59

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
60 S. Janezic et al.

evolution. The C. difficile population structure population structure (i.e. mutational evolu-
contains hundreds of strain types organized in tion) of C. difficile species (Lemee et al.
phylogenetic clades (Dingle et al. 2014; Elliott 2004). The study by Griffiths et al. (2010)
et al. 2014; Janezic et al. 2016). using a different MLST scheme (with different
The first complete genome sequence of a set of genes included) confirmed the clonal
C. difficile strain was published in 2006 (Sebaihia population structure of C. difficile and
et al. 2006) enabling the development of compar- identified two additional lineages, one
ative genomics. Initially, microarray compara- represented by the ST-22 (PCR ribotype
tive genome hybridization (CGH) were used in 023, toxinotype IV) and the genetically distant
global studies to estimate the diversity and evo- outlier of ST-11 (PCR ribotype 078, toxinotype
lution of strains (Table 1). However, many V). In 2012, Knetsch et al. (2012) described a
laboratories over the world can now afford the putative sixth lineage, using MLST,
frequent and even the routine sequencing of represented by a single sequence type
C. difficile strains. The availability of genome (ST-122; PCR ribotype 131). However phylo-
sequences in large numbers, mainly due to the genetic analysis based on core genome com-
use of next-generation sequencing (NGS) parison did not confirm the topology of the tree
methods, have now undoubtedly shown their and placed this strain as an outlier within the
immense advantages in the determination of the clade 1, possibly a recombinant between clade
C. difficile population structure. The implemen- 1 and clade 2 (Dingle et al. 2014). The popula-
tation of fine-scale comparative genomic tion structure composed of clades was later
approaches have paved the way to global trans- confirmed by comparisons of whole genome
mission and recurrence studies, but also more sequences on more diverse collection of
targeted studies such as the PaLoc or the strains (He et al. 2010; Dingle et al. 2014;
CRISPR/Cas systems (Table 1). Janezic and Rupnik 2015; Knight et al. 2015).
Here we provide an overview of the recent High concordance of MLST and core genome
and significant findings on C. difficile using com- phylogeny demonstrated that MLST could be
parative genomics studies. Those researches shed used as a good proxy to whole genome
new light on the epidemiology, the evolution and comparisons (Griffiths et al. 2010; Didelot
the clinical practice used for C. difficile. et al. 2012; Dingle et al. 2014).
There are currently eight distinct phylogenetic
clades described, which are designated from one
2 Global Comparative Genomics to five and three cryptic clades C-I, -II, -III
(Fig. 1) (Janezic et al. 2016). The cryptic clades
2.1 Population Structure represent highly divergent groups of strains, thus
of C. difficile Species it is speculated that these groups of strains might
even represent novel species or a subspecies
Lemee et al. (2004) conducted the first analy- (Dingle et al. 2014). These clades were initially
sis of genetic relationship and population associated only with non-toxigenic strains. How-
structure of C. difficile isolates using ever, in a recent publication toxigenic strains
multilocus sequence typing (MLST). They were characterized in one of these clades,
identified 34 different MLST sequence types i.e. the clade C-I (Monot et al. 2015).
(MLST-ST) among the 72 isolates. The phylo- The population structure composed of the first
genetic analysis demonstrated three distinct six clades (1–5 and C-I) was defined mainly on
phylogenetic clades with no specific associa- isolates originating from humans and in lesser
tion between a particular clade and hosts or extent from animals. Recently, a study on MLST
geographic origins. Furthermore, they showed analysis of isolates originating from the environ-
that loci included in the MLST scheme were in ment (mainly soil) demonstrated two highly diver-
linkage disequilibrium demonstrating a clonal gent clades (C-II and C-III) comprising mainly
Comparative Genomics of Clostridium difficile 61

Table 1 C. difficile comparative genomic studies

Year Strains References Topics Summary
Hybridization: microArrays
2006 8 Sebaihia et al. (2006) Comparison Core genome
– 75 Stabler et al. (2006) Evolution Phylogenomics
2009 73 Janvilisri et al. (2009) Comparison Core and divergence between host
2010 167 Scaria et al. (2010) Comparison Core genome
– 94 Marsden et al. (2010) Comparison UK and European ribotype 027
Sequencing: Sanger and NGS
2009 2 Stabler et al. (2009) Comparison Historic and modern ribotype 027
2010 29 He et al. (2010) Evolution Short and long time scales
– 15 Scaria et al. (2010) Comparison Core genome
2011 14 Forgetta et al. (2011) Comparison Severe disease associated genomic markers
2012 15 Eyre et al. (2012) Transmission WGS for outbreak detection
– 486 Didelot et al. (2012) Transmission Micro-evolution
2013 151 He et al. (2013) Evolution Emergence and global spread of ribotype
– 1 Eyre et al. (2013d) Evolution Short-term stability of a single ribotype
– 1223 Eyre et al. (2013b) Transmission Identification of diverse source of infection
– 15 Eyre et al. (2013a) Transmission Detection of mixed infection
– 176 Eyre et al. (2013c) Transmission Role of asymptomatic carriage in
2014 1693 Dingle et al. (2014) Evolution History of the pathogenicity locus
– 48 Kurka et al. (2014) Typing Ribotype and MLST correlation
– 185 Eyre et al. (2014) Antibiotics Fidaxomicin in relapse and reinfection
– 3 Moura et al. (2014) Antibiotics Metronidazole resistance
– 31a Moura et al. (2014), Hargreaves et al. CRISPR Distribution and diversity
2015 53 Mac Aogain et al. (2015) Recurrence Discrimination between relapses and
– 18a Boudry et al. (2015) CRISPR Mechanistic and physiology
– 3 Monot et al. (2015) Evolution Model of the pathogenicity locus evolution
2016 96 Quesada-Gomez et al. (2016) Toxins Specificity of hypervirulent Clade 2 TcdB
– 5 Chowdhury et al. (2016) Toxins Toxin-negative strains in human and
– 108 Kumar et al. (2016) Transmission Relapse and reinfection of ribotype 027
2017 35 Sim et al. (2017) Recurrence Rate of relapses and reinfections
– 277 Cairns et al. (2017) Evolution Phylogeny of ribotype 017
– 265 Mawer et al. (2017) Transmission Symptomatic patients but fecal toxin
– 971 Eyre et al. (2017) Transmission WGS as hospital surveillance tools
C. difficile phage or prophage

non-toxigenic isolates (Janezic et al. 2016). The sporadic isolation from clinical samples, these
topology of the MLST based tree was also con- strains could represent native environmental
firmed by whole genome comparisons (Janezic isolates, which are not primarily associated with
and Rupnik, unpublished). It was hypothesized humans and/or animals (Janezic et al. 2016).
that, due to high abundance of isolates from these The most heterogeneous clade, in terms of num-
two clades in the environmental samples and only ber of MLST-STs and PCR ribotypes, is clade
62 S. Janezic et al.

Fig. 1 Maximum
likelihood phylogenetic
tree showing the eight
currently described
C. difficile clades
(Reproduced from Janezic
et al. 2016)

1, where more than 200 different MLST-STs are e.g. PCR ribotypes 014, 002, 001, 015,
present (data from PubMLST C. difficile database, 018, which are among the ten most prevalent
accessed 21.7.2017) (Table 2). Many of the strains PCR ribotypes isolated from CDI (C. difficile infec-
from this clade are of clinical significance, tion) patients in Europe (Davies et al. 2016). Clade
Comparative Genomics of Clostridium difficile 63

Table 2 Overview of heterogeneity within clades and correlation between main MLST-ST and PCR ribotypes
Clade Nr. MLST STa Most known PCR ribotypes/MLST-ST(s)b
1 241 001 ST-3
002 ST-8, ST-35, ST-48, ST-146
012 ST-54
014 ST-2, ST-13:14, ST-49:50, ST-132
015 ST-10, ST-44
018 ST-17
2 61 027 ST-1
176 ST-1
244 ST-41
3 13 023 ST-5, ST-22, ST-25
4 53 017 ST-37, ST-86
5 30 033 ST-11
126 ST-11
078 ST-11
C-Ic 6 ND ST-177:181
C-IId 3 ND ST-200, ST-337:338
C-IIId 9 ND ST-336, ST-339, ST-341:347
Data from PubMLST C. difficile database (accessed 21.7.2017)
Data from Griffiths et al. 2010; Stabler et al. 2012; Knetsch et al. 2012; Dingle et al. 2014
Data from Dingle et al. 2014
Data from Janezic et al. 2016

2 is the second most heterogeneous clade, WGS) demonstrated the opposite, since there are
containing 61 different MLST-STs, including currently 30 STs found in clade 5 (Table 2).
PCR ribotype 027 (ST-1), a well-known epidemic Large scale analyses of strains from diverse
strain, and two emerging ribotypes 176 (ST-1) and sources and geographic origins also revealed that
244 (ST-41) (Valiente et al. 2012; Lim et al. 2014). significant microdiversity exist within clades and
In the clade 3, 13 different STs are present and the that C. difficile is continuously evolving
most known representative is PCR ribotype (Table 2) (Griffiths et al. 2010; Dingle et al.
023 (represented with ST-5, ST-22 and ST-25), 2011, 2014; Knetsch et al. 2012; Janezic et al.
which is also often isolated from humans in 2016).
European countries (Davies et al. 2016). The
clade 4, composed of 53 STs, is also known as
A-B+ clade due to its association with PCR
2.2 Worldwide Evolution
ribotype 017 strains (A-B+CDT-). Despite the
of Important C. difficile PCR
altered toxin expression this strain is widespread,
especially in Asia (Shin et al. 2008; Collins et al.
2013). One of the best known representative of
2.2.1 Epidemic C. difficile PCR Ribotype
clade 5 is PCR ribotype 078, which has in recent
years emerged in human CDI (Rupnik et al. 2008),
C. difficile PCR ribotype 027 has in the last two
while before it was thought to be primarily an
decades gained much interest because of its rapid
animal pathogen (Jhung et al. 2008). Although in
emergence worldwide. The strain has been
the first studies it seemed that this clade was more
associated with large CDI outbreaks and
homogeneous, consisting primarily of ST-11
increased morbidity and mortality, which have
strains (Griffiths et al. 2010; Dingle et al. 2011;
first started to appear in USA and Canada (Pepin
Knetsch et al. 2012), later studies (MLST and
et al. 2004, 2005; Loo et al. 2005; McDonald
et al. 2005). The strain was later also introduced
64 S. Janezic et al.

in Europe, with first outbreaks documented in the outbreaks that gained attention also in media
United Kingdom, and in following years also in (He et al. 2013).
continental Europe (Kuijper et al. 2008).
Although the prevalence of PCR ribotype
2.2.2 Toxin Variant C. difficile PCR
027 declined markedly in Europe, the strain still
Ribotype 017
remains one of the most common strains causing
Another important strain that has gained much
CDI (Bauer et al. 2011; Davies et al. 2016). To
attention is PCR ribotype 017 (toxinotype VIII,
explore global population structure and genetic
MLST ST-37). Despite producing only one of the
changes associated with its rapid emergence and
three C. difficile toxins (A-B+CDT-) PCR
global spread, He et al. (2013) sequenced
ribotype 017 strains are causing clinically signif-
genomes of 151 strains, representing the global
icant infections worldwide (Drudy et al. 2007;
population of ribotype 027 strains, collected
Collins et al. 2013; Cairns et al. 2015). Initially,
between 1985 and 2010. They showed that
ribotype 017 strains have been identified in
ribotype 027 population consists of two geneti-
outbreaks in Asia where they were responsible
cally distinct fluoroquinolone resistant (FQR1
for the majority of CDIs (Collins et al. 2013). It
and FQR2) epidemic lineages. Both lineages
was hypothesized that this strain has spread from
have independently acquired the same mutation
Asia throughout the world (Drudy et al. 2007;
in DNA gyrase, conferring fluoroquinolone resis-
Collins et al. 2013; Cairns et al. 2015). To define
tance, and a novel conjugative transposon
population structure and patterns of global
(CTn5-like element, Tn6192). These were the
spread, Cairns et al. (2017) conducted a phyloge-
only two genetic traits differentiating FQR1 and
netic analysis on a global collection of PCR
FQR2 lineages from the historic 027 isolates, and
ribotype 017 strains. Two hundred and seventy
were most likely key genetic changes associated
seven PCR ribotype 017 strains, including
with the rapid emergence of ribotype 027. Also,
human, animal and environmental isolates were
low level of genomic diversity within the core
obtained from all six continents and were
genome of the 151 PCR ribotype 027 strains
isolated between 1990 and 2013. The phyloge-
analysed was demonstrated, with only
netic analysis based of the core SNPs
536 SNPs identified. Only two of these SNPs
demonstrated presence of two genetically diverse
(limited to a single isolate) were discovered in
lineages (SL1 and SL2) which are geographically
the PaLoc region of historic and epidemic
and temporally widespread. In both lineages
isolates (He et al. 2013) which contrast with the
multiple clonal expansions were revealed.
earlier assumptions that genetic changes in the
Phylogeographic analysis also revealed, contrary
PaLoc were the cause of emergence of C. difficile
to current Asia-origin hypothesis, that origin of
027 (McDonald et al. 2005; Warny et al. 2005).
ribotype 017 is in North America, from where the
Although both lineages emerged in North
strain has been introduced first to Europe and
America, they showed different global spread
then from Europe to Asia and Australia and
and limited geographic clustering. FQR1
from there it was then spread worldwide. Further
originated in Pittsburgh (Pennsylvania, USA),
genetic analysis, based on the SNPs present in
and was subsequently transmitted to Switzerland
gyrA, gyrB and rpoB genes, predicted that
and South Korea. The FQR2 lineage which
ribotype 017 strains are commonly resistant to
contains majority of epidemic strains was trans-
fluoroquinolone (76%) and rifampicin (35%)
mitted to continental Europe and United King-
classes of antibiotics. Due to different clusters
dom on several different occasions, and a single
of genes inserted in the same genomic locations
introduction to Australia was demonstrated. The
the authors also identified hot-spot regions for
phylogenetic analysis of UK collection of epi-
DNA uptake (Cairns et al. 2015).
demic FQR2 strains further demonstrated fre-
quent long-range transmissions within the UK,
some of them associated with large scale
Comparative Genomics of Clostridium difficile 65

2.3 C. difficile Transmissions these cases shared the same ward with at least
and Epidemiology one other case or had some sort of hospital con-
of Recurrent CDI tact, which is much lower than expected. Almost
half (45%) of isolates were genetically unrelated
2.3.1 C. difficile Transmissions (10 SNPs) to any other previous case and could
in Hospital Environment not be linked by transmission (direct or indirect),
In the past, assessment of genetic relatedness of meaning that they were likely acquired from
C. difficile isolates has been hampered by the use sources other than symptomatic patients. Identi-
of sub-optimal genotyping methods that do not fication of a rather diverse pool of C. difficile
have sufficient discriminatory power (e.g. PCR strains indicate existence of substantial
ribotyping, MLST) to distinguish closely related reservoirs of C. difficile and that transmission
strains. Whole genome sequence analysis which routes other than those due to symptomatic CDI
enables comparison at the highest level of patients should be considered (e.g. asymptomatic
genetic resolution has been widely adopted for patients and environment) (Eyre et al. 2013b).
global and national C. difficile surveillances and Role of asymptomatic carriage of C. difficile
has revealed some novel insights about in transmission was explored in a small study
transmissions dynamics and recurrent infections including 132 participants (Eyre et al. 2013c).
(Table 1). The authors demonstrated that even though
Estimating the rates at which bacterial asymptomatic carriage is common, onward trans-
genomes evolve (e.g. within-host diversity and mission from asymptomatic case is relatively
short-term evolution) is critical for understand- rare. The same group has also described novel
ing transmission patterns (Duchene et al. 2016). approach using WGS that enables assessment of
For C. difficile, rates of short-term evolution and the extent of infection transmission within
within-host diversity have been explored in sev- healthcare institutions by measuring the propor-
eral studies, using serial samples from patients tion of cases that are acquired from a previous
with recurrent or on-going CDI and in vitro gut case (i.e. linked cases) (Eyre et al. 2017).
model of CDI. In all these studies similar
estimations of evolutionary rates were obtained, 2.3.2 C. difficile Recurrence:
1–2 SNPs/genome/year and within-host diversity Reinfections Versus Relapses
0.30 SNPs/genome/year (Didelot et al. 2012; WGS has also been shown to be a valuable tool in
Eyre et al. 2013b, d). By using these estimations understanding the epidemiology of CDI
two isolates obtained less than 124 days apart recurrences with greater accuracy, especially
would be expected to have 0–2 SNPs differences within hospital settings with endemic strains
and isolates obtained 124–364 days apart should (Eyre et al. 2014; Mac Aogain et al. 2015;
exhibit 0–3 SNPs differences (Eyre et al. 2013b). Kumar et al. 2016). Recurrent C. difficile
This definition of genetically related isolates, infections occurs in up to 25% of patients after
i.e. isolates that are most probably a result of the first CDI episode and discriminating between
transmission, has now been widely adopted. reinfections (infection with newly acquired
It is traditionally believed that most cases of strain) and relapses (recurrent episode due to
C. difficile infections are acquired within hospital original strain) is important for CDI manage-
settings, where they are being transmitted from ment; infection prevention and treatment, respec-
person to person (Vonberg et al. 2008; Khanna tively (Kelly 2012).
and Pardi 2012). Eyre and colleagues (Eyre et al. Similar methodology that was used in trans-
2013b) compared genomic sequences of 1223 mission studies can also be applied in studies
C. difficile isolates and demonstrated that only resolving the contribution of relapses and
35% of cases were acquired from another known reinfections in recurrent CDI. Relapse is defined
case within a hospital setting and only a subset of as a recurrent infection with an isolate differing
2 SNPs from the isolate from initial episode
66 S. Janezic et al.

and reinfection involving pairs of isolates differ- during the sub-culturing. Genetic variations
ing 10 SNPs (Eyre et al. 2014). Mac Aogain between the strains were found responsible for
et al. (2015) applied this methodology to the phenotypic differences observed in both
19 patients with recurrent CDI to resolve the mutants (growth rate, motility, sporulation and
nature of the recurrences and demonstrated that virulence), explaining different outcomes of both
majority of recurrences (16 out of 19) were due studies. Since 630Δerm strain more closely
to relapse with endogenous strain. Similar resembles the progenitor strain, the authors
findings were also found in a study by Eyre concluded that this strain should be favored
et al. (2014) that used WGS to determine whether over 630E and that re-sequencing of genomes
recurrences of CDI in 93 patients (28 were of mutant strains should become a routine prac-
treated with fidaxomicin and 65 were treated tice (Collery et al. 2016).
with vancomycin) were due to reinfection or
relapse. Overall 79.6% (74 of 93) recurrent
CDIs were due to relapse. Reinfection accounted 2.5 Comparative Genomic Analysis
for just one fifth of recurrences. of Non-toxigenic Strains

Comparative genomic studies demonstrated that

2.4 Influence of SNPs on Virulence non-toxigenic C. difficile strains are represented
and Phenotype of CD630 in all clades, alongside toxigenic isolates (Dingle
Derivatives et al. 2014; Monot et al. 2015). Although toxin-
negative C. difficile strains can be isolated from
The C. difficile strain CD630 was isolated in patients and animals suffering from gastrointes-
1982 in Zurich, Switzerland from a patient with tinal diseases, they are not considered to play a
pseudomembranous colitis (Sebaihia et al. 2006). role in disease (Vedantam et al. 2012).
This is the first strain of C. difficile which Chowdhury et al. (2016) undertook a compara-
genome has been sequenced and which tive genomic analysis of five toxin negative
derivatives were used as a model strain for gen- strains isolated from faeces from humans and
eration of mutants in different studies exploring animals with symptoms of gastrointestinal
the importance of C. difficile toxins in pathogen- (GI) disease. Even though the authors stated
esis. Two groups that used isogenic mutants that GI symptoms were likely due to
(in which production of one of both toxins was non-toxigenic C. difficile, this could also be due
ablated) of erythromycin sensitive derivatives to undetected co-infection with toxigenic
(630E and 630Δerm) from the strain CD630, C. difficile or to infection with a yet unknown
got contradictory outcomes on virulence poten- or un-cultivable organism. Phylogenetic analysis
tial of toxin A (TcdA) (Collery et al. 2016). In a demonstrated that all five isolates clustered with
study by Lyras et al. (2009) the outcome was that toxigenic isolates (belonging also to the same
tcdB mutant, producing only toxin TcdA (A+B-), MLST-ST) and had also a similar virulence
was unable to cause disease in hamster model, associated gene repertoire as those found in toxi-
whereas in a study by Kuehne et al. (2010), the genic strains (e.g. genes required for sporulation
authors demonstrated that both toxins, TcdA and (spo0A) and adhesion (groEL, fliC), genes coding
TcdB, are capable of causing disease in a hamster for surface proteins (slpA and cwp) necessary for
model. Both strains possessed the same deletion colonization of the gut and different serine-
of ermB gene and were isolated in two different proteases and metalloproteases).
laboratories by repeated sub-culturing of strain Recently a transfer of the PaLoc from toxi-
CD630 (Collery et al. 2016). Re-sequencing of genic to non-toxigenic strain, that was able to
both strains revealed that both strains had a num- produce toxins after acquiring the PaLoc, has
ber of SNPs, compared to the published genome been demonstrated (Brouwer et al. 2013) further
of CD630, which were most likely accumulated suggesting that non-toxigenic isolates could
Comparative Genomics of Clostridium difficile 67

represent reservoir for toxigenic strains, as was possible by plotting the distribution of indels
already suggested (Dingle et al. 2014). Current and SNPs at the chromosomal scale and by
knowledge of the pathogenicity of non-toxigenic analyzing in more details the SNP plots for the
strains is still limited and therefore further regions around the PaLoc. Distinctively, specific
research is required to explore potential of recombination mediated by an integrase supplied
non-toxigenic strains to cause diseases as well in trans appears to be the mechanism involved in
as the PaLoc exchange between toxigenic and the initial PaLoc acquisition. The reason for this
non-toxigenic strains (Chowdhury et al. 2016). is the absence of recombination signatures on
DNA sequences distant from the PaLoc in
non-toxigenic strains (Dingle et al. 2014).
3 Targeted Comparative Brouwer et al. (2013) demonstrated experi-
Genomics mentally that non-toxigenic C. difficile strains
could be converted into toxin producers by hori-
3.1 Evolution of the C. difficile zontal gene transfer and genetic recombination.
Pathogenicity Locus It is worrying to think that different versions of
the PaLoc can be acquired and transferred seem-
The pathogenicity locus encodes the exotoxins ingly at any time by any strain because this
TcdA and TcdB, the two main virulence factors makes all the non-toxigenic strains possible
involved in CDI. Bacterial strains completely candidates for becoming toxin producers
lacking this genomic region are unable to cause (Brouwer et al. 2013). The possible acquisition
the disease and its associated symptoms, so it of the PaLoc by non-toxigenic strains that
appears of outmost importance to understand already exhibit high resistance to antibiotics
how this locus has been acquired and how it widely used in clinics for the treatment of CDI,
can evolve over time (Cohen et al. 2000). Com- (e.g. strains belonging to ribotype 010 highly
parative genomics, which is a very powerful resistant to metronidazole Moura et al. 2013,
approach to elucidate the evolutionary history 2014), is a very concerning scenario. All those
of the PaLoc, have shown that this locus has recent findings concerning the PaLoc are of out-
undergone a very complex and intriguing event- most importance and can have profound
ful history (Dingle et al. 2014; Elliott et al. 2014; repercussions on the evolution of the disease in
Monot et al. 2015). However, the conclusions clinics. It is highly conceivable that the events
drawn from such analyses are likely in constant reported here and the related mechanisms might
evolution as they depend on the strains available. be more prevalent than first thought and may be
relevant to other commensal and pathogenic bac-
3.1.1 PaLoc Acquisition and Exchange teria as well.

Dingle et al. (2014) have estimated that the most

3.1.2 PaLoc Organization and Evolution
recent acquisition of the PaLoc would have
occurred some 500 years ago. The latest
The evolutionary history of the PaLoc was first
exchange of the PaLoc between C. difficile
studied by performing comparative genomics on
strains have been calculated to about 300 years
C. difficile genomes from a collection of 1693
and the most recent PaLoc loss from the genome
toxigenic and non-toxigenic strains (Dingle et al.
would have happened in recent times (~30 years
2014). Thereafter, further studies have refined
ago). Because of the very long genomic
the established model by adding new PaLoc
fragments concurrently swapped during those
variants (Elliott et al. 2014; Janezic et al. 2015;
recent PaLoc losses and exchanges, it is thought
Monot et al. 2015) leading to the actual known
that host-mediated homologous recombination is
gene contents and organizations of the PaLoc
the mechanism by which those recent events
detailed in Fig. 2a.
have arisen. Those observations were made
68 S. Janezic et al.

Fig. 2 PaLoc diversity and evolution. (a) Known types of C. difficile PaLoc and (b) Model of evolution from “Mono-
Toxin” to “Bi-Toxin” PaLoc (Adapted from Fig. 6B and S7 of Monot et al. 2015)

Monot et al. (2015) found two types of geno- The PaLoc could also be altered during evolu-
mic organization of the PaLoc that each tion by insertion of mobile elements. These strains
contained only one of the two toxins (A+B- and have been associated with milder clinical
A-B+). These two “Mono-Toxins PaLocs” were phenotypes and the presence of the transposable
located at different positions in the C. difficile element Tn6218 is believably responsible for this
genome far from the usual PaLoc integration site, change in the bacterial phenotype (Dingle et al.
which was not described before. Based on 2014). This specific genetic region has probably
sequences similarity analysis, the authors undergone many different exchanges or separate
detected two gene remnants of these PaLoc acquisition events, as many accessory genes were
variants in the classical PaLoc, i.e. “Bi-Toxin noticed in several variants widely spread in the
PaLoc”. Altogether, this work supports a sce- C. difficile population. It is important to carefully
nario in which the “Bi-Toxin PaLoc” was study and follow this type of transposable region
generated by a fusion of two “Mono-Toxin such as Tn6218, since it carries, among others, a set
PaLoc” from ancestral C. difficile strains through of genes providing high-level resistance to
multiple independent PaLoc acquisitions antibiotics used in clinical settings (Spigaglia
(Fig. 2b) (Monot et al. 2015). et al. 2011; Kelly 2012; Deshpande et al. 2013).
Comparative Genomics of Clostridium difficile 69

Elements related to Tn6218 have been found in 3.2 Advances in CRISPR/Cas Systems
other various genomes such as Bifidobacterium and Phage-Host Interaction
breve, Ruminococcus, Lachnospiraceae and
Coprobacillus sp., suggesting that the transfer of Mobile genetic elements (MGE) and especially
this element between different species is also bacteriophages are major contributors and
highly probable and should undoubtedly be further facilitators of genetic evolution in bacteria,
investigated (Dingle et al. 2014). including C. difficile. It has been suggested that
First identified and described by Braun et al. C. difficile is exhibiting a complex, highly mobile
in 2000, IStrons represent another type of mobile and mosaic genome because it is striving in an
genetic element that has been shown to create environment where it is constantly being
variations inside the C. difficile genome and confronted to numerous interacting bacteria and
inside the PaLoc region (Rupnik 2008). It has phages also struggling to survive (Sebaihia et al.
been hypothesized that the original IStron 2006). Therefore, C. difficile is incessantly
(CdISt1-0) was the result of a fusion event incorporating favorable genetic material useful
between an insertion element (IS) and a group I for its adaptation while simultaneously develop-
intron, generating a novel class of chimeric ing defense mechanisms in order to limit the
ribozymes adapted to propagate in eubacterial incorporation and influence of harmful genetic
genomes (Hasselmayer et al. 2004). Widely material (Boudry et al. 2015). A myriad of
spread in C. difficile genomes, four variants of defense mechanisms against foreign MGE and
IStrons have now been identified, all exhibiting a phages are now better known, but the CRISPR/
self-splicing ribozyme activity and which trans- Cas system has only recently been more actively
position was found to be harmless for the explored in C. difficile. CRISPR/Cas systems
interrupted gene (i.e. does not affect TcdA toxin have been defined in three majors types (I, II,
production in C. difficile). Braun et al. III), further divided in 12 different subtypes
(2000) have hypothesized that this particular chi- (Makarova et al. 2011, 2013). C. difficile only
meric element might be more efficient and more harbors the subtype I-B, a system probably
adapted, as the risk of mutation usually observed acquired by mean of horizontal gene transfer
during transposition of an IS-elements is signifi- (HGT) from Archaea (Richter et al. 2012; Peng
cantly reduced by the precise splicing activity et al. 2014).
provided by the group I intron (Braun et al.
3.2.1 CRISPR Mechanism
The complex relationship between C. difficile
and Physiology
and the PaLoc, and also the multiple ways by
The analogy between the mammal acquired
which it is able to evolve, can ostensibly lead to
immunity and the bacterial CRISPR/Cas system
concrete repercussions on its virulence and epi-
is often used, since bacteria can become
demiology. This is illustrated by the characteri-
protected against genetically akin phages after
zation of a clinical strain RA09-70 exhibiting a
exposition, in a fashion reminiscent of vaccina-
new major variant of the PaLoc producing only
tion. The bacteria memorize previous unsuccess-
the toxin A, the A+B- strain RA09-70 (Fig. 2a)
ful infections by acquiring small sequences of the
(Eckert et al. 2013; Monot et al. 2015). This type
assailants and integrating them to its own
of strain would go completely undetected by
genome, inside a specific region or array
cytotoxicity assays, which successfully confirm
containing other similar protective sequences.
CDIs only when TcdB is present. Dissemination
Those sequences, called “spacers” in the
of this type of strain could lead to a problematic
CRISPR/Cas array system, are used by the bac-
under-diagnostic scenario, since this assay is
teria to scan and recognize the identical or near
commonly used as a sole method for the diagno-
identical sequences, called “protospacer”, in the
sis of CDI (Monot et al. 2015).
genome of a future potentially more lethal phage
70 S. Janezic et al.

invader. When the sequence is recognized, a pyogenes in which the CRISPR loci are barely
functional CRISPR system is able to neutralize expressed or event completely silent (Pougach
the infecting agent by cutting and digesting its et al. 2010; Deltcheva et al. 2011).
DNA, interrupting the infection cycle, which
may also result in the acquisition of additional 3.2.2 CRISPR Distribution and Diversity
protective sequences.
Recently, important findings have been made Hargreaves et al. (2014) determined the distribu-
for this system in C. difficile using comparative tion and diversity of the CRISPR/Cas system in
genomics associated with laboratory procedures, C. difficile. To do this, they examined the
such as transcriptome sequencing (RNA-Seq) relationships between spacers and 31 C. difficile
and plasmid conjugation efficiency assays phages and prophage genomes. The spacer con-
(Hargreaves et al. 2014; Boudry et al. 2015). tent is thought to bring a good perception of the
Those analyses have allowed to conclude that predominant and relatively recent phage preda-
the CRISPR/Cas system in C. difficile was func- tion history (Diez-Villasenor et al. 2010). How-
tional and used in this species, since many genes ever, a large number of spacers match sequences
and arrays coding for important components of of unknown nature, possibly targeting unknown
the CRISPR arrays were actively transcribed. C. difficile phages or even non-clostridial phages.
Nine different CRISPR arrays were found to be They also found, in several strains of C. difficile,
present and transcribed in the epidemic strain CRISPR arrays inside prophage genomes, which
R20291, and reference strain 630 exhibited is considered an unusual situation for this system
12 expressed arrays (Boudry et al. 2015). The (Hargreaves et al. 2014; Boudry et al. 2015).
analysis of the targets for the identified spacers Those phages carried spacers that were found to
showed that a unique phage could be targeted by match sequences of other bacteriophages. Once
numerous different spacers, surely to increase the they have successfully integrated the bacterial
efficiency of phage neutralization by the system genome, prophages could plausibly use those
(Boudry et al. 2015). This could also be an indi- spacers in order to give them an advantage over
cation that phage has the ability to evade the other phages by blocking their capacity to infect
CRISPR system using a mutational process. Con- the same strain (Hargreaves et al. 2014).
trastively, a single spacer can have the ability to Prophages possessing CRISPR arrays are
target conserved genes present in multiple thought to rely on the bacterial host for the proper
related phages, thus bestowing them with an functioning of the system, since the cas operon
efficient and inexpensive defense against multi- containing the set of genes necessary to process
ple potential invaders at once. Boudry et al. the arrays were always absent (Boudry et al.
(2015) concluded that there is a good correlation 2015).
between the real and predicted phage susceptibil- To obtain a global view of the distribution of
ity according to the spacer content of the bacte- the CRISPR/Cas system in C. difficile, Boudry
rial strains and the theoretical predicted phage et al. (2015) tested the presence of cas operons in
targets. Remarkably, the spacer sequences found 2207 C. difficile published and available
in C. difficile strain 630 were anticipated to genomes. Nearly 90% of them possessed a com-
target all known and isolated clostridial plete cas operon, making the CRISPR/Cas a
bacteriophages. Experimentally, this strain common system in this bacteria.
exhibited resistance to infection by all the phages
that could be tested.
The CRISPR/Cas system seems particularly 4 Conclusions
active and meaningful in C. difficile as numerous
highly active CRISPR arrays are found, which The evolution of comparative genomics of
moreover greatly contrast with what is observed in C. difficile strains from molecular typing and
other bacteria such as E. coli and Streptococcus microarrays to whole genome sequence enabled
Comparative Genomics of Clostridium difficile 71

significant improvements in the determination of ribotype 017 strain in London, England. J Clin
the population structure of C. difficile. Beyond a Microbiol 53(10):3141–3147.
deeper understanding of the diversity of strains, Cairns MD, Preston MD, Hall CL, Gerding DN, Hawkey
WGS also makes possible the emergence of new PM, Kato H, Kim H, Kuijper EJ, Lawley TD, Pituch
area of research such as transmission or reinfec- H, Reid S, Kullin B, Riley TV, Solomon K, Tsai PJ,
tion studies. Weese JS, Stabler RA, Wren BW (2017) Comparative
genome analysis and global phylogeny of the toxin
Another aspect to be taken into account is the variant Clostridium difficile PCR ribotype 017 reveals
availability of massive sequence data allowing the evolution of two independent sublineages. J Clin
the analysis of specific loci. Due to its impor- Microbiol 55(3):865–876.
tance in virulence, the PaLoc was extensively JCM.01296-16
Chowdhury G, Joshi S, Bhattacharya S, Sekar U, Birajdar
explored and it has been concluded that this B, Bhattacharyya A, Shinoda S, Ramamurthy T
locus is in constant evolution. (2016) Extraintestinal infections caused by non-toxi-
This leads us to conclude that findings of genic Vibrio cholerae non-O1/non-O139. Front
comparative genomics highly depend on the Microbiol 7:144.
strains available, thus making the availability of Cohen SH, Tang YJ, Silva J Jr (2000) Analysis of the
raw data in public database of primordial pathogenicity locus in Clostridium difficile strains. J
importance. Infect Dis 181(2):659–663.
Collery MM, Kuehne SA, McBride SM, Kelly ML,
Acknowledgements JG was supported by a discovery Monot M, Cockayne A, Dupuy B, Minton NP (2016)
grant from the Natural Sciences and Engineering What’s a SNP between friends: the influence of single
Research Council of Canada (NSERC #341450-2010). nucleotide polymorphisms on virulence and
SJ was supported by Slovenian Research Agency grant phenotypes of Clostridium difficile strain 630 and
J4-8224. derivatives. Virulence 8(6):767–781.
Collins DA, Hawkey PM, Riley TV (2013) Epidemiology
of Clostridium difficile infection in Asia. Antimicrob
References Resist Infect Control 2(1):21.
Bauer MP, Notermans DW, van Benthem BH, Brazier JS, Davies KA, Ashwin H, Longshaw CM, Burns DA, Davis
Wilcox MH, Rupnik M, Monnet DL, van Dissel JT, GL, Wilcox MH, group Es (2016b) Diversity of Clos-
Kuijper EJ, Group ES (2011) Clostridium difficile tridium difficile PCR ribotypes in Europe: results from
infection in Europe: a hospital-based survey. Lancet the European, multicentre, prospective, biannual,
377(9759):63–73. point-prevalence study of Clostridium difficile infec-
6736(10)61266-4 tion in hospitalised patients with diarrhoea (EUCLID),
Boudry P, Semenova E, Monot M, Datsenko KA, 2012 and 2013. Euro Surveill 21(29).
Lopatina A, Sekulovic O, Ospina-Bedoya M, Fortier 10.2807/1560-7917.ES.2016.21.29.30294
LC, Severinov K, Dupuy B, Soutourina O (2015) Deltcheva E, Chylinski K, Sharma CM, Gonzales K,
Function of the CRISPR-Cas system of the human Chao Y, Pirzada ZA, Eckert MR, Vogel J, Charpentier
pathogen Clostridium difficile. MBio 6(5):e01112– E (2011) CRISPR RNA maturation by trans-encoded
e01115. small RNA and host factor RNase III. Nature 471
Braun V, Mehlig M, Moos M, Rupnik M, Kalt B, Mahony (7340):602–607.
DE, von Eichel-Streiber C (2000) A chimeric ribo- Deshpande A, Pasupuleti V, Thota P, Pant C, Rolston DD,
zyme in Clostridium difficile combines features of Sferra TJ, Hernandez AV, Donskey CJ (2013) Com-
group I introns and insertion elements. Mol Microbiol munity-associated Clostridium difficile infection and
36(6):1447–1459 antibiotics: a meta-analysis. J Antimicrob Chemother
Brouwer MS, Roberts AP, Hussain H, Williams RJ, Allan 68(9):1951–1961.
E, Mullany P (2013) Horizontal gene transfer converts Didelot X, Eyre DW, Cule M, Ip CL, Ansari MA,
non-toxigenic Clostridium difficile strains into toxin Griffiths D, Vaughan A, O’Connor L, Golubchik T,
producers. Nat Commun 4:2601. Batty EM, Piazza P, Wilson DJ, Bowden R, Donnelly
1038/ncomms3601 PJ, Dingle KE, Wilcox M, Walker AS, Crook DW,
Cairns MD, Preston MD, Lawley TD, Clark TG, Stabler Peto TE, Harding RM (2012) Microevolutionary anal-
RA, Wren BW (2015) Genomic epidemiology of a ysis of Clostridium difficile genomes to investigate
protracted hospital outbreak caused by a toxin transmission. Genome Biol 13(12):R118. https://doi.
A-negative Clostridium difficile sublineage PCR org/10.1186/gb-2012-13-12-r118
72 S. Janezic et al.

Diez-Villasenor C, Almendros C, Garcia-Martinez J, Peto TE, Walker AS (2013b) Diverse sources of C.

Mojica FJ (2010) Diversity of CRISPR loci in difficile infection identified on whole-genome
Escherichia coli. Microbiology 156(Pt 5):1351–1361. sequencing. N Engl J Med 369(13):1195–1205.
Dingle KE, Griffiths D, Didelot X, Evans J, Vaughan A, Eyre DW, Griffiths D, Vaughan A, Golubchik T, Acharya
Kachrimanidou M, Stoesser N, Jolley KA, Golubchik M, O’Connor L, Crook DW, Walker AS, Peto TE
T, Harding RM, Peto TE, Fawley W, Walker AS, (2013c) Asymptomatic Clostridium difficile
Wilcox M, Crook DW (2011) Clinical Clostridium colonisation and onward transmission. PLoS One 8
difficile: clonality and pathogenicity locus diversity. (11):e78445.
PLoS One 6(5):ee19993. 0078445
nal.pone.0019993 Eyre DW, Walker AS, Freeman J, Baines SD, Fawley
Dingle KE, Elliott B, Robinson E, Griffiths D, Eyre DW, WN, Chilton CH, Griffiths D, Vaughan A, Crook
Stoesser N, Vaughan A, Golubchik T, Fawley WN, DW, Peto TE, Wilcox MH (2013d) Short-term
Wilcox MH, Peto TE, Walker AS, Riley TV, Crook genome stability of serial Clostridium difficile
DW, Didelot X (2014) Evolutionary history of the ribotype 027 isolates in an experimental gut model
Clostridium difficile pathogenicity locus. Genome and recurrent human disease. PLoS One 8(5):e63540.
Biol Evol 6(1):36–52.
evt204 Eyre DW, Babakhani F, Griffiths D, Seddon J, Del Ojo
Drudy D, Harnedy N, Fanning S, Hannan M, Kyne L EC, Gorbach SL, Peto TE, Crook DW, Walker AS
(2007) Emergence and control of fluoroquinolone- (2014) Whole-genome sequencing demonstrates that
resistant, toxin A-negative, toxin Bpositive Clostrid- fidaxomicin is superior to vancomycin for preventing
ium difficile. Infect Control Hosp Epidemiol 28 reinfection and relapse of infection with Clostridium
(8):932–940. difficile. J Infect Dis 209(9):1446–1451. https://doi.
Duchene S, Holt KE, Weill FX, Le Hello S, Hawkey J, org/10.1093/infdis/jit598
Edwards DJ, Fourment M, Holmes EC (2016) Eyre DW, Fawley WN, Rajgopal A, Settle C, Mortimer K,
Genome-scale rates of evolutionary change in bacte- Goldenberg SD, Dawson S, Crook DW, Peto TEA,
ria. Microb Genome 2(11):e000094. Walker AS, Wilcox MH (2017) Comparison of control
10.1099/mgen.0.000094 of Clostridium difficile infection in six english
Eckert C, Coignard B, Hebert M, Tarnaud C, Tessier C, hospitals using whole-genome sequencing. Clin Infect
Lemire A, Burghoffer B, Noel D, Barbut F, Group IC- Dis 65:433.
RW (2013) Clinical and microbiological features of Forgetta V, Oughton MT, Marquis P, Brukner I,
Clostridium difficile infections in France: the ICD- Blanchette R, Haub K, Magrini V, Mardis ER,
RAISIN 2009 national survey. Med Mal Infect 43 Gerding DN, Loo VG, Miller MA, Mulvey MR,
(2):67–74. Rupnik M, Dascal A, Dewar K (2011) Fourteen-
004 genome comparison identifies DNA markers for
Elliott B, Dingle KE, Didelot X, Crook DW, Riley TV severe-disease-associated strains of Clostridium diffi-
(2014) The complexity and diversity of the pathoge- cile. J Clin Microbiol 49(6):2230–2238. https://doi.
nicity locus in Clostridium difficile clade 5. Genome org/10.1128/JCM.00391-11
Biol Evol 6(12):3159–3170. Griffiths D, Fawley W, Kachrimanidou M, Bowden R,
gbe/evu248 Crook DW, Fung R, Golubchik T, Harding RM,
Eyre DW, Golubchik T, Gordon NC, Bowden R, Piazza P, Jeffery KJ, Jolley KA, Kirton R, Peto TE, Rees G,
Batty EM, Ip CL, Wilson DJ, Didelot X, O’Connor L, Stoesser N, Vaughan A, Walker AS, Young BC,
Lay R, Buck D, Kearns AM, Shaw A, Paul J, Wilcox Wilcox M, Dingle KE (2010) Multilocus sequence
MH, Donnelly PJ, Peto TE, Walker AS, Crook DW typing of Clostridium difficile. J Clin Microbiol 48
(2012) A pilot study of rapid benchtop sequencing of (3):770–778.
Staphylococcus aureus and Clostridium difficile for Hargreaves KR, Flores CO, Lawley TD, Clokie MR
outbreak detection and surveillance. BMJ Open 2(3): (2014) Abundant and diverse clustered regularly
e001124. interspaced short palindromic repeat spacers in Clos-
001124 tridium difficile strains and prophages target multiple
Eyre DW, Cule ML, Griffiths D, Crook DW, Peto TE, phage types within this pathogen. MBio 5(5):e01045–
Walker AS, Wilson DJ (2013a) Detection of mixed e01013.
infection from bacterial whole genome sequence data Hasselmayer O, Braun V, Nitsche C, Moos M, Rupnik M,
allows assessment of its role in Clostridium difficile von Eichel-Streiber C (2004) Clostridium difficile
transmission. PLoS Comput Biol 9(5):e1003059. IStron CdISt1: discovery of a variant encoding two complete transposase-like proteins. J Bacteriol 186
Eyre DW, Cule ML, Wilson DJ, Griffiths D, Vaughan A, (8):2508–2510
O’Connor L, Ip CL, Golubchik T, Batty EM, Finney He M, Sebaihia M, Lawley TD, Stabler RA, Dawson LF,
JM, Wyllie DH, Didelot X, Piazza P, Bowden R, Martin MJ, Holt KE, Seth-Smith HM, Quail MA,
Dingle KE, Harding RM, Crook DW, Wilcox MH, Rance R, Brooks K, Churcher C, Harris D, Bentley
Comparative Genomics of Clostridium difficile 73

SD, Burrows C, Clark L, Corton C, Murray V, Rose G, Kuehne SA, Cartman ST, Heap JT, Kelly ML, Cockayne
Thurston S, van Tonder A, Walker D, Wren BW, A, Minton NP (2010) The role of toxin A and toxin B
Dougan G, Parkhill J (2010) Evolutionary dynamics in Clostridium difficile infection. Nature 467
of Clostridium difficile over short and long time (7316):711–713.
scales. Proc Natl Acad Sci U S A 107(16):7527– Kuijper EJ, Barbut F, Brazier JS, Kleinkauf N, Eckmanns
7532. T, Lambert ML, Drudy D, Fitzpatrick F, Wiuff C,
He M, Miyajima F, Roberts P, Ellison L, Pickard DJ, Brown DJ, Coia JE, Pituch H, Reichert P, Even J,
Martin MJ, Connor TR, Harris SR, Fairley D, Mossong J, Widmer AF, Olsen KE, Allerberger F,
Bamford KB, D’Arc S, Brazier J, Brown D, Coia JE, Notermans DW, Delmee M, Coignard B, Wilcox M,
Douce G, Gerding D, Kim HJ, Koh TH, Kato H, Patel B, Frei R, Nagy E, Bouza E, Marin M, Akerlund
Senoh M, Louie T, Michell S, Butt E, Peacock SJ, T, Virolainen-Julkunen A, Lyytikainen O, Kotila S,
Brown NM, Riley T, Songer G, Wilcox M, Ingebretsen A, Smyth B, Rooney P, Poxton IR,
Pirmohamed M, Kuijper E, Hawkey P, Wren BW, Monnet DL (2008) Update of Clostridium difficile
Dougan G, Parkhill J, Lawley TD (2013b) Emergence infection due to PCR ribotype 027 in Europe, 2008.
and global spread of epidemic healthcareassociated Euro Surveill 13(31):18942
Clostridium difficile. Nat Genet 45(1):109–113. Kumar N, Miyajima F, He M, Roberts P, Swale A, Ellison L, Pickard D, Smith G, Molyneux R, Dougan G,
Janezic S, Rupnik M (2015) Genomic diversity of Clos- Parkhill J, Wren BW, Parry CM, Pirmohamed M,
tridium difficile strains. Res Microbiol 166(4):353– Lawley TD (2016) Genome-based infection tracking
360. reveals dynamics of Clostridium difficile transmission
Janezic S, Marin M, Martin A, Rupnik M (2015) A new and disease recurrence. Clin Infect Dis 62(6):746–
type of toxin A-negative, toxin B-positive Clostridium 752.
difficile strain lacking a complete tcdA gene. J Clin Kurka H, Ehrenreich A, Ludwig W, Monot M, Rupnik M,
Microbiol 53(2):692–695. Barbut F, Indra A, Dupuy B, Liebl W (2014) Sequence
JCM.02211-14 similarity of Clostridium difficile strains by analysis of
Janezic S, Potocnik M, Zidaric V, Rupnik M (2016) conserved genes and genome content is reflected by
Highly divergent Clostridium difficile strains isolated their ribotype affiliation. PLoS One 9(1):e86535.
from the environment. PLoS One 11(11):e0167101. Lemee L, Dhalluin A, Pestel-Caron M, Lemeland JF,
Janvilisri T, Scaria J, Thompson AD, Nicholson A, Pons JL (2004) Multilocus sequence typing analysis
Limbago BM, Arroyo LG, Songer JG, Grohn YT, of human and animal Clostridium difficile isolates of
Chang YF (2009) Microarray identification of Clos- various toxigenic types. J Clin Microbiol 42(6):2609–
tridium difficile core components and divergent 2617.
regions associated with host origin. J Bacteriol 191 2004
(12):3881–3891. Lim SK, Stuart RL, Mackin KE, Carter GP, Kotsanas D,
Jhung MA, Thompson AD, Killgore GE, Zukowski WE, Francis MJ, Easton M, Dimovski K, Elliott B, Riley
Songer G, Warny M, Johnson S, Gerding DN, TV, Hogg G, Paul E, Korman TM, Seemann T, Stinear
McDonald LC, Limbago BM (2008) Toxinotype V TP, Lyras D, Jenkin GA (2014) Emergence of a
Clostridium difficile in humans and food animals. ribotype 244 strain of Clostridium difficile associated
Emerg Infect Dis 14(7):1039–1045. with severe disease and related to the epidemic
10.3201/eid1407.071641 ribotype 027 strain. Clin Infect Dis 58(12):1723–
Kelly CP (2012) Can we identify patients at high risk of 1730.
recurrent Clostridium difficile infection? Clin Loo VG, Poirier L, Miller MA, Oughton M, Libman MD,
Microbiol Infect 18(Suppl 6):21–27. Michaud S, Bourgault AM, Nguyen T, Frenette C, Kelly
10.1111/1469-0691.12046 M, Vibien A, Brassard P, Fenn S, Dewar K, Hudson TJ,
Khanna S, Pardi DS (2012) Clostridium difficile Horn R, Rene P, Monczak Y, Dascal A (2005) A pre-
infection: new insights into management. Mayo Clin dominantly clonal multi-institutional outbreak of Clos-
Proc 87(11):1106–1117. tridium difficile associated diarrhea with high morbidity
mayocp.2012.07.016 and mortality. N Engl J Med 353(23):2442–2449. https://
Knetsch CW, Terveer EM, Lauber C, Gorbalenya AE,
Harmanus C, Kuijper EJ, Corver J, van Leeuwen HC Lyras D, O’Connor JR, Howarth PM, Sambol SP, Carter GP,
(2012) Comparative analysis of an expanded Clostrid- Phumoonna T, Poon R, Adams V, Vedantam G, Johnson
ium difficile reference strain collection reveals genetic S, Gerding DN, Rood JI (2009) Toxin B is essential for
diversity and evolution through six lineages. Infect virulence of Clostridium difficile. Nature 458
Genet Evol 12(7):1577–1585. (7242):1176–1179.
1016/j.meegid.2012.06.003 Mac Aogain M, Moloney G, Kilkenny S, Kelleher M,
Knight DR, Elliott B, Chang BJ, Perkins TT, Riley TV Kelleghan M, Boyle B, Rogers TR (2015) Whole-
(2015) Diversity and evolution in the genome of Clos- genome sequencing improves discrimination of
tridium difficile. Clin Microbiol Rev 28(3):721–741. relapse from reinfection and identifies transmission events among patients with recurrent Clostridium
74 S. Janezic et al.

difficile infections. J Hosp Infect 90(2):108–116. Pepin J, Valiquette L, Cossette B (2005) Mortality attrib- utable to nosocomial Clostridium difficile-associated
Makarova KS, Haft DH, Barrangou R, Brouns SJ, disease during an epidemic caused by a hypervirulent
Charpentier E, Horvath P, Moineau S, Mojica FJ, strain in Quebec. CMAJ 173(9):1037–1042. https://
Wolf YI, Yakunin AF, van der Oost J, Koonin EV
(2011) Evolution and classification of the CRISPR- Pougach K, Semenova E, Bogdanova E, Datsenko KA,
Cas systems. Nat Rev Microbiol 9(6):467–477. https:// Djordjevic M, Wanner BL, Severinov K (2010) Tran- scription, processing and function Of CRISPR
Makarova KS, Wolf YI, Koonin EV (2013) The basic cassettes in Escherichia coli. Mol Microbiol 77
building blocks and evolution of CRISPR-CAS (6):1367–1379.
systems. Biochem Soc Trans 41(6):1392–1400. 2010.07265.x Quesada-Gomez C, Lopez-Urena D, Chumbler N, Kroh
Marsden GL, Davis IJ, Wright VJ, Sebaihia M, Kuijper HK, Castro-Pena C, Rodriguez C, Orozco-Aguilar J,
EJ, Minton NP (2010) Array comparative Gonzalez-Camacho S, Rucavado A, Guzman-Verri C,
hybridisation reveals a high degree of similarity Lawley TD, Lacy DB, Chaves-Olarte E (2016) Anal-
between UK and European clinical isolates of hyper- ysis of TcdB proteins within the hypervirulent clade
virulent Clostridium difficile. BMC Genomics 11:389. 2 reveals an impact of RhoA glucosylation on Clos- tridium difficile proinflammatory activities. Infect
Mawer DPC, Eyre DW, Griffiths D, Fawley WN, Martin Immun 84(3):856–865.
JSH, Quan TP, Peto TEA, Crook DW, Walker AS, 01291-15
Wilcox MH (2017) Contribution to Clostridium diffi- Richter H, Zoephel J, Schermuly J, Maticzka D, Backofen
cile transmission of symptomatic patients with toxi- R, Randau L (2012) Characterization of CRISPR RNA
genic strains who are fecal toxin negative. Clin Infect processing in Clostridium thermocellum and
Dis 64(9):1163–1170. Methanococcus maripaludis. Nucleic Acids Res 40
cix079 (19):9887–9896.
McDonald LC, Killgore GE, Thompson A, Owens RC Jr, Rupnik M (2008) Heterogeneity of large clostridial
Kazakova SV, Sambol SP, Johnson S, Gerding DN toxins: importance of Clostridium difficile
(2005) An epidemic, toxin genevariant strain of Clos- toxinotypes. FEMS Microbiol Rev 32(3):541–555.
tridium difficile. N Engl J Med 353(23):2433–2441. Rupnik M, Widmer A, Zimmermann O, Eckert C, Barbut
Monot M, Eckert C, Lemire A, Hamiot A, Dubois T, F (2008) Clostridium difficile toxinotype V, ribotype
Tessier C, Dumoulard B, Hamel B, Petit A, Lalande 078, in animals and humans. J Clin Microbiol 46
V, Ma L, Bouchier C, Barbut F, Dupuy B (2015) (6):2146.
Clostridium difficile: new insights into the evolution Scaria J, Ponnala L, Janvilisri T, Yan W, Mueller LA,
of the pathogenicity locus. Sci Rep 5:15023. https:// Chang YF (2010) Analysis of ultra low genome con- servation in Clostridium difficile. PLoS One 5(12):
Moura I, Spigaglia P, Barbanti F, Mastrantonio P (2013) e15147.
Analysis of metronidazole susceptibility in different Sebaihia M, Wren BW, Mullany P, Fairweather NF,
Clostridium difficile PCR ribotypes. J Antimicrob Minton N, Stabler R, Thomson NR, Roberts AP,
Chemother 68(2):362–365. Cerdeno-Tarraga AM, Wang H, Holden MT, Wright
jac/dks420 A, Churcher C, Quail MA, Baker S, Bason N, Brooks
Moura I, Monot M, Tani C, Spigaglia P, Barbanti F, K, Chillingworth T, Cronin A, Davis P, Dowd L,
Norais N, Dupuy B, Bouza E, Mastrantonio P (2014) Fraser A, Feltwell T, Hance Z, Holroyd S, Jagels K,
Multidisciplinary analysis of a nontoxigenic Clostrid- Moule S, Mungall K, Price C, Rabbinowitsch E, Sharp
ium difficile strain with stable resistance to metronida- S, Simmonds M, Stevens K, Unwin L, Whithead S,
zole. Antimicrob Agents Chemother 58(8):4957– Dupuy B, Dougan G, Barrell B, Parkhill J (2006) The
4960. multidrug-resistant human pathogen Clostridium diffi-
Peng L, Pei J, Pang H, Guo Y, Lin L, Huang R (2014) cile has a highly mobile, mosaic genome. Nat Genet
Whole genome sequencing reveals a novel CRISPR 38(7):779–786.
system in industrial Clostridium acetobutylicum. J Ind Shin BM, Kuak EY, Yoo SJ, Shin WC, Yoo HM (2008)
Microbiol Biotechnol 41(11):1677–1685. https://doi. Emerging toxin A-B+ variant strain of Clostridium
org/10.1007/s10295-014-1507-3 difficile responsible for pseudomembranous colitis at
Pepin J, Valiquette L, Alary ME, Villemure P, Pelletier A, a tertiary care hospital in Korea. Diagn Microbiol
Forget K, Pepin K, Chouinard D (2004) Clostridium Infect Dis 60(4):333–337.
difficile-associated diarrhea in a region of Quebec diagmicrobio.2007.10.022
from 1991 to 2003: a changing pattern of disease Sim JH, Truong C, Minot SS, Greenfield N, Budvytiene I,
severity. CMAJ 171(5):466–472. Lohith A, Anikst V, Pourmand N, Banaei N (2017)
1503/cmaj.1041104 Determining the cause of recurrent Clostridium diffi-
cile infection using whole genome sequencing. Diagn
Comparative Genomics of Clostridium difficile 75

Microbiol Infect Dis 87(1):11–16. Valiente E, Dawson LF, Cairns MD, Stabler RA, Wren
1016/j.diagmicrobio.2016.09.023 BW (2012) Emergence of new PCR ribotypes from
Spigaglia P, Barbanti F, Mastrantonio P, European Study the hypervirulent Clostridium difficile 027lineage. J
Group on Clostridium difficile (2011) Multidrug resis- Med Microbiol 61(Pt 1):49–56.
tance in European Clostridium difficile clinical 1099/jmm.0.036194-0
isolates. J Antimicrob Chemother 66(10):2227–2234. Vedantam G, Clark A, Chu M, McQuade R, Mallozzi M, Viswanathan VK (2012) Clostridium difficile infec-
Stabler RA, Gerding DN, Songer JG, Drudy D, Brazier tion: toxins and non-toxin virulence factors, and their
JS, Trinh HT, Witney AA, Hinds J, Wren BW (2006) contributions to disease establishment and host
Comparative phylogenomics of Clostridium difficile response. Gut Microbes 3(2):121–134. https://doi.
reveals clade specificity and microevolution of hyper- org/10.4161/gmic.19399
virulent strains. J Bacteriol 188(20):7297–7305. Vonberg RP, Kuijper EJ, Wilcox MH, Barbut F, Tull P, Gastmeier P, European Cd-ICG, European Centre for
Stabler RA, He M, Dawson L, Martin M, Valiente E, Disease P, Control, van den Broek PJ, Colville A,
Corton C, Lawley TD, Sebaihia M, Quail MA, Rose Coignard B, Daha T, Debast S, Duerden BI, van den
G, Gerding DN, Gibert M, Popoff MR, Parkhill J, Hof S, van der Kooi T, Maarleveld HJ, Nagy E,
Dougan G, Wren BW (2009a) Comparative genome Notermans DW, O’Driscoll J, Patel B, Stone S,
and phenotypic analysis of Clostridium difficile 027 Wiuff C (2008) Infection control measures to limit
strains provides insight into the evolution of a hyper- the spread of Clostridium difficile. Clin Microbiol
virulent bacterium. Genome Biol 10(9):R102. https:// Infect 14(Suppl 5):2–20. 1469-0691.2008.01992.x
Stabler RA, Dawson LF, Valiente E, Cairns MD, Martin Warny M, Pepin J, Fang A, Killgore G, Thompson A,
MJ, Donahue EH, Riley TV, Songer JG, Kuijper EJ, Brazier J, Frost E, McDonald LC (2005) Toxin pro-
Dingle KE, Wren BW (2012) Macro and micro diver- duction by an emerging strain of Clostridium difficile
sity of Clostridium difficile isolates from diverse associated with outbreaks of severe disease in North
sources and geographical locations. PLoS One 7: America and Europe. Lancet 366(9491):1079–1084.
Cellular Uptake and Mode-of-Action
of Clostridium difficile Toxins

Panagiotis Papatheodorou, Holger Barth, Nigel Minton,

and Klaus Aktories

Abstract proportion of studies on C. difficile toxins were

Research on the human gut pathogen Clostrid- performed in European laboratories. In this
ium difficile and its toxins has gained much chapter we will highlight important recent
attention, particularly as a consequence of the advances in C. difficile toxins research.
increasing threat to human health presented by
emerging hypervirulent strains. Toxin A (TcdA) Keywords
and B (TcdB) are the two major virulence Clostridium difficile · Bacterial disease ·
determinants of C. difficile. Both are single- Bacterial toxins · Toxin uptake · Toxin
chain proteins with a similar multidomain archi- receptor
tecture. Certain hypervirulent C. difficile strains
also produce a third toxin, namely binary toxin
CDT (Clostridium difficile transferase). As
C. difficile toxins are the causative agents of
C. difficile-associated diseases (CDAD), such
as antibiotics-associated diarrhea and 1 Introduction
pseudomembranous colitis, considerable efforts
have been expended to unravel their molecular The human gut pathogen Clostridium difficile is
mode-of-action and the cellular mechanisms capable of producing at least three exotoxins,
responsible for their uptake. Notably, a high namely toxin A (TcdA), toxin B (TcdB) and the

H. Barth
P. Papatheodorou (*) Institute of Pharmacology and Toxicology, University of
Institute of Experimental and Clinical Pharmacology and Ulm Medical Center, Ulm, Germany
Toxicology, Albert Ludwig University of Freiburg,
Freiburg im Breisgau, Germany N. Minton
BBSRC/EPSRC Synthetic Biology Research Centre,
Institute of Pharmacology and Toxicology, University of University of Nottingham, Nottingham, UK
Ulm Medical Center, Ulm, Germany
K. Aktories
Faculty of Natural Sciences, University of Ulm, Ulm, Institute of Experimental and Clinical Pharmacology and
Germany Toxicology, Albert Ludwig University of Freiburg,
e-mail: Freiburg im Breisgau, Germany

# Springer International Publishing AG 2018 77

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
78 P. Papatheodorou et al.

binary toxin CDT (Clostridium difficile transfer- glucosylating toxins (CGTs)), such as Clostrid-
ase). The number of toxins and the quantities ium sordellii lethal toxin and hemorrhagic toxin,
produced vary between different C. difficile Clostridium novyi α-toxin and Clostridium
strains. Certain hypervirulent strains release all perfringens TpeL toxin (Voth and Ballard 2005;
three toxins during infection. Others produce Aktories et al. 2017). The large size of toxin A
strain-specific isoforms of toxin A and B (Rupnik and B led quite early to the assumption that both
and Janezic 2016). Toxin A and toxin B are toxins contain several domains with specific
related but they differ in structure and function functions during the intoxication process. Eventu-
from the binary toxin CDT. However, the three ally, a number of fundamental findings confirmed
toxins share some fundamental similarities dur- the modular composition of toxin A and B, which is
ing the intoxication process. All three toxins are also true for the other LCTs (Fig. 2).
released by the bacteria and enter into host cells
via receptor-mediated endocytosis. An enzymat- 2.1.1 The CROP Domain
ically active portion of the toxins then escapes
from acidified endosomes into the host cell cyto- At first, a region consisting of series of com-
sol in order to reach and modify its specific target bined, repetitive oligopeptides (CROP) was
proteins. In the case of toxin A and B, the identified and characterized in the C-terminal
enzyme portion is a glucosyltransferase that part of toxin A (Von Eichel-Streiber and
inactivates small GTPases of the Rho family. Sauerborn 1990; Von Eichel-Streiber et al.
The enzyme portion of CDT is an 1992b). In toxin A, the CROP domain makes up
ADP-ribosyltransferase that modifies monomeric nearly one-third of the complete protein and
G-actin. In the following sections, we will sum- consists of 7 long repeats of 30 residues and
marize the current knowledge about C. difficile 31 short repeats of 15–21 residues. In toxin B,
toxins´ cellular uptake and mode-of-action which the CROP domain contains 7 long repeats of
is fundamental for understanding their patho- 30 residues and only 21 short repeats of 20–23
physiological role in C. difficile infections residues and thus is significantly shorter than in
(CDI). A model of C. difficile toxins´ uptake toxin A. The number and length of the repeating
process and mode-of-action is depicted in Fig. 1. CROP modules have been found to vary between
toxins from different C. difficile isolates (Rupnik
et al. 1998). Historically, the CROP domain was
2 Structure, Uptake and Mode- considered to start around residue 1849 of toxin
of-Action of C. difficile Toxin A A and residue 1852 of toxin B, respectively.
and B However, according to more recent structural
studies by Orth et al. (2014), the CROP domain
2.1 Modular Composition starts at glycine-1832 for toxin A and at glycine-
of C. difficile Toxin A and B 1834 for toxin B.
A series of studies including monoclonal
Toxin A and B are large, single-chain protein antibodies or recombinant toxin fragments have
toxins that comprise several functional domains. provided evidence for a role of the CROP domain
The two toxins exhibit a high sequence homol- of toxin A in receptor binding (Frey and Wilkins
ogy (~50% amino acid identity) and an identical 1992; Sauerborn et al. 1997; Frisch et al. 2003).
multidomain architecture indicating that a gene In further studies, crystal structures of two
duplication event led to the existence of two C-terminal fragments (terminal 127 and
nearly-identical toxins in C. difficile (Von 255 residues) of toxin A were obtained, thus
Eichel-Streiber et al. 1992a). Both toxins are providing new insights into the overall structure
also highly similar to other large clostridial of the CROP domain (Ho et al. 2005; Greco et al.
toxins (LCTs; also denoted as clostridial 2006). The CROP domain of toxin A adopts a
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 79

Fig. 1 Model of the uptake process and mode-of-action receptors at the cell surface, are then taken up by
of toxin A/B and CDT. On the left part, the uptake and receptor-mediated endocytosis, form pores in endosomes
cellular action of toxin A and B are shown paradigmati- after acidification of the endosomal lumen and translocate
cally with the toxin B-specific receptor Frizzled. On the an enzyme domain into the cytosol. The detailed mecha-
right part, the uptake and cellular action of CDT are nism for the uptake process and the mode-of-action is
shown. Both types of toxins bind to their specific described in the main text of this review

Fig. 2 Multidomain
architecture of toxin A and
B. Shown is a schematic
representation of the
multidomain architecture
of toxin A and B and below
a 3D model of toxin A
obtained with negative
stain electron microscopy
(Pruitt et al. 2010) overlaid
with the crystal structure of
toxin A lacking the CROP
domain (Chumbler et al.
2016). EM structure of
toxin A was obtained with
publisher’s permission
from the following original
article: Pruitt et al. (2010)
80 P. Papatheodorou et al.

solenoid-like (screw-like) fold (Greco et al. co-substrate specificity (Jank et al. 2005); a
2006; Ho et al. 2005; Jank and Aktories 2008). four-helical-bundle subdomain at the
One of the two CROP structures was obtained by N-terminus of the glucosyltransferase is required
co-crystallization with the trisaccharide for the interaction with the inner plasma mem-
Galα1–3Galβ1–4GlcNAc, which was found to brane (Geissler et al. 2010). Additional essential
interact with toxin A in earlier reports (Greco amino acids for substrate binding were identified
et al. 2006; Krivan et al. 1986; Tucker and by Jank et al. (2007). In 2012, D’Urzo and
Wilkins 1991). However, this carbohydrate co-workers presented the crystal structure of the
structure is not present on human cells and thus glucosyltransferase domain of toxin A bound to
is unlikely to be part of intestinal receptors of Mn2+ and UDP-glucose (D’Urzo et al. 2012). In
toxin A in humans. The carbohydrate-binding the same year, Pruitt and colleagues succeeded in
properties of the CROP domain of toxin A were solving the structure of the glucosyltransferase
also supported by a study from Dingle et al. domain of toxin A in the presence and absence of
(2008). Notably, the CROP domain of toxin A its co-substrate UDP-glucose (Pruitt et al. 2012).
and B is similar to certain saccharide-binding Very recently, Alvin and Lacy reported new
proteins from Streptococcus downei and Strepto- crystal structures of the glucosyltransferase
coccus mutans (Wren 1991). domains of toxin A and B in complex with a
non-hydrolysable UDP-glucose analogue and an
2.1.2 The Glucosyltransferase Domain apo-like structure of the glucosyltransferase
domain of toxin B (Alvin and Lacy 2017).
In 1995, the group of Klaus Aktories (Freiburg,
Germany) found that toxin A and B modify the 2.1.3 The Cysteine Protease Domain
small GTPase Rho and other members of the Rho
subfamily via transfer of the glucose moiety from In 2003, Barth and colleagues showed with toxin
the co-substrate UDP-glucose to threonine-37 of B that only the N-terminal glucosyltransferase
the GTPase (Just et al. 1995a, b). Thus, it became domain reaches the cytosol after completion of
apparent that toxin A and B are bacterial the uptake process (Pfeifer et al. 2003). Thus, it
glucosyltransferases capable of inactivating was feasible that processing of toxin A and B is a
small GTPases of host cells. Deletion analyses prerequisite of the intoxication process. The
from Hofmann et al. with toxin B revealed cleavage site of toxin B was identified between
glucosyltransferase activity in the N-terminal leucine-543 and glycine-544 (Rupnik et al.
part of the toxin (Hofmann et al. 1997). In 2005). Yet it was not clear whether the
2005, the crystal structure of the glucosyl- processing of toxin A and B occurs by a host
transferase domain of toxin B in the presence of protease or an internal domain of the toxins.
UDP-glucose and Mn2+ was determined (Reinert Eventually, the group of Eichel-Streiber
et al. 2005). It became obvious from the 3D (Mainz, Germany) identified a small cytosolic
structure that the glucosyltransferase domain of compound, namely inositol hexakisphosphate
toxin B belongs to the glucosyltransferase type A (InsP6), which is capable of inducing autocata-
family. Subsequent biochemical studies revealed lytic processing of toxin A and B (Reineke et al.
important residues that are crucial for the enzy- 2007). However, it was still not clear how
matic activity: residues 364–516 are important processing occurred. This question was
for substrate recognition (Hofmann et al. 1998); answered, when a cysteine protease domain,
an essential and highly conserved DXD motif which is located adjacent to the glucosyl-
between amino acids 286 and 288 is involved in transferase domain, was identified by Egerer
binding Mn2+ (Busch et al. 1998); residue et al. (2007) in toxin A and B. A fragment of
tryptophan-102 is involved in UDP-glucose toxin B comprising only the glucosyltransferase
binding (Busch et al. 2000); isoleucine-383 and and the cysteine protease domain is cleaved in
glutamine-385 are crucial residues for the the presence of InsP6, indicating that InsP6
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 81

induces autocatalytic processing of toxin A and such as leucine-1106. By a series of C-terminal

B by activating the cysteine protease domain. deletions of toxin B that were fused to the
Lysine-600 of the cysteine protease domain is receptor-binding domain of the diphtheria toxin
essential for InsP6-binding, whereas cysteine- (DTRD), Genisyuerek et al. identified that amino
698, histidine-653, or aspartate-587 of toxin B acids 830–1550 of the toxin is sufficient for
represent the catalytic triad of the protease translocation of the enzyme portion into the cyto-
(Egerer et al. 2007, 2009). A first 3D structure sol, assuming that the region between amino
of the cysteine protease domain (bound to InsP6) acids 1551 and 1834 (start of the CROP domain)
was provided for toxin A in 2009 by the group of is not part of the translocation domain.
Borden Lacy. The crystal structure uncovered a
highly basic pocket that is required for InsP6- 2.1.5 Additional Receptor-Binding
binding, which is separated from the active site Domains
by a beta-flap structure (Pruitt et al. 2009). Later, Given the fact that the translocation domain of
the 3D structure of the InsP6-bound cysteine pro- toxin A and B is much shorter than previously
tease domain of toxin B was presented either in assumed, the question remains about the function
the absence (Shen et al. 2011) or in the presence a of the remaining toxin segment between the
specific small molecule inhibitor, respectively translocation domain and the CROP domain.
(Puri et al. 2010). It became apparent from these Recent discoveries suggest that this domain is
studies that InsP6-binding allosterically improves involved in binding of toxin A and B to the cell
the access of the active site to its substrate. Very surface (Gerhard 2016). Already in 1994,
recently, a structural study from Chumbler et al. Barroso et al. tested various C-terminally
(2016) revealed the requirement for zinc in the truncated toxin B variants in intoxication assays
mechanism of autoprocessing of toxin A and B. and found that removal of the CROP domain did
not fully diminish cytotoxicity (Barroso et al.
2.1.4 The Translocation Domain 1994). In this study, the authors did not use
purified proteins but lysates from E. coli that
During cellular uptake, toxin A and B are trapped expressed the various toxin B variants. Later,
in endosomes and presumably form pores, which Frisch et al. (2003) observed that an
allow the translocation of the glucosyltransferase N-terminally extended CROP domain of toxin
domain into the cytosol. A relatively large region A competitively inhibited intoxication of cells
between the cysteine protease domain and the by toxin A more efficiently than the CROP
CROP domain of toxin A and B, denoted as domain alone. Eventually, two German
translocation domain, was initially suggested to laboratories from Freiburg (Aktories and
be involved in these processes (Dove et al. 1990; Papatheodorou) and Hanover (Just and Gerhard)
Von Eichel-Streiber et al. 1992a; Barroso et al. confirmed in 2011 with purified recombinant
1994). In 2011, Genisyuerek et al. aimed to more proteins that the CROP domain is not absolutely
precisely narrow down the pore-forming region required for binding and uptake of toxin A and B
and the translocation domain of toxin B. They into host cells (Genisyuerek et al. 2011; Olling
found that a small segment reaching from amino et al. 2011). The concept of CROP-independent
acid residues 830–990 of toxin B is already suf- binding and uptake of toxin A and B was further
ficient for pore formation, at least in artificial supported by the identification of the homolo-
lipid bilayers (Genisyuerek et al. 2011). In addi- gous TpeL toxin from C. perfringens, which is
tion, the authors found that the residues naturally devoid of a CROP domain (Amimoto
glutamate-970 and glutamate-976 of toxin B et al. 2007). Schorch et al. (2014) substantiated
were crucial for pore formation by acting as pH that the C-terminus of TpeL represents its
sensors for membrane insertion. Zhang et al. receptor-binding domain by identifying the
(2014) identified additional amino acids that are LDL-related lipoprotein receptor 1 (LRP1) as
crucially involved in pore formation of toxin B, host receptor for TpeL and by showing direct
82 P. Papatheodorou et al.

binding between the TpeL C-terminus and an of toxin A and B had already become evident in
extracellular portion of LRP1. In the same earlier attempts to obtain low resolution structures
study, the authors also proved independent cell of the holotoxins by small-angle X-ray scattering
surface-binding of a fragment of toxin B cover- (SAXS) and negative stain electron microscopy,
ing residues 1349–1811, which virtually respectively (Albesa-Jove et al. 2010; Pruitt et al.
corresponds to the proposed receptor-binding 2010). In 2016, the group of Borden Lacy reported
domain of TpeL. Furthermore, the authors were the long-sought crystal structure of toxin
able to competitively inhibit cell binding of A. Despite the fact that the structure of toxin A
CROP-deficient toxin B by co-incubation with obtained in this study did not include the CROP
this fragment. These data argued strongly for a domain, it showed for the first time how the other
two-receptor model of toxin A and B, where the domains are organized within the holotoxin. In
toxins independently bind host receptors via the addition, the structure included additional
CROP domain or the newly defined receptor- domains of toxin A whose structure had not been
binding domain. Recently, Lambert and Baldwin solved so far, such as the translocation domain and
provided additional direct evidence for dual the newly discovered, second receptor-binding
receptor-binding sites in toxin A (Lambert and domain (Chumbler et al. 2016).
Baldwin 2016). Confusingly enough, experimen-
tal data from a recent work by Manse and
Baldwin suggested at least three independent 2.2 Binding and Uptake of C. difficile
binding sites in toxin B (Manse and Baldwin Toxin A and B
2015). Beside the CROP domain, the newly
defined receptor-binding domain, which 2.2.1 Host Receptors of Toxin A and B
precedes the CROP domain, was shown to harbor
two independent regions (residues 1372–1493 Toxin A was found to interact with different cell
and 1493–1848) with cell binding-capability. surface carbohydrate structures and with two
However, it is not clear yet whether two indepen- proteins, namely the sucrase-isomaltase and the
dent binding sites are also present in the CROP- glycoprotein gp96 (Gerhard 2016). More
preceding receptor-binding domain of toxin A or recently, powerful genetic screens were
other LCTs. Eventually, the recent identification established that finally allowed the discovery of
of toxin B receptors that bind to the newly host receptors of toxin B, such as CSPG4 (chon-
defined receptor-binding domain (described in a droitin sulphate proteoglycan-4), PVRL3 (polio-
following section of this chapter) constitutes the virus receptor-like 3) and members of the Wnt
strongest evidence for the existence of additional receptor frizzled family, such as FZD2 (Yuan
binding sites outside of the CROP domain. et al. 2015; LaFrance et al. 2015; Tao et al.
2016). Recently, two binding sites were
2.1.6 Modular Structure (ABCD Model) postulated within the newly defined receptor-
binding domain of toxin B. Toxin B region
On the basis of the different domains of toxin A 1372–1493 is bound by PVRL3 and toxin B
and B that have been described above, the modu- region 1501–1830 by FZD proteins, respectively,
lar composition of toxin A and B is best described whereas CSPG4 is a CROP-dependent receptor
with the so-called ABCD model already (Manse and Baldwin 2015; Tao et al. 2016).
suggested by Jank and Aktories in 2008. In the
ABCD model, A stands for biological activity 2.2.2 Endocytic Pathways
(glucosyltransferase domain), B for binding for the Cellular Uptake of Toxin
(CROP domain and preceding additional binding A and B
sites), C for cutting (cysteine protease domain), Upon binding to a cell surface receptor, toxin A
and D for delivery (translocation domain) (Jank and B are taken up into host cells via receptor-
and Aktories 2008). The multidomain architecture mediated endocytosis. For many years, the exact
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 83

endocytic pathway for the uptake of toxin A and for the insertion of the toxins into the endosomal
B remained unclear. At first glance, Kushnaryov membrane (Qa’Dan et al. 2000, 2001). Low
and Sedmark provided evidence for endocytosis pH-dependent pore formation of toxin A and B
of C. difficile toxin A via coated pits, by in cellular and artificial membranes was con-
visualizing colloidal gold labelled toxin A in firmed by the Aktories group (Barth et al. 2001;
CHO cells by electron microscopy (Kushnaryov Giesemann et al. 2006). Formation of a pore in
and Sedmark 1989). In 2010, Papatheodorou the endosomal membrane by the toxins´ translo-
et al. aimed to study the endocytic uptake of cation domain might be an essential step for the
toxin A and B in more detail by the use of delivery of the glucosyltransferase domain into
pharmacological and genetic inhibitors of dis- the cytosol. It is generally assumed that toxin A
tinct endocytic pathways (Papatheodorou et al. and B are able to form membrane pores as
2010). Their findings indicated that the endocytic monomers and independent of host cell proteins.
uptake of toxin A and B involves a dynamin- Pore formation of toxin A and B can be forced to
dependent process that is mainly governed by occur also at the plasma membrane by artificially
clathrin (Papatheodorou et al. 2010). Gerhard acidifying the extracellular medium of cultured
and colleagues confirmed that clathrin and cells (Barth et al. 2001; Giesemann et al. 2006;
dynamin are substantially involved in endocyto- Qa’Dan et al. 2000). Giesemann et al. could
sis of toxin A and toxin A1–1874 (lacking almost show that the efficacy of pore formation by toxin
the entire CROP domain). However, as inhibition A and B was dependent on membrane cholesterol
or knockdown of clathrin did not completely (Giesemann et al. 2006). The glucosyltransferase
prevent uptake of toxin A and toxin A1–1874, the domain is not required for pore formation of toxin
authors suggested alternative endocytic routes A and B at the plasma membrane or in artificial
for the toxin (Gerhard et al. 2013). Indeed, lipid bilayers (Barth et al. 2001; Genisyuerek et al.
Chandrasekaran et al. (2016) reported very 2011). Black lipid bilayer experiments with
recently that the uptake of toxin A into CaCo- purified toxins revealed that the pores formed by
2 and MEF cells is clathrin-independent but toxin A and B are more of a temporary nature and,
requires dynamin and the Fer-CIP4 homology- presumably, less stable than pores formed by other
BAR (F-BAR) domain-containing protein classical pore-forming bacterial toxins (Barth et al.
PACSIN2. 2001; Genisyuerek et al. 2011). Most likely, the
glucosyltransferase domain of toxin A and B needs
2.2.3 Delivery to be unfolded during the translocation process.
of the Glucosyltransferase However, it remains an open question as to how
Domain into the Cytosol unfolding of the glucosyltransferase domain is
Toxin A and B are so-called ‘short-trip’ toxins, initiated and whether the unfolded glucosyl-
which deliver their enzymatic portion into the transferase domain dips into the membrane pore
cytosol directly after reaching endosomal via its N- or C-terminus. In addition, it is not clear if
compartments via receptor-mediated endocyto- the glucosyltransferase domain translocates across
sis. The translocation of the glucosyltransferase the membrane pore alone or together with the
domain across the endosomal membrane is by far adjacent cysteine protease domain.
the least understood step of the intoxication pro-
cess of toxin A and B, respectively. This is
mainly due to the lack of structural information 2.3 Mode-of-Action of Toxin A and B
of membrane-embedded conformations of the
toxins, either prior or directly after the transloca- C. difficile toxin A and B were the first toxins to
tion event. Acidification of endosomal vesicles be shown to modify target proteins by glycosyla-
by vacuolar H+-ATPases triggers conformational tion (Just et al. 1995a, b). Meanwhile, it is clear
changes within toxin A and B, leading to the that this type of post-translational modification is
exposure of hydrophobic segments responsible used by many toxins to interfere with eukaryotic
84 P. Papatheodorou et al.

cell functions, including various types of large Rho proteins are extracted from membranes by
clostridial glucosylating toxins (Just et al. 1996; GDIs (guanine nucleotide dissociation inhibitors)
Jank et al. 2015a; Jank and Aktories 2008) but and are in a GDI-Rho complex in the cytosol.
also toxins from Legionella (Belyi et al. 2006), C. difficile toxins glucosylate Rho proteins in
Photorhabdus (Jank et al. 2013), Yersinia (Jank threonine37, and Rac and Cdc42 in threonine35,
et al. 2015b) and E. coli (EPEC) (Li et al. 2013) which is the equivalent residue (Just et al. 1995a,
species. Toxin A and B catalyze the b). This modification blocks the signal/switch
glucosylation of Rho GTPases by utilizing functions of Rho proteins, because they are no
UDP-glucose as a co-substrate (Just et al. longer able to interact with effectors.
1995a, b). Other related clostridial glycosyl- Glucosylation inhibits the activation of Rho
transferases (e.g., C. novyi α-toxin and GTPases by GEF proteins, and completely
C. perfringens TpeL) prefer UDP-N-acetylglu- blocks the interaction with GAPs (Sehr et al.
cosamine (UDP-GlcNAc) (Selzer et al. 1996; 1998). Moreover, glucosylation fixes Rho
Guttenberg et al. 2012; Nagahama et al. 2011). proteins in their inactive conformation (Vetter
Primary substrates of toxin A and B are RhoA,B, et al. 2000; Geyer et al. 2003). Additionally, it
C, Rac1,2 and Cdc42 but also other isoforms of was shown that glucosylated Rho proteins
the Rho family such as TC10 and RhoG are remain attached to the cell membrane and are
modified. Secondary substrates are also some not extracted from membranes by GDI proteins
Ras proteins like Rap1,2, Ral, and Ras (Just and (Genth et al. 1999).
Gerhard 2004; Zeiser et al. 2013). Rho proteins Because glucosylation of Rho proteins blocks
are 21–25 kDa GTP-binding proteins and all functions of the switch proteins, C. difficile
members of the Ras superfamily. The ~20 Rho toxins A and B affect numerous cellular
family members are switch proteins governed by functions. Therefore, important questions are:
a GTPase cycle and act as master regulators of How is the action of the toxins related to their
the actin cytoskeleton and of numerous cellular pathophysiological effects? What kind of actions
processes, such as cell migration, phagocytosis of toxins A and B result in diarrhea, inflamma-
and intracellular traffic, cell cycle progression tion and enterocolitis, which are the major
and apoptosis (Nobes and Hall 1994; Burridge symptoms of C. difficile infection?
and Wennerberg 2004; Jaffe and Hall 2005; Cytopathological effects of toxins A and B are
Aktories 2011; Lemichez and Aktories 2013). characterized by gross changes in cell morphol-
Rho proteins are inactive in the GDP-bound ogy, redistribution of the actin cytoskeleton, loss
state and become activated after nucleotide of stress fibers and retraction of the cell body
exchange and GTP-binding (Cherfils and Zeghouf with remaining irregular cell extensions, a pro-
2013; Bishop and Hall 2000). This GDP/GTP cess, which was called arborisation (Fiorentini
exchange is mediated by numerous guanine nucle- and Thelestam 1991; Ottlinger and Lin 1988).
otide exchange factors (GEFs) (Garcia-Mata and All these effects can be referred to inhibition of
Burridge 2007). Active Rho proteins interact with Rho protein functions. Especially, glucosylation
various effector proteins to elicit cellular functions of Rac appears to be essential for the cytopathic
(Bishop and Hall 2000; Burridge and Wennerberg effects of toxins A and B (Halabi-Cabezon et al.
2004). This active state is blocked by GTP 2008). The RacQ61L mutant, which is hardly
hydrolyses, which is stimulated by various modified by the toxins, prevents cytopathic
GTPase-activating proteins (GAPs) (Tcherkezian effects. The toxins alter cell-cell contacts and
and Lamarche-Vane 2007; Cherfils and Zeghouf cell adhesion, which also depend on Rho
2013). Active GTP-bound Rho proteins are cell proteins, thereby barrier functions of enterocytes
membrane associated, which is caused by are disabled (Hecht et al. 1988, 1992; Nusrat
N-terminal isoprenylation. Inactive, GDP-bound et al. 2001; Nusrat et al. 1995). The functional
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 85

consequences are paracellular fluxes as a conse- massive release of cellular content that can induce
quence of alteration of tight junctions, which strong inflammation (Miao et al. 2010; Jorgensen
depend on Rho and actin (Nusrat et al. 1995; and Miao 2015).
Hirase et al. 2001). While the above mentioned toxin actions
Toxin A and B were shown to induce apoptosis depend on the glucosyltransferase activity of
in several types of cells (Mahida et al. 1996; Brito toxin A and B, toxin effects have been described
et al. 2002; Qa’Dan et al. 2005; Fiorentini et al. which reportedly occur with “glucosyl-
1998). Induction of apoptosis (at least at low and transferase-dead” toxins. For example it has
moderate toxin concentrations) essentially depends been reported that toxin-induced production of
on the glucosyltransferase activity of the toxins reactive oxygen species (ROS) participate
(Brito et al. 2002; Gerhard et al. 2008). enteritis and necrosis caused by C. difficile
Ng and coworkers reported that toxin A and B toxins (Qiu et al. 1999; Farrow et al. 2013;
induce inflammasome activation in an ASC Wohlan et al. 2014; Donald et al. 2013). How-
(apoptosis-associated speck-like protein)- ever, these toxin effects occurred at very high
dependent manner, thereby causing the release concentrations of toxins (often 100–1000 times
of IL-1β (Ng et al. 2010). More recently, the higher than that necessary for cytopathic
group of Feng Shao showed that Pyrin, which is effects). Therefore, the pathophysiological rele-
encoded by the Mediterranean fever gene vance is not clear.
MEFV, acts as an intracellular “sensor” for
toxin-modified RhoA-dependent inflammasome
activation (Xu et al. 2014). Pyrin associates
2.4 Relative Importance of Toxin A
with the ASC adaptor protein thereby activating
and B in Clostridium difficile
pro-caspase 1 (Lu and Wu 2015). Caspase-1 is a
key enzyme to activate IL-1β and IL-18, the final
common path of inflammasome activation.
Historically, symptoms of CDI were mainly
Inflammasome formation appears to be regulated
attributed to the action of toxin A, due to the
by phosphorylation of Pyrin and binding to 14-3-
fact that only purified toxin A but not toxin B
3 proteins that keeps Pyrin in an inactive state
was able to cause disease symptoms in hamsters
(Gao et al. 2016). Moreover, it was reported that
when applied intragastrically (Lyerly et al.
Pyrin is phosphorylated by Rho effector protein
1988). However, C. difficile strains have been
kinase N (PKN), resulting in binding to 14-3-3
isolated from symptomatic patients that produce
proteins and inhibition of inflammasome activa-
only toxin B (Lyerly et al. 1992; Kim et al. 2012).
tion (Park et al. 2016). Toxin-induced activation
Thus, two previous studies from the laboratories
and release of IL-1β can induce release of IL-6,
of Nigel Minton (Nottingham, UK) and Julian
interferon-γ (IFN-γ) and IL-8, respectively. IL-8
Rood (Melbourne, Australia) have attempted to
is a highly potent neutrophil attractant. This is in
more precisely determine in the hamster infec-
line with the strong neutrophil invasion into
tion model the in vivo relevance of toxin A and
colon mucosa that occurs during C. difficile infec-
B. To this end, both laboratories generated iso-
tion and which is probably essentially involved in
genic C. difficile mutants in the same strain
mucosal damage (Linevsky et al. 1997; Warny
(C. difficile 630) defective in the production of
et al. 2000; Ishida et al. 2004; Jafari et al. 2013;
either toxin A or toxin B. Whereas both studies
Steiner et al. 1997; Mahida et al. 1996). An addi-
showed that toxin B alone causes disease
tional recent finding is of interest, where it was
symptoms in hamsters, contradictory results
shown that the pyrin inflammasome triggers
were obtained in terms of the importance of
pyroptosis (Russo et al. 2016). Pyroptosis is fea-
toxin A. Whereas a toxin B mutant created in
tured by cell swelling followed by cell lysis with
the Rood group and which was capable of
86 P. Papatheodorou et al.

producing only toxin A did not cause disease in ADP-ribosyltransferases that resemble anthrax
hamsters (Lyras et al. 2009), the equivalent toxin of Bacillus anthracis with respect to their
mutant from the Minton group remained virulent binding components. For instance, CDTb
(Kuehne et al. 2010). Compelling evidence has exhibits a 36% identity to protective antigen
been provided recently by the Minton group that (PA), the binding component of anthrax toxin
the reason for the observed contradiction resides (Young and Collier 2007). Much that we know
in the use of two different erythromycin-sensitive about the structure-to-function relationship of
derivatives of strain 630 for mutagenesis, which CDTb was learned from previous extensive stud-
are genetically and phenotypically distinct. ies on the binding components of the anthrax
Unique Single Nucleotide Polymorphisms toxin (PA) and, in part, the C2 toxin (C2II).
(SNPs) were identified in both strains that dra- From the already available structures of PA
matically affected certain phenotypes, as well (Schleberger et al. 2006; Petosa et al. 1997), it
having marked effects on the transcriptome, was possible to deduce that CDTb consists of
which most likely impact on virulence (Collery four domains (I to IV) with distinct functions.
et al. 2017). The recent isolation of a toxin Domain I at the N-terminus forms the activation
A-positive, toxin B-negative C. difficile strain domain and is followed by Domain II, which is
from a clinical case of CDI further supports the involved in membrane insertion and pore forma-
in vivo relevance of toxin A (Monot et al. 2015). tion. Domain III is responsible for pore formation
and oligomerization. The C-terminal Domain IV
corresponds to the receptor-binding domain of
3 Structure, Uptake and Mode- CDTb (Barth et al. 2004). Domain IV is highly
of-Action of CDT (C. difficile similar among the binding components of CDT
Transferase) (CDTb), CST (CSTb) and iota toxin (Ib). Inter-
estingly, binding and enzymatic components are
3.1 Bipartite Composition of CDT mutually interchangeable among CDT, CST and
iota-toxin, but not among the latter toxins and the
In contrast to toxin A and B, CDT is an AB-type C2 or anthrax toxin (Considine and Simpson
binary toxin composed of a binding and translo- 1991; Popoff and Boquet 1988).
cation component (CDTb) and a separate enzyme CDTb is expressed as a precursor protein of
component (CDTa). CDTb mediates binding to 876 amino acids (~90 kD) including an
the host cell surface, internalization of CDTa into N-terminal signal peptide. Serine-type proteases
endocytic vesicles and pore formation in activate the CDTb precursor by removal of a
endosomes for the translocation of CDTa into 20 kD peptide from the N-terminus (Perelle
the cytosol of host cells. Pore formation of et al. 1997). The activated binding component
CDTb is accomplished by oligomerization of has as a size of ~75 kD and is now able to form
CDTb into heptamers that are capable of heptamers. It is unclear, whether the activation
integrating into the endosomal membrane. and oligomerization process occurs prior or after
CDTa is an ADP-ribosyltransferase that is spe- binding of the CDTb precursor to host cells
cific for monomeric G-actin. (Gerding et al. 2014).

3.1.1 The Binding Component of CDT 3.1.2 The Enzyme Component of CDT

CDT is most similar to other clostridial binary The enzyme component of CDT (CDTa) has a
toxins, such as Clostridium perfringens iota- size of ~53 kD and consists of 463 amino acids,
toxin and Clostridium spiroforme toxin CST, including an N-terminal signal sequence of
and more distantly related to Clostridium botuli- 43 amino acids, which is probably cleaved by
num C2 toxin. All those toxins are actin proteolysis (Perelle et al. 1997). The mature
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 87

CDTa finally has a size of ~48 kD (420 amino receptor in the liver for the clearance of chylomi-
acids) and is most similar to the enzyme cron remnants from the blood, but is also
components of iota-toxin (Ia; 84% sequence expressed in various other tissues, including the
identity) and CST (CSTa; 82% sequence iden- intestine (Yen et al. 1994, 1999; Mesli et al.
tity). CDTa consists of two domains with similar 2004). Later studies identified a role of LSR in
folding, which might originate from a duplication the formation of tricellular tight junctions
process of an ancient ADP-ribosyltransferase gene (Masuda et al. 2011; Furuse et al. 2012; Czulkies
(Han et al. 1999). Amino acids 1–215 of mature et al. 2017). Another recent study found that LSR
CDTa are probably involved in the interaction with is critically required for proper blood-brain bar-
CDTb, whereas amino acids 224–420 harbour the rier formation (Sohet et al. 2015). Eventually,
catalytically active ADP-ribosyltransferase portion. several studies found a role of LSR in cancer
CDTa belongs to the R-S-E class of progression and metastasis (Papatheodorou and
ADP-ribosyltransferases, which are characterized Aktories 2016). As shown by Hemmasi et al.
by the presence of a typical arginine residue (R), (2015), amino acids 757–866 at the C-terminal
an STS motif (S) and an EXE motif (E). So far, the end of CDTb interact with an immunoglobulin
enzyme component of the iota-toxin has been (Ig)-like, V-type domain of LSR present in its
crystallized either in the presence of a stable NAD N-terminal, extracellular part.
analogue (Tsuge et al. 2008) or in complex with
actin (Tsurumura et al. 2013). Recently, NMR 3.2.2 Endocytic Pathways
assignments were reported for the CDTb- for the Cellular Uptake of CDT
interacting and the active portion of CDTa (Roth Until now, the endocytic route of CDT (and other
et al. 2016a, b). iota-like toxins) has not been entirely clarified.
However, it was shown by the group of Michel
Popoff (Paris, France) that dynamin, but not
3.2 Binding and Uptake of CDT clathrin, is required for cellular uptake of iota-
toxin (Gibert et al. 2011). In this study,
3.2.1 The Lipolysis-Stimulated colocalisation of iota-toxin with the interleukin-
Lipoprotein Receptor 2 receptor in endocytic vesicles was observed,
As for toxin A and B, binding to a specific struc- indicating a similar endocytic route for both
ture at the cell surface of host cells is a prerequi- proteins (Gibert et al. 2011). The endocytic
site of the intoxication process of CDT. CDT uptake of the interleukin-2 receptor is negatively
belongs to the iota-like toxins, a subfamily of regulated by RhoGDI (RhoGDP-dissociation
the family of clostridial, binary actin inhibitor) (Lamaze et al. 2001). Strikingly, iota-
ADP-ribosylating toxins. It was already known toxin entry into Cos-1 cells was inhibited upon
from a previous study that iota-like toxins use a overexpression of RhoGDI (Gibert et al. 2011).
proteinaceous receptor for cell entry (Stiles et al. Endocytic uptake of CDT and other iota-like
2000). Eventually, in 2011, the LSR (lipolysis- toxins might involve lipid rafts, since oligomers
stimulated lipoprotein receptor) was identified as of the binding components have been identified
host receptor for iota-like toxins by the help of a in detergent-resistant, cholesterol-rich membrane
novel genetic screen (haploid genetic screen), microdomains (Nagahama et al. 2004; Hale et al.
which is based on the human haploid cell line 2004). Importantly, Papatheodorou and
Hap1 (Papatheodorou et al. 2011). Interestingly, colleagues observed clustering of LSR into lipid
it turned out that LSR is the host receptor also for rafts after binding of CDTb (Papatheodorou et al.
the CDT-related C. perfringens iota-toxin and 2013). LSR-clustering into lipid rafts occurred
C. spiroforme toxin but not for the more distantly also after binding of the RBD of CDTb, which
related C. botulinum C2 toxin (Papatheodorou is not able to oligomerize by itself
et al. 2011, 2012). LSR acts as a lipoprotein (Papatheodorou et al. 2013). Wigelsworth et al.
88 P. Papatheodorou et al.

found that the lipid rafts-protein CD44 (cluster of by other binary actin-ADP-ribosylating toxins,
differentiation 44) is required for cellular uptake including C. botulinum C2 toxin and
of CDT (Wigelsworth et al. 2012). Interestingly, C. perfringens iota toxin (Vandekerckhove
CD44 was found in lipid rafts from Ib-treated et al. 1987, 1988). In contrast to monomeric
Vero cells (Blonder et al. 2005). It might be G-actin, polymerized F-actin is not a substrate
possible that CD44 interacts with LSR-CDT of CDT and of any other related binary toxin,
complexes in lipid rafts, thus facilitating the because arginine-177 is not available for modifi-
endocytic uptake of the toxin. cation in the double helix of F-actin (Holmes
et al. 1990; Margarit et al. 2006). Essential for
3.2.3 Role of Chaperones During actin functions is the ability of the microfilament
the Cellular Uptake of CDT protein to reversibly polymerize from G- to
The delivery of CDTa into the host cell cytosol F-actin, a process that is tightly regulated by
depends on CDTb, which under acidic conditions numerous actin binding proteins (Dominguez
likely forms pores in endosomal membranes that and Holmes 2011). Early studies obtained with
serve as translocation channels for the trans- C. botulinum C2 toxin and C. perfringens iota
membrane transport of CDTa (Roeder et al. toxin showed that modification of actin in
2014; Ernst et al. 2016). The pH-driven transport arginine-177 inhibits actin polymerization
of CDTa across endosomal membranes requires (Aktories et al. 1986; Schering et al. 1988).
the activities of certain host cell chaperones This holds also true for CDT-induced
(Roeder et al. 2014). In vitro, CDTa directly ADP-ribosylation of actin. Moreover, all previ-
and specifically binds to the heat shock proteins ous results obtained with other types of binary
Hsp90 and Hsp70, as well as to some peptidyl- actin-ADP-ribosylating toxins that modify argi-
prolyl cis/trans isomerases (PPIases) of the nine177 of actin can be reliably referred to the
cyclophilin (Cyp) and FK506 binding protein action of CDT. This includes the early finding
(FKBP) families (Kaiser et al. 2011; Ernst et al. that ADP-ribosylated actin binds to plus ends of
2015, 2017). The current model suggests that F-actin filaments and acts as a capping protein to
these host cell factors specifically and selectively block F-actin elongation by inhibition of the
facilitate the intracellular trans-membrane trans- binding of non-ADP-ribosylated actin (Aktories
port of ADP-ribosylating toxins by interacting and Wegner 1989; Perieteanu et al. 2010; Weigt
with the ADP-ribosyltransferase domain of the et al. 1989; Wegner and Aktories 1988). Also the
A subunits. These findings were mainly obtained interaction of actin with actin binding proteins
by the group of Holger Barth (Ulm, Germany) (for example gelsolin) that is largely affected by
and contribute to a better understanding of the toxin-induced ADP-ribosylation (Wille et al.
cellular uptake of CDT into human cells and to 1992), is similarly relevant for CDT.
the development of novel pharmacological Binary toxin-induced F-actin depolymerization
strategies against infections with hypervirulent, has typical cytotoxic effects in cell culture
CDT-producing C. difficile strains. Host cyto- (Wiegers et al. 1991), resulting in rounding-up of
solic factors that might assist during refolding cells and loss of cell adherence followed by apo-
of the translocated glucosyltransferase domain ptosis (Heine et al. 2008). Notably, not only the
of toxin A and B have yet to be described. actin cytoskeleton but also microtubules are
affected by binary actin-depolymerizing toxins.
CDT and other actin-depolymerizing toxins induce
3.3 Mode-of-Action of CDT long microtubule-based protrusions (Schwan et al.
2009). These cell membrane protrusions form a
CDT ADP-ribosylates monomeric G- actin in network of long tentacle-like structures on the sur-
arginine-177. Thus, modification of actin occurs face of epithelial cells. Microtubule-based
at the same residue of actin that is also modified protrusions are dynamic structures. They grow
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 89

and retract. CDT-induced depolymerization causes formation of a network of microtubule-based

the mislocalization of capture proteins like ACF7 protrusions, which facilitates adherence of
and Clasp2, which are involved in stabilization of C. difficile bacteria. In the same direction points
growing microtubules at the actin cell cortex the finding that CDT causes the redistribution of
(Kodama et al. 2003; Drabek et al. 2006). Without fibronectin from the basolateral membrane of
appropriate capture proteins at the cell membrane, epithelial cells to the apical side, where it acts as
microtubule growth is no longer stopped, resulting a receptor for C. difficile. Moreover, it is of interest
in protrusion formation (Schwan et al. 2009). More that CDT was shown to efficiently induce apoptosis
recent studies indicate that septins, which are of protective colonic eosinophils in a TLR2-
GTP-binding proteins that can reversibly dependent manner (Cowardin et al. 2016). More-
oligomerize (Mostowy and Cossart 2012), are cru- over, it should be considered that actin and
cially involved in toxin-induced protrusion forma- microtubules play a crucial role in activation of
tion (Nolke et al. 2016). Moreover, these findings the inflammasome (Gao et al. 2016). Also this
also show that septin-dependent protrusion forma- could be an important functional connection even-
tion is regulated by the Rho protein family member tually leading to increase in virulence of C. difficile
Cdc42 and its effectors Borg (binder of Rho in the presence of CDT, and toxins A and B.
GTPases) (Nolke et al. 2016).
CDT-induced partial depolymerization of
F-actin disturbs re-cycling of vesicles at the
4 Conclusions
basolateral side of epithelial cells. Thereby, the
vesicles, which contain extracellular matrix
It is well-accepted that C. difficile diseases are
(ECM) proteins like fibronectin and vitronectin,
mainly governed by the production of protein
are re-routed from the basolateral side to the
toxins, including C. difficile toxins A (TcdA) and
apical membrane, where microtubules form
B (TcdB). The third toxin, CDT, appears to be an
protrusions. Here, fibronectin and other ECM
important enhancing virulence factor. Therefore,
proteins are released (Schwan et al. 2014).
recent progress in our knowledge about the mode-
of-actions of these toxins is key for the understand-
ing of the pathophysiology of C. difficile infections
3.4 Role of CDT During C. difficile and the development of novel therapeutic strategies
Infection against the diseases caused by the pathogen. How-
ever, many open questions remain. In respect to
Although CDT is a very potent and efficient TcdA and TcdB, the membrane translocation of
cytotoxin, its role in C. difficile infection is not these toxins into target cells is still largely enig-
well understood. Only in extremely few cases matic. Moreover, C. difficile enterocolitis is
C. difficile-dependent enterocolitis could be characterized by severe inflammation and cell
traced back to CDT in the absence of necrosis. The precise pathophysiological pathways
C. difficile toxins A and B. What is then is its caused by the toxins leading to inflammation and
role in disease? The group of Nigel Minton necrosis are still not satisfactorily understood and
(Nottingham, UK) assessed the virulence of all explained. The great success of fecal transplanta-
possible combinations of isogenic C. difficile tion in therapy of C. difficile diseases indicate that
toxin mutants in the hamster infection model the microbiome is crucially involved in the patho-
and found that CDT is a factor that increases genesis of C. difficile infections. This also indicates
the virulence of C. difficile in the presence of a pivotal role of the immune system of the host.
toxins A and B (Kuehne et al. 2014). Several Therefore, the actions of C. difficile toxins on vari-
mechanisms are discussed. First, CDT may ous types of immune cells in context of intestinal
increase the adherence of bacteria due to the tissue should be studied in detail.
90 P. Papatheodorou et al.

References Busch C, Hofmann F, Gerhard R, Aktories K (2000)

Involvement of a conserved tryptophan residue in the
UDP-glucose binding of large clostridial cytotoxin
Aktories K (2011) Bacterial protein toxins that modify
glycosyltransferases. J Biol Chem 275:13228–13234
host regulatory GTPases. Nat Rev Microbiol
Chandrasekaran R, Kenworthy AK, Lacy DB (2016)
Clostridium difficile toxin A undergoes clathrin-
Aktories K, Wegner A (1989) ADP-ribosylation of actin
independent, PACSIN2-dependent endocytosis.
by clostridial toxins. J Cell Biol 109:1385–1387
PLoS Pathog 12:e1006070
Aktories K, Bärmann M, Ohishi I, Tsuyama S, Jakobs
Cherfils J, Zeghouf M (2013) Regulation of small
KH, Habermann E (1986) Botulinum C2 toxin
GTPases by GEFs, GAPs, and GDIs. Physiol Rev
ADP-ribosylates actin. Nature 322:390–392
Aktories K, Schwan C, Jank T (2017) Clostridium difficile
Chumbler NM, Rutherford SA, Zhang Z, Farrow MA,
toxin biology. Annu Rev Microbiol 71:281–307
Lisher JP, Farquhar E et al (2016) Crystal structure
Albesa-Jove D, Bertrand T, Carpenter EP, Swain GV,
of Clostridium difficile toxin A. Nat Microbiol
Lim J, Zhang J et al (2010) Four distinct structural
domains in Clostridium difficile toxin B visualized
Collery MM, Kuehne SA, McBride SM, Kelly ML,
using SAXS. J Mol Biol 396:1260–1270
Monot M, Cockayne A et al (2017) What’s a SNP
Alvin JW, Lacy DB (2017) Clostridium difficile toxin
between friends: the influence of single nucleotide
glucosyltransferase domains in complex with a
polymorphisms on virulence and phenotypes of Clos-
non-hydrolyzable UDP-glucose analogue. J Struct
tridium difficile strain 630 and derivatives. Virulence
Biol 198:203–209
Amimoto K, Noro T, Oishi E, Shimizu M (2007) A novel
Considine RV, Simpson LL (1991) Cellular and molecu-
toxin homologous to large clostridial cytotoxins found
lar actions of binary toxins possessing
in culture supernatant of Clostridium perfringens type
ADP-ribosyltransferase activity. Toxicon 29:913–936
C. Microbiology 153:1198–1206
Cowardin CA, Buonomo EL, Saleh MM, Wilson MG,
Barroso LA, Moncrief JS, Lyerly DM, Wilkins TD (1994)
Burgess SL, Kuehne SA et al (2016) The binary
Mutagenesis of the Clostridium difficile toxin B gene
toxin CDT enhances Clostridium difficile virulence
and effect on cytotoxic activity. Microb Pathog
by suppressing protective colonic eosinophilia. Nat
Microbiol 1:16108
Barth H, Pfeifer G, Hofmann F, Maier E, Benz R,
Czulkies BA, Mastroianni J, Lutz L, Lang S, Schwan C,
Aktories K (2001) Low pH-induced formation of ion
Schmidt G et al (2017) Loss of LSR affects epithelial
channels by Clostridium difficile toxin B in target
barrier integrity and tumor xenograft growth of CaCo-
cells. J Biol Chem 276:10670–10676
2 cells. Oncotarget 8:37009–37022
Barth H, Aktories K, Popoff MR, Stiles BG (2004) Binary
D’Urzo N, Malito E, Biancucci M, Bottomley MJ,
bacterial toxins: biochemistry, biology, and
Maione D, Scarselli M, Martinelli M (2012) The struc-
applications of common Clostridium and Bacillus
ture of Clostridium difficile toxin A glucosyl-
proteins. Microbiol Mol Biol Rev 68:373–402
transferase domain bound to Mn2þ and UDP
Belyi Y, Niggeweg R, Opitz B, Vogelsgesang M,
provides insights into glucosyltransferase activity
Hippenstiel S, Wilm M, Aktories K (2006) Legionella
and product release. FEBS J 279:3085–3097
pneumophila glucosyltransferase inhibits host elonga-
Dingle T, Wee S, Mulvey GL, Greco A, Kitova EN, Sun J
tion factor 1A. Proc Natl Acad Sci U S A
et al (2008) Functional properties of the carboxy-
terminal host cell-binding domains of the two toxins,
Bishop AL, Hall A (2000) Rho GTPases and their effector
TcdA and TcdB, expressed by Clostridium difficile.
proteins. Biochem J 348:241–255
Glycobiology 18:698–706
Blonder J, Hale ML, Chan KC, Yu LR, Lucas DA,
Dominguez R, Holmes KC (2011) Actin structure and
Conrads TP et al (2005) Quantitative profiling of the
function. Annu Rev Biophys 40:169–186
detergent-resistant membrane proteome of iota-b
Donald RG, Flint M, Kalyan N, Johnson E, Witko SE,
toxin induced vero cells. J Proteome Res 4:523–531
Kotash C et al (2013) A novel approach to generate a
Brito GAC, Fujji J, Carneiro-Filho BA, Lima AAM,
recombinant toxoid vaccine against Clostridium diffi-
Obrig T, Guerrant RL (2002) Mechanism of Clostrid-
cile. Microbiology 159:1254–1266
ium difficile toxin A – induced apoptosis in T84 cells. J
Dove CH, Wang SZ, Price SB, Phelps CJ, Lyerly DM,
Infect Dis 186:1438–1447
Wilkins TD, Johnson JL (1990) Molecular characteri-
Burridge K, Wennerberg K (2004) Rho and Rac take
zation of the Clostridium difficile toxin A gene. Infect
center stage. Cell 116:167–179
Immun 58:480–488
Busch C, Hofmann F, Selzer J, Munro J, Jeckel D,
Drabek K, van HM, Stepanova T, Draegestein K, van HR,
Aktories K (1998) A common motif of eukaryotic
Sayas CL et al (2006) Role of CLASP2 in microtubule
glycosyltransferases is essential for the enzyme activ-
stabilization and the regulation of persistent motility.
ity of large clostridial cytotoxins. J Biol Chem
Curr Biol 16:2259–2264
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 91

Egerer M, Giesemann T, Jank T, Satchell KJ, Aktories K and translocation of Clostridium difficile toxin B. Mol
(2007) Auto-catalytic cleavage of Clostridium difficile Microbiol 79:1643–1654
toxins A and B depends on a cysteine protease activ- Genth H, Aktories K, Just I (1999) Monoglucosylation of
ity. J Biol Chem 282:25314–25321 RhoA at threonine-37 blocks cytosol-membrane
Egerer M, Giesemann T, Herrmann C, Aktories K (2009) cycling. J Biol Chem 274:29050–29056
Autocatalytic processing of Clostridium difficile toxin Gerding DN, Johnson S, Rupnik M, Aktories K (2014)
B. Binding of inositol hexakisphosphate. J Biol Chem Clostridium difficile binary toxin CDT: mechanism,
284:3389–3395 epidemiology, and potential clinical importance. Gut
Ernst K, Langer S, Kaiser E, Osseforth C, Michaelis J, Microbes 5:15–27
Popoff MR et al (2015) Cyclophilin-facilitated mem- Gerhard R (2016) Receptors and binding structures for
brane translocation as pharmacological target to pre- Clostridium difficile toxins A and B. Curr Top
vent intoxication of mammalian cells by binary Microbiol Immunol. [Epub ahead of print]
clostridial actin ADP-ribosylated toxins. J Mol Biol Gerhard R, Nottrott S, Schoentaube J, Tatge H, Olling A,
427:1224–1238 Just I (2008) Glucosylation of Rho GTPases by Clos-
Ernst K, Schnell L, Barth H (2016) Host cell chaperones tridium difficile toxin A triggers apoptosis in intestinal
Hsp70/Hsp90 and peptidyl-prolyl cis/trans isomerases epithelial cells. J Med Microbiol 57:765–770
are required for the membrane translocation of bacte- Gerhard R, Frenzel E, Goy S, Olling A (2013) Cellular
rial ADP-ribosylating toxins. Curr Top Microbiol uptake of Clostridium difficile TcdA and truncated
Immunol. May 20. [Epub ahead of print] TcdA lacking the receptor binding domain. J Med
Ernst K, Schmid J, Beck M, Hagele M, Hohwieler M, Microbiol 62:1414–1422
Hauff P et al (2017) Hsp70 facilitates trans-membrane Geyer M, Wilde C, Selzer J, Aktories K, Kalbitzer HR
transport of bacterial ADP-ribosylating toxins into the (2003) Glucosylation of Ras by Clostridium sordellii
cytosol of mammalian cells. Sci Rep 7:2724 lethal toxin: consequences for the effector loop
Farrow MA, Chumbler NM, Lapierre LA, Franklin JL, conformations observed by NMR spectroscopy. Bio-
Rutherford SA, Goldenring JR, Lacy DB (2013) Clos- chemistry 42:11951–11959
tridium difficile toxin B-induced necrosis is mediated Gibert M, Monier MN, Ruez R, Hale ML, Stiles BG,
by the host epithelial cell NADPH oxidase complex. Benmerah A et al (2011) Endocytosis and toxicity of
Proc Natl Acad Sci U S A 110:18674–18679 clostridial binary toxins depend on a clathrin-
Fiorentini C, Thelestam M (1991) Clostridium difficile independent pathway regulated by Rho-GDI. Cell
toxin A and its effects on cells. Toxicon 29:543–567 Microbiol 13:154–170
Fiorentini C, Fabbri A, Falzano L, Fattorossi A, Giesemann T, Jank T, Gerhard R, Maier E, Just I, Benz R,
Matarrese P, Rivabene R, Donelli G (1998) Clostrid- Aktories K (2006) Cholesterol-dependent pore forma-
ium difficile toxin B induces apoptosis in intestinal tion of Clostridium difficile toxin A. J Biol Chem
cultured cells. Infect Immun 66:2660–2665 281:10808–10815
Frey SM, Wilkins TD (1992) Localization of two epitopes Greco A, Ho JG, Lin SJ, Palcic MM, Rupnik M, Ng KK
recognized by monoclonal antibody PCG-4 on Clos- (2006) Carbohydrate recognition by Clostridium diffi-
tridium difficile toxin A. Infect Immun 60:2488–2492 cile toxin A. Nat Struct Mol Biol 13:460–461
Frisch C, Gerhard R, Aktories K, Hofmann F, Just I Guttenberg G, Hornei S, Jank T, Schwan C, Lu W, Einsle
(2003) The complete receptor-binding domain of O et al (2012) Molecular characteristics of Clostrid-
Clostridium difficile toxin A is required for endocyto- ium perfringens TpeL toxin and consequences of
sis. Biochem Biophys Res Commun 300:706–711 mono-O-GlcNAcylation of Ras in living cells. J Biol
Furuse M, Oda Y, Higashi T, Iwamoto N, Masuda S Chem 287:24929–24940
(2012) Lipolysis-stimulated lipoprotein receptor: a Halabi-Cabezon I, Huelsenbeck J, May M, Ladwein M,
novel membrane protein of tricellular tight junctions. Rottner K, Just I, Genth H (2008) Prevention of the
Ann N Y Acad Sci 1257:54–58 cytopathic effect induced by Clostridium difficile
Gao W, Yang J, Liu W, Wang Y, Shao F (2016) Site- toxin B by active Rac1. FEBS Lett 582:3751–3756
specific phosphorylation and microtubule dynamics Hale ML, Marvaud JC, Popoff MR, Stiles BG (2004)
control Pyrin inflammasome activation. Proc Natl Detergent-resistant membrane microdomains facili-
Acad Sci U S A 113:E4857–E4866 tate Ib oligomer formation and biological activity of
Garcia-Mata R, Burridge K (2007) Catching a GEF by its Clostridium perfringens iota-toxin. Infect Immun
tail. Trends Cell Biol 17:36–43 72:2186–2193
Geissler B, Tungekar R, Satchell KJ (2010) Identification Han S, Craig JA, Putnam CD, Carozzi NB, Tainer JA
of a conserved membrane localization domain within (1999) Evolution and mechanism from structures of an
numerous large bacterial protein toxins. Proc Natl ADP-ribosylating toxin and NAD complex. Nat Struct
Acad Sci U S A 107:5581–5586 Biol 6:932–936
Genisyuerek S, Papatheodorou P, Guttenberg G, Hecht G, Pothoulakis C, LaMont JT, Madara JL (1988)
Schubert R, Benz R, Aktories K (2011) Structural Clostridium difficile toxin A pertubs cytoskeletal
determinants for membrane insertion, pore formation
92 P. Papatheodorou et al.

structure and tight junction permeability of cultured tyrosine glycosylation of Rho and deamidation of Gq
human intestinal epithelial monolayers. J Clin Invest and Gi proteins. Nat Struct Mol Biol 20:1273–1280
82:1516–1524 Jank T, Belyi Y, Aktories K (2015a) Bacterial glycosyl-
Hecht G, Koutsouris A, Pothoulakis C, LaMont JT, transferase toxins. Cell Microbiol 17:1752–1765
Madara JL (1992) Clostridium difficile toxin B Jank T, Eckerle S, Steinemann M, Trillhaase C,
disrupts the barrier function of T84 monolayers. Gas- Schimpl M, Wiese S et al (2015b) Tyrosine glycosyl-
troenterology 102:416–423 ation of Rho by Yersinia toxin impairs blastomere cell
Heine K, Pust S, Enzenmuller S, Barth H (2008) behaviour in zebrafish embryos. Nat Commun 6:7807
ADP-ribosylation of actin by the Clostridium botuli- Jorgensen I, Miao EA (2015) Pyroptotic cell death
num C2 toxin in mammalian cells results in delayed defends against intracellular pathogens. Immunol
caspase-dependent apoptotic cell death. Infect Immun Rev 265:130–142
76:4600–4608 Just I, Gerhard R (2004) Large clostridial cytotoxins. Rev
Hemmasi S, Czulkies BA, Schorch B, Veit A, Aktories K, Physiol Biochem Pharmacol 152:23–47
Papatheodorou P (2015) Interaction of the Clostridium Just I, Selzer J, Wilm M, Von Eichel-Streiber C, Mann M,
difficile binary toxin CDT and its host cell receptor, Aktories K (1995a) Glucosylation of Rho proteins by
lipolysis-stimulated lipoprotein receptor (LSR). J Biol Clostridium difficile toxin B. Nature 375:500–503
Chem 290:14031–14044 Just I, Wilm M, Selzer J, Rex G, Von Eichel-Streiber C,
Hirase T, Kawashima S, Wong EY, Ueyama T, Mann M, Aktories K (1995b) The enterotoxin from
Rikitake Y, Tsukita S et al (2001) Regulation of tight Clostridium difficile (ToxA) monoglucosylates the
junction permeability and occludin phosphorylation Rho proteins. J Biol Chem 270:13932–13936
by Rhoa-p160ROCK-dependent and -independent Just I, Selzer J, Hofmann F, Green GA, Aktories K (1996)
mechanisms. J Biol Chem 276:10423–10431 Inactivation of Ras by Clostridium sordellii lethal
Ho JG, Greco A, Rupnik M, Ng KK (2005) Crystal toxin-catalyzed glucosylation. J Biol Chem
structure of receptor-binding C-terminal repeats from 271:10149–10153
Clostridium difficile toxin A. Proc Natl Acad Sci U S Kaiser E, Kroll C, Ernst K, Schwan C, Popoff M, Fischer
A 102:18373–18378 G et al (2011) Membrane translocation of binary actin-
Hofmann F, Busch C, Prepens U, Just I, Aktories K ADP-ribosylating toxins from Clostridium difficile
(1997) Localization of the glucosyltransferase activity and Clostridium perfringens is facilitated by
of Clostridium difficile toxin B to the N-terminal part cyclophilin A and Hsp90. Infect Immun
of the holotoxin. J Biol Chem 272:11074–11078 79:3913–3921
Hofmann F, Busch C, Aktories K (1998) Chimeric clos- Kim J, Pai H, Seo MR, Kang JO (2012) Clinical and
tridial cytotoxins: identification of the N-terminal microbiologic characteristics of tcdA-negative variant
region involved in protein substrate recognition. Infect Clostridium difficile infections. BMC Infect Dis
Immun 66:1076–1081 12:109
Holmes KC, Popp D, Gebhard W, Kabsch W (1990) Kodama A, Karakesisoglou I, Wong E, Vaezi A, Fuchs E
Atomic model of the actin filament. Nature 347:44–49 (2003) ACF7: an essential integrator of microtubule
Ishida Y, Maegawa T, Kondo T, Kimura A, Iwakura Y, dynamics. Cell 115:343–354
Nakamura S, Mukaida N (2004) Essential involve- Krivan HC, Clark GF, Smith DF, Wilkins TD (1986) Cell
ment of IFN-gamma in Clostridium difficile toxin surface binding site for Clostridium difficile entero-
A-induced enteritis. J Immunol 172:3018–3025 toxin: evidence for a glycoconjugate containing the
Jafari NV, Kuehne SA, Bryant CE, Elawad M, Wren BW, sequence Gal alpha 1-3Gal beta 1-4GlcNAc. Infect
Minton NP et al (2013) Clostridium difficile modulates Immun 53:573–581
host innate immunity via toxin-independent and Kuehne SA, Cartman ST, Heap JT, Kelly ML,
dependent mechanism(s). PLoS One 8:e69846 Cockayne A, Minton NP (2010) The role of toxin A
Jaffe AB, Hall A (2005) Rho GTPases: biochemistry and and toxin B in Clostridium difficile infection. Nature
biology. Annu Rev Cell Dev Biol 21:247–269 467:711–713
Jank T, Aktories K (2008) Structure and mode of action of Kuehne SA, Collery MM, Kelly ML, Cartman ST,
clostridial glucosylating toxins: the ABCD model. Cockayne A, Minton NP (2014) Importance of
Trends Microbiol 16:222–229 toxin A, toxin B, and CDT in virulence of an epidemic
Jank T, Reinert DJ, Giesemann T, Schulz GE, Aktories K Clostridium difficile strain. J Infect Dis 209:83–86
(2005) Change of the donor substrate specificity of Kushnaryov VM, Sedmark JJ (1989) Effect of Clostrid-
Clostridium difficile toxin B by site-directed mutagen- ium difficile enterotoxin A on ultrastructure of chinese
esis. J Biol Chem 280:37833–37838 hamster ovary cells. Infect Immun 57(12):3914–3921
Jank T, Giesemann T, Aktories K (2007) Clostridium La France ME, Farrow MA, Chandrasekaran R, Sheng JS,
difficile glucosyltransferase toxin B – essential amino Rubin DH, Lacy DB (2015) Identification of an epi-
acids for substrate-binding. J Biol Chem thelial cell receptor responsible for Clostridium diffi-
282:35222–35231 cile TcdB-induced cytotoxicity. Proc Natl Acad Sci U
Jank T, Bogdanovic X, Wirth C, Haaf E, Spoerner M, S A 112:7073–7078
Bohmer KE et al (2013) A bacterial toxin catalyzing
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 93

Lamaze C, Dujeancourt A, Baba T, Lo CG, Benmerah A, insights into the evolution of the pathogenicity locus.
Dautry-Varsat A (2001) Interleukin 2 receptors and Sci Rep 5:15023
detergent-resistant membrane domains define a Mostowy S, Cossart P (2012) Septins: the fourth compo-
clathrin-independent endocytic pathway. Mol Cell nent of the cytoskeleton. Nat Rev Mol Cell Biol
7:661–671 13:183–194
Lambert GS, Baldwin MR (2016) Evidence for dual Nagahama M, Yamaguchi A, Hagiyama T, Ohkubo N,
receptor-binding sites in Clostridium difficile toxin Kobayashi K, Sakurai J (2004) Binding and internali-
A. FEBS Lett 590:4550–4563 zation of Clostridium perfringens iota-toxin in lipid
Lemichez E, Aktories K (2013) Hijacking of Rho rafts. Infect Immun 72:3267–3275
GTPases during bacterial infection. Exp Cell Res Nagahama M, Ohkubo A, Oda M, Kobayashi K,
319:2329–2336 Amimoto K, Miyamoto K, Sakurai J (2011) Clostrid-
Li S, Zhang L, Yao Q, Li L, Dong N, Rong J et al (2013) ium perfringens TpeL glycosylates the Rac and Ras
Pathogen blocks host death receptor signalling by subfamily proteins. Infect Immun 79:905–910
arginine GlcNAcylation of death domains. Nature Ng J, Hirota SA, Gross O, Li Y, Ulke-Lemee A, Potentier
501:242–246 MS et al (2010) Clostridium difficile toxin-induced
Linevsky JK, Pothoulakis C, Keates S, Warny M, Keates inflammation and intestinal injury are mediated by
AC, LaMont JT, Kelly CP (1997) IL-8 release and the inflammasome. Gastroenterology 139:542–552
neutrophil activation by Clostridium difficile toxin- Nobes C, Hall A (1994) Regulation and function of the
exposed human monocytes. Am J Phys 273:G1333– Rho subfamily of small GTPases. Curr Opin Genet
G1340 Dev 4:77–81
Lu A, Wu H (2015) Structural mechanisms of Nolke T, Schwan C, Lehmann F, Ostevold K, Pertz O,
inflammasome assembly. FEBS J 282:435–444 Aktories K (2016) Septins guide microtubule
Lyerly DM, Krivan HC, Wilkins TD (1988) Clostridium protrusions induced by actin-depolymerizing toxins
difficile: its disease and toxins. Clin Microbiol Rev like Clostridium difficile transferase (CDT). Proc
1:1–18 Natl Acad Sci U S A 113:7870–7875
Lyerly DM, Barroso LA, Wilkins TD, Depitre C, Corthier Nusrat A, Giry M, Turner JR, Colgan SP, Parkos CA,
G (1992) Characterization of a toxin A-negative, toxin Carnes D et al (1995) Rho protein regulates tight
B-positive strain of Clostridium difficile. Infect junctions and perijunctional actin organization in
Immun 60:4633–4639 polarized epithelia. Proc Natl Acad Sci U S A
Lyras D, O’Connor JR, Howarth PM, Sambol SP, Carter 92:10629–10633
GP, Phumoonna T et al (2009) Toxin B is essential for Nusrat A, Von Eichel-Streiber C, Turner JR, Verkade P,
virulence of Clostridium difficile. Nature Madara JL, Parkos CA (2001) Clostridium difficile
458:1176–1179 toxins disrupt epithelial barrier function by altering
Mahida YR, Makh S, Hyde S, Gray T, Borriello SP (1996) membrane microdomain localization of tight junction
Effect of Clostridium difficile toxin A on human intes- proteins. Infect Immun 69:1329–1336
tinal epithelial cells: induction of interleukin 8 produc- Olling A, Goy S, Hoffmann F, Tatge H, Just I, Gerhard R
tion and apoptosis after cell detachment. Gut (2011) The repetitive oligopeptide sequences modu-
38:337–347 late cytopathic potency but are not crucial for cellular
Manse JS, Baldwin MR (2015) Binding and entry of uptake of Clostridium difficile toxin A. PLoS One 6:
Clostridium difficile toxin B is mediated by multiple e17623
domains. FEBS Lett 589:3945–3951 Orth P, Xiao L, Hernandez LD, Reichert P, Sheth PR,
Margarit SM, Davidson W, Frego L, Stebbins CE (2006) Beaumont M et al (2014) Mechanism of action and
A steric antagonism of actin polymerization by a sal- epitopes of Clostridium difficile toxin B-neutralizing
monella virulence protein. Structure 14:1219–1229 antibody bezlotoxumab revealed by X-ray crystallog-
Masuda S, Oda Y, Sasaki H, Ikenouchi J, Higashi T, raphy. J Biol Chem 289:18008–18021
Akashi M et al (2011) LSR defines cell corners for Ottlinger ME, Lin S (1988) Clostridium difficile toxin B
tricellular tight junction formation in epithelial cells. J induces reorganization of actin, vinculin, and talin in
Cell Sci 124:548–555 cultures cells. Exp Cell Res 174:215–229
Mesli S, Javorschi S, Berard AM, Landry M, Priddle H, Papatheodorou P, Aktories K (2016) Receptor-binding
Kivlichan D et al (2004) Distribution of the lipolysis and uptake of binary actin-ADP-ribosylating toxins.
stimulated receptor in adult and embryonic murine Curr Top Microbiol Immunol. Nov 6. [Epub ahead of
tissues and lethality of LSR/ embryos at 12.5 to print]
14.5 days of gestation. Eur J Biochem 271:3103–3114 Papatheodorou P, Zamboglou C, Genisyuerek S,
Miao EA, Leaf IA, Treuting PM, Mao DP, Dors M, Sarkar Guttenberg G, Aktories K (2010) Clostridial
A et al (2010) Caspase-1-induced pyroptosis is an glucosylating toxins enter cells via clathrin-mediated
innate immune effector mechanism against intracellu- endocytosis. PLoS One 5:e10673
lar bacteria. Nat Immunol 11:1136–1142 Papatheodorou P, Carette JE, Bell GW, Schwan C,
Monot M, Eckert C, Lemire A, Hamiot A, Dubois T, Guttenberg G, Brummelkamp TR, Aktories K (2011)
Tessier C et al (2015) Clostridium difficile: new Lipolysis-stimulated lipoprotein receptor (LSR) is the
94 P. Papatheodorou et al.

host receptor for the binary toxin Clostridium difficile Qa’Dan M, Christensen KA, Zhang L, Roberts TM, Col-
transferase (CDT). Proc Natl Acad Sci U S A lier RJ (2005) Membrane insertion by anthrax protec-
108:16422–16427 tive antigen in cultured cells. Mol Cell Biol
Papatheodorou P, Wilczek C, Nolke T, Guttenberg G, 25:5492–5498
Hornuss D, Schwan C, Aktories K (2012) Identifica- Qiu B, Pothoulakis C, Castagliuolo I, Nikulasson S, La
tion of the cellular receptor of Clostridium spiroforme Mont JT (1999) Participation of reactive oxygen
toxin. Infect Immun 80:1418–1423 metabolites in Clostridium difficile toxin A-induced
Papatheodorou P, Hornuss D, Nolke T, Hemmasi S, enteritis in rats. Am J Phys 276:G485–G490
Castonguay J, Picchianti M, Aktories K (2013) Clos- Reineke J, Tenzer S, Rupnik M, Koschinski A,
tridium difficile binary toxin CDT induces clustering Hasselmayer O, Schrattenholz A et al (2007) Autocat-
of the lipolysis-stimulated lipoprotein receptor into alytic cleavage of Clostridium difficile toxin B. Nature
lipid rafts. MBio 4:e00244–e00213 446:415–419
Park YH, Wood G, Kastner DL, Chae JJ (2016) Pyrin Reinert DJ, Jank T, Aktories K, Schulz GE (2005) Struc-
inflammasome activation and RhoA signaling in the tural basis for the function of Clostridium difficile
autoinflammatory diseases FMF and HIDS. Nat toxin B. J Mol Biol 351:973–981
Immunol 17:914–921 Roeder M, Nestorovich EM, Karginov VA, Schwan C,
Perelle S, Gibert M, Bourlioux P, Corthier G, Popoff MR Aktories K, Barth H (2014) Tailored cyclodextrin pore
(1997) Production of a complete binary toxin (actin- blocker protects mammalian cells from Clostridium
specific ADP-ribosyltransferase) by Clostridium diffi- difficile binary toxin CDT. Toxins (Basel)
cile CD196. Infect Immun 65:1402–1407 6:2097–2114
Perieteanu AA, Visschedyk DD, Merrill AR, Dawson JF Roth BM, Godoy-Ruiz R, Varney KM, Rustandi RR,
(2010) ADP-ribosylation of cross-linked actin Weber DJ (2016a) 1H, 13C, and 15N resonance
generates barbed-end polymerization-deficient assignments of an enzymatically active domain from
F-actin oligomers. Biochemistry 49:8944–8954 the catalytic component (CDTa, residues 216-420) of
Petosa C, Collier RJ, Klimpel KR, Leppla SH, a binary toxin from Clostridium difficile. Biomol
Liddingtom RC (1997) Crystal structure of the anthrax NMR Assign 10:213–217
toxin protective antigen. Nature 385:833–838 Roth BM, Varney KM, Rustandi RR, Weber DJ (2016b)
Pfeifer G, Schirmer J, Leemhuis J, Busch C, Meyer DK, (1)H(N), (13)C, and (15)N resonance assignments of
Aktories K, Barth H (2003) Cellular uptake of Clos- the CDTb-interacting domain (CDTaBID) from the
tridium difficile toxin B: translocation of the Clostridium difficile binary toxin catalytic component
N-terminal catalytic domain into the cytosol of (CDTa, residues 1–221). Biomol NMR Assign
eukaryotic cells. J Biol Chem 278:44535–44541 10:335–339
Popoff MR, Boquet P (1988) Clostridium spiroforme Rupnik M, Janezic S (2016) An update on Clostridium
toxin is a binary toxin which ADP- ribosylates cellular difficile toxinotyping. J Clin Microbiol 54:13–18
actin. Biochem Biophys Res Commun 152:1361–1368 Rupnik M, Avesani V, Janc M, Von Eichel-Streiber C,
Pruitt RN, Chagot B, Cover M, Chazin WJ, Spiller B, Delmée M (1998) A novel toxinotyping scheme and
Lacy DB (2009) Structure-function analysis of inosi- correlation of toxinotypes with serogroups of Clostrid-
tol hexakisphosphate-induced autoprocessing in Clos- ium difficile isolates. J Clin Microbiol 36:2240–2247
tridium difficile toxin A. J Biol Chem Rupnik M, Pabst S, Rupnik M, Von Eichel-Streiber C,
284:21934–21940 Urlaub H, Soling HD (2005) Characterization of the
Pruitt RN, Chambers MG, Ng KK, Ohi MD, Lacy DB cleavage site and function of resulting cleavage
(2010) Structural organization of the functional fragments after limited proteolysis of Clostridium dif-
domains of Clostridium difficile toxins A and B. Proc ficile toxin B (TcdB) by host cells. Microbiology
Natl Acad Sci U S A 107:13467–13472 151:199–208
Pruitt RN, Chumbler NM, Rutherford SA, Farrow MA, Russo HM, Rathkey J, Boyd-Tressler A, Katsnelson MA,
Friedman DB, Spiller B, Lacy DB (2012) Structural Abbott DW, Dubyak GR (2016) Active caspase-1
determinants of Clostridium difficile toxin A glucosyl- induces plasma membrane pores that precede
transferase activity. J Biol Chem 287:8013–8020 pyroptotic lysis and are blocked by lanthanides. J
Puri AW, Lupardus PJ, Deu E, Albrow VE, Garcia KC, Immunol 197:1353–1367
Bogyo M, Shen A (2010) Rational design of inhibitors Sauerborn M, Leukel P, Von Eichel-Streiber C (1997)
and activity-based probes targeting Clostridium diffi- The C-terminal ligand-binding domain of Clostridium
cile virulence factor TcdB. Chem Biol 17:1201–1211 difficile toxin A (TcdA) abrogates TcdA-specific bind-
Qa’Dan M, Spyres LM, Ballard JD (2000) pH-induced ing to cells and prevents mouse lethality. FEMS
conformational changes in Clostridium difficile toxin Microbiol Lett 155:45–54
B. Infect Immun 68:2470–2474 Schering B, Bärmann M, Chhatwal GS, Geipel U,
Qa’Dan M, Spyres LM, Ballard JD (2001) pH-enhanced Aktories K (1988) ADP-ribosylation of skeletal mus-
cytopathic effects of Clostridium sordellii lethal toxin. cle and non- muscle actin by Clostridium perfringens
Infect Immun 69:5487–5493 iota toxin. Eur J Biochem 171:225–229
Cellular Uptake and Mode-of-Action of Clostridium difficile Toxins 95

Schleberger C, Hochmann H, Barth H, Aktories K, Schulz snapshots of iota-toxin and actin complex. Proc Natl
GE (2006) Structure and action of the binary C2 toxin Acad Sci U S A 110:4267–4272
from Clostridium botulinum. J Mol Biol 364:705–715 Tucker KD, Wilkins TD (1991) Toxin A of Clostridium
Schorch B, Song S, van Diemen FR, Bock HH, May P, difficile binds to the human carbohydrate antigens I, X,
Herz J et al (2014) LRP1 is a receptor for Clostridium and Y. Infect Immun 59:73–78
perfringens TpeL toxin indicating a two-receptor Vandekerckhove J, Schering B, Bärmann M, Aktories K
model of clostridial glycosylating toxins. Proc Natl (1987) Clostridium perfringens iota toxin
Acad Sci U S A 111:6431–6436 ADP-ribosylates skeletal muscle actin in Arg-177.
Schwan C, Stecher B, Tzivelekidis T, Van HM, Rohde M, FEBS Lett 225:48–52
Hardt WD et al (2009) Clostridium difficile toxin CDT Vandekerckhove J, Schering B, Bärmann M, Aktories K
induces formation of microtubule-based protrusions (1988) Botulinum C2 toxin ADP-ribosylates cytoplas-
and increases adherence of bacteria. PLoS Pathog 5: mic b/g-actin in arginine 177. J Biol Chem
e1000626 263:696–700
Schwan C, Kruppke AS, Nolke T, Schumacher L, Koch- Vetter IR, Hofmann F, Wohlgemuth S, Herrmann C, Just I
Nolte F, Kudryashev M et al (2014) Clostridium diffi- (2000) Structural consequences of mono-
cile toxin CDT hijacks microtubule organization and glucosylation of Ha-Ras by Clostridium sordellii
reroutes vesicle traffic to increase pathogen adher- lethal toxin. J Mol Biol 301:1091–1095
ence. Proc Natl Acad Sci U S A 111:2313–2318 Von Eichel-Streiber C, Sauerborn M (1990) Clostridium
Sehr P, Joseph G, Genth H, Just I, Pick E, Aktories K difficile toxin A carries a C-terminal repetitive struc-
(1998) Glucosylation and ADP-ribosylation of Rho ture homologous to the carbohydrate binding region of
proteins – effects on nucleotide binding, GTPase streptococcal glycosyltransferases. Gene 96:107–113
activity, and effector-coupling. Biochemistry Von Eichel-Streiber C, Laufenberg-Feldmann R,
37:5296–5304 Sartingen S, Schulze J, Sauerborn M (1992a) Compar-
Selzer J, Hofmann F, Rex G, Wilm M, Mann M, Just I, ative sequence analysis of the Clostridium difficile
Aktories K (1996) Clostridium novyi alpha-toxin- toxins A and B. Mol Gen Genet 233:260–268
catalyzed incorporation of GlcNAc into Rho subfam- Von Eichel-Streiber C, Sauerborn M, Kuramitsu HK
ily proteins. J Biol Chem 271:25173–25177 (1992b) Evidence for a modular structure of the
Shen A, Lupardus PJ, Gersch MM, Puri AW, Albrow VE, homologous repetitive C-terminal carbohydrate-bind-
Garcia KC, Bogyo M (2011) Defining an allosteric ing sites of Clostridium difficile toxins and Streptococ-
circuit in the cysteine protease domain of Clostridium cus mutans glucosyltransferases. J Bacteriol
difficile toxins. Nat Struct Mol Biol 18:364–371 174:6707–6710
Sohet F, Lin C, Munji RN, Lee SY, Ruderisch N, Soung A Voth DE, Ballard JD (2005) Clostridium difficile toxins:
et al (2015) LSR/angulin-1 is a tricellular tight junc- mechanism of action and role in disease. Clin
tion protein involved in blood-brain barrier formation. Microbiol Rev 18:247–263
J Cell Biol 208:703–711 Warny M, Keates AC, Keates S, Castagliuolo I, Zacks JK,
Steiner TS, Flores CA, Pizarro TT, Guerrant RL (1997) Aboudola S et al (2000) p38 MAP kinase activation by
Fecal lactoferrin, interleukin-1beta, and interleukin- Clostridium difficile toxin A mediates monocyte
8 are elevated in patients with severe Clostridium necrosis, IL-8 production, and enteritis. J Clin Invest
difficile colitis. Clin Diagn Lab Immunol 4:719–722 105:1147–1156
Stiles BG, Hale ML, Marvaud J-C, Popoff M (2000) Wegner A, Aktories K (1988) ADP-ribosylated actin caps
Clostridium perfringens iota toxin: binding studies the barbed ends of actin filaments. J Biol Chem
and characterization of cell surface receptor by 263:13739–13742
fluorescence-activated cytometry. Infect Immun Weigt C, Just I, Wegner A, Aktories K (1989) Nonmuscle
68:3475–3484 actin ADP-ribosylated by botulinum C2 toxin caps
Tao L, Zhang J, Meraner P, Tovaglieri A, Wu X, Gerhard actin filaments. FEBS Lett 246:181–184
R et al (2016) Frizzled proteins are colonic epithelial Wiegers W, Just I, Müller H, Hellwig A, Traub P,
receptors for C. difficile toxin B. Nature 538:350–355 Aktories K (1991) Alteration of the cytoskeleton of
Tcherkezian J, Lamarche-Vane N (2007) Current knowl- mammalian cells cultured in vitro by Clostridium bot-
edge of the large RhoGAP family of proteins. Biol ulinum C2 toxin and C3 ADP-ribosyltransferase. Eur J
Cell 99:67–86 Cell Biol 54:237–245
Tsuge H, Nagahama M, Oda M, Iwamoto S, Utsunomiya H, Wigelsworth DJ, Ruthel G, Schnell L, Herrlich P,
Marquez VE et al (2008) Structural basis of actin recog- Blonder J, Veenstra TD et al (2012) CD44 promotes
nition and arginine ADP-ribosylation by Clostridium intoxication by the clostridial iota-family toxins. PLoS
perfringens iota-toxin. Proc Natl Acad Sci U S A One 7:e51356
105:7399–7404 Wille M, Just I, Wegner A, Aktories K (1992)
Tsurumura T, Tsumori Y, Qiu H, Oda M, Sakurai J, ADP-ribosylation of the gelsolin-actin complex by
Nagahama M, Tsuge H (2013) Arginine clostridial toxins. J Biol Chem 267:50–55
ADP-ribosylation mechanism based on structural Wohlan K, Goy S, Olling A, Srivaratharajan S, Tatge H,
Genth H, Gerhard R (2014) Pyknotic cell death
96 P. Papatheodorou et al.

induced by Clostridium difficile TcdB: chromatin con- cloning of a lipolysis-stimulated remnant receptor
densation and nuclear blister are induced indepen- expressed in the liver. J Biol Chem 274:13390–13398
dently of the glucosyltransferase activity. Cell Young JA, Collier RJ (2007) Anthrax toxin: receptor-
Microbiol 16:1678–1692 binding, internalization, pore formation, and translo-
Wren BW (1991) A family of clostridial and streptococcal cation. Annu Rev Biochem 76:243–265
ligand-binding proteins with conserved C-terminal Yuan PF, Zhang HM, Cai CZ, Zhu SY, Zhou YX, Yang
repeat sequences. Mol Microbiol 5:797–803 XZ et al (2015) Chondroitin sulfate proteoglycan
Xu H, Yang J, Gao W, Li L, Li P, Zhang L et al (2014) 4 functions as the cellular receptor for Clostridium
Innate immune sensing of bacterial modifications of difficile toxin B. Cell Res 25:157–168
Rho GTPases by the pyrin inflammasome. Nature Zeiser J, Gerhard R, Just I, Pich A (2013) Substrate
513:237–241 specificity of clostridial glucosylating toxins and
Yen FT, Mann CJ, Guermani LM, Hannouche NF, their function on colonocytes analyzed by proteomics
Hubert N, Hornick CA et al (1994) Identification of techniques. J Proteome Res 12:1604–1618
a lipolysis-stimulated receptor that is distinct from the Zhang Z, Park M, Tam J, Auger A, Beilhartz GL, Lacy
LDL receptor and the LDL receptor-related protein. DB, Melnyk RA (2014) Translocation domain
Biochemistry 33:1172–1180 mutations affecting cellular toxicity identify the Clos-
Yen FT, Masson M, Clossais-Besnard N, Andre P, tridium difficile toxin B pore. Proc Natl Acad Sci U S
Grosset JM, Bougueleret L et al (1999) Molecular A 111:3721–3726
Clostridium difficile Biofilm

Claudia Vuotto, Gianfranco Donelli, Anthony Buckley,

and Caroline Chilton

Abstract Findings on C. difficile biofilm, possible

Clostridium difficile infection (CDI) is an implications in CDI pathogenesis and treatment,
important healthcare-associated disease efficacy of currently available antibiotics in
worldwide, mainly occurring after antimicro- treating biofilm-forming C. difficile strains, and
bial therapy. Antibiotics administered to treat some antimicrobial alternatives under investiga-
a number of infections can promote C. difficile tion will be discussed here.
colonization of the gastrointestinal tract and,
thus, CDI. A rise in multidrug resistant clini- Keywords
cal isolates to multiple antibiotics and their Biofilm · Clostridium difficile · Genetic
reduced susceptibility to the most commonly factors · EPS matrix · Adhesion
used antibiotic molecules have made the treat-
ment of CDI more complicated, allowing the
persistence of C. difficile in the intestinal 1 Introduction
Gut colonization and biofilm formation Microbial biofilms are considered as the ‘true’
have been suggested to contribute to the path- habitat for many causative agents of infection
ogenesis and persistence of C. difficile. In fact, and disease. These microbial communities grow-
biofilm growth is considered as a serious ing on biotic and abiotic surfaces are embedded
threat because of the related increase in bacte- in a matrix of extracellular polymeric substances
rial resistance that makes antibiotic therapy (EPS) (Heydorn et al. 2000), offering to
often ineffective. However, although the microorganisms an efficacious protection from
involvement of the C. difficile biofilm in the antibiotics (Goldberg 2002) and disinfectants
pathogenesis and recurrence of CDI is (Peng et al. 2002), as well as the possibility to
attracting more and more interest, the survive in conditions of nutrient deficiency
mechanisms underlying biofilm formation of (Koch et al. 2001). Biofilm formation is
C. difficile as well as the role of biofilm in CDI characterised by several phases, starting from
have not been extensively described. reversible and irreversible attachment to the

A. Buckley · C. Chilton
C. Vuotto (*) · G. Donelli Healthcare Associated Infection Research Group,
Microbial Biofilm Laboratory, IRCCS Fondazione Santa Section of Molecular Gastroenterology, Institute for
Lucia, Rome, Italy Biomedical and Clinical Sciences, University of Leeds,
e-mail: Leeds, UK

# Springer International Publishing AG 2018 97

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
98 C. Vuotto et al.

surface, passing through the development of a 2016; Rossi et al. 2017; Owrangi et al. 2017).
single-species community or a polymicrobial Furthermore, other intestinal isolates, belonging
one, and then ending with the dispersion of to the anaerobic species Bacteroides, Clostrid-
cells from the biofilm (Percival et al. 2015). ium, Fusobacterium, Finegoldia, Prevotella, and
The architecture of a mature biofilm can vary Veillonella, have been demonstrated to be able to
depending on the microorganisms that constitute develop as in vitro mono-species biofilms, and to
it, forming flat or mushroom-shaped structures interact with each other by forming dual-species
(Klausen et al. 2003), with the latter generally biofilms (Donelli et al. 2012).
observed within in vitro biofilms only. A growing interest in the potential biofilm
Intracellular and intercellular communication growth of C. difficile has been recorded in recent
within a biofilm is supported by signals released years, due to the prominence of this microorgan-
when cell density reaches a critical level, a phe- ism as etiologic agent of nosocomial diarrhoea
nomenon known as quorum sensing (QS) (Lindsay worldwide.
and Von Holy 2006; Li and Tian 2012). QS is cell CDI is one of the principal threats to
density-dependent gene regulation through the pro- hospitalized and immunocompromised patients,
duction of signalling molecules, termed mainly when antibiotics are administered to them
autoinducers (AI), that activate the maturation and in order to treat a number of infections. In fact,
disassembly of the biofilm in a coordinate manner, antibiotic molecules, by disrupting the protective
with dispersal of microbial cells into the intestinal microbiota, can promote C. difficile colo-
surrounding environment increasing the dissemina- nization of the gastrointestinal tract and, thus, CDI.
tion risk and the colonisation of new niches The resistance of an increased number of clinical
(Donelli 2006). This “lifestyle” allows pathogenic isolates to multiple antibiotics, such as clindamycin
microorganisms to acquire numerous advantages in and fluoroquinolones, and the reduced susceptibil-
terms of survivability and spread in hostile ity to antibiotics commonly used against milder
environments (Hall-Stoodley et al. 2004). cases of CDI, e.g. metronidazole (Dupont 2013),
The human gut is a clear example of a rich and allow C. difficile to persist after treatment. The
diverse microbial ecosystem, consisting of a selective advantage for their dissemination is
huge number of microbial species that play a mainly gained through the acquisition of mobile
crucial role in maintaining metabolic and immu- genetic elements involved in antibiotic resistance
nologic homeostasis (Cummings et al. 2004). and alterations of the antibiotic target sites
Despite this, few studies have been published (Spigaglia 2016).
on microbial biofilms growing in the gut, where Even if the two main C. difficile virulence
different bacterial species coexist in association factors, toxin A and toxin B (Carter et al. 2012),
with the mucosal membrane as well as the intes- and the actin-ADP-ribosylating toxin, play the
tinal luminal particles (Macfarlane and Dillon major role in clinical manifestation of CDI, also
2007). These mucosal communities show differ- adherence and motility have to be taken into
ent fermentation profiles (Macfarlane and account. In fact, the surface layer proteins
Macfarlane 2006), that may be important in (SLPs) coded by slpA are involved in adherence
modulating the host’s immune system and and inflammatory stimulation, the extracellular
contributing to some inflammatory bowel matrix-binding domain, the surface anchor pro-
diseases (ulcerative colitis, Crohn’s disease), tein needed for covalent attachment to peptido-
due to their proximity to the epithelial surface glycan, the fimbriae, and the extracellular
(Macfarlane et al. 2011). Single species biofilms polysaccharides must be all considered as addi-
of gut pathogens, such as Escherichia coli, Sal- tional factors involved in C. difficile pathogene-
monella, and Vibrio spp., are the most studied as sis (Sebaihia et al. 2006).
their extremely adhesive and invasive features The importance of adhesive properties as key
can modify the dynamics of the gut and cause virulence factor lies in the fact that adherence is
infections (Azriel et al. 2015; Sengupta et al. the first and most essential step of the biofilm
Clostridium difficile Biofilm 99

growth cycle (R€ omling and Balsalobre 2012; on flat bottomed plastic tissue culture plates
Percival et al. 2015). (Donelli et al. 2012). Afterwards, the hyperviru-
In this framework, the complex multifactorial lent strain R20291 was revealed to be a strong
process leading to the C. difficile biofilm forma- biofilm producer, identifying a link between
tion (Dawson et al. 2012; Ðapa et al. 2013; Dapa sporulation and biofilm formation with a biofilm
and Unnikrishnan 2013) should be taken into due reduction in a spo0A mutant (Dawson et al.
consideration and interventions should be also 2012). Further analysis, by Ðapa and
focused on this mode of infection, mainly in co-workers on the massive biofilm formation of
light of the recurrent CDI in ~20% of patients R20291 strain, confirmed the involvement of
(Barbut et al. 2000). A better understanding of virulence-associated proteins, Cwp84, flagella,
the process of C. difficile biofilm formation as and a putative quorum-sensing regulator, LuxS.
well as its contribution to CDI recurrence could In the same conditions, the strain 630 formed a
significantly improve disease prevention and weak biofilm (Ðapa et al. 2013).
treatment. Biofilm formation by hypervirulent and other
Findings on C. difficile biofilm, possible C. difficile strains showed differences in terms of
implications of biofilm formation in CDI patho- ability to form weak, moderate or strongly adher-
genesis, treatment efficacy of currently available ent biofilms, with the hypervirulent strains
antibiotics, and some antimicrobial alternatives always producing greater biofilms (Hammond
under investigation will be here discussed. et al. 2014; Mathur et al. 2016; Piotrowski et al.
Biofilm structure is supported by the EPS matrix,
2 Main Features of C. difficile mainly composed of proteins, extracellular DNA
Biofilm (eDNA) and polysaccharides, that provides the scaf-
fold by which bacteria adhere to each other and to
The mechanisms underlying biofilm formation in surfaces. EPS matrix is responsible for the
Clostridium species, particularly C. difficile impenetrability of bacterial biofilms, thus
(Pantaléon et al. 2014), as well as the role of contributing to the antibiotic resistance in vivo as
biofilm in CDI have not been extensively well as to the escape from immune responses during
analysed with respect to other bacterial species the infection. Specifically, C. difficile biofilm is
(Hall-Stoodley and Stoodley 2009). However, composed of a multi-component matrix (Fig. 1)
C. difficile biofilm may develop either associated made of proteins, extracellular DNA and polysac-
with intestinal microbiota or during gut charide II (PSII) (Dawson et al. 2012; Ðapa et al.
infections, by growing as mono-species or 2013). The latter is an antigen commonly found on
being part of a complex multi-species biofilm. the surface of all C. difficile species (Ganeshapillai
Therefore, biofilm mode of growth may play a et al. 2008) and detected in the matrix of several
key role in the gut colonization and bacterial C. difficile strains (Ðapa et al. 2013; Semenyuk
survival of C. difficile, affecting its pathogenesis et al. 2014). Semenyuk and colleagues found, in
and persistence, and possibly contributing to the the C. difficile biofilm matrix extract and in the
recurrence of CDI. whole cell extracts, six proteins involved in metab-
For this reason, research on the ability of olism: formate-tetrahydrofolate ligase, acetyl-CoA
C. difficile to form a biofilm has attracted consid- acetyltransferase, 2-hydroxyisocaproate
erable interest, with a number of in vitro studies CoA-transferase, NAD-specific glutamate dehydro-
being carried out in this regard. Donelli and genase, 3-hydroxybutyryl-CoA dehydrogenase,
co-workers, by using crystal violet staining and fructose-bisphosphate aldolase. On the contrary,
Field Emission Scanning Electron Microscopy cell wall-associated proteins were revealed in cell-
(FESEM), first showed that a clinical isolate of surface extracts only, the matrix proteins not arising
C. difficile (CdiBs21) formed a moderate biofilm from the cell surface. These proteins, possibly
100 C. Vuotto et al.

and lysis caused by TA systems in a small per-

centage of the bacterial cells could contribute to
the assembly of the matrix during biofilm forma-
tion for the ‘greater good’ of the population (Gil
et al. 2015). The C. difficile genome encodes a
number of putative TA systems (Gil et al. 2015)
with the MazE-MazFTA system best described
(Rothenbacher et al. 2012). However their con-
tribution towards biofilm formation has not been
Additionally, toxins and spores were discov-
ered in the biofilm matrix embedding toxigenic
C. difficile cells (Semenyuk et al. 2014). Interest-
ingly, toxins resulted to be at low concentrations
in biofilms after 24 h and at higher level in 3 day-
old biofilms, while spores have reduced germina-
tion efficiency in mature biofilms, thus presum-
Fig. 1 CLSM analysis of C. difficile in vitro biofilm after
ably facilitating the preservation of a dormant
48 h. The red-fluorescent propidium iodide stain labels
bacteria, while the lectin Concanavalin A binds to population ready to cause recurrent infections
residues of the exopolysaccharides matrix (Semenyuk et al. 2014). Remarkably, by indirect
immunofluorescence analysis, the presence of
two exosporium proteins (i.e., CdeC and the
originated from the cell lysis, most likely con- N-terminal domain of BclA1) have been detected
tribute in some way to biofilm formation on spores in C. difficile biofilms (Pizarro-
(Semenyuk et al. 2014). Guajardo et al. 2016a). By transmission electron
Other biofilm forming Gram-positive bacteria microscopy, it has been also demonstrated that
display heterogeneity within biofilms, where two exosporium morphotypes, one with a thick
vegetative cells, sporulating cells and matrix- outermost exosporium layer and another with a
producing cells coexist with different spatial thin outermost exosporium layer, were formed
localisation (Vlamakis et al. 2008). These hetero- during biofilm development (Pizarro-Guajardo
geneous populations imply differential gene et al. 2016b). Dormant spores located within
expression, and genetic regulation occurs within biofilms were detected for the duration of the
a biofilm. Electron micrographs of C. difficile experiment within a triple-stage chemostat gut
biofilms show the composition to be vegetative model inoculated with indigenous gut microbiota
cells, sporulating cells and cell debris (Donelli and C. difficile cells (Crowther et al. 2014a, b).
et al. 2012; Dawson et al. 2012). eDNA is an Sessile spores displaying increased recalcitrance
essential component of the C. difficile biofilm to germination may be compared to
matrix, as incubation with DNase I reduces the superdormant spores of Bacillus spp. (Ghosh
biofilm biomass produced (Dawson et al. 2012; et al. 2009), resulting persister cells.
Ðapa et al. 2013; Semenyuk et al. 2014). One The complex biofilm architecture of
way to explain the presence of eDNA and cell C. difficile strains has been analysed in different
debris seen within C. difficile biofilms could be in vitro studies by FESEM (Fig. 2) and Confocal
through the differential expression of toxin- Laser Scanning Microscopy (CLSM). FESEM
antitoxin (TA) systems. TA systems comprise a micrographs of C. difficile grown on glass
stable toxin, which is intracellular and only coverslips revealed wide mats of rod-shaped veg-
affects an essential cellular process, and an etative cells, spores, and sporulating cells
unstable antitoxin, which sequesters the effect interconnected by a network of extracellular
of the toxin (Wen et al. 2014). The cell death material constituted by cell debris and string-
Clostridium difficile Biofilm 101

Fig. 2 FESEM analysis of

C. difficile biofilm formed
in vitro after 48 h. Biofilms
micrographs were obtained
at an accelerating voltage
of 2 kV with magnifications
of 1000 (a) and 5000

like material connecting the cells (Fig. 3). The Maldarelli et al. 2016), and also the amount of
appearance seems to be consistent with their matrix constituting biofilm increases proportion-
being biofilms and with other SEM observations ally (Dapa and Unnikrishnan 2013).
on plastics (Dawson et al. 2012; Semenyuk et al. Whilst the so far reported in vitro evidence on
2014) or agar (Lipovsek et al. 2013). the ability of C. difficile to form a biofilm, in vivo
CLSM analysis describes more accurately the confirmation needs further investigation.
biofilm architecture, allowing one to define the C. difficile adhesion to epithelial mucosa of ani-
thickness and to visualize cells inside the biofilm mal models, including mice and hamsters, has
(Fig. 4). Semenyuk and colleagues explored the been demonstrated (Borriello et al. 1988;
evolution of biofilm structure and composition Spigaglia et al. 2013), but scarce and conflicting
over the time, identifying, after 24 h, regions proofs exist on C. difficile adherence to human
with a high concentration of apparently gut tissues (Borriello 1979; Lyra et al. 2012).
proliferating cells and cell debris as well as More specifically, regarding C. difficile biofilm
small colonies, distant from the main biofilm formation in vivo, clumps of C. difficile cells
colony, interpreted as sites of new growth formed have been observed in a mouse model associated
by cells migrated from the larger colony edge. with damaged tissue (Lawley et al. 2009), while
After 3 days, together with rod-shaped cells and aggregation or clusters of C. difficile cells were
apparent cell debris, authors detected ovoid cells observed in hamster and monoxenic mouse,
in the biofilm that were identified as spores by respectively (Spencer et al. 2014;
phase contrast microscopy. At 6 days, most of the Soavelomandroso et al. 2015). More recently,
cells in the biofilm had become spores with multispecies communities associated with the
isolated regions of vegetative cells (Semenyuk mucus of the cecum and colon have been
et al. 2014). detected, with C. difficile present as a minority
As already demonstrated for other bacterial member of communities in the outer mucus layer
species, the C. difficile biofilm thickness tends (Semenyuk et al. 2015).
to increase every day, even if the depth varied Although the in vivo data at our disposal are
according to the areas (Dawson et al. 2012; limited and results obtained in vitro might not
102 C. Vuotto et al.

Fig. 3 FESEM analysis of

C. difficile biofilm formed
on glass coverslips after
5 days; mushroom-like
structures formed by
rod-shaped vegetative
cells, spores, sporulating
cells and cell debris.
Biofilms micrographs were
obtained at an accelerating
voltage of 2 kV with
magnifications of 5000

3 Genetic Factors Behind

C. difficile Biofilm Formation

The formation of C. difficile biofilms is a multi-

factorial process involving many virulence-
associated proteins and potentially several com-
plex networks to regulate biofilm formation. The
cell surface of C. difficile plays a pivotal role
throughout the whole biofilm process, from the
initial adherence of a cell to the dispersal of
biofilm. Thus, structures directly involved in bio-
film formation have been identified by
investigating proteins and macromolecules pres-
ent on the cell surface. Flagella, Type IV pili
(T4P) and the S-layers are all implicated in
C. difficile biofilm formation.
Fig. 4 Three-dimensional CLSM image of C. difficile In the closely related bacterium, Clostridium
biofilm grown in vitro for 5 days. The red-fluorescent perfringens, T4P plays an important role in
propidium iodide stain labels bacteria, while the lectin twitching motility, biofilm formation and disease
Concanavalin A binds to residues of the pathogenesis (Varga et al. 2006). The T4P fila-
ment in Clostridium spp. is typically made up of
a major pilin subunit, PilA, and minor pilin
reflect the in vivo situation, it is likely that the
subunits, PilJ, with further genes putatively
presence of large microcolonies of C. difficile, or
involved in the retraction of the pilus to provides
biofilm communities including this species, play
the twitching motility (Varga et al. 2006;
a pivotal role in its gut colonization and survival,
Piepenbrink et al. 2014, 2015; Melville and
biofilm formation in vivo possibly being another
Craig 2013). T4P were once thought only to be
factor contributing to recurrence of CDI.
Clostridium difficile Biofilm 103

present in Gram negative bacteria, but Varga for Pseudomonas aeruginosa early biofilm
et al. (2006) first identified several putative development (Klausen et al. 2003). The role of
pilin genes within the genome of C. difficile T4P during C. difficile colonisation and persis-
strain 630, and Goulding et al. (2009) used tence remains to be investigated.
immunogold labelling to show that pili structures Recent work on how T4P is regulated in
are present on the cell surface of C. difficile dur- C. difficile has identified the bacterial secondary
ing infection in hamsters. Analysis of pilin gene messenger molecule Bis-(30 -50 )-cyclic dimeric
transcripts from in vitro C. difficile biofilm guanosine monophosphate (c-di-GMP) as a key
cultures, shows an upregulation of pilA1 component to the regulatory pathway. In Gram
transcripts compared to planktonic cultures negative bacteria, c-di-GMP modulates virulence
(Maldarelli et al. 2016), which is even more attributes, such as biofilm formation in Vibrio
prominent in C. difficile strain R20291 compared cholerae (Tischler and Camilli 2005) and
to strain 630 (Purcell et al. 2016). The impor- P. aeruginosa (Kulasakara et al. 2006),
tance of T4P in C. perfringens can be seen in decreased flagella-mediated motility in
mutants that are defective in T4P formation as Escherichia coli, and cell differentiation in
these mutants display abnormal biofilm forma- Caulobacter crescentus (Aldridge et al. 2003).
tion compared to the wild-type strain (Varga Two enzymes, diguanylatecyclases (DGCs) and
et al. 2008). In C. difficile, mutants that have a phosphodiesterases (PDEs) that either synthesise
disrupted pilA1 gene lack T4P structures on the or degrade c-di-GMP (R€omling and Amikam
cell surface under laboratory conditions 2006), tightly control the intracellular levels of
(Bordeleau et al. 2015). Interestingly, T4P play c-di-GMP. C. difficile is unusual among Gram
an important role in the early stages of C. difficile positive organisms by the number of DGCs &
biofilm formation, as mutants with a pilA1 dis- PDEs encoded on the genome; strain 630 has
ruption show a reduced biofilm biomass com- 37 putative c-di-GMP metabolising enzymes.
pared to wild-type (Maldarelli et al. 2016; Ectopic expression of 31 of these enzymes in
Purcell et al. 2016). However, T4P seem to play the surrogate organism, V. cholerae, confirmed
little role in the maturation of a biofilm, as these these genes as either having DGC or PDE activ-
mutants showed no difference in biofilm biomass ity (Bordeleau et al. 2011). Interestingly, heter-
compared to wild-type when grown over 7 days ologous and homologous expression of
(Maldarelli et al. 2016). Up to nine putative pilin- C. difficile 630 CD1420 (dccA) in either
like proteins are encoded on the C. difficile V. cholerae or C. difficile, respectively, increased
genome (Melville and Craig 2013; Maldarelli cellular levels of c-di-GMP and induced biofilm
et al. 2014), three of these being designated as formation (Bordeleau et al. 2011; Purcell et al.
major pilin subunits (pilA1–3). The biological 2012). Through overexpression of dccA, high
function for each of these pilin genes remains intracellular levels of c-di-GMP resulted in
unclear, with current hypotheses suggesting increased expression of the genes in the T4P
T4P made from these different pilin subunits operon and a greater number of pili observed on
could perform different functions, or pilin the cell surface (Bordeleau et al. 2015). In other
switching could be a mechanism for immune bacteria, c-di-GMP controls the transcription and
evasion, or, as many of these are not located in translation of many genes by direct binding to c-
T4P operons, these could be non-functional. In di-GMP riboswitches (Sudarsan et al. 2008).
other bacteria, T4P plays a pivotal role in biofilm Riboswitches are mRNA molecules that bind
formation and disease pathogenesis; T4P is small molecules (such as c-di-GMP) resulting
essential for passage of Neisseria meningitidis in the transcription of downstream genes
to cross the blood-brain barrier (Nassif et al. (Winkler and Breaker 2005). In this way, the
1994), whilst T4P-mediated motility is important same small molecule can coordinate multiple
104 C. Vuotto et al.

genetic pathways. RNA-seq experiments first significantly less biofilm biomass compared to
identified a Type II c-di-GMP riboswitch located wild-type (Ðapa et al. 2013). The genetic
up stream of the start of the C. difficile major T4P organisation of the C. difficile flagella operon
operon (pilA1) (Soutourina et al. 2013), called can be split into three parts, however the F2
Cdi2_4, which is switched ‘ON’ via a conforma- locus is the most divergent between the genomes
tional change upon binding c-di-GMP to the of different C. difficile strains (Stabler et al.
riboswitch to relieve a predicted Rho-independent 2009; Stevenson et al. 2015). The F2 locus
transcription terminator (Bordeleau et al. 2015). encodes genes involved in glycosylation of the
Between different strains of C. difficile there flagella with sugar moieties, and the disruption of
appears to be subtle variations in pilA1 expression these genes resulted in the production of flagella
patterns during biofilm formation with strains on the cell surface even though most of these
630 and R20291 (Purcell et al. 2016), which could mutants were non-motile (Twine et al. 2009;
be due to differences in the total c-di-GMP levels. Faulds-Pain et al. 2014; Valiente et al. 2016).
Research on other regulatory proteins within Interestingly, these mutants produced more bio-
C. difficile suggests that its pathogenesis is inti- film biomass compared to the wild-type strain
mately linked to the metabolic state of the bacte- (Faulds-Pain et al. 2014; Valiente et al. 2016).
rium (Bouillaut et al. 2015). CodY is a pleotropic In the closely related bacterium, Bacillus subtilis,
regulator involved in the adaptive response of inhibition of flagella rotation acts as a mechani-
Gram-positive bacteria to low nutrient levels, cal trigger to activate the DegS-DegU
and in C. difficile, an estimated 52 genes are two-component signal transduction system,
directly regulated by CodY (Dineen et al. 2010; which regulates biofilm formation and matrix
Bouillaut et al. 2015). One of these genes is pdcA production (Cairns et al. 2013, 2014). Although
(CD1515), which is a PDE enzyme that affects no DegS/DegU homologues have been identified
the regulation of flagella biosynthesis by in C. difficile, this could be why these flagellate,
influencing c-di-GMP levels (Purcell et al. non-motile mutants produced more biofilm bio-
2012; Purcell et al. 2017). Thus, through this mass, although more work is needed to under-
regulatory pathway, C. difficile biofilm formation stand the regulatory mechanisms behind this
is connected to the nutrient availability of the phenotype.
bacterium. c-di-GMP acts as a signalling mole- Using riboswitches is one way C. difficile
cule coordinating the transition from a plank- regulates the change from motility to biofilm,
tonic, motile lifestyle to a sessile, biofilm however other regulatory RNA molecules appear
lifestyle in many bacterial pathogens. In to play a role. Small non-coding RNAs (sRNAs)
C. difficile, high c-di-GMP levels directly repress act by base pairing with their target mRNAs,
the major flagella operon flgB through a Type I c- leading to modulation of mRNA stability or
di-GMP riboswitch, Cdi1_3, located 496 bp translation (Chao and Vogel 2010; Soutourina
upstream of the flgB start codon (Sudarsan et al. 2017). Some sRNAs require an RNA chaperone
2008; Soutourina et al. 2013). Through the two protein called Hfq to help the base pair binding of
types of riboswitches, one family of signalling the sRNA and mRNA molecules. In other bacte-
molecules can regulate the expression of T4P and ria, mutating Hfq has pleotropic effects on cell
flagella biosynthesis during C. difficile biofilm physiology, ranging from increased sensitivity to
formation. external stresses (detergents, iron limitation and
A decrease in flagella transcripts would indi- oxidative stress), to increased biofilm formation,
cate a limited role for flagella during biofilm or reduced virulence (Chao and Vogel 2010).
formation, and targeted disruption of fliC gene The creation of a C. difficile hfq gene disruption
in strains 630 or R20291 had no effect on biofilm has been unsuccessful to date, so Boudry et al.
formation compared to the wild-type strains (2014), used a knockdown approach to decrease
(Faulds-Pain et al. 2014; Valiente et al. 2016). Hfq protein levels fivefold compared to wild-
However, one report has shown a fliC mutant had type to determine its contribution toward cell
Clostridium difficile Biofilm 105

physiology. Using this approach, the authors interspecies communication molecule.

observed an increase in biofilm formation in C. difficile encodes a luxS homologue and
the Hfq depleted strain, indicating that sRNAs produces a chemically active AI-2 molecule
play a role in negatively regulating biofilm that can induce homologous and heterologous
formation. Alongside this, the authors gene expression (Carter et al. 2005; Lee and
observed a decrease in flagella present on the Song 2005). Biofilm formation in a C. difficile
cell surface and increased expression of cell luxS mutant was severely diminished compared
wall/membrane proteins, all of which could to wild-type strain, where not even a bacterial
have contributed to the increase in biofilm for- monolayer was able to form (Ðapa et al. 2013;
mation (Boudry et al. 2014). Slater and Unnkrishnan 2015).The regulatory
Another cell surface organelle that has been pathway behind AI-2 induced biofilm formation
implicated in C. difficile biofilm formation is the is currently unknown. In the gut mucosa,
S-layer. The C. difficile S-layer (Cerquetti et al. C. difficile interacts with members of the sessile
2000) is a two-dimensional paracrystalline pro- community (Lawley et al. 2009; Buckley et al.
tein array coating the cell and is made up of SlpA 2011; Donelli et al. 2012; Crowther et al.
subunits that are post-translational cleaved by 2014a, b; Semenyuk et al. 2015), where such
another protein called Cwp84 (de la Riva et al. interspecies signalling could play an important
2011; Fagan and Fairweather 2014). The S-layer role for disease progression.
harbours up to 28 different cell wall proteins that Based on current research, it seems clear that
are anchored to the cell wall by CWB2 protein the genetic regulation behind C. difficile biofilm
domains (Fagan and Fairweather 2014; Willing formation is extremely complex and several dif-
et al. 2015). Disruption of cwp84 results in ferent global regulators that link various meta-
uncleaved SlpA in the cell wall, which in turn bolic pathways influence it. The C. difficile
results in aberrant retention of other cell wall sporulation master regulator, spo0A, besides
proteins at the cell surface (Kirby et al. 2009; coordinating sporulation by undergoing post-
de la Riva et al. 2011). The effect of cwp84 translational phosphorylation (Spo0A-P) in
disruption on biofilm formation was dependent order to activate the sigma factor cascade (Pettit
on the strain background. In strain R20291 a et al. 2014; Al-Hinai et al. 2015), also plays a role
cwp84 mutant showed reduced biofilm formation in biofilm formation. Disrupting the spo0A gene
(Ðapa et al. 2013), whereas in strain 630 this resulted in a reduced biofilm phenotype that
mutant showed an increase in biofilm formation could be restored by complementation (Dawson
(Pantaléon et al. 2015). As these strains encode a et al. 2012; Ðapa et al. 2013). In Bacillus spp.,
different array of proteins that are predicted to the intracellular concentration of Spo0A-P is crit-
associate with the S-layer (Biazzo et al. 2013), an ical to determining if the cell proceeds down
immature S-layer may contain different surface- either the sporulation pathway (high Spo0A-P
associated proteins between the two strains. levels), or biofilm pathway (low Spo0A-P levels)
Whether the S-layer per se is involved in biofilm (Mhatre et al. 2014). During the early stages of
formation or if this effect is due to the proteins biofilm formation Spo0A-P induces the expres-
associated with the S-layer remains uncertain. sion of sinI, which inhibits a protein that
Regarding quorum sensing, bacteria detect a represses the biofilm matrix genes, SinR
threshold level of autoinducer (AI) molecules (Vlamakis et al. 2013; Cairns et al. 2014).
and activate a signal cascade that leads to altered C. difficile encodes homologues of sinI and sinR
gene expression. The AI-2 molecule is (Edwards et al. 2014), however their role in
synthesised by LuxS and is produced by Gram- regulating biofilms, and the regulon of SinR,
positive and Gram-negative bacteria. Due to the are unknown and deserves further investigation.
number of bacteria that can produce and detect It has also been demonstrated that biofilm
AI-2 molecules, this quorum signalling mecha- formation in C. difficile may be stress-inducible;
nism is thought to function as an intra- and exposure of cells to sub-inhibitory
106 C. Vuotto et al.

concentrations of antibiotics, such as metronida- defined as ‘closed’ or ‘open’. Closed (or static)
zole and vancomycin, induce biofilm formation biofilm models, such as the popular microtiter
(Ðapa et al. 2013; Vuotto et al. 2016). In other tray based models, are based on batch culture,
bacteria, this stress-induced biofilm formation is in which there is limited nutrient availability and
induced by the SOS regulatory network in aeration, as well as a build-up of metabolic
response to DNA damage, through activation of products. Open (or dynamic) biofilm model
the transcriptional repressor, LexA, by the systems are based on continuous flow models,
recombinase protein, RecA (Butala et al. 2009). whereby fresh media replace metabolic products
Mutation of lexA in C. difficile caused pleotropic and waste constantly. Whilst open biofilm
effects to the cell: elongated cell morphology, models may be better able to simulate sheer
decreased sporulation and motility and increased forces and flow, they often require more techni-
biofilm formation (Walter et al. 2015). In silico cal expertise and complex equipment than closed
analysis of predicted LexA binding sites within systems, and so are less amenable to high
the C. difficile genome suggests LexA could reg- throughput workloads.
ulate up to 29 loci (Walter et al. 2014). How this The use of microtiter trays is one of the sim-
regulatory pathway contributes to C. difficile bio- plest methods used to investigate both mono- and
film formation is unclear and warrants further poly-microbial species biofilm formation. This
investigation. method has been used to determine some of the
genetic mechanisms behind C. difficile biofilm
formation, as well as its interaction with other
4 In Vitro and In Vivo Models gut microbiota. Donelli et al. (2012) found that
to Study the Interactions several gastrointestinal residing bacteria were
of Sessile Microorganisms able to cooperatively form a biofilm when
co-cultured together and in addition, highlighting
The mammalian intestinal mucosa is home to a a positive interaction between C. difficile and
complex mixture of microbial communities, Finegoldia magna. This method has been also
which can aggregate to form mats or biofilm used to characterise the inter-kingdom
structures over the epithelial cells. C. difficile interactions between C. difficile and Candida
cells can associate with these microbial albicans. Biofilm formation of C. albicans was
communities during CDI (Lawley et al. 2009; reduced when co-incubated with filter sterilized
Goulding et al. 2009; Spencer et al. 2014; C. difficile growth media, which was attributed to
Semenyuk et al. 2015). The interactions between the production of p-cresol by C. difficile (van
microbial species within a biofilm can vary Leeuwen et al. 2016), although the direct inter-
depending on the associated microbial species, action was not reported. The interactions
and this can affect the spatial organisation of between fungi and other intestinal microflora
cells within biofilm. Sessile microbes can form are probably more complex than we assume and
synergistic, exploitive or competitive others have found a correlation between CDI and
relationships with other biofilm-forming the presence of Candida spp. (Raponi et al.
microorganisms (Liu et al. 2016). Since under- 2014). The inhibition of C. difficile grown in a
standing the interactions between C. difficile and planktonic culture by probiotic strains Lactoba-
the sessile community could be key to designing cillus and Bifidobacteria has been documented
defined microbial treatments for recurrent CDI, previously (Plummer et al. 2004; Trejo et al.
several in vitro and in vivo models have been 2010), but recent unpublished studies by
developed to study these kind of interactions. Normington and coworkers have shown these
The use of in vitro models allows researchers to probiotic organisms inhibit C. difficile biofilm
manipulate and control certain factors and/or formation (Normington et al. 2017).
conditions, thus providing a valuable tool for Regarding the open (or dynamic) biofilm
biofilm research. Systems can generally be model systems, Crowther et al. (2014b)
Clostridium difficile Biofilm 107

developed a modified version of the continuous

triple chemostat system (Macfarlane et al. 1998)
to monitor the sessile populations, by using glass
rods suspended from the lid. During simulated
CDI, the authors observed consistent sessile
populations formed upon the different rods sam-
pled at the same time. The composition of the
sessile communities in these experiments was
Bacteroides spp., Bifidobacteria spp., Lactoba-
cillus spp. and Enterococcus spp., however as
determined by total viable counts, many more
bacterial species must be associated within the
biofilms (Crowther et al. 2014a). Currently,
Buckley and colleagues are using bacterial 16S
rRNA sequencing analysis to identify the com-
position of the sessile community from these
rods, and in addition, they recently identified
Fig. 5 Scanning electron micrograph of an in vitro
fungal species within the biofilm structure polymicrobial biofilm. A biofilm containing C. difficile
(Unpublished results). Upon instillation of (red cells), Candida spp. (green cells) and Staphylococ-
C. difficile spores, these spores became cus spp. (blue cells) was grown anaerobically for 3 days.
White scale bar indicates 20 μm (SEM image taken from
associated with the biofilm, and both sessile
Normington et al. 2017)
spore and vegetative populations were isolated
during the CDI phase (Crowther et al. 2014a, b;
Unpublished results). The interactions between the biofilm, or allows extra-intestinal invasion
C. difficile on the other sessile populations are (Ng et al. 2013), as seen by (Goulding et al.
currently under investigation (Fig. 5). 2009; Lawley et al. 2009), remains to be
In vivo models of CDI have been used to determined.
specifically identify the bacterial populations
associated with the mucus layer during disease.
Using paraffin embedded sections, to preserve
the mucus layer, and fluorescent in situ 5 Effects of Antibiotics
hybridisation (FISH), Semenyuk et al. (2015) on C. difficile Biofilm
identified C. difficile vegetative cells within the
outer mucus layer. Microbial taxonomy analysis Biofilm formation has been demonstrated to be
from 16S rRNA sequences recognized other bac- an important factor enhancing the antimicrobial
terial genera residing within the mucus layer, resistance (Ciofu et al. 2017). In fact, during
from several families belonging to Bacteroidetes infection the biofilm mode of growth protects
and Firmicutes (Lactobacillaceae, cells from antibiotic treatment, their resistance
Lachnospiraceae and Clostridium cluster XVII often increasing from 10- to 1000-fold compared
and XIV). Those microbial species that directly with the same cells growing planktonically (Mah
interact with C. difficile in vivo are still and O’Toole 2001; Hoiby et al. 2010). Several
unknown. Interestingly, during the early phase mechanisms can contribute to antibiotic resis-
of CDI an increase in Enterbacteriaceae was tance in biofilm; including the biofilm matrix,
observed within the mucosal populations acting as a physical barrier that affects penetra-
(Semenyuk et al. 2015). Whether such an tion of antimicrobial agents (Flemming and
increase enhances C. difficile recruitment into Wingender 2010), the presence of persister cells
108 C. Vuotto et al.

(Shah et al. 2006) and the genetic mutations with metronidazole resulted to be able to only
occurring within bacteria in biofilm (Tyerman temporarily suppress Gardnerella vaginalis
et al. 2013). biofilms but not completely eradicate it, in most
Tolerance mechanisms have been proposed in cases rapidly regaining activity after treatment
C. difficile biofilm (Ðapa et al. 2013), so the ending (Swidsinski et al. 2008, 2014). Another
effect of antibiotics most commonly used to study also showed that 30 BV-associated bio-
treat CDI, such as metronidazole and vancomy- film-forming bacteria were resistant to metroni-
cin (Peng et al. 2017), has been assessed against dazole (Alves et al. 2014).
biofilm-growing cells and pre-formed biofilms. Vancomycin, compared to metronidazole,
Semenyuk and colleagues determined that demonstrates a higher clinical cure rate in adults
630 and VPI 10463 C. difficile cells grown as with severe CDI and a similar clinical cure rate in
biofilm for 20 h had greater resistance to metro- moderate CDI cases, thus becoming the
nidazole than planktonic cells, with 1 μg/ml of recommended therapy for more severe cases
antibiotic inhibiting liquid cell growth by about (Ofosu 2016). However, regarding its ability to
100-fold and 100 μg/ml reducing only about a act against mature biofilms, a number of papers
tenfold of the sessile cells. These data have been published on staphylococcal species
demonstrated that biofilms conferred a 100-fold (Meeker et al. 2016; Ozturk et al. 2016; Hashem
increase in metronidazole resistance (Semenyuk et al. 2017; Jimi et al. 2017) but limited and not
et al. 2014). encouraging data are so far available for
In addition to being ineffective to counteract C. difficile.
in vitro C. difficile biofilm, it has been demonstrated Ðapa and co-workers first analysed the influ-
that, at sub-inhibitory concentrations, metronida- ence of vancomycin on biofilms of a C. difficile
zole can even enhance biofilm formation in specific strain belonging to the PCR-ribotype 027, by
cases. In particular, three clinical strains belonging examining the effects of different concentrations
to PCR-ribotype 010, non-toxigenic and showing of antibiotic. High concentrations of vancomycin
different metronidazole susceptibility profiles, (20 μg/mL) failed to kill bacteria within biofilms
exhibited variation in biofilm-forming ability. In while sub-inhibitory and inhibitory concentrations
the presence of metronidazole, a susceptible strain of vancomycin (0.25 μg/mL and 0.5 μg/mL,
and a strain with reduced-susceptibility revealed a respectively) induced C. difficile biofilm formation.
significant increase in biofilm biomass, due to a This suggests that increased antibiotic resistance in
more abundant EPS matrix production, while the C. difficile may be mediated by the thick biofilm
biofilm-forming ability of the stable-resistant strain matrix and/or by the physiological state of bacteria
was not affected by the antibiotic pressure (Vuotto within biofilms (Ðapa et al. 2013). These results
et al. 2016). This study highlights the possibility that were corroborated by Mathur et al. (2016), whom
the exposure of C. difficile to low concentrations of observed low efficacy of vancomycin against vari-
antibiotic present in the gut at the beginning or end ous PCR-ribotypes.
of antibiotic therapy for CDI could serve as stress Using a triple-stage human gut model,
signal and, thus, stimulate biofilm production, with Crowther and colleagues simulated CDI and
severe clinical implications in the treatment failure determined the effect of vancomycin on the
and recurrence of CDI. When similar experiments motile and sessile C. difficile populations. Van-
were carried out by using Bacteroides fragilis, comycin exposure reduced the C. difficile plank-
opposed results were obtained. In fact, sub- inhibi- tonic populations to below the limit of detection,
tory concentrations of metronidazole were able to however the sessile populations were unaffected.
inhibit biofilm formation (Silva et al. 2014). This could be due to the levels of vancomycin
Aside from C. difficile biofilms, metronida- that were detected within the biofilms [mean
zole efficacy has been evaluated on biofilm- 40.4 mg/L (range 38.7–43.4 mg/L)] compared
related bacterial vaginosis (BV). Monotherapy to those (54.7 mg/L) of the vessel lumen
Clostridium difficile Biofilm 109

(Crowther et al. 2014a). A reduced level of van- The conventional antibiotics used in CDI ther-
comycin within the biofilm could prevent a criti- apy are often unsuccessful and recurrent
cal level of vancomycin from being achieved, or infections may occur, perhaps due to its ability
even further enhance matrix production. We to grow as a biofilm thus impairing antimicrobial
clearly observed a differential response of sessile activity. Different approaches, which are an
bacteria to antimicrobial administration, with alternative to the use of antibiotics, have been
C. difficile spores being largely unresponsive proposed to decrease C. difficile biofilm forma-
either to clindamycin instillation. tion or disrupt mature biofilm.
The effect of tigecycline, teicoplanin, rifam- Among the huge number of antimicrobial
picin and nitazoxanide was also evaluated on the compounds today at our disposal, relatively few
biofilm of five different C. difficile strains, noting have been tested so far against C. difficile bio-
that the sensitivities of these biofilms to different film. The first one tested was Manuka honey, its
antimicrobials were strain-dependent, regardless anti-biofilm properties on other species being
of the produced biomass (Mathur et al. 2016). already demonstrated (Badet and Quero 2011).
Biofilms formed by two C. difficile strains, a
ribotype 027 strain and a ribotype 106 strain,
were used to test the effect of Manuka honey at
6 Alternatives to Counteract
varying concentrations of 1–50% (w/v). A dose-
Biofilm-Growing C. difficile
dependent response was observed for both test
strains, with the optimum Manuka honey activity
Antibiotic administration, although carried out at
obtained at 40–50% (v/v) (Hammond et al.
higher doses over a prolonged period, often fails
2014). Consistent results were also obtained by
to counteract biofilm-related infections. In addi-
evaluating its efficacy on clinical C. difficile
tion, antibiotic overuse and misuse are key
strains belonging to four prominent PCR
factors contributing to the global increase of
ribotypes (R017, R023, R027 and R046)
antibiotic resistance. Alternative therapeutic
(Piotrowski et al. 2017).
agents with antibacterial properties that prevent,
The antimicrobial agent thuricin CD, a
disrupt, weaken or kill the microbial community
sactibiotic produced by a bacterial strain derived
within a biofilm, are becoming increasingly
from a human faecal sample, was also assessed
attractive. In particular, anti-biofilm compounds:
against biofilms of R027, Liv022 R106 and
(i) may prevent biofilm formation by killing
DPC6350, alone or in combination with some
planktonic cells or blocking bacterial adhesion;
antibiotics commonly used to treat CDI. Results
(ii) may counteract mature biofilms by
underlined the effectiveness of thuricin CD
destabilising the matrix or by making the micro-
against all the tested strains and its ability to
bial cells susceptible to antimicrobial and/or host
significantly potentiate the efficacy of the
defence mechanisms; (iii) may undo virulence
antibiotics rifampicin, tigecycline, vancomycin
factors involved in biofilm formation or may affect
and teicoplanin against R027 biofilms (Mathur
quorum sensing; (iv) may have a bactericidal effect
et al. 2016).
on biofilm-growing cells (Roy et al. 2017).
More innovative proposals to avoid treatment
Efforts to fight these microbial communities
failure and recurrent CDI infection have been
include the use of different compounds, alone or
sort through the use of bacteriophages and pho-
in combination, to target different phases of bio-
todynamic therapy.
film, drug repurposing, peptides, nanomaterials,
It has been demonstrated that some
and medical device coatings refractory to micro-
bacteriophages have good activity against biofilms
bial adhesion or functionalised with anti-biofilm
of different species by invading it and significantly
compounds (Ribeiro et al. 2016).
110 C. Vuotto et al.

reducing the viable numbers of cells. Accordingly, 2015) or by employing specific QS inhibitors
bacteriophages appear to be a highly promising able to interfere with biofilm maturation (Ðapa
therapeutic option for eradicating CDI by replacing et al. 2013).
antibiotics or supplementing them (Azeredo and
Sutherland 2008). Nale and colleagues evaluated
the impact of a four-phage cocktail on C. difficile
7 Conclusions
ribotype 014/020 biofilm, in vitro alone or in com-
bination with vancomycin treatment in Galleria
Biofilms are the most representative form of bac-
mellonella larva CDI model. Phages were able to
terial growth in the large intestine, with biofilm
prevent in vitro biofilm formation, to penetrate
formation being known to influence the ability of
established biofilms, and also to reduce colonization
pathogens to colonize and establish during
and/or prevent disease in the Galleria mellonella
model, when used alone or in combination with
Clinically relevant strains of C. difficile have
vancomycin (Nale et al. 2016).
been proven to be able to form biofilms in vitro
Photodynamic therapy, more frequently
that appear as complex cellular processes involving
applied to determine its usefulness to treat peri-
an array of different regulating proteins, intracellu-
odontal (Sculean et al. 2015) and wound
lar chemical signals and effector proteins, all having
(Percival et al. 2014) infections, has also been
a role in different aspects of bacterial physiology.
tested against planktonic and sessile-growing
Although much has been done to understand the
C. difficile strains. This approach exploits the
regulatory signals governing biofilm formation, this
ability of light-activated photosensitisers (PS) to
picture is still incomplete and the details of the
produce reactive oxygen species (ROS) lethal to
precise function and regulation of each of these
cells. Three of thirteen PS screened were able to
proteins/pathways remain to be studied. This inter-
kill 99.9% of the tested C. difficile strains both in
twinement is likely to allow an accurate modulation
planktonic and biofilm states, after exposure to
of the differentiation pathways for motility, biofilm
red laser light (0.2 J/cm2) (De Sordi et al. 2015).
formation or sporulation at a spatio-temporal
Although PS are an interesting perspective for
biofilm eradication, as they work by producing
Even if the C. difficile colonization in vivo has
free radical species, their use in the human gas-
yet to be analysed in deep, the demonstrated ability
trointestinal tract remains limited without further
to form a mature biofilm in vitro seems to be
development of the technology.
predictive of the in vivo colonization mode. In
Recent discoveries of alternative C. difficile
CDI, the establishment of persistent biofilms
treatments include rhodanine derivatives
in vivo, in addition to the formation of spores,
(AbdelKhalek et al. 2016) and acyldespiptides (Gil
could potentially explain the occurrence of recur-
and Paredes-Sabja 2016), that exhibit in vitro activ-
rent infections. Thus, a potential infection model
ity against planktonic populations, while their effi-
involving the colonization of the colon by
cacy against the sessile populations remains to be
C. difficile through the formation of microcolonies
or biofilms, which is followed by toxin production.
In addition to the antimicrobial compounds
This in vivo biofilm mode of growth possibly
already tested and the other approaches above
protects the bacterium from the cellular immune
mentioned, further possibilities to interfere
responses triggered by the toxins and from the
with C. difficile biofilm could presumably
antibiotic treatment. In light of the above, a deeper
come from the discovery of novel compounds
knowledge of the factors involved in the C. difficile
that bind c-di-GMP riboswitches (Furukawa
biofilm development during infection might provide
et al. 2012), from the use of DNase as enhancer
an advanced understanding of the role of biofilm
of the effect of metronidazole (Machado et al.
in CDI.
Clostridium difficile Biofilm 111

References Buckley AM, Spencer J, Candlish D et al (2011) Infection

of hamsters with the UK Clostridium difficile ribotype
027 outbreak strain R20291. J Med Microbiol
AbdelKhalek A, Ashby CR, Patel BA et al (2016) In vitro
antibacterial activity of rhodanine derivatives against
Butala M, Žgur-Bertok D, Busby SJW (2009) The bacte-
pathogenic clinical isolates. PLoS One 11(10):
rial LexA transcriptional repressor. Cell Mol Life Sci
Aldridge P, Paul R, Goymer P et al (2003) Role of the
Cairns LS, Marlow VL, Bissett E et al (2013) A mechani-
GGDEF regulator PleD in polar development of
cal signal transmitted by the flagellum controls signal-
Caulobacter crescentus. Mol Microbiol
ling in Bacillus subtilis. Mol Microbiol 90:6–21
Cairns LS, Hobley L, Stanley-Wall NR (2014) Biofilm
Al-Hinai MA, Jones SW, Papoutsakis ET (2015) The
formation by Bacillus subtilis: new insights into regu-
Clostridium sporulation programs: diversity and pres-
latory strategies and assembly mechanisms. Mol
ervation of endospore differentiation. Microbiol Mol
Microbiol 93:587–598
Biol Rev 79:19–37
Carter GP, Purdy D, Williams P et al (2005) Quorum
Alves P, Castro J, Sousa C et al (2014) Gardnerella
sensing in Clostridium difficile: analysis of a luxS-
vaginalis outcompetes 29 other bacterial species
type signalling system. J Med Microbiol 54:119–127
isolated from patients with bacterial vaginosis, using
Carter GP, Rood JI, Lyras D (2012) The role of toxin A
in an in vitro biofilm formation model. J Infect Dis
and toxin B in the virulence of Clostridium difficile.
Trends Microbiol 20:21–29
Azeredo J, Sutherland IW (2008) The use of phages for
Cerquetti M, Molinari A, Sebastianelli A et al (2000)
the removal of infectious biofilms. Curr Pharm
Characterization of surface layer proteins from differ-
Biotechnol 9:261–266
ent Clostridium difficile clinical isolates. Microb
Azriel S, Goren A, Rahav G et al (2015) The stringent
Pathog 28:363–372
response regulator DksA is required for Salmonella
Chao Y, Vogel J (2010) The role of Hfq in bacterial
enteric Serovar Typhimurium growth in minimal
pathogens. Curr Opin Microbiol 13:24–33
medium, motility, biofilm formation, and intestinal
Chilton CH, Crowther GS, Freeman J et al (2014) Suc-
colonization. Infect Immun 84:375–384
cessful treatment of simulated Clostridium difficile
Badet C, Quero F (2011) The in vitro effect of manuka
infection in a human gut model by fidaxomicin first
honeys on growth and adherence of oral bacteria.
line and after vancomycin or metronidazole failure. J
Anaerobe 17:19–22
Antimicrob Chemother 69:451–462
Barbut F, Richard A, Hamadi K et al (2000) Epidemiol-
Ciofu O, Rojo-Molinero E, Macià MD et al (2017) Anti-
ogy of recurrences or reinfections of Clostridium diffi-
biotic treatment of biofilm infections. APMIS
cile-associated diarrhea. J Clin Microbiol
Crowther GS, Chilton CH, Todhunter SL et al (2014a)
Biazzo M, Cioncada R, Fiaschi L et al (2013) Diversity of
Comparison of planktonic and biofilm-associated
cwp loci in clinical isolates of Clostridium difficile. J
communities of Clostridium difficile and indigenous
Med Microbiol 62:1444–1452
gut microbiota in a triple-stage chemostat gut model. J
Bordeleau E, Fortier LC, Malouin F et al (2011) c-di-
Antimicrob Chemother 69:2137–2147
GMP turn-over in Clostridium difficile is controlled
Crowther GS, Chilton CH, Todhunter SL et al (2014b)
by a plethora of diguanylatecyclases and
Development and validation of a chemostat gut model
phosphodiesterases. PLoS Genet 7:e1002039
to study both planktonic and biofilm modes of growth
Bordeleau E, Purcell EB, Lafontaine DA et al (2015)
of Clostridium difficile and human microbiota. PLo
Cyclic di-GMP riboswitch-regulated type IV pili con-
SONE 9:e88396
tribute to aggregation of Clostridium difficile. J
Cummings JH, Antoine JM, Azpiroz F et al (2004)
Bacteriol 197:819–832
PASSCLAIM: gut health and immunity. Eur J Nutr
Borriello SP (1979) Clostridium difficile and its toxin in
the gastrointestinal tract in health and disease. Res
Dapa T, Unnikrishnan M (2013) Biofilm formation by
Clin Forums 1:33–35
Clostridium difficile. Gut Microbes 4:397–402
Borriello SP, Welch AR, Barclay FE et al (1988) Mucosal
Ðapa T, Leuzzi R, Baban ST et al (2013) Multiple factors
association by Clostridium difficile in the hamster
modulate biofilm formation by the anaerobic pathogen
gastrointestinal tract. J Med Microbiol 25:191–19629
Clostridium difficile. J Bacteriol 195:545–555
Boudry P, Gracia C, Monot M et al (2014) Pleiotropic role
Dawson LF, Valiente E, Faulds-Pain A et al (2012)
of the RNA chaperone protein Hfq in the human
Characterisation of Clostridium difficile biofilm for-
pathogen Clostridium difficile. J Bacteriol
mation, a role for Spo0A. PLoS One 7:e50527
de la Riva L, Willing SE, Tate EW et al (2011) Roles of
Bouillaut L, Dubois T, Sonenshein AL et al (2015) Inte-
cysteine proteases Cwp84 and Cwp13 in biogenesis of
gration of metabolism and virulence in Clostridium
the cell wall of Clostridium difficile. J Bacteriol
difficile. Res Microbiol 166:375–383
112 C. Vuotto et al.

De Sordi L, Butt MA, Pye H et al (2015) Development of Manuka honey. BMC Complement Altern Med
Photodynamic Antimicrobial Chemotherapy (PACT) 14:329
for Clostridium difficile. PLoS One 10:e0135039 Hashem AA, Abd El Fadeal NM et al (2017) In vitro
Dineen SS, McBride SM, Sonenshein AL (2010) Integra- activities of vancomycin and linezolid against
tion of metabolism and virulence by Clostridium diffi- biofilm-producing methicillin-resistant staphylococci
cile CodY. J Bacteriol 192:5350–5362 species isolated from catheter-related bloodstream
Donelli G (2006) Vascular catheter-related infection and infections from an Egyptian tertiary hospital. J Med
sepsis. Surg Infect 7:S25–S27 Microbiol 66:744–752
Donelli G, Vuotto C, Cardines R et al (2012) Biofilm- Heydorn A, Ersboll B, Hentzer M et al (2000) Experimen-
growing intestinal anaerobic bacteria. FEMS Immunol tal reproducibility in flow-chamber biofilms. Microbi-
Med Microbiol 65:318–325 ology 146:2409–2415
Dupont HL (2013) Diagnosis and management of Clos- Hoiby N, Bjarnsholt T, Givskov M et al (2010) Antibiotic
tridium difficile infection. Clin Gastroenterol Hepatol resistance of bacterial biofilms. Int J Antimicrob
11:1216–1223 Agents 35:322–332
Edwards AN, Nawrocki KL, McBride SM (2014) Jimi S, Miyazaki M, Takata T et al (2017) Increased drug
Conserved oligopeptide permeases modulate sporula- resistance of meticillin-resistant Staphylococcus
tion initiation in Clostridium difficile. Infect Immun aureus biofilms formed on a mouse dermal chip
82:4276–4291 model. J Med Microbiol 66:542–550
Fagan RP, Fairweather NF (2014) Biogenesis and Kirby JM, Ahern H, Roberts AK et al (2009) Cwp84, a
functions of bacterial S-layers. Nat Rev Microbiol surface-associated cysteine protease, plays a role in
12:211–222 the maturation of the surface layer of Clostridium
Faulds-Pain A, Twine SM, Vinogradov E et al (2014) The difficile. J Biol Chem 284:34666–34673
post-translational modification of the Clostridium dif- Klausen M, Aaes-Jørgensen A, Molin S et al (2003)
ficile flagellin affects motility, cell surface properties Involvement of bacterial migration in the development
and virulence. Mol Microbiol 94:272–289 of complex multicellular structures in Pseudomonas
Flemming HC, Wingender J (2010) The biofilm matrix. aeruginosa biofilms. Mol Microbiol 50:61–68
Nat Rev Microbiol 8:623–633 Koch B, Worm J, Jensen LE et al (2001) Carbon limita-
Furukawa K, Gu H, Sudarsan N et al (2012) Identification tion induces s-dependent gene expression in Pseudo-
of ligand analogues that control c-di-GMP monas fluorescens in soil. Appl Environ Microbiol
riboswitches. ACS ChemBiol 7:1436–1443 67:3363–3370
Ganeshapillai J, Vinogradov E, Rousseau J et al (2008) Kulasakara H, Lee V, Brencic A et al (2006) Analysis of
Clostridium difficile cell-surface polysaccharides Pseudomonas aeruginosa diguanylatecyclases and
composed of pentaglycosyl and hexaglycosyl phos- phosphodiesterases reveals a role for bis-(30 -50 )-
phate repeating units. Carbohydr Res 343:703e10 cyclic-GMP in virulence. Proc Natl Acad Sci
Ghosh S, Zhang P, Li YQ et al (2009) Superdormant 103:2839–2844
spores of Bacillus species have elevated wet-heat Lawley TD, Clare S, Walker AW et al (2009) Antibiotic
resistance and temperature requirements for heat acti- treatment of Clostridium difficile carrier mice triggers
vation. J Bacteriol 191:5584–5591 a supershedder state, spore-mediated transmission,
Gil F, Paredes-Sabja D (2016) Acyldepsipeptide and severe disease in immunocompromised hosts.
antibiotics as a potential therapeutic agent against Infect Immun 77:3661–3669
Clostridium difficile recurrent infections. Future Lee ASY, Song KP (2005) LuxS/autoinducer-2 quorum
Microbiol 11:1179–1189 sensing molecule regulates transcriptional virulence
Gil F, Pizarro-Guajardo M, Álvarez R (2015) Clostridium gene expression in Clostridium difficile. Biochem
difficile recurrent infection: possible implication of Biophys Res Commun 335:659–666
TA systems. Future Microbiol 10:1649–1657 Li YH, Tian X (2012) Quorum sensing and bacterial
Goldberg J (2002) Biofilms and antibiotic resistance: a social interactions in biofilms. Sensors (Basel)
genetic linkage. Trends Microbiol 10:264 12:2519–2538
Goulding D, Thompson H, Emerson J et al (2009) Dis- Lindsay D, von Holy A (2006) Bacterial biofilms within
tinctive profiles of infection and pathology in hamsters the clinical setting: what healthcare professionals
infected with Clostridium difficile strains 630 and B1. should know. J Hosp Infect 64:313–325
Infect Immun 77:5478–5485 Lipovsek S, Leitinger G, Rupnik M (2013) Ultrastructure
Hall-Stoodley L, StoodleyP (2009) Evolving concepts in of Clostridium difficile colonies. Anaerobe 24:66e70
biofilm infections. Cell Microbiol 11:1034–1043 Liu W, Røder HL, Madsen JS et al (2016) Interspecific
Hall-Stoodley L, Costerton JW, Stoodley P (2004) Bacte- bacterial interactions are reflected in multispecies bio-
rial biofilms: from the natural environment to infec- film spatial organization. Front Microbiol 7:1366
tious diseases. Nat Rev Microbiol 2:95–108 Lyra A, Forssten S, Rolny P et al (2012) Comparison of
Hammond EN, Donkor ES, Brown CA (2014) Biofilm bacterial quantities in left and right colon biopsies and
formation of Clostridium difficile and susceptibility to faeces. World J Gastroenterol 18:4404–4411
Clostridium difficile Biofilm 113

Macfarlane S, Dillon JF (2007) Microbial biofilms in the formation. In: Microbiology society annual confer-
human gastrointestinal tract. J Appl Microbiol ence, p P418
102:1187–1196 Ofosu A (2016) Clostridium difficile infection: a review of
Macfarlane S, Macfarlane GT (2006) Composition and current and emerging therapies. Ann Gastroenterol
metabolic activities of bacterial biofilms colonizing 29:147–154
food residues in the human gut. Appl Environ Owrangi B, Masters N, Vollmerhausen TL et al (2017)
Microbiol 72:6204–6211 Comparison between virulence characteristics of dom-
Macfarlane GT, Macfarlane S, Gibson GR (1998) Valida- inant and non-dominant Escherichia coli strains of the
tion of a three-stage compound continuous culture gut and their interaction with Caco-2 cells. Microb
system for investigating the effect of retention time Pathog 105:171–176
on the ecology and metabolism of bacteria in the Ozturk B, Gunay N, Ertugrul BM et al (2016) Effects of
human colon. Microb Ecol 35:180–187 vancomycin, daptomycin, and tigecycline on
Macfarlane S, Bahrami B, Macfarlane GT (2011) Muco- coagulase-negative staphylococcus biofilm and bacte-
sal biofilm communities in the human intestinal tract. rial viability within biofilm: an in vitro biofilm model.
Adv Appl Microbiol 75:111–143 Can J Microbiol 62:735–743
Machado D, Castro J, Palmeira-de-Oliveira A et al (2015) Pantaléon V, Bouttier S, Soavelomandroso AP et al
Bacterial vaginosis biofilms: challenges to current (2014) Biofilms of Clostridium species. Anaerobe
therapies and emerging solutions. Front Microbiol 30:193–198
6:152 Pantaléon V, Soavelomandroso AP, Bouttier S et al
Mah TF, O’Toole GA (2001) Mechanisms of biofilm (2015) The Clostridium difficile protease Cwp84
resistance to antimicrobial agents. Trends Microbiol modulates both biofilm formation and cell- surface
9:34–39 properties. PLoS One 10:1–20
Maldarelli GA, De Masi L, von Rosenvinge EC et al Peng JS, Tsai WC, Chou CC (2002) Inactivation and
(2014) Identification, immunogenicity and cross- removal of Bacillus cereus by sanitizer and detergent.
reactivity of Type IV pilin and pilin-like proteins Int J Food Microbiol 77:11–18
from Clostridium difficile. Pathog Dis 71:302–314 Peng Z, Jin D, Kim HB et al (2017) Update on antimicro-
Maldarelli GA, Piepenbrink KH, Scott AJ et al (2016) bial resistance in Clostridium difficile: resistance
Type IV pili promote early biofilm formation by Clos- mechanisms and antimicrobial susceptibility testing.
tridium difficile. Pathog Dis 74:ftw061 J Clin Microbiol 55:1998–2008
Mathur H, Rea MC, Cotter PD et al (2016) The efficacy of Percival SL, Suleman L, Francolini I et al (2014) The
thuricin CD, tigecycline, vancomycin, teicoplanin, effectiveness of photodynamic therapy on planktonic
rifampicin and nitazoxanide, independently and in cells and biofilms and its role in wound healing. Future
paired combinations against Clostridium difficile Microbiol 9:1083–1094
biofilms and planktonic cells. Gut Pathog 8:20 Percival SL, Suleman L, Vuotto C et al (2015)
Meeker DG, Beenken KE, Mills WB et al (2016) Evalua- Healthcare-associated infections, medical devices
tion of antibiotics active against methicillin-resistant and biofilms: risk, tolerance and control. J Med
Staphylococcus aureus based on activity in an Microbiol 64:323–334
established biofilm. Antimicrob Agents Chemother Pettit LJ, Browne HP, Yu L et al (2014) Functional
60:5688–5694 genomics reveals that Clostridium difficile Spo0A
Melville S, Craig L (2013) Type IV pili in Gram-Positive coordinates sporulation, virulence and metabolism.
bacteria. Microbiol Mol Biol Rev 77:323–341 BMC Genomics 15:160
Mhatre E, Monterrosa RG, Kovács ÁT (2014) From envi- Piepenbrink KH, Maldarelli GA, de la Peña CF et al
ronmental signals to regulators: modulation of biofilm (2014) Structure of Clostridium difficile PilJ exhibits
development in Gram-positive bacteria. J Basic unprecedented divergence from known Type IV
Microbiol 54:616–632 pilins. J Biol Chem 289:4334–4345
Nale JY, Chutia M, Carr P et al (2016) ‘Get in Early’; Piepenbrink KH, Maldarelli GA, Martinez de la Peña CF
biofilm and wax moth (Galleria mellonella) models et al (2015) Structural and evolutionary analyses show
reveal new insights into the therapeutic potential of unique stabilization strategies in the Type IV pili of
Clostridium difficile bacteriophages. Front Microbiol Clostridium difficile. Structure 23:385–396
7:1383 Piotrowski M, Karpiński P, Pituch H, van Belkum A,
Nassif X, Beretti JL, Lowy J et al (1994) Roles of pilin Obuch-Woszczatyński P (2017) Antimicrobial effects
and PilC in adhesion of Neisseria meningitidis to of Manuka honey on in vitro biofilm formation by
human epithelial and endothelial cells. Proc Natl Clostridium difficile. Eur J Clin Microbiol Infect Dis.
Acad Sci U S A 91:3769–3773
Ng KM, Ferreyra JA, Higginbottom SK et al (2013) Pizarro-Guajardo M, Calderón-Romero P, Castro-
Microbiota-liberated host sugars facilitate post- Córdova P et al (2016a) Ultrastructural variability of
antibiotic expansion of enteric pathogens. Nature the exosporium layer of Clostridium difficile spores.
502:96–99 Appl Environ Microbiol 82:2202–2209
Normington C, Chilton C, Buckley A, et al (2017) Influ- Pizarro-Guajardo M, Calderón-Romero P, Paredes-Sabja
ence of gut microflora on C. difficile biofilm D (2016b) Ultrastructure variability of the exosporium
114 C. Vuotto et al.

layer of Clostridium difficile spores from sporulating Shah D, Zhang Z, Khodursky A et al (2006) Persisters: a
cultures and biofilms. Appl Environ Microbiol distinct physiological state of E. coli. BMC Microbiol
82:5892–5898 6:53
Plummer S, Weaver MA, Harris JC et al (2004) Clostrid- Silva JO, Martins Reis AC, Quesada-Gómez C et al
ium difficile pilot study: effects of probiotic supple- (2014) In vitro effect of antibiotics on biofilm forma-
mentation on the incidence of C .difficile. Int tion by Bacteroides fragilis group strains isolated from
Microbiol 7:59–62 intestinal microbiota of dogs and their antimicrobial
Purcell EB, McKee RW, McBride SM et al (2012) Cyclic susceptibility. Anaerobe 28:24–28
diguanylate inversely regulates motility and aggrega- Slater, Unnkrishnan M (2015) Characterisation of LuxS
tion in Clostridium difficile. J Bacteriol dependent biofilm formation by Clostridium difficile.
194:3307–3316 In: 5th international Clostridium difficile symposium,
Purcell EB, McKee RW, Bordeleau E et al (2016) Regu- p P76
lation of Type IV pili contributes to surface Soavelomandroso AP, Bouttier S, Hoys S, Candela T,
behaviours of historical and epidemic strains of Clos- Janoir C (2015). Spatial organization of tissue-
tridium difficile. J Bacteriol 198:565–577 associated bacteria in a Clostridium difficile
Purcell EB, McKee RW, Courson DS et al (2017) A monoxenic mouse model.P95, 5th International Clos-
nutrient-regulated cyclic diguanylate phosphodiester- tridium difficile Symposium. Bled, Slovenia
ase controls Clostridium difficile biofilm and toxin Soutourina O (2017) RNA-based control mechanisms of
production during stationary phase. Infect Immun 85: Clostridium difficile. Curr Opin Microbiol 36:62–68
IAI.00347–IAI.00317 Soutourina OA, Monot M, Boudry P et al (2013) Genome-
Raponi G, Visconti V, Brunetti G et al (2014) Clostridium wide identification of regulatory RNAs in the human
difficile infection and Candida colonization of the gut: pathogen Clostridium difficile. PLoS Genet 9:
is there a correlation? Clin Infect Dis 59:1648–1649 e1003493
Ribeiro SM, Felı́cio MR, Boas EV et al (2016) New Spencer J, Leuzzi R, Buckley A et al (2014) Vaccination
frontiers for anti-biofilm drug development. against Clostridium difficile using toxin fragments:
Pharmacol Ther 160:133–144 observations and analysis in animal models. Gut
R€omling U, Amikam D (2006) Cyclic di-GMP as a sec- Microbes 5:23–22
ond messenger. Curr Opin Microbiol 9:218–228 Spigaglia P (2016) Recent advances in the understanding
R€omling U, Balsalobre C (2012) Biofilm infections, their of antibiotic resistance in Clostridium difficile infec-
resilience to therapy and innovative treatment tion. Ther Adv Infect Dis 3:23–42
strategies. J Intern Med 272:541–561 Spigaglia P, Barketi-Klai A, Collignon A et al (2013)
Rossi E, Cimdins A, Lüthje P et al (2017) “It’s a gut Surface-layer (S-layer) of human and animal Clostrid-
feeling” – Escherichia coli biofilm formation in the ium difficile strains and their behaviour in adherence to
gastrointestinal tract environment. Crit Rev Microbiol epithelial cells and intestinal colonization. J Med
9:1–30 Microbiol 62:1386–1393
Rothenbacher FP, Suzuki M, Hurley JM et al (2012) Stabler RA, He M, Dawson L et al (2009) Comparative
Clostridium difficile MazF toxin exhibits selective, genome and phenotypic analysis of Clostridium diffi-
not global, mRNA cleavage. J Bacteriol cile 027 strains provides insight into the evolution of a
194:3464–3474 hypervirulent bacterium. Genome Biol 10(9):R102
Roy R, Tiwari M, Donelli G et al (2017) Strategies for Stevenson E, Minton NP, Kuehne SA (2015) The role of
combating bacterial biofilms: a focus on anti-biofilm flagella in Clostridium difficile pathogenicity. Trends
agents and their mechanisms of action. Virulence. Microbiol 23:1–8 Sudarsan N, Lee ER, Weinberg Z et al (2008)
Sculean A, Aoki A, Romanos G et al (2015) Is photody- Riboswitches in eubacteria sense the second messen-
namic therapy an effective treatment for periodontal ger cyclic di-GMP. Science 321:411–413
and peri-implant infections? Dent Clin N Am Swidsinski A, Mendling W, Loening-Baucke V et al
59:831–858 (2008) An adherent Gardnerella vaginalis biofilm
Sebaihia M, Wren BW, Mullany P et al (2006) The persists on the vaginal epithelium after standard ther-
multidrug resistant pathogen Clostridium difficile has apy with oral metronidazole. Am J Obstet Gynecol
a highly mobile mosaic genome. Nat Genet 198(97):e1–e6
38:779–786 Swidsinski A, Loening-Baucke V, Mendling W et al
Semenyuk EG, Laning ML, Foley J et al (2014) Spore (2014) Infection through structured polymicrobial
formation and toxin production in Clostridium difficile Gardnerella biofilms (StPM-GB). Histol Histopathol
biofilms. PLoS One 9:e87757 29:567–587
Semenyuk EG, Poroyko VA, Johnston PF et al (2015) Tischler AD, Camilli A (2005) Cyclicdiguanylate
Analysis of bacterial communities during Clostridium regulates Vibrio cholera virulence gene expression.
difficile infection in the mouse. Infect Immun Infect Immun 73:5873–5882
83:4383–4391 Trejo FM, Pérez PF, De Antoni GL (2010) Co-culture
Sengupta C, Mukherjee O, Chowdhury R (2016) Adher- with potentially probiotic microorganisms antagonises
ence to intestinal cells promotes biofilm formation in virulence factors of Clostridium difficile in vitro.
Vibrio cholerae. J Infect Dis 214:1571–1578 Antonie Van Leeuwenhoek 98:19–29
Clostridium difficile Biofilm 115

Twine SM, Reid CW, Aubry A et al (2009) Motility and Vlamakis H, Chai Y, Beauregard P et al (2013) Sticking
flagellar glycosylation in Clostridium difficile. J together: building a biofilm the Bacillus subtilis way.
Bacteriol 191:7050–7062 Nat Rev Microbiol 11:157–168
Tyerman JG, Ponciano JM, Joyce P et al (2013) The Vuotto C, Moura I, Barbanti F et al (2016) Sub-inhibitory
evolution of antibiotic susceptibility and resistance concentrations of metronidazole increase biofilm for-
during the formation of Escherichia coli biofilms in mation in Clostridium difficile strains. Pathog Dis 74:
the absence of antibiotics. BMC Evol Biol 13:22 ftv114
Valiente E, Bouché L, Hitchen P et al (2016) Role of Walter BM, Rupnik M, Hodnik V et al (2014) The LexA
glycosyltransferases modifying type B flagellin of regulated genes of the Clostridium difficile. BMC
emerging hypervirulent Clostridium difficile lineages Microbiol 14:88
and their impact on motility and biofilm formation. J Walter BM, Cartman ST, Minton NP et al (2015) The
Biol Chem 291:25450–25461 SOS response master regulator LexA is associated
van Leeuwen PT, van der Peet JM, Bikker FJ et al (2016) with sporulation, motility and biofilm formation in
Interspecies Interactions between Clostridium difficile Clostridium difficile. PLoS One 10:1–17
and Candida albicans. mSphere 1:e00187–e00116 Wen Y, Behiels E, Devreese B (2014) Toxin-Antitoxin
Varga JJ, Nguyen V, O’Brien DK et al (2006) Type IV systems: their role in persistence, biofilm formation,
pili-dependent gliding motility in the Gram-positive and pathogenicity. Pathog Dis 70:240–249
pathogen Clostridium perfringens and other Willing SE, Candela T, Shaw HA et al (2015) Clostridium
Clostridia. Mol Microbiol 62:680–694 difficile surface proteins are anchored to the cell wall
Varga JJ, Therit B, Melville SB (2008) Type IV pili and using CWB2 motifs that recognise the anionic poly-
the CcpA protein are needed for maximal biofilm mer PSII. Mol Microbiol 96:596–608
formation by the gram-positive anaerobic pathogen Winkler WC, Breaker RR (2005) Regulation of bacterial
Clostridium perfringens. Infect Immun 76:4944–4951 gene expression by riboswitches. Ann Rev Microbiol
Vlamakis H, Aguilar C, Losick R et al (2008) Control of 59:487–517
cell fate by the formation of an architecturally com-
plex bacterial community. Genes Dev 22:945–953
European Practice for CDI Treatment

Fidelma Fitzpatrick, Mairead Skally, Melissa Brady,

Karen Burns, Christopher Rooney, and Mark H. Wilcox

Abstract surgical intervention and for

Clostridium difficile infection (CDI) remains a non-antimicrobial management (e.g., faecal
significant cause of morbidity and mortality microbiota transplantation, FMT). A 2017
worldwide. Historically, two antibiotics (met- survey of 20 European countries found that
ronidazole and vancomycin) and a recent third while the majority (n ¼ 14) have national CDI
(fidaxomicin) have been used routinely for guidelines that provide a variety of
CDI treatment; convincing data are now avail- recommendations for CDI treatment, only
able showing that metronidazole is the least five have audited guideline implementation.
efficacious agent. The European Society of A variety of restrictions are in place in
Clinical Microbiology and Infectious 13 (65%) countries prior to use of new anti-
Diseases CDI treatment guidelines outline CDI treatments, including committee/infec-
the treatment options for a variety of CDI tion specialist approval or economic review/
clinical scenarios, including use of the more restrictions. Novel anti-CDI agents are being
traditional anti-CDI therapies (e.g., metroni- evaluated in Phase III trials; it is not yet clear
dazole, vancomycin), the role of newer anti- what will be the roles of these agents. Prophy-
CDI agents (e.g., fidaxomicin), indications for laxis is an optimum approach to reduce the

F. Fitzpatrick (*)
Department of Clinical Microbiology, The Royal College
of Surgeons in Ireland, Dublin, Ireland
Department of Clinical Microbiology, Beaumont
Hospital, Dublin, Ireland
M. Skally · M. Brady C. Rooney
Department of Clinical Microbiology, Beaumont Microbiology, Leeds Teaching Hospitals and University
Hospital, Dublin, Ireland of Leeds, Leeds, UK
e-mail:; e-mail:
M. H. Wilcox (*)
K. Burns Microbiology, Leeds Teaching Hospitals and University
Department of Clinical Microbiology, Beaumont of Leeds, Leeds, UK
Hospital, Dublin, Ireland
Leeds Teaching Hospitals and University of Leeds,
Health Protection Surveillance Centre, Dublin, Ireland Leeds, UK
e-mail: e-mail:

# Springer International Publishing AG 2018 117

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
118 F. Fitzpatrick et al.

impact of CDI especially in high-risk When discussing European practice for CDI
populations; monoclonal antibodies, antibi- treatment, variability between countries is inevi-
otic blocking approaches and multiple table for a number of reasons. Treatment of
vaccines are currently in advanced clinical patients with CDI begins with making the diag-
trials. The treatment of recurrent CDI is par- nosis, specifically having a high index of clinical
ticularly troublesome, and several different suspicion if a patient has a combination of signs
live bio therapeutics are being developed, in and symptoms and/or CDI risk factors and there-
addition to FMT. after confirmation by microbiological testing or
colonoscopic/histopathological findings. Clini-
Keywords cian awareness of CDI as part of the differential
C. difficile treatment · Anti-CDI agents · CDI diagnosis and access to timely laboratory
guidelines · Novel C. difficile agents · diagnostics is therefore crucial for appropriate
C. difficile prophylaxis patient management. However, there remains
considerable variability across countries with an
estimated 40,000 inpatients potentially undiag-
nosed annually in European hospitals (Davies
1 Introduction et al. 2014). Mnemonic checklists can be useful
tools to reduce clinician error and promote
The European Society of Clinical Microbiology awareness (Chew et al. 2016). Albeit potentially
and Infectious Diseases (ESCMID) first more useful when English is the commonly spo-
published guidelines for Clostridium difficile ken language, the SIGHT mnemonic is a useful
infection (CDI) treatment in 2009, which were aide memoire for clinicians when managing
revised in 2014 (Debast et al. 2014). These patients with suspected potentially infectious
evidence-based guidelines outline the treatment diarrhoea (Fig. 1) (Public Health England 2013).
options for a variety of CDI clinical scenarios, Once CDI is diagnosed, variability in anti-
including recommendations for use of the more CDI treatment practices may be due to individual
traditional anti-CDI therapies (e.g., metronida- judgement and/or knowledge, individualised
zole, vancomycin), the role of newer anti-CDI patient factors and national regulatory or eco-
agents (e.g., fidaxomicin), indications for surgi- nomic issues, e.g., the availability of newer
cal intervention and for non-antimicrobial man- (more expensive) anti-CDI agents. Lastly, the
agement (e.g., faecal microbiota transplantation, ESCMID (and national) guidelines recommend
FMT). Many European countries have published a number of potential treatment options for simi-
their own national CDI treatment guidelines, lar CDI clinical scenarios, so individual clinician
which are broadly similar to the ESCMID preference will likely be a potential cause of
guidelines, though contextualised to the local variability. This variability in anti-CDI treatment
setting (ECDC 2017).

S Suspect that a case may be infective where there is no clear alternative cause for diarrhoea.
I Isolate the patient/resident. Consult with the infection prevention and control team where available
while determining the cause of the diarrhoea.

G Gloves and aprons must be used for all contacts with the patient/resident and their environment.
H Hand washing with soap and water should be carried out after each contact with the
patient/resident and the patient/resident’s environment.

T Test the stool for C. difficile toxin, by sending a specimen immediately.

Fig. 1 SIGHT Mnemonic protocol (Adapted with permission from SIGHT Mnemonic UK protocol (DH and HPA
European Practice for CDI Treatment 119

preferences has previously been described in non-severe CDI is fidaxomicin (Grade B–I).
Ireland (Prior et al. 2017). In the United States Fidaxomicin is also recommended later in the
(US) almost half of patients with severe CDI guidelines for treatment of severe/complicated
were treated with metronidazole, despite vanco- and first recurrent CDI (Grade B–I) and multiple
mycin being recommended in national guidelines recurrent CDI (Grade B–II). The non-inferiority
at that time (Stevens et al. 2017). of fidaxomicin to vancomycin for treatment of
In this chapter, we firstly review the ESCMID CDI with lower recurrences rate and superior
CDI guideline recommendations and include an sustained clinical response has been reported,
update as relevant of subsequent publications, though patients with severe CDI were not
present the findings of a 2017 survey of evaluated (Louie et al. 2011; Cornely et al.
European CDI national experts regarding CDI 2012). Subsequently, superiority of fidaxomicin
guidelines and their implementation and lastly to vancomycin in patients with non-NAP1/BI/
look to the future as we summarise promising 027 strains was reported (Crook et al. 2012).
new therapies for CDI treatment. However, in patients infected with the NAP1/
B1/027 strain, there was no significant difference
in recurrence rates between the two drugs. What
2 ESCMID Guidelines for CDI implication this particular finding has for clinical
Treatment practice in Europe will depend on the current
prevalence rate of this strain in a country. How-
The ESCMID guidelines provide a number of ever fidaxomicin is considerably more expensive
definitions to guide clinical management of than metronidazole or vancomycin, therefore
patients with CDI, including diagnosis, treatment economic factors may come into play in
response, severity and recurrence (Debast et al. European countries regarding its availability
2014). A number of CDI scenarios are consid- and use (Nelson et al. 2017).
ered including the initial management of CDI in Since publication of the ESCMID guidelines,
addition to the management of recurrent and the superiority of vancomycin over metronida-
severe CDI. (Table 1) For all scenarios the timely zole for treatment of mild-to-moderate primary
implementation of appropriate infection preven- or recurrent CDI has been reported. (Johnson
tion and control measures to prevent further et al. 2014) and numerous publications have
cross-infection is highlighted, in addition to the examined the benefits of fidaxomicin in a number
discontinuation of antimicrobial therapy of patient populations. A recent Cochrane review
(if clinically indicated), fluid and electrolyte evaluated anti-CDI treatment options and
replacement, review of proton pump inhibitor reported that vancomycin is superior to metroni-
use and avoidance of anti-motility medications. dazole and fidaxomicin is superior to vancomy-
cin for achieving symptomatic cure (Nelson et al.
2017). The authors noted that the lack of any ‘no
2.1 Non-severe CDI treatment’ control studies does not allow for any
conclusions regarding the need for specific anti-
Three potential options are recommended for CDI treatment in patients with mild CDI and
treatment of non-severe CDI, namely metronida- pointed to the economic advantage of
zole, vancomycin or fidaxomicin. Metronida- metronidazole.
zole, which is a relatively safe and inexpensive
antimicrobial is the treatment of choice (grade
A–I), once there is no contraindications for its 2.2 Definition and Treatment
use. However, adverse effects such as metallic of Severe CDI
taste and nausea may limit its use/compliance in
certain patient populations. Another Classification of CDI by severity can be prob-
recommended option for treatment of lematic, as patients with severe ileus may not
120 F. Fitzpatrick et al.

have diarrhoea. In practice, the clinical spectrum 20,000 cells/μL) may be useful for clinicians
of severe CDI varies considerably and the diag- to identify high-risk patients likely to benefit
nosis is usually reached using a combination of from more aggressive therapy (e.g., early admin-
findings. The ESCMID guidelines summarise the istration of oral vancomycin).(Na et al. 2015).
range of patient, laboratory, endoscopic and The recommended treatment of choice for
radiological factors associated with severity of severe CDI in the ESCMID guidelines is oral
CDI colitis and recommend three unfavourable vancomycin (Grade A–I) which achieves high
prognostic factors, namely raised leukocyte intracolonic concentrations with minimal sys-
count >15 109/L, decreased albumin <30 g/L temic adverse effects (Debast et al. 2014). Intra-
and rise in serum creatinine level (>1.5 times the venous metronidazole combined with
premorbid level or >133 μM) (Debast et al. vancomycin retention enema or oral/NG vanco-
2014). A recently validated clinical prediction mycin at the higher 500 mg dose is provided as
rule to identify patients at risk of severe an alternative (Grade B–III). A recent retrospec-
outcomes (age 60 years, peak serum creatinine tive study comparing vancomycin and metroni-
1.5 mg/dL and peak leukocyte count of dazole reported superiority of vancomycin for

Table 1 Overview of ESCMID recommendations for CDI treatment (Debast et al. 2014)
Clinical Non-antibiotic Not
scenario Oral antibiotic treatment Oral treatment not possible treatment recommended
Non-severe Metronidazole 500 mg IV Metronidazole 500 mg Stop inducing Probiotics (D–I)
CDI TDS (A–I) TDS 10 days (A–II) antibiotic (s) and
Or Vancomycin 125 mg 48 h clinical Toxin binding
QDS (B–I) observation (C-II) (D–I)
Or Fidaxomicin 200 mg BD
All 10 days
First Fidaxomicin 200 mg BD
recurrence (B-I)
Or Vancomycin 125 mg
Or Metronidazole 500 mg
All 10 days
Multiple Fidaxomicin 200 mg BD: Faecal transplantation in Metronidazole
recurrences 10 days (B-II) combination with oral 500 mg TDS
antibiotic treatment (A–I) (D-II)
Or Vancomycin 125 mg Probiotics (D–I)
QDS: 10 days followed by Passive
pulse or taper strategy (B-II) immunotherapy
with immune
whey (D–I)
Severe CDI Vancomycin 125 mg QDS IV Metronidazole 500 mg Surgery: Total Metronidazole
or (A–I) Consider increasing TDS 10 days (A-II) colectomy and 500 mg TDS (D–
complicated to 500 mg QDS (B-III) combined with either ileostomy I)
course Or Fidaxomicin 200 mg BD Vancomycin retention Fidaxomicin
(B–I) enema (500 mg in 100 mL (D-III)
normal saline QDS
All 10 days or Vancomycin 500 mg
QDS by oral/nasogastric
tube for 10 days (B-III)
PO oral, IV intravenous, BD twice daily, TDS three times daily, QDS four times daily
European Practice for CDI Treatment 121

severe CDI, though no difference in CDI recur- of recurrences (Hu et al. 2009). The ESCMID
rence rates (Stevens et al. 2017). At the time of guidelines recommendation for the first recur-
publication of ESCMID guidelines, it was noted rence of non-severe CDI is either vancomycin
that there was insufficient data available for or fidaxomicin (both B–I recommendations).
fidaxomicin. While there have been subsequent For subsequent recurrences, while a variety of
reports of fidaxomicin use in critical care patients strategies are recommended (Table 1), FMT is
with CDI and case reports of salvage use after allocated an A–I recommendation.
failure of standard therapy, (Penziner et al. 2015; Recent surveys have highlighted the interest
Arends et al. 2017), as most studies exclude of European clinicians in FMT as a therapeutic
patients with severe CDI the role of fidaxomicin option for patients with CDI; though note its
in these patients has yet to be fully elucidated. potential underutilisation (Porter and Fogg
(Nelson et al. 2017). 2015; Prior et al. 2017). Since publication of
The precise role of surgical management in the ESCMID guidelines, a recent two-centred
severe CDI is a topic of debate (Fitzpatrick randomized controlled trial of FMT via colonos-
2008). There are no clear guidelines or protocols copy for recurrent CDI reported a 91% cure rate
to guide the timing of surgical intervention. Cer- with donor FMT (63% with autologous FMT –
tainly, the decision that surgical management is though this varied significantly between the two
required for CDI should be taken by the multi- centres at 43% and 90% cure rates respectively)
disciplinary team, surgeons consulted at an (Kelly et al. 2016). Notably patients with recur-
‘early’ stage (though there is no clear definition rence after autologous FMT resolved after a
as to when this is) and an interdisciplinary risk/ subsequent donor FMT. Severe and severe-
benefit analysis of surgery individualised for that complicated indication, inpatient status during
patient. The ESCMID guidelines recommend FMT, and the number of previous CDI-related
total colectomy, ‘before colitis becomes very hospitalizations are strongly associated with
severe’, if colonic perforation or if there is sys- early failure of a single FMT for CDI (Fischer
temic inflammation and the patient’s condition et al. 2016).
has deteriorated and is not responding to anti-
CDI therapy (Table 1) (Debast et al. 2014).
Because of the morbidity (and mortality)
3 Survey of European CDI
associated with colectomy in a systemically
Experts on CDI Treatment
unwell patient, there is increased interest in
evaluating options that avoid colon resection
Though European and National CDI guidelines
(Kautza and Zuckerbraun 2016; Sartelli et al.
exist and variability in practice for treatment of
2015). The potential role of FMT as an alterna-
patients with CDI is likely as previously
tive to emergency bowel surgery has also been
discussed, to our knowledge there has been no
recently highlighted (van Beurden et al. 2017).
recent assessment of CDI treatment guideline
recommendations and their implementation in
European countries. We designed an interactive
2.3 Recurrent CDI
online survey in this regard using Demographix®
software (57 Chestnut Road, London SE27 9EZ
Recurrent CDI itself is a significant risk factor
UK). The purpose was to describe the practice
with the risk of recurrence increasing signifi-
for CDI management and treatment in Europe.
cantly with each episode of recurrence.
National experts from European countries were
Predicting which patients will develop recurrent
invited by email to complete the online survey,
CDI would enable clinicians to minimise recur-
during the period 07th June 2017 to 28th July
rence risk (e.g., avoid concomitant
2017. Data was analysed using an Excel® data-
antimicrobials) and also by heightening aware-
base (Microsoft Corp., Redmond, WA, USA).
ness, facilitates prompt diagnosis and treatment
122 F. Fitzpatrick et al.

Eighty-three CDI experts from 35 European • Two were conducted in the past 5 years and
countries were invited to take the survey with three more than 5 years ago. No surveys were
34 respondents, representing 20 (57%) countries. conducted in the last year.
Respondents included experts in the fields of • CDI treatment was included in one national
microbiology, public health and infection pre- survey only.
vention and control, who were working in • Facilities surveyed included hospitals only
hospitals (n ¼ 10), laboratories (n ¼ 2), health (n ¼ 4) or diagnostic microbiology
protection, public health or infectious diseases laboratories only (n ¼ 1)
agencies (n ¼ 4) or other organisations (n ¼ 4).
To avoid study bias arising from multiple Of the six countries that did not have national
respondents from the same country, data from guidelines, a previous survey or audit of some
one respondent per country was included in the (but not all) aspects of local CDI guidelines was
analysis. conducted for one and five (83%) did not previ-
National guidelines for managing patients ously conduct a survey or audit. For the survey
with CDI were available in 14 (70%) countries that was conducted, CDI treatment was not
with guideline revisions undertaken during the included and facilities surveyed included
last 5 years (n ¼ 7), 1 year (n ¼ 2), or were hospitals only. Information on when the survey
presently under revision (n ¼ 1). Revisions had took place was not provided.
not been undertaken in four countries with these Severe CDI was defined as a variable combi-
guidelines published in 2007, 2011 (n ¼ 2) and nation of factors, as outlined in Fig. 2. The
2013. Of the six countries that did not have commonest being leucocytosis of 15,000 cells
national guidelines, guidance was sought from per μL (n ¼ 17; 85%). A variety of anti-CDI
the ESCMID CDI guidelines (n ¼ 5) or local regimens were recommended as summarised in
guidelines (n ¼ 1). The recommendations Table 3. In addition, a number of other factors
provided in national guidelines varied by coun- were reported to influence choice of the
try, as outlined in Table 2. Of the options recommended anti-CDI therapy including:
provided in the survey, the commonest recom-
mendation was treatment of patients with CDI • C. difficile ribotype.
(93%; n ¼ 13) and the least common were CDI • Patient factors:
key performance indicators (KPIs) and audit of – Risk factors for recurrence.
guideline implementation (21%; n ¼ 3). Other – Patient tolerance/ability to take oral
recommendations were provided in national medications/response to treatment.
guidelines of 36% (n ¼ 5) countries, including: • Fidaxomicin use:
essential elements of a CDI prevention – Approval required from microbiology/
programme, use of tools such as checklists, infectious diseases for use.
C. difficile reference laboratory requirements, – Economic considerations because of
access to infection specialists in the non-acute high cost.
sector, healthcare facility infrastructure – Reservations about its use as lack of sur-
requirements, environmental and equipment vival benefit.
decontamination, epidemiology, clinical diagno- • FMT:
sis of CDI, antimicrobial stewardship, FMT and – Availability of facilities for a FMT service.
defining roles and responsibilities to support the – Use as an option for severe CDI when
implementation of the guidance. surgery is not possible.
In total, 36% (n ¼ 5) of countries previously • Immunoglobulin therapy recommended in
surveyed or audited some (but not all) aspects of case of severe protein loss.
the implementation of national CDI guidelines
though the majority, 64% (n ¼ 9), had not. Of the
five surveys/audits conducted:
European Practice for CDI Treatment 123

Table 2 Recommendations for CDI management in 14 European countries with national CDI guidelines
Included in guideline Not included in guideline
Recommendation Number (n) and percentage (%) of countries
Surveillance of CDI, n (%) 11 (79) 3 (21)
Laboratory diagnosis of CDI, n (%) 12 (86) 2 (14)
Treatment of patients with CDI, n (%) 13 (93) 1 (7)
Management of outbreaks and clusters of CDI, n (%) 11 (79) 3 (21)
CDI key performance indicators (KPIs), n (%) 3 (21) 11 (79)
Audit of guideline implementation, n (%) 3 (21) 11 (79)
Other recommendations, n (%) 5 (36) 9 (64)
One country with local guidelines is not included as applicable data was not available for this country

Percentage (%) of countries

Serum Serum Pseudo Evidence of
Abdominal creatinine of creatinine membranous colitis or
Fevers Rigors of ≥15,000 Other
pain ≥50% above >133 μmol colitis on ascites on
cells per μL
baseline per L endoscopy CT imaging
% 55 40 45 85 60 20 80 70 35

Fig. 2 Definition of severe CDI in 20 European to the ICU for treatment of CDI, colectomy due to CDI,
Countries as a percentage (%) of countries surveyed mortality within 30 days of diagnosis of CDI, suspicion of
‘Other’ defining factors were included for 35% (n ¼ 7 pseudomembranous colitis, diarrhoea, positive stool test,
countries), and were a combination of: toxic megacolon, hemodynamic instability, signs of septic shock, signs of
ileus, colonic dilation in CT scan >6 cm, immunosup- peritonitis, decreased bowel sounds, vomiting, lack of
pression, shock, hypotension, admission to hospital for bowel movements, left shift, hypoproteinemia, anaemia
treatment of CDI acquired outside the hospital, admission and increased serum lactate

Table 3 Recommendations for CDI treatment in 15 European countries with national (n ¼ 14) or local (n ¼ 1) CDI
vancomycin Immunoglobulin
MTZ Vancomycin Fidaxomicin regimen therapy FMT
Number (n) and percentage (%) of countries surveyed
New CDI, n (%) 13 (87) 9 (60) 3 (20) 0 (0) 0 (0) 0 (0)
Recurrence (1st), n (%) 4 (27) 13 (87) 6 (40) 1 (7) 0 (0) 1 (7)
Recurrence (2nd), n (%) 0 (0) 8 (53) 7 (47) 6 (40) 1 (7) 3 (20)
Three or more 0 (0) 6 (40) 4 (27) 9 (60) 3 (20) 12 (80)
recurrences, n (%)
Severe CDI, n (%) 3 (20) 11 (73) 4 (27) 1 (7) 1 (7) 1 (7)
Other, n (%) 3 (20) 2 (13) 1 (7) 0 (0) 1 (7) 1 (7)
MTZ metronidazole, FMT faecal microbiota transplantation
124 F. Fitzpatrick et al.

Of the 20 countries, a variety of restrictions et al. 2017; Cornely et al. 2012; Crook et al.
were in place in 13 (65%) countries before new 2012). High acquisition cost of fidaxomicin has
anti-CDI therapies could be used including: inhibited uptake in some settings and was
observed in our survey of European countries as
• Reimbursement restrictions (n ¼ 1). outlined above. However, a recent real world
• Health technology assessment (n ¼ 1). study suggested a reduction in mortality
• Pharmacoeconomic review (n ¼ 3). associated with fidaxomicin use and that this
• Committee approval either national (n ¼ 6) or was therapy was cost-effective (Goldenberg
local (n ¼ 4). et al. 2016). In the phase 3 trials, bezlotoxumab
• Microbiology or infectious diseases approval was associated with a significant reduction in
(n ¼ 2). CDI readmissions.
• CEO/financial director approval (n ¼ 2). The ideal antimicrobial agent for CDI should
• Cost and access issues re monoclonal therapy reduce vegetative C. difficile cells, toxins and
(n ¼ 1). spores in the host gut lumen without perturbation
• Antimicrobial resistance (n ¼ 1). of the host microbiota, both to avoid creating an
environment that is conducive to C. difficile
expansion or to select for resistant potential
pathogens (e.g. vancomycin resistant enterococci
4 Clostridium difficile Pipeline [VRE] or multi-resistant Gram-negative bacilli)
Prophylactic and Therapeutic (Chang et al. 2008). This is a very challenging
Agents profile for an antibiotic and indeed recent
‘failures’ of two antimicrobial agents in late-
The four current approved therapeutic agents for stage clinical trials emphasise how difficult it is
CDI vary markedly in efficacy. Whilst metroni- to improve on current CDI therapies.
dazole has historically been the most commonly
used option for treating CDI, as previously
discussed, it is now known that this antibiotic is 4.1 Surotomycin and Cadazolid
inferior to vancomycin (Johnson et al. 2014;
Nelson et al. 2017). Concern regarding treatment Surotomycin, an oral lipopeptide derivative of
failures with metronidazole remains (Vardakas daptomycin, was examined in two phase 3 trials
et al. 2012). Metronidazole achieves poor intra- (NCT01598311 and NCT01597505) but did not
luminal colonic concentrations, especially as demonstrate non-inferiority compared with van-
mucosal inflammation subsides, such that the comycin (Boix et al. 2017). Notably,
antibiotic may be undetectable as diarrhoea surotomycin dosing caused an overgrowth of
resolves. Also, some C. difficile isolates show Gram-negative bacilli in both in mice and in a
reduced susceptibility to metronidazole, which gut model of CDI that is highly predictive of
may be relevant given the sub-optimal pharma- human disease; recurrent CDI was also seen in
cokinetics for this antibiotic in CDI. Laboratory the latter model (Deshpande et al. 2016; Chilton
detection of reduced metronidazole susceptibil- et al. 2014b). Most recently, a press release
ity is itself problematic with variations in meth- announced that cadazolid (Actelion), which is a
odology and MIC interpretation limiting analysis novel hybrid oxazolidinone-fluroquinolone anti-
of trends and comparisons with published data biotic that inhibits C. difficile protein synthesis
(Moura et al. 2013). and, to a lesser extent, DNA synthesis, did not
Fidaxomicin and bezlotoxumab, a monoclo- meet its primary endpoint in comparison with
nal anti-toxin B antibody and the most recently vancomycin in one of two phase 3 trials
approved therapeutic agent, have been shown to (ActelionLtd. 2017; Gehin et al. 2015; Chilton
reduce the risk of recurrent CDI by 40–50% in et al. 2014a; Baldoni et al. 2014). It is too early to
comparison with vancomycin alone (Wilcox determine why this result was obtained, but may
European Practice for CDI Treatment 125

relate to activity of cadazolid on the gut response rates were 67% and 42%, respectively
microbiome in vivo, and/or persistence of (n ¼ 69 mITT population); CDI recurrence
C. difficile spores (Chilton et al. 2014a). occurred in 14% of ridinilazole recipients com-
pared with 35% of vancomycin subjects; this
difference meant that ridinilazole achieved a
4.2 Ridinilazole sustained response rate of 66.7% vs. 42.4% for
vancomycin, which met pre-set statistical superi-
Ridinilazole (SMT19969) is a novel, ority criteria (Vickers et al. 2017). Microbiome
non-absorbable, very narrow-spectrum antimi- analyses of faecal samples from subjects in this
crobial with minimal activity against host gut phase 2 study showed that vancomycin recipients
microbiota (Goldstein et al. 2013). While its had a marked loss of diversity and replacement
mode of action has not been fully determined, it of the predominant phyla of healthy stool
does not appear to act through classical antibiotic (Bacteroides and Firmicutes) by Enterobac-
pathways, such as inhibition of cell wall, protein, teriaceae. These disruptions were still present
lipid, RNA or DNA synthesis (Vickers et al. 2 weeks after the end of treatment, even in
2016). Bassiere et al. described the effects of subjects who had not had a recurrence at that
ridinilazole on C. difficile cell morphology, as point. By contrast, ridinilazole, had a minimal
visualised by scanning electron microscopy and effect on gut microbiota (Chang et al. 2016).
confocal microscopy (Basseres et al. 2016). Fol-
lowing exposure to sub-lethal concentrations of
ridinilazole, bacterial cell division was halted 4.3 CDI Prophylaxis
and there was an absence of septum formation;
this resulted in marked cell elongation. It has not 4.3.1 Ribaxamase
been confirmed whether these observations are a
direct effect of ridinilazole, or a downstream Ribaxamase (SYN-004, synthetic biologics) is a
response to the antibiotic. Ridinilazole has good recombinant beta-lactamase that has been
activity against some but not all clostridia; it is 7- formulated to be administered orally in patient
to 17-fold more active in vitro than metronida- receiving beta-lactam antibiotic therapy (Kaleko
zole and vancomycin and has similar potency to et al. 2016; Connelly et al. 2015). Ribaxamase
fidaxomicin against C. difficile (Baines et al. degrades unmetabolised antibiotic in the colon to
2015; Weiss et al. 2014; Sattar et al. 2015; reduce the deleterious effects on the gut
Corbett et al. 2015). Notably, in vitro, in vivo microbiota (Roberts et al. 2016). Animal studies
and gut model data confirm that ridinilazole has have demonstrated safety, and notably no reduc-
little antimicrobial activity against indigenous tion in the systemic concentration of
gut microflora groups, except selected clostridia co-administered ceftriaxone (Connelly et al.
(Freeman et al. 2015; Goldstein et al. 2013; 2015). A phase 2 double-blind placebo-con-
Baines et al. 2015; Corbett et al. 2015; Chang trolled study has examined the potential of
et al. 2016). ribaxamase to prevent CDI, antibiotic-associated
Safety and tolerability of ridinilazole was diarrhoea and the emergence of antimicrobial
established in healthy subjects and in a recently resistant potential pathogens in patients
reported phase II randomised double-blind trial hospitalized with a lower respiratory tract infec-
(CoDIFy) (Vickers et al. 2015; Vickers et al. tion treated with IV ceftriaxone (Synthetic
2017). CoDIFy was designed as a Biologics 2017). Patients who received
non-inferiority study and compared 10 days ther- ribaxamase had a 71.4% relative risk reduction
apy of either oral ridinilazole 200 mg BD or oral for CDI (p ¼ 0.045). There was also a significant
vancomycin 125 mg QDS. Sustained clinical reduction in new colonisation by VRE in
126 F. Fitzpatrick et al.

ribaxamase versus placebo recipients 4.4 Active C. difficile Immunisation

(p ¼ 0.0002). Adverse events were similar in
active and placebo patients. Vaccination to boost host antibody-mediated
immunity is an attractive strategy to prevent
CDI. The relative importance of C. difficile
4.3.2 DAV132
toxins A and B to human infection remains con-
troversial, but host immune response to these
Another novel approach to CDI prophylaxis is
toxins likely influences the likelihood of infec-
DAV132 (DaVolterra), which is an activated
tion, clinical severity and outcome of CDI (Solo-
charcoal based product that is administered as
mon et al. 2013; Kuehne et al. 2010). Higher
an enteric coated capsule. DAV132 irreversibly
serum IgG levels to toxin A have been shown
captures antibiotics in the intestine whilst
in patients with asymptomatic colonisation
avoiding interruption of antibiotic absorption.
compared with those with CDI, and recurrent
DAV132 has been examined in a proof-of-con-
infection is associated with poor IgG and IgM
cept study involving 18 healthy subjects who had
responses (Kyne et al. 2000, 2001). Interestingly,
received DAV132, uncoated formulated
the effectiveness of the anti-toxin B monoclonal
activated charcoal (FAC) or water 16 and 8 h
antibody bezlotoxumab at reducing the risk of
before, alongside the probe drugs, and 8 h there-
CDI recurrence was not enhanced by the addition
after. The AUC0-96 h of amoxicillin was
of an anti-toxin A monoclonal antibody,
reduced by more than 70% when it was taken
actoxumab; also, actoxumab alone was not effi-
with FAC, but was not adversely affected when
cacious at preventing recurrence. Nevertheless, it
taken with water or DAV132. By contrast, the
remains logical to design a vaccine around the
AUC0-96 h of sulfapyridine was reduced by
augmentation of the host response to both toxins
>90% when administered with either FAC or
A and B (Kuehne et al. 2010). Other C. difficile
DAV132 in comparison with water. Hence,
antigens may also be important, noting for exam-
DAV132 can selectively adsorb drugs in the
ple that antibodies to surface proteins are greater
proximal colon, without interfering with their
in colonised versus infected patients (Pechine
et al. 2005).
A further healthy volunteer trial examined the
Three vaccines that use C. difficile toxin
efficacy of DAV132 to protects the gut
targets have progressed to phase 2 or 3 clinical
microbiome and prevent CDI during
development. The first to reach a phase 3 clinical
moxifloxacin (MOX) treatment (de Gunzburg
trial is a formalin-inactivated toxoid-based vac-
et al. 2015). DAV132 decreased free faecal
cine developed by Sanofi Pasteur (Foglia et al.
MOX concentration by >99% compared with
2012). Following vaccination, seroconversion to
MOX alone, but MOX plasma PK did not change
toxin A was more pronounced than to toxin B
significantly. Alterations of the faecal
(but took up to 70 days) and notably was less
microbiome observed with MOX were prevented
common in elderly subjects; three vaccine doses
by co-administration of DAV132. In a human gut
were required to achieve an adequate
model DAV132 protected the microbiota and
neutralising-antibody response (Foglia et al.
prevented C. difficile overgrowth and toxin pro-
2012; Kotloff et al. 2001). A 100 μg dose
duction (de Gunzburg et al. 2015). Hamsters
(given with an AlOH adjuvant) was found to
were also fully protected by DAV132 against
yield the best immunogenic response, and a
MOX-induced CDI (de Gunzburg et al. 2015).
phase 3 trial of this vaccine in the prevention of
Such results warrant further clinical development
primary CDI in at-risk subjects aged >50 years
of DAV132 to protect the lower gut microbiota,
commenced in 2013 (NCT01887912). Another
and so prevent CDI associated with antibiotic
formalin-inactivated toxoid based vaccine, but
with alterations in both toxins A and B to reduce
European Practice for CDI Treatment 127

toxigenicity, has recently commenced a phase protective effectiveness of specific components

3 primary CDI prevention trial (Pfizer; of the gut microbiota, but possibly with greater
NCT03090191), also based on a three dose strat- reassurance on safety. In the US, Openbiome is
egy (Donald et al. 2013; Sheldon et al. 2016). A aiming to overcome some of the practical
third C. difficile vaccine candidate (VLA84, barriers to FMT, and safety concerns, by
Valneva) has completed a phase II trial with facilitating access to screened faecal transplant
500 subjects (Valneva 2016). VLA84 uses a dif- material and by collecting longer term follow up
ferent antigen approach to either of the two data. (
toxoid-based vaccines that are currently The first randomised (sham procedure con-
undergoing phase 3 evaluation. VLA84 is a sin- trolled) trial of FMT to treat recurrent CDI
gle recombinant fusion protein consisting of demonstrated an intention-to treat (ITT) efficacy
portions of the C-terminal cell binding domains rate of 81% to prevent further recurrences; nota-
of toxins A and B. The developers claim that bly, however, the study contained only
production and characterization of VLA84 16 patients in the FMT arm (van Nood et al.
could be simpler and less costly compared with 2013). In a randomised but non-blinded clinical
toxoid-based vaccines. The phase 2 study of trial, 39 subjects with recurrent CDI were given
VLA84 met its primary endpoint in terms of FMT (preceded by vancomycin 125 mg QDS for
identifying the dose and formulation with the 3 days), comprising at least one infusion of
highest seroconversion rate against both toxins faeces via colonoscopy, or vancomycin 125 mg
A and B (subjects were followed up to day 210) QDS for 10 days and then 125–500 mg/day every
and confirmed the favourable safety profile that 2–3 days for at least 3 weeks. The primary end
was seen in Phase I. A phase 3 programme for point was the resolution of diarrhoea related to
VLA84 is being planned. CDI at week 10; surprisingly, a positive
C. difficile test was not required to define recur-
rence post-study treatment (Cammarota et al.
4.5 Microbiome Based Therapeutics 2015). The study was stopped after a 1-year
interim analysis, at which point 18/20 (90%)
4.5.1 Faecal Microbiota Transplantation vs. 5/19 (26%) patients in the FMT
(FMT) vs. vancomycin treatment groups, respectively
The evidence base concerning the effectiveness had resolution of C. difficile diarrhoea
of FMT continues to grow, but it remains a (P < 0.0001). There were no significant adverse
non-regulated product, with many different events in either of the study groups.
versions reported. FMT comprises the adminis- Adults with recurrent or refractory CDI were
tration of a complex live faeces-derived mixture enrolled in a randomised, double-blind,
of micro-organisms, including some of uncertain non-inferiority study in six Canadian centres of
significance (some beneficial, others possibly free-thawed (n ¼ 114) vs. fresh (n ¼ 118) FMT
harmful or neither) and so (particularly longer via enema. Clinical resolution without recur-
term) safety remains unproven. Of particular rence up to 13 weeks did not differ significantly
concern here is the increasing use of FMT when in the per-protocol (83.5% vs. 85.1%) and mITT
licensed CDI therapeutics has not been tried. (75.0% vs. 70.3%) populations (Lee et al. 2016).
Hence, different regulatory authorities have These results suggest that using freeze-thawed
taken varied stances on FMT to safeguard patient faecal material is a practicable alternative to
interests. Requirements for consenting subjects, fresh donor material. All patients received sup-
screening of donors and recipients, faecal mate- pressive antibiotics for the most recent episode of
rial preparation and delivery via either rectal or CDI, and these were discontinued 24–48 h before
nasogastro/duodenal routes, mean that there are FMT; this probably explains why only 38% of
intensive endeavours to develop alternatives to the subjects were positive for toxin or toxin gene
FMT that can still harness the restorative and immediately prior to FMT administration.
128 F. Fitzpatrick et al.

Notably, about one third of FMT recipients in idiopathic thrombocytopenic purpura, micro-
both groups, who were ultimately, classified as scopic colitis, contact dermatitis, rheumatoid
resolved, required two FMTs, which is a rela- arthritis, obesity, bacteraemia, and ulcerative
tively common observation. A non-blinded, colitis flare after FMT (Tariq et al. 2016; De
non-randomised study of encapsulated (and Leon et al. 2013; Quera et al. 2014; Alang and
freeze-thawed) faeces was performed in Kelly 2015). Institutions need to ensure they are
20 subjects with at least three episodes of mild- working within their national and European
to-moderate CDI and failure of a 6- to 8-weeks of frameworks and regulations. Where national
vancomycin therapy, or 2 episodes of severe regulations are absent, comparisons should be
CDI requiring hospitalization (Youngster et al. made to international standards to ensure the
2014). Diarrhoea resolution occurred in highest level of safety. In Europe, the regulation
14 patients (70%; 95% CI, 47%–85%) after a of FMT is currently at the discretion of the EU
single capsule-based FMT; 4/6 re-treated member states, though in many countries no such
non-responders had resolution of diarrhoea, giv- national regulation exists. Future planned EU
ing an overall 90% (95% CI, 68%–98%) regulation of FMT donor material may hinder
response rate. No serious adverse events were its widespread use, depending on whether it is
attributed to FMT. regulated as a drug or bodily tissue. A recent
The six randomised controlled trials of FMT European Consensus paper provided
have been recently reviewed; three that com- recommendations on a number of areas pertinent
pared FMT to antibiotic management; the to FMT implementation, including regulatory,
remainder compared FMT to various ‘types’ of administrative and laboratory guidelines
FMT in terms of preparation, source and delivery (Cammarota et al. 2017).
(Johnson and Gerding 2017). It is important to
note that, unlike prior uncontrolled studies that 4.5.2 Live Bio Therapeutic Microbiota
reported FMT efficacy rates of at least 90%, Preparations
efficacy (for one FMT) in these RCTs was
44–91%, with four recording success rates of RBX2660
65%. These include a randomized controlled RBX2660 is a live bio therapeutic microbiota
trial of FMT versus a 6-week vancomycin taper- suspension that aims to harness the effectiveness
ing regimen (VAN-TP) (Hota et al. 2017). of FMT, but within a standardised, regulated
VAN-TP was stopped early for futility; 56% of product, for the treatment of recurrent CDI. It
patients randomized to FMT by enema devel- has been studied in three phase 2 clinical trials.
oped recurrent CDI, compared with 42% PUNCH CD (NCT01925417) was a safety
VAN-TP recipients. focussed, prospective multi-centre, open-label
There are many important factors for study; 34 subjects (with 2 recurrent CDI
European clinicians to consider when episodes or 2 severe episodes resulting in hos-
establishing or using a FMT service. Factors pitalization) received at 1 dose of RBX2660
that should be taken into account at an institu- and 31 completed 6 months follow up (Orenstein
tional level when commencing an FMT service et al. 2016). Following a 10–14 day course of
are the national regulatory frameworks that FMT anti-CDI antibiotics and a 24–48 h washout
falls under (i.e. as a drug or biological material), period, RBX2660 was administered as a single
donor selection and screening practices, stool dose via enema. Further recurrent CDI occurred
preparation techniques and long term safety of in 48% of subjects after one dose of RBX2660,
microbiome manipulation in these patients. with 15/31 patients receiving a second enema; of
Concerns regarding the long term safety of these, 78.6% were considered to be treatment
FMT are not unfounded, especially in patients successes, contributing to an overall success
with inflammatory bowel disease. Reports of rate of 27/31 (87.1%). No serious adverse events
peripheral neuropathy, Sj€ogren syndrome, were related to RBX2660.
European Practice for CDI Treatment 129

PUNCH CD 2 (NCT02299570) was a phase from 42 subjects treated with RBX2660 treat-
2b multi-centre randomized double-blind, pla- ment arm and for 19 RBX2660 drug lots. The
cebo-controlled trial with 2 year follow-up RBX2660 microbial profiles had similar taxo-
(Dubberke et al. 2016). The primary efficacy nomic distributions, with a group mean that was
objective was assessment of response (defined highly divergent and significantly different from
as no CDI recurrence) to RBX2660 versus pla- those of patients at baseline. However, after
cebo at 8 weeks. A total of 127 patients formed RBX2660 treatment, patients’ microbiomes pro-
the ITT population (enrolled at 21 sites in the gressively resembled those of RBX2660.
U.S. and Canada); patients were randomized into
three treatment arms: two doses of RBX2660 SER-109
(Group A, n ¼ 41); two doses of placebo SER-109 (Seres) is also a live biotheraputic that
(Group B, n ¼ 44); or one dose of RBX2660 comprises an encapsulated mixture of purified
and one dose of placebo (Group C, n ¼ 42) via Firmicutes spores, obtained from the faeces of
enema with doses 7 days apart. Efficacy for healthy humans, which were effective at
Group A was 61% vs. 45.5% for Group B, preventing CDI in animal models. The resilience
P ¼ 0.152. Efficacy for Group C was 66.7% of the spores means that an ethanol based purifi-
compared with Group B (45.5%), P ¼ 0.048; cation process can be applied to reduce the risk
efficacy of Group A and C (63.9%) vs. B that transmissible infectious agents contaminate
(45.5%), P ¼ 0.046. For subjects who developed the therapeutic product. Also, resistance to gas-
recurrent CDI after receipt of study drug, open- tric acid facilitates oral dosing. Two phase 2 stud-
label treatment success was Group A (68.8%, ies of SER-109 have been completed. The first
11/16); Group B (87.5%, 21/24); Group C was a non-comparative study in patients with 3
(71.4%; 10/14) for an overall open label success CDI episodes during 12 months (Khanna et al.
rate of 77.8%. Adverse events at 56 days were 2016). Following standard of care CDI antibiotic
primarily gastrointestinal, with no significant dif- treatment, patients received SER-109 either on
ference in the proportion of adverse or serious two consecutive days (geometric mean dose,
adverse events among the treatment groups. As 1.7  109 spores), or on 1 day (geometric mean
the two doses of RBX2660 treatment arm was dose, 1.1  108 spores). The primary end point
not superior to two doses of placebo, the primary was absence of C. difficile-positive diarrhoea
efficacy endpoint was not met. during 8 weeks of follow-up. In total, 26/30
The third phase 2 study, PUNCH Open Label patients (86.7%) across the two dosing groups
(NCT02589847) had 31 active treatment sites met the primary efficacy end point. Three
and four control sites in the US and Canada. patients with early, self-limiting C. difficile-posi-
One hundred thirty-two RBX2660 and 110 histor- tive diarrhoea did not require antibiotic treat-
ical control subjects were included; follow up ment, and were C. difficile-negative on
results at 8 weeks have been reported, although re-testing at 8 weeks; thus, 29/30 (96.7%) were
there is a 2-year assessment point also (Rebiotix considered to have achieved clinical resolution.
Inc 2017). RBX2660 met its primary efficacy Notably, gut microbiome analyses showed that
endpoint at 8 weeks, preventing CDI recurrence, baseline loss of microbiota diversity was rapidly
with a success rate of 78.8% compared with reversed after receipt of SER-109, with persis-
51.8% in historical controls treated with tence of Firmicutes spores. There were no safety
antibiotics alone (p < 0.0001). No new safety concerns in the study.
concerns were identified. Analyses of faecal A recently completed, phase 2 (ECOSPORE)
microbiomes shows that these became more study of SER-109 enrolled 89 subjects with 3
diverse and aligned to a ‘healthy’ microbiome recurrences who were randomized (2:1 ratio) in a
after treatment with RBX2660 (Blount et al. placebo-controlled, double-blind, 24-week trial
2017; Ray et al. 2017). 16S rRNA sequencing (Trucksis et al. 2017). SER-109 was
was also performed on stool samples collected administered orally as a single dose (1  108
130 F. Fitzpatrick et al.

bacterial spores), after CDI antibiotic treatment. of SER-109 (14 of 31) recipients, and in 80% of
Recurrence was defined as diarrhoea for 2 con- those who received placebo (12 of 15). A
secutive days, a positive CDI test, and the need re-analysis showed that the disappointing results
for antibiotic treatment. The study’s primary may be because cases were included and
endpoint of reducing the relative risk of CDI recurrences diagnosed without the most stringent
recurrence at 8 weeks was not achieved, despite requirement for free faecal toxin to be present.
a (non-significant) reduction in the relative risk Also, while SER-109 was biologically active, a
of CDI recurrence. In the ITT population, recur- higher dose may be necessary. Further clinical
rence occurred in 44% (26/59) vs. 53% (16/30) trials are now in progress.
of subjects who received SER-109 vs. placebo,
respectively. A pre-specified sub-group analysis
Non-toxigenic C. difficile
showed that the lack of efficacy of SER-109 to
Non-toxigenic C. difficile (NTCD) strains are
prevent recurrence occurred in subjects aged
avirulent. Theoretically, it may be possible to
<65 years old. However, in subjects aged
displace toxigenic strains in colonised
65 years old, CDI recurrence occurred in 45%
(or infected) individuals. A randomized,

Table 4 Anti-CDI agents in the pipeline agents that have completed at least a phase 2 clinical trial for treatment or
prevention of CDI
trial phase Drug/product (developer) Indication Notes
Phase III C. difficile vaccine (Sanofi Primary prevention of CDI
Pasteur) NCT01887912: efficacy of vaccine (3 doses) containing toxin A and B
C. difficile vaccine Primary prevention of CDI
(Pfizer) Vaccine containing toxoids of toxin A and B. 3 doses
NCT03090191: efficacy of vaccine (3 doses) containing toxin A and B
SER-109 (Seres) Treatment of recurrent CDI
Oral microbiome therapeutic (mixture of bacterial spores) tested in a
single-arm, open-label clinical trial
NCT03183128: Is SER-109 superior vs placebo to reduce recurrence of
Phase II Ridinilazole (SMT 19969, Treatment of CDI
Summit) Ridinilazole is a novel, small molecule, highly selective antibiotic.
Successful phase 2 trial completed; phase 3 initiation expected 2018
RBX2660 (Rebiotix) Treatment of recurrent CDI
Microbiota Suspension. 3 completed phase 2 trials
Expected to enter Phase 3 in 2017/18
SYN-004 (Synthetic Prevention of CDI. SYN-004 is a class A b-lactamase
Biologics) Successful phase 2 trial completed; phase 3 initiation expected 2017/18
VLA84 (Valneva) Primary prevention of CDI
Vaccine consisting of a fusion protein with portions of toxins A and B
Successful phase 2 trial completed in 2016
Non-toxigenic C. difficile Prevention of recurrent CDI
(Viropharma) Biological therapy. Completed successful phase 2 trial in 2013
Ramoplanin Treatment of CDI
(Nanotherapeutics) No new clinical efficacy data published since a phase 2 study was
completed in 2004
Development plans/potential is therefore unclear. No clinical studies
listed in
European Practice for CDI Treatment 131

double-blind, placebo-controlled, dose-ranging Chemother 70(1):182–189.

study examined the efficacy of a NTCD strain jac/dku324
Baldoni D, Gutierrez M, Timmer W, Dingemanse J
to prevent recurrent CDI in patients with either (2014) Cadazolid, a novel antibiotic with potent activ-
primary (>80%) or recurrent CDI who had ity against Clostridium difficile: safety, tolerability
completed treatment with metronidazole, vanco- and pharmacokinetics in healthy subjects following
mycin, or both (Gerding et al. 2015). Approxi- single and multiple oral doses. J Antimicrob
Chemother 69(3):706–714.
mately two thirds (69%) of recipients became jac/dkt401
colonised by NTCD. CDI recurrence rates were Basseres E, Endres BT, Khaleduzzaman M, Miraftabi F,
2% in colonized subjects, compared with 31% Alam MJ, Vickers RJ, Garey KW (2016) Impact on
(similar to placebo) in those not colonised toxin production and cell morphology in Clostridium
difficile by ridinilazole (SMT19969), a novel treat-
(p < 0.001), highlighting the correlation between ment for C. difficile infection. J Antimicrob
engraftment and clinical efficacy. Interestingly, Chemother 71(5):1245–1251.
no subjects who were colonised at week six 1093/jac/dkv498
remained so at week 26. It remains unclear Blount K, Jones C, Shannon W, Carter S (2017) Changing
the microbiome: patients with a successful outcome
whether this successful proof of concept phase following microbiota-based RBX2660 treatment trend
2 clinical trial will lead to commercial develop- toward human microbiome project healthy subjects’
ment of the NTCD strain. profile. Paper presented at the Americal Society of
In summary, there are varied approaches in Microbiology (ASM) Microbe, New Orleans, USA.
Abstract 212
advanced clinical trials for the primary preven- Boix V, Fedorak RN, Mullane KM, Pesant Y,
tion, treatment and/or secondary prevention of Stoutenburgh U, Jin M, Adedoyin A, Chesnel L,
CDI (Table 4). Unfortunately, however, recent Guris D, Larson KB, Murata Y (2017) Primary
experience shows us that developing new man- outcomes from a phase 3, randomized, double-blind,
active-controlled trial of surotomycin in subjects with
agement options for CDI is very challenging. Clostridium difficile infection. Open Forum Infect Dis
Well-designed trials with clearly defined patient 4(1):ofw275.
populations are key to delivering new therapeutic Cammarota G, Masucci L, Ianiro G, Bibbo S, Dinoi G,
and preventative options. Costamagna G, Sanguinetti M, Gasbarrini A (2015)
Randomised clinical trial: faecal microbiota transplan-
tation by colonoscopy vs. vancomycin for the treat-
Acknowledgements We wish to thank Mr. Myles ment of recurrent Clostridium difficile infection.
Houlden, Health Protection Surveillance Centre for assis- Aliment Pharmacol Ther 41(9):835–843. https://doi.
tance with Demographix® software and the National CDI org/10.1111/apt.13144
experts who took the time to complete the online survey Cammarota G, Ianiro G, Tilg H, Rajilic-Stojanovic M,
of CDI treatment in Europe. Kump P, Satokari R, Sokol H, Arkkila P, Pintus C,
Hart A, Segal J, Aloi M, Masucci L, Molinaro A,
Scaldaferri F, Gasbarrini G, Lopez-Sanroman A,
Link A, de Groot P, de Vos WM, Hogenauer C,
References Malfertheiner P, Mattila E, Milosavljevic T,
Nieuwdorp M, Sanguinetti M, Simren M, Gasbarrini
ActelionLtd. (2017) Actelion provides an update on the A (2017) European consensus conference on faecal
Phase III IMPACT program with cadazolid in CDAD microbiota transplantation in clinical practice. Gut
Alang N, Kelly CR (2015) Weight gain after fecal 66(4):569–580.
microbiota transplantation. Open Forum Infect Dis 2 313017
(1):ofv004. Chang JY, Antonopoulos DA, Kalra A, Tonelli A, Khalife
Arends S, Defosse J, Diaz C, Wappler F, Sakka SG (2017) WT, Schmidt TM, Young VB (2008) Decreased diver-
Successful treatment of severe Clostridium difficile sity of the fecal Microbiome in recurrent Clostridium
infection by administration of crushed fidaxomicin difficile-associated diarrhea. J Infect Dis 197
via a nasogastric tube in a critically ill patient. Int J (3):435–438.
Infect Dis 55:27–28. Chang J, Kane A, McDermott L, Vickers R, Snydman D,
2016.12.020 Thorpe C (2016) Ridinilazole preserves major
Baines SD, Crowther GS, Freeman J, Todhunter S, components of the intestinal microbiota during treat-
Vickers R, Wilcox MH (2015) SMT19969 as a treat- ment of Clostridium difficile infection. Paper
ment for Clostridium difficile infection: an assessment presented at the ECCMID Amsterdam, Netherlands.
of antimicrobial activity using conventional suscepti- Abstract LB-116
bility testing and an in vitro gut model. J Antimicrob
132 F. Fitzpatrick et al.

Chew KS, van Merrienboer J, Durning SJ (2016) A porta- Chachaty E, Sayah-Jeanne S, Doré J, Le Chatelier E,
ble mnemonic to facilitate checking for cognitive Levenez F, Kennedy S, Pons N, Weiss W, Wilcox M,
errors. BMC Res Notes 9(1):445. Mentré F, Andremont A, Dusko Ehrlich S (2015)
1186/s13104-016-2249-2 DAV132, an adsorbent-based product, protects the
Chilton CH, Crowther GS, Baines SD, Todhunter SL, gut microbiome and prevents Clostridium difficile
Freeman J, Locher HH, Athanasiou A, Wilcox MH infections during moxifloxacin treatments. Paper
(2014a) In vitro activity of cadazolid against clinically presented at the IDWeek 2015, San Diego, USA
relevant Clostridium difficile isolates and in an in vitro De Leon LM, Watson JB, Kelly CR (2013) Transient flare
gut model of C. difficile infection. J Antimicrob of ulcerative colitis after fecal microbiota transplanta-
Chemother 69(3):697–705. tion for recurrent Clostridium difficile infection. Clin
jac/dkt411 Gastroenterol Hepatol 11(8):1036–1038. https://doi.
Chilton CH, Crowther GS, Todhunter SL, Nicholson S, org/10.1016/j.cgh.2013.04.045
Freeman J, Chesnel L, Wilcox MH (2014b) Efficacy Debast SB, Bauer MP, Kuijper EJ (2014) European Soci-
of surotomycin in an in vitro gut model of Clostridium ety of Clinical Microbiology and Infectious Diseases:
difficile infection. J Antimicrob Chemother 69 update of the treatment guidance document for Clos-
(9):2426–2433. tridium difficile infection. Clin Microbiol Infect 20
Connelly S, Widmer G, Mukherjee J, Huynh K, Bristol (Suppl 2):1–26.
JA, Hubert S, Sliman J, Tzipori S, Kaleko M (2015) 12418
Tu2054 SYN-004, a clinical stage oral beta-lactamase Department of Health, England and the Health Protection
therapy, protects the intestinal microflora from Agency (HPA) (2008) Clostridium difficile infection:
antibiotic-mediated damage in humanized pigs. Gas- how to deal with the problem
troenterology 148(4):S-1195. Deshpande A, Hurless K, Cadnum JL, Chesnel L, Gao L,
S0016-5085(15)34082-8 Chan L, Kundrapu S, Polinkovsky A, Donskey CJ
Corbett D, Wise A, Birchall S, Warn P, Baines SD, (2016) Effect of surotomycin, a novel cyclic
Crowther G, Freeman J, Chilton CH, Vernon J, lipopeptide antibiotic, on intestinal colonization with
Wilcox MH, Vickers RJ (2015) In vitro susceptibility vancomycin-resistant enterococci and Klebsiella
of Clostridium difficile to SMT19969 and pneumoniae in mice. Antimicrob Agents Chemother
comparators, as well as the killing kinetics and post- 60(6):3333–3339.
antibiotic effects of SMT19969 and comparators 15
against C. difficile. J Antimicrob Chemother 70 Donald RG, Flint M, Kalyan N, Johnson E, Witko SE,
(6):1751–1756. Kotash C, Zhao P, Megati S, Yurgelonis I, Lee PK,
Cornely OA, Crook DW, Esposito R, Poirier A, Somero Matsuka YV, Severina E, Deatly A, Sidhu M, Jansen
MS, Weiss K, Sears P, Gorbach S (2012) Fidaxomicin KU, Minton NP, Anderson AS (2013) A novel
versus vancomycin for infection with Clostridium dif- approach to generate a recombinant toxoid vaccine
ficile in Europe, Canada, and the USA: a double-blind, against Clostridium difficile. Microbiology 159
non-inferiority, randomised controlled trial. Lancet (Pt 7):1254–1266.
Infect Dis 12(4):281–289. 066712-0
s1473-3099(11)70374-7 Dubberke E, Lee C, Orenstein R, Khanna S, Hecht G,
Crook DW, Walker AS, Kean Y, Weiss K, Cornely OA, Fraiz J (2016) Efficacy and safety of RBX2660 for the
Miller MA, Esposito R, Louie TJ, Stoesser NE, Young prevention of recurrent Clostridium difficile infection:
BC, Angus BJ, Gorbach SL, Peto TE (2012) results of the PUNCH CD 2 trial. Paper presented at
Fidaxomicin versus vancomycin for Clostridium diffi- the IDWeek 2016, New Orleans, LA, USA. Abstract
cile infection: meta-analysis of pivotal randomized 1341
controlled trials. Clin Infect Dis 55(Suppl 2): ECDC (2017) Clostridium difficile. https://ecdc.europa.
S93–103. eu/en/publications-data/directory-guidance-preven
Davies KA, Longshaw CM, Davis GL, Bouza E, tion-and-control/clostridium-difficile. Accessed
Barbut F, Barna Z, Delmee M, Fitzpatrick F, 30 June 2017
Ivanova K, Kuijper E, Macovei IS, Mentula S, Fischer M, Kao D, Mehta SR, Martin T, Dimitry J,
Mastrantonio P, von Muller L, Oleastro M, Keshteli AH, Cook GK, Phelps E, Sipe BW, Xu H,
Petinaki E, Pituch H, Noren T, Novakova E, Nyc O, Kelly CR (2016) Predictors of early failure after fecal
Rupnik M, Schmid D, Wilcox MH (2014) Underdiag- microbiota transplantation for the therapy of Clostrid-
nosis of Clostridium difficile across Europe: the ium Difficile infection: a multicenter study. Am J
European, multicentre, prospective, biannual, point- Gastroenterol 111(7):1024–1031.
prevalence study of Clostridium difficile infection in 1038/ajg.2016.180
hospitalised patients with diarrhoea (EUCLID). Lan- Fitzpatrick F (2008) Management of Clostridium difficile
cet Infect Dis 14(12):1208–1219. infection – medical or surgical? Surgeon 6
1016/s1473-3099(14)70991-0 (6):325–328
de Gunzburg J, Ghozlane A, Ducher A, Duval X, Foglia G, Shah S, Luxemburger C, Pietrobon PJ (2012)
Ruppé E, Pulse M, Chilton C, Armand-Lefevre L, Clostridium difficile: development of a novel
European Practice for CDI Treatment 133

candidate vaccine. Vaccine 30(29):4307–4309. Kaleko M, Bristol JA, Hubert S, Parsley T, Widmer G, Tzipori S, Subramanian P, Hasan N, Koski P, Kokai-
Freeman J, Vernon J, Vickers R, Wilcox MH (2015) Kun J, Sliman J, Jones A, Connelly S (2016) Develop-
Susceptibility of Clostridium difficile isolates of vary- ment of SYN-004, an oral beta-lactamase treatment to
ing antimicrobial resistance phenotypes to SMT19969 protect the gut microbiome from antibiotic-mediated
and 11 comparators. Antimicrob Agents Chemother damage and prevent Clostridium difficile infection.
60(1):689–692. Anaerobe 41:58–67.
Gehin M, Desnica B, Dingemanse J (2015) Minimal sys- obe.2016.05.015
temic and high faecal exposure to cadazolid in patients Kautza B, Zuckerbraun BS (2016) The surgical manage-
with severe Clostridium difficile infection. Int J ment of complicated Clostridium difficile infection:
Antimicrob Agents 46(5):576–581. alternatives to colectomy. Surg Infect 17(3):337–342.
Gerding DN, Meyer T, Lee C, Cohen SH, Murthy UK, Kelly CR, Khoruts A, Staley C, Sadowsky MJ, Abd M,
Poirier A, Van Schooneveld TC, Pardi DS, Ramos A, Alani M, Bakow B, Curran P, McKenney J, Tisch A,
Barron MA, Chen H, Villano S (2015) Administration Reinert SE, Machan JT, Brandt LJ (2016) Effect of
of spores of nontoxigenic Clostridium difficile strain fecal microbiota transplantation on recurrence in mul-
M3 for prevention of recurrent C. difficile infection: a tiply recurrent Clostridium difficile infection: a
randomized clinical trial. JAMA 313(17):1719–1727. randomized trial. Ann Intern Med 165(9):609–616.
Goldenberg SD, Brown S, Edwards L, Gnanarajah D, Khanna S, Pardi DS, Kelly CR, Kraft CS, Dhere T, Henn
Howard P, Jenkins D, Nayar D, Pasztor M, Oliver S, MR, Lombardo MJ, Vulic M, Ohsumi T, Winkler J,
Planche T, Sandoe JA, Wade P, Whitney L (2016) The Pindar C, McGovern BH, Pomerantz RJ, Aunins JG,
impact of the introduction of fidaxomicin on the man- Cook DN, Hohmann EL (2016) A novel microbiome
agement of Clostridium difficile infection in seven therapeutic increases gut microbial diversity and
NHS secondary care hospitals in England: a series of prevents recurrent Clostridium difficile infection. J
local service evaluations. Eur J Clin Microbiol Infect Infect Dis 214(2):173–181.
Dis 35(2):251–259. infdis/jiv766
015-2538-z Kotloff KL, Wasserman SS, Losonsky GA, Thomas W Jr,
Goldstein EJ, Citron DM, Tyrrell KL, Merriam CV Nichols R, Edelman R, Bridwell M, Monath TP
(2013) Comparative in vitro activities of SMT19969, (2001) Safety and immunogenicity of increasing
a new antimicrobial agent, against Clostridium diffi- doses of a Clostridium difficile toxoid vaccine
cile and 350 gram-positive and gram-negative aerobic administered to healthy adults. Infect Immun 69
and anaerobic intestinal flora isolates. Antimicrob (2):988–995.
Agents Chemother 57(10):4872–4876. https://doi. 995.2001
org/10.1128/aac.01136-13 Kuehne SA, Cartman ST, Heap JT, Kelly ML,
Hota SS, Sales V, Tomlinson G, Salpeter MJ, McGeer A, Cockayne A, Minton NP (2010) The role of toxin A
Coburn B, Guttman DS, Low DE, Poutanen SM and toxin B in Clostridium difficile infection. Nature
(2017) Oral vancomycin followed by fecal transplan- 467(7316):711–713.
tation versus tapering oral vancomycin treatment for nature09397
recurrent Clostridium difficile infection: an open- Kyne L, Warny M, Qamar A, Kelly CP (2000) Asymp-
label, randomized controlled trial. Clin Infect Dis 64 tomatic carriage of Clostridium difficile and serum
(3):265–271. levels of IgG antibody against toxin A. N Engl J
Hu MY, Katchar K, Kyne L, Maroo S, Tummala S, Med 342(6):390–397.
Dreisbach V, Xu H, Leffler DA, Kelly CP (2009) nejm200002103420604
Prospective derivation and validation of a clinical Kyne L, Warny M, Qamar A, Kelly CP (2001) Associa-
prediction rule for recurrent Clostridium difficile tion between antibody response to toxin A and protec-
infection. Gastroenterology 136(4):1206–1214. tion against recurrent Clostridium difficile diarrhoea. Lancet 357(9251):189–193.
Johnson S, Gerding DN (2017) Fecal fixation: fecal s0140-6736(00)03592-3
microbiota transplantation for Clostridium difficile Lee CH, Steiner T, Petrof EO, Smieja M, Roscoe D,
infection. Clin Infect Dis 64(3):272–274. https://doi. Nematallah A, Weese JS, Collins S, Moayyedi P,
org/10.1093/cid/ciw735 Crowther M, Ropeleski MJ, Jayaratne P, Higgins D,
Johnson S, Louie TJ, Gerding DN, Cornely OA, Chasan- Li Y, Rau NV, Kim PT (2016) Frozen vs fresh fecal
Taber S, Fitts D, Gelone SP, Broom C, Davidson DM microbiota transplantation and clinical resolution of
(2014) Vancomycin, metronidazole, or tolevamer for diarrhea in patients with recurrent Clostridium difficile
Clostridium difficile infection: results from two multi- infection: a randomized clinical trial. JAMA 315
national, randomized, controlled trials. Clin Infect Dis (2):142–149.
59(3):345–354. 18098
134 F. Fitzpatrick et al.

Louie TJ, Miller MA, Mullane KM, Weiss K, Lentnek A, presented at the ASM Microbe, New Orleans, USA.
Golan Y, Gorbach S, Sears P, Shue YK (2011) Abstract 262
Fidaxomicin versus vancomycin for Clostridium diffi- Rebiotix Inc (2017) Rebiotix reports positive top line data
cile infection. N Engl J Med 364(5):422–431. https:// from open-label phase 2 trial of RBX2660 in recurrent Clostridium difficile
Moura I, Spigaglia P, Barbanti F, Mastrantonio P (2013) Roberts T, Kokai-Kun JF, Coughlin O, Lopez BV,
Analysis of metronidazole susceptibility in different Whalen H, Bristol JA, Hubert S, Longstreth J,
Clostridium difficile PCR ribotypes. J Antimicrob Lasseter K, Sliman J (2016) Tolerability and pharma-
Chemother 68(2):362–365. cokinetics of SYN-004, an orally administered beta-
jac/dks420 lactamase for the prevention of Clostridium difficile-
Na X, Martin AJ, Sethi S, Kyne L, Garey KW, Flores SW, associated disease and antibiotic-associated diarrhea,
Hu M, Shah DN, Shields K, Leffler DA, Kelly CP in two phase 1 studies. Clin Drug Investig 36
(2015) A multi-center prospective derivation and val- (9):725–734.
idation of a clinical prediction tool for severe Clos- 0420-0
tridium difficile infection. PLoS One 10(4):e0123405. Sartelli M, Malangoni MA, Abu-Zidan FM, Griffiths EA, Di Bella S, McFarland LV, Eltringham I, Shelat VG,
Nelson RL, Suda KJ, Evans CT (2017) Antibiotic treat- Velmahos GC, Kelly CP, Khanna S, Abdelsattar ZM,
ment for Clostridium difficile-associated diarrhoea in Alrahmani L, Ansaloni L, Augustin G, Bala M,
adults. Cochrane Database Syst Rev 3:Cd004610. Barbut F, Ben-Ishay O, Bhangu A, Biffl WL, Brecher SM, Camacho-Ortiz A, Cainzos MA, Canterbury LA,
Orenstein R, Dubberke E, Hardi R, Ray A, Mullane K, Catena F, Chan S, Cherry-Bukowiec JR, Clanton J,
Pardi DS, Ramesh MS (2016) Safety and durability of Coccolini F, Cocuz ME, Coimbra R, Cook CH, Cui Y,
RBX2660 (microbiota suspension) for recurrent Clos- Czepiel J, Das K, Demetrashvili Z, Di Carlo I, Di
tridium difficile infection: results of the PUNCH CD Saverio S, Dumitru IM, Eckert C, Eckmann C, Eiland
study. Clin Infect Dis 62(5):596–602. EH, Enani MA, Faro M, Ferrada P, Forrester JD, Fraga
10.1093/cid/civ938 GP, Frossard JL, Galeiras R, Ghnnam W, Gomes CA,
Pechine S, Gleizes A, Janoir C, Gorges-Kergot R, Barc Gorrepati V, Ahmed MH, Herzog T, Humphrey F,
MC, Delmee M, Collignon A (2005) Immunological Kim JI, Isik A, Ivatury R, Lee YY, Juang P, Furuya-
properties of surface proteins of Clostridium difficile. J Kanamori L, Karamarkovic A, Kim PK, Kluger Y, Ko
Med Microbiol 54(Pt 2):193–196. WC, LaBarbera FD, Lee JG, Leppaniemi A,
1099/jmm.0.45800-0 Lohsiriwat V, Marwah S, Mazuski JE, Metan G,
Penziner S, Dubrovskaya Y, Press R, Safdar A (2015) Moore EE, Moore FA, Nord CE, Ordonez CA, Junior
Fidaxomicin therapy in critically ill patients with GA, Petrosillo N, Portela F, Puri BK, Ray A, Raza M,
Clostridium difficile infection. Antimicrob Agents Rems M, Sakakushev BE, Sganga G, Spigaglia P,
Chemother 59(3):1776–1781. Stewart DB, Tattevin P, Timsit JF, To KB, Trana C,
1128/aac.04268-14 Uhl W, Urbanek L, van Goor H, Vassallo A, Zahar JR,
Porter RJ, Fogg C (2015) Faecal microbiota transplanta- Caproli E, Viale P (2015) WSES guidelines for man-
tion for Clostridium difficile infection in the United agement of Clostridium difficile infection in surgical
Kingdom. Clin Microbiol Infect 21(6):578–582. patients. World J Emerg Surg: World J Emerg Surg 10:38.
Prior AR, Kevans D, McDowell L, Cudmore S, Sattar A, Thommes P, Payne L, Warn P, Vickers RJ
Fitzpatrick F (2017) Treatment of Clostridium difficile (2015) SMT19969 for Clostridium difficile infection
infection: a national survey of clinician (CDI): in vivo efficacy compared with fidaxomicin
recommendations and the use of faecal microbiota and vancomycin in the hamster model of CDI. J
transplantation. J Hosp Infect 95(4):438–441. https:// Antimicrob Chemother 70(6):1757–1762. https://doi. org/10.1093/jac/dkv005
Public Health England (2013) Updated guidance on the Sheldon E, Kitchin N, Peng Y, Eiden J, Gruber W,
management and treatment of Clostridium difficile Johnson E, Jansen KU, Pride MW, Pedneault L
infection (2016) A phase 1, placebo-controlled, randomized
Quera R, Espinoza R, Estay C, Rivera D (2014) Bacter- study of the safety, tolerability, and immunogenicity
emia as an adverse event of fecal microbiota trans- of a Clostridium difficile vaccine administered with or
plantation in a patient with Crohn’s disease and without aluminum hydroxide in healthy adults. Vac-
recurrent Clostridium difficile infection. J Crohns cine 34(18):2082–2091.
Colitis 8(3):252–253. cine.2016.03.010
crohns.2013.10.002 Solomon K, Martin AJ, O’Donoghue C, Chen X,
Ray A, Jones C, Shannon W, Carter S (2017) Resetting Fenelon L, Fanning S, Kelly CP, Kyne L (2013) Mor-
the microbial landscape: donor microbiome engraft- tality in patients with Clostridium difficile infection
ment in patients treated with RBX2660 for multi- correlates with host pro-inflammatory and humoral
recurrent Clostridium difficile infection. Paper immune responses. J Med Microbiol 62
European Practice for CDI Treatment 135

(Pt 9):1453–1460. and recurrence of Clostridium difficile infection fol-

058479-0 lowing treatment with vancomycin or metronidazole:
Stevens VW, Nelson RE, Schwab-Daugherty EM, a systematic review of the evidence. Int J Antimicrob
Khader K, Jones MM, Brown KA, Greene T, Croft Agents 40(1):1–8.
LD, Neuhauser M, Glassman P, Goetz MB, Samore ijantimicag.2012.01.004
MH, Rubin MA (2017) Comparative effectiveness of Vickers R, Robinson N, Best E, Echols R, Tillotson G,
vancomycin and metronidazole for the prevention of Wilcox M (2015) A randomised phase 1 study to
recurrence and death in patients with Clostridium dif- investigate safety, pharmacokinetics and impact on
ficile infection. JAMA Intern Med 177(4):546–553. gut microbiota following single and multiple oral doses in healthy male subjects of SMT19969, a
Synthetic Biologics I (2017) SYN-004 (Ribaxamase) novel agent for Clostridium difficile infections. BMC
receives breakthrough therapy designation from Infect Dis 15:91.
U.S. Food and Drug Administration for prevention of 0759-5
Clostridium difficile infection Vickers RJ, Tillotson G, Goldstein EJ, Citron DM, Garey
Tariq R, Smyrk T, Pardi DS, Tremaine WJ, Khanna S KW, Wilcox MH (2016) Ridinilazole: a novel therapy
(2016) New-onset microscopic colitis in an ulcerative for Clostridium difficile infection. Int J Antimicrob
colitis patient after fecal microbiota transplantation. Agents 48(2):137–143.
Am J Gastroenterol 111(5):751–752. ijantimicag.2016.04.026
10.1038/ajg.2016.67 Vickers RJ, Tillotson GS, Nathan R, Hazan S, Pullman J,
Trucksis M, Baird I, Cornely O, Golan Y, Hecht G, Lucasti C, Deck K, Yacyshyn B, Maliakkal B,
Pardi D, Pullman J, Polage C, Wilcox M, Pesant Y, Tejura B, Roblin D, Gerding DN, Wilcox
Bernardo P, Ford C, O’Brien E, Vetro R, Wortman J, MH (2017) Efficacy and safety of ridinilazole com-
Weston J, Henn M (2017) An analysis of results from pared with vancomycin for the treatment of Clostrid-
the first placebo-controlled trial of single-dose ium difficile infection: a phase 2, randomised, double-
SER-109, an investigational oral microbiome thera- blind, active-controlled, non-inferiority study. Lancet
peutic to reduce the recurrence of Clostridium difficile Infect Dis 17(7):735–744.
infection (CDI). Paper presented at the ECCMID, s1473-3099(17)30235-9
Vienna, Austria. Abstract OS0250C Weiss W, Pulse M, Vickers R (2014) In vivo assessment
Valneva S (2016) Valneva announces successful comple- of SMT19969 in a hamster model of Clostridium
tion of Phase II for Clostridium difficile vaccine difficile infection. Antimicrob Agents Chemother 58
candidate (10):5714–5718.
van Beurden YH, Nieuwdorp M, van de Berg P, Mulder 14
CJJ, Goorhuis A (2017) Current challenges in the Wilcox MH, Gerding DN, Poxton IR, Kelly C, Nathan R,
treatment of severe Clostridium difficile infection: Birch T, Cornely OA, Rahav G, Bouza E, Lee C,
early treatment potential of fecal microbiota trans- Jenkin G, Jensen W, Kim YS, Yoshida J,
plantation. Therap Adv Gastroenterol 10(4):373–381. Gabryelski L, Pedley A, Eves K, Tipping R, Guris D, Kartsonis N, Dorr MB (2017) Bezlotoxumab for pre-
van Nood E, Vrieze A, Nieuwdorp M, Fuentes S, vention of recurrent Clostridium difficile infection. N
Zoetendal EG, de Vos WM, Visser CE, Kuijper EJ, Engl J Med 376(4):305–317.
Bartelsman JF, Tijssen JG, Speelman P, Dijkgraaf NEJMoa1602615
MG, Keller JJ (2013) Duodenal infusion of donor Youngster I, Russell GH, Pindar C, Ziv-Baran T, Sauk J,
feces for recurrent Clostridium difficile. N Engl J Hohmann EL (2014) Oral, capsulized, frozen fecal
Med 368(5):407–415. microbiota transplantation for relapsing Clostridium
NEJMoa1205037 difficile infection. JAMA 312(17):1772–1778. https://
Vardakas KZ, Polyzos KA, Patouni K, Rafailidis PI,
Samonis G, Falagas ME (2012) Treatment failure
Antibiotic Resistances of Clostridium

Patrizia Spigaglia, Paola Mastrantonio, and Fabrizio Barbanti

Abstract maintained in C. difficile regardless of the

The rapid evolution of antibiotic resistance in burden imposed on fitness, and therefore
Clostridium difficile and the consequent resistances may persist in C. difficile popula-
effects on prevention and treatment of tion in absence of antibiotic selective
C. difficile infections (CDIs) are matter of pressure.
concern for public health. Antibiotic resis-
tance plays an important role in driving Keywords
C. difficile epidemiology. Emergence of new C. difficile · Antibiotic susceptibility
types is often associated with the emergence methods · Mechanisms of resistance · Multi-
of new resistances and most of epidemic drug resistance (MDR)
C. difficile clinical isolates is currently resis-
tant to multiple antibiotics. In particular, it is
to worth to note the recent identification of 1 Introduction
strains with reduced susceptibility to the first-
line antibiotics for CDI treatment and/or for Clostridium difficile is recognized as the major
relapsing infections. Antibiotic resistance in cause of healthcare antibiotic-associated diarrhea
C. difficile has a multifactorial nature. Acqui- (Lessa et al. 2015; European Centre for Disease
sition of genetic elements and alterations of Prevention and Control (ECDC) 2013). Poten-
the antibiotic target sites, as well as other tially, all antibiotic classes may promote
factors, such as variations in the metabolic C. difficile infection (CDI) by disrupting intesti-
pathways and biofilm production, contribute nal microflora and allowing C. difficile, ingested
to the survival of this pathogen in the presence or resident, to proliferate, colonize the gastroin-
of antibiotics. Different transfer mechanisms testinal tract, and infect the host. Therefore,
facilitate the spread of mobile elements resistance to multiple agents represents a selec-
among C. difficile strains and between tive advantage for C. difficile strains to enhance
C. difficile and other species. Furthermore, their survival and spread.
recent data indicate that both genetic elements An alarming increase in incidence of CDI has
and alterations in the antibiotic targets can be been observed worldwide over the last 15 years,

P. Spigaglia (*) · P. Mastrantonio · F. Barbanti

Department of Infectious Diseases, Istituto Superiore di
Sanità, Rome, Italy

# Springer International Publishing AG 2018 137

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
138 P. Spigaglia et al.

with a significant financial burden on the 2 C. difficile Antibiotic

healthcare system (Redelings et al. 2007; Susceptibility
Burckhardt et al. 2008; Bauer et al. 2011; Gravel
et al. 2009; Miller et al. 2011a; Dubberke and C. difficile susceptibility is usually evaluated for
Olsen 2012; Lessa et al. 2012). The increased antibiotics known to be significantly associated
number of infections has been mainly associated to CDI or used for CDI treatment. Among the
with the emergence of highly virulent C. difficile first group, clindamycin (CLI) and
strains. In particular, strains PCR-ribotype cephalosporins (CFs) are historically recognized
(RT) 027/North American pulsed field gel elec- as high-risk agents for CDI (Bartlett et al. 1977;
trophoresis type I (NAPI)/restriction endonucle- Bignardi 1998). Although a decreased number of
ase analysis group B1, have been recognized infections has been observed in the hospitals that
responsible for severe CDI, characterized by have curtailed the use of these antibiotics
high rate of recurrences, mortality and refractory (de Lalla et al. 1989; Khan and Cheesbrough
to traditional therapy (Pépin et al. 2004; 2005b; 2003; Wistrom et al. 2001), the risk of hospital
McDonald et al. 2005; Muto et al. 2005; acquired-CDI remains high after CLI or CFs
Goorhuis et al. 2007; Clements et al. 2010). therapy, so their importance as promoting agents
Despite the wide diffusion of RT 027, recent should not be minimized. More recently, a rise in
European surveillances indicated the emergence the FQs-associated CDI has been observed in
of highly virulent RTs different from RT concomitant with the increasing incidence of
027 (Davies et al. 2014; Freeman et al. 2015a). C. difficile RT 027. Current strains RT
Several types, such as RT 014/020, RT 001/072 027 show high-level resistance to FQs, never
and RT078 are endemic in almost all European observed in historical isolates of the same type
countries, whereas others RTs have a regional (McDonald et al. 2005). Infection control
spread (Freeman et al. 2015a). procedures and antimicrobial stewardship have
Antibiotic have a central role in driving the led to a significant reduction in the incidence of
emergence of new C. difficile types. The global infections caused by RT 027 but this type is still
spread of C. difficile RT 027 has been associated globally widespread (Muto et al. 2007; Lessa
with the massive use of fluoroquinolones (FQs) et al. 2015; Freeman et al. 2015a). Furthermore,
and the acquisition of resistance to these resistance to FQs has become very common also
antibiotics by strains of this type (He et al. in strains belonging to other epidemic types
2013). Actually, the majority of epidemic and (Freeman et al. 2015a, b; Spigaglia et al. 2011).
emergent strains, RT 027 or not, show resistance Standard CDI therapies include metronida-
to multiple antibiotics (Spigaglia et al. 2011). zole (MTZ) and vancomycin (VAN) as first
Genetic analysis have demonstrated that choice for mild and severe CDI, respectively
C. difficile has a versatile genome content, with (Debast et al. 2014; Jarrad et al. 2015; Lyras
a wide range of mobile elements, many of them and Cooper 2015). In addition, rifamycins
encoding for predicted antibiotic resistances (RFs), in particular rifaximin (RFX), have
(Sebaihia et al. 2006; He et al. 2010, 2013). recently been prosed as “chaser therapy” for
Besides horizontal gene transfer, other treatment of relapsing CDI (Iv et al. 2014),
mechanisms may contribute to promote antibi- while fidaxomicin (FDX), a bactericidal new
otic resistance in C. difficile, which appears to be narrow spectrum macrocyclic antibiotic, is used
a multifactorial phenomenon. for the management of CDI with high risk for
In this chapter, antibiotic resistances of recurrences (Chaparro-Rojas and Mullane 2013).
C. difficile will be discussed taking in consider-
ation the most recent published data.
Antibiotic Resistances of Clostridium difficile 139

2.1 Antibiotics Associated to CDI observed an increase in the geometric mean of

MICs for isolates RT 027 (1.1–1.42 mg/L), RT
Although rates of antibiotic resistance varies 001/072 (0.65 mg/L), RT 106 (0.65 mg/L), RT
considerably depending on the geographic 356 (0.61 mg/L) and in the non-toxigenic RT
regions and local/national antibiotic policy, data 010 (1.5 mg/L), compared to other RTs
extrapolated from studies recently published (0.13–0.41 mg/L) (Moura et al. 2013; Freeman
indicate that the majority of C. difficile clinical et al. 2015a, b). In addition, several recent papers
isolates are resistant to CFs, FQs, ERY and CLI have reported the isolation of strains with MICs
(Table 1). In recent studies, performed on a large >2 mg/L, the EUCAST epidemiological cut-off
number of C. difficile strains, is reported that (ECOFF) for MTZ (
resistance to CFs of second generation is more cal_breakpoints) (Table 2). Although the clinical
commonly observed compared to resistance to relevance of strains with reduced susceptibility to
CFs of third generation (95% vs. 38%) (Dong MTZ is still unclear, it has been suggested a
et al. 2013; Pirs et al. 2013; Norman et al. 2014; potential impact of strains RT 027 with reduced
Oka et al. 2012; Karlowsky et al. 2012; Buchler susceptibility to MTZ on the pathophysiology of
et al. 2014; Kuwata et al. 2015; Knight et al. recurrent CDIs (Richardson et al. 2015). In addi-
2015; Knight and Riley 2016). Similarly, resis- tion, strains RT 027 with this characteristic have
tance to ciprofloxacin (CIP), a FQ of second recently been identified as cause of severe
generation, is very common in C. difficile infections in Israel (Adler et al. 2015; Miller-
(99%) (Rodriguez-Pardo et al. 2013; Lee et al. Roll et al. 2016). In particular, a wide outbreak
2014; Norman et al. 2014; Lachowicz et al. 2015; caused by a strain RT 027 with high MIC values
Kuwata et al. 2015; Shayganmehr et al. 2015), for MTZ has been reported in Jerusalem in 2013
while resistance to FQs of fourth generation such (Adler et al. 2015). Besides RT 027, reduced
as moxifloxacin (MXF) and gatifloxacin (GAT) susceptibility to MTZ has also been observed in
has been detected in 36% and 68% of the strains other important epidemic types, such as RT
analyzed, respectively (Karlowsky et al. 2012; 078 and RT 126 (Table 2).
Tenover et al. 2012; Eckert et al. 2013; C. difficile colonies with increased MICs to
Rodriguez-Pardo et al. 2013; Lee et al. 2014; MTZ can be isolated in presence of
Kim et al. 2012; Liao et al. 2012; Terhes et al. sub-inhibitory concentrations of antibiotic
2014; Weber et al. 2013; Pirs et al. 2013; (Peláez et al. 2008; Moura et al. 2013). Heterore-
Varshney et al. 2014; Freeman et al. 2015a, b; sistance, that is the capacity of a part of bacterial
Senoh et al. 2015; Adler et al. 2015; Kociolek population to acquire resistance and grow in
et al. 2016; Putsahit et al. 2017; Gao et al. 2016; presence of an antibiotic, could be considered a
Santos et al. 2016; Knight et al. 2015; Kullin pre-resistance stage in C. difficile (Falagas et al.
et al. 2017). 2008; Peláez et al. 2008). Mean concentrations of
MTZ in the feces of treated patients are not so
high (from 0.8 to 24.2 μg/g) (Bolton and Culshaw
1986), therefore it is possible that the
2.2 Antibiotics for CDI Treatment
concentrations achieved in the colon may be
insufficient for the treatment of infections due
2.2.1 Metronidazole
to strains with higher MIC values for MTZ (Bra-
zier et al. 2001; Baines et al. 2008; Moura et al.
Although percentage of C. difficile strains resis-
tant to MTZ is low (Table 1), several
studies have reported high rate of treatment
2.2.2 Vancomycin
failures in patients that received this antibiotic
(Musher et al. 2005; Pépin et al. 2005a; Vardakas
Reduced susceptibility to VAN in C. difficile is
et al. 2012). Furthermore, it has recently been
not largely diffused as in Enterococci and
140 P. Spigaglia et al.

Table 1 Antibiotic susceptibility of C. difficile clinical isolates as reported in 46 papers published between 2012 and
Antibiotica Number of strains analyzed Number of resistant strains % of resistance
CTT 212 24 11.2
FOX 423 404 95.5
CRO 1252 393 31.4
CTX 95 95 100
CAZ 86 65 76.0
ERY 2316 1138 49.1
CLI 5839 2982 51.1
CIP 1326 1312 99.0
MXF 6053 2161 35.7
GAT 199 136 68.3
MTZ 6724 114 1.7
VAN 5760 134 2.3
RIF 3450 525 15.2
CFs cephalosporins, CTT cefotetan, FOX cefoxitin, CRO ceftriaxone, CTX cefotaxime, CAZ ceftazidime, MLSB
macrolide-lincosamide-streptogramin B, ERY erythromycin, CLI clindamycin, FQs fluoroquinolones, CIP ciprofloxa-
cin, MXF moxifloxacin, GAT gatifloxacin, MTZ metronidazole, VAN vancomycin, RIF rifampin
References: Karlowsky et al. (2012), Liao et al. (2012), Reil et al. (2012), Kim et al. (2012), Oka et al. (2012), Tenover
et al. (2012), Dong et al. (2013), Gouderzi et al. (2013), Pirs et al. (2013), Eckert et al. (2013), Obuch-Woszczatynski
et al. (2013, 2014), Rodriguez-Pardo et al. (2013), Weber et al. (2013), Lee et al. (2014), Simango and Uladi (2014),
Buchler et al. (2014), Norman et al. (2014), Novak et al. (2014), Terhes et al. (2014), Varshney et al. (2014), Zhou et al.
(2014), Kuwata et al. (2015), Shayganmehr et al. (2015), Mackin et al. (2015), Freeman et al. (2015a), Knight et al.
(2015), Knight and Riley (2016), Adler et al. (2015), Eitel et al. (2015), Krutova et al. (2015), Lachowicz et al. (2015),
Senoh et al. (2015), Seugendo et al. (2015), Kociolek et al. (2016), Gao et al. (2016), Kouzegaran et al. (2016), López-
Ureña et al. (2016), Santos et al. (2016), Jamal and Rotimi (2016), Kullin et al. (2016, 2017), Putsahit et al. (2017),
Alvarez-Perez et al. (2017), Nyc et al. (2017), and Ramı́rez-Vargas et al. (2017)

Staphylococci, although an increased number of 2.2.3 Rifamycins

strains with higher MICs to this antibiotic (MICs
range >2-16 mg/L) have recently been isolated Recent data indicate that 15% of C. difficile clin-
(Tables 1 and 2). The clinical significance of ical isolates are resistant to rifampin (RIF)
strains with reduced susceptibility to VAN (Table 1) and the rate of overall resistance
remains to be determined due to the high appears to be rising (Huang et al. 2013;
concentrations that this antibiotic reaches in the Rodrı́guez-Pardo et al. 2013; Eitel et al. 2015;
gastrointestinal tract (Young et al. 1985). Any- Terhes et al. 2014). C. difficile strains resistant to
way, it is noteworthy that reduced susceptibilities rifamycines (RFs) have been detected in almost
to VAN and to MTZ are reported in several RTs, all the countries (17/22) participating in a recent
including RT 027, RT 001, RT 017, RT 078 and C. difficile pan-European surveillance and, in
RT 356/607 (Chia et al. 2013; Goudarzi et al. particular, higher percentages of resistance, rang-
2013; Adler et al. 2015; Freeman et al. 2015a, b; ing from 57% to 64%, have been reported in
Miller-Roll et al. 2016). Strains with these Italy, Czech Republic, Denmark and Hungary
characteristics could represent a potentially seri- (Freeman et al. 2015a, b). Selective pressure
ous problem for first-line treatment of CDI in the after exposure to antibiotic seems to have a role
future. in selecting C. difficile colonies resistant to RFs
Antibiotic Resistances of Clostridium difficile 141

Table 2 C. difficile susceptibility to metronidazole and vancomycin as reported in 14 papers published between 2012
and 2017
Year of of strains % of MIC values Susceptibility Prevalent
Antibiotica publication analyzed resistance (n. of strains) methodb PCR-ribotype References
MTZ 2013 110 3.6 >2(4) AD ndc Chia et al.
2013 75 5.3 32(3); 64(1) AD nd Gouderzi
et al. (2013)
2014 271 13.3 32 (36) ET nd Norman et al.
2015 916 0.1 8(1) AI 106 Freeman et al.
2015 208 18.3 >2 (38) ET 027 Adler et al.
2015 86 4.7 32(4) AD nd Shayganmehr
et al. (2015)
2016 35 28.6 >5 (10) DD nd Kouzegaran
et al. (2016)
2016 457 3.5 >2(16) ET 027, Santos et al.
126,203,651 (2016)
2016 146 2.6 8 ET nd Jamal and
Rotimi (2016)
2016 166 4.2 >2(7) ET 027 Miller-Roll
et al. (2016)
2017 50 4.0 256 (2) ET 078, 126 Alvarez-Perez
VAN 2012 403 0.5 4(2) AD nd Liao et al.
2013 110 9.1 >2(10) AD nd Chia et al.
2013 75 8.0 2 (4); 4 (2) AD nd Gouderzi
et al. (2013)
2014 86 1.2 8(1) ET nd Buchler et al.
2015 918 0.9 >8(8) AI 001/072, Freeman et al.
018, 027, (2015a)
126, 356
2015 208 47.1 >2 (98) ET 017, 027 Adler et al.
2016 196 3.1 4(6) AD nd Kociolek et al.
2016 35 20.0 >5(7) DD nd Kouzegaran
et al. (2016)
2016 457 0.4 3(2) ET 001 Santos et al.
2016 166 15 >2 (25) ET 027 Miller-Roll
et al. (2016)
MTZ metronidazole, VAN vancomycin
AD agar dilution, ET epsilometer test, AI agar incorporation, DD disk diffusion
nd not determined
142 P. Spigaglia et al.

(Curry et al. 2009; Miller et al. 2011b). There- MXF and RIF characterized most of the strains
fore, resistant C. difficile strains might emerge belonging to RT 356/607 and RT 018, two genet-
even during therapy (Johnson et al. 2009; ically correlated types recently emerged in Italy
Carman et al. 2012). RFs are commonly used as (Spigaglia et al. 2010, 2015). Interestingly, RT
anti-tuberculosis (TB) agents. Interestingly, in 018 strains isolated in Korea and Japan show
Poland, all strains belonging to the emergent resistance only to CLI, ERY and MXF (Kim
RT 046 isolated from patients affected by TB et al. 2012; Senoh et al. 2015). The 20-years of
and treated with prolonged RIF therapy, showed use of RIFs in Italy (Salix Pharmaceuticals, Ltd.
high MICs to these antibiotics (Obuch- 10 December 2003, posting date), could explain
Woszczatyński et al. 2013). Susceptibility to the spread of this resistance in Italian C. difficile
RIF correlated completely with susceptibility to isolates. Strains RT 018 are highly virulent and
RFX (Miller et al. 2011b). Thus, susceptibility of transmissible, with a transmission index that has
the rifamycin class in C. difficile can be assessed been demonstrated tenfold higher compared to
by testing susceptibility to RIF. that of strains RT 078 (Baldan et al. 2015). Old
age (65 years), severe pulmonary comorbidity,
previous use of FQs, and infection by RT
3 Multi-drug Resistance (MDR) 018 have been associated as significant risk
in C. difficile factors for complicated infections (Bauer et al.
Many of the most common epidemic RTs,
including the high virulent RT 027 and RT
078, are associated to MDR (Table 3). The first 4 C. difficile Antibiotic
European prospective survey of C. difficile Susceptibility Methods
infections in 2005 showed that 55% of resistant
clinical isolates were MDR (Spigaglia et al. Susceptibility testing is usually performed by
2011). Data from papers published in the last clinical microbiology laboratories to determine
6 years, indicate that about 60% of the analyzed antimicrobial resistance profiles of C. difficile
strains are MDR and the MDR patterns mainly isolates recovered from patients, but it is also
include resistance to CLI, FQs, ERY and CFs used to monitor resistant patterns of strains
(Table 3). Resistance to other antibiotic classes, isolated during epidemiological studies and sur-
such as tetracycline (TET), chloramphenicol veillance networks.
(CHL), imipenem (IMP) is less commonly The most common antibiotic susceptibility
detected in MDR C. difficile isolates. In general, methods used for C. difficile are the agar dilution
percentage of TET-resistant strains ranged (AD) and the epsilometer test, a commercially
between 2.4% and 41.7% (Dong et al. 2013; available gradient diffusion system for quantita-
Pirš et al. 2013; Lachowicz et al. 2015; Norman tive antibiotic susceptibility testing (Fig. 1).
et al. 2014; Simango and Uladi 2014; Zhou et al. The AD is indicated as the reference method
2014), while resistance to CHL and IMP is found for C. difficile by the Clinical and Laboratory
in about 3% and 7.1% of the European clinical Standards Institute (CLSI) (Clinical and Labora-
isolates (Freeman et al. 2015a, b). tory Standards Institute 2012). The AD assay
Interestingly, resistance to multiple shows some advantages for epidemiological stud-
antibiotics characterized recently emerged epi- ies because it is an accurate method, the choice of
demic RTs. In particular, strains RT 176, a type antibiotics to be tested is flexible and can be
closely related to RT 027, recently circulating in modified according to investigational necessity
Poland and the Czech Republic, are and, finally, it is suitable for large number of
characterized by resistance to ERY, MXF, CIP isolates. The disadvantages of the AD approach
and RIF (Obuch-Woszczatyński et al. 2014; are the laborious, time-consuming steps required
Krutova et al. 2015). Resistant to CLI, ERY, to prepare testing plates, particularly when the
Table 3 Antibiotic susceptibility patterns most frequently observed in MDR C. difficile clinical isolates as reported in 19 papers published between 2012 and 2017
Number of % of
Year of strains MDR PCR-
publication analyzed strains Main antibiotic susceptibility patterns (n. of strains)a ribotype References
2012 80 28 CLI MXF RIF (22) 027 Tenover et al. (2012)
2013 145 60 ERY CLI CTX (30) ndb Obuch-Woszczatyński et al. (2013),
CLI CIP FOX (14) DTMc Goudarzi et al. (2013) and Dong et al.
Antibiotic Resistances of Clostridium difficile

2014 183 77 ERY CLI CIP (94) 001, Lee et al. (2014), Obuch-Woszczatyński
017, 018 et al. (2014), Novak et al. (2014) and
ERY CLI CIP CTX (23) nd Simango and Uladi (2014)
ERY CLI MXF CIP GAT (17) 027, 176
2015 525 66 ERY CLI MXF GAT (85) 018, 369 Lachowicz et al. (2015), Senoh et al.
CLI CIP CRO (51) DTM (2015), Kuwata et al. (2015), Spigaglia
ERY CLI MXF RIF (48) 018, et al. (2015), Krutova et al. (2015) and
027, 356/607 Shayganmehr et al. (2015)
ERY MXF CIP IMP (34) 027
ERY MXF RIF (15) 027
CIP CAZ IMP (14) nd
ERY MXF CIP RIF (13) 176
ERY CLI GAT (11) 018, 369
ERY CLI MXF (11) 046,
078, 126
CIP CAZ AMK (10) nd
Table 3 (continued)

Number of % of
Year of strains MDR PCR-
publication analyzed strains Main antibiotic susceptibility patterns (n. of strains)a ribotype References
2016 159 35 CLI MXF RIF CIP (32) 012, 017 Knight and Riley (2016) and López-
ERY CLI MXF (23) 078, 126 Ureña et al. (2016)
2017 276 62 ERY MXF RIF (81) 017 Alvarez-Perez et al. (2017), Kullin et al.
CLI MXF CIP LVX RIF TET CHL TGC LZD (12) 012 (2017) and Ramı́rez-Vargas et al. (2017)
ERY MXF LVX TET (5) 078, 126
ERY LVX TET (5) 078, 126
MXF LVX TET (4) 078, 126
TET LVX ETP (4) 126
ERY erythromycin, CLI clindamycin, MXF moxifloxacin, CIP ciprofloxacin, LVX levofloxacin, GAT gatifloxacin, RIF rifampin, MTZ metronidazole, VAN vancomycin, TET
tetracycline, CHL chloramphenicol, CTX cefotaxime, FOX cefoxitin, IMP imipenem, ETP ertapenem, CAZ ceftazidime, AMK amikacin, CRO ceftriaxone, LZD linezolid, TGC
nd not determined
DTM different typing method
P. Spigaglia et al.
Antibiotic Resistances of Clostridium difficile 145

30 29


n° of papers



5 4
Epsilometer test Agar Dilution Agar Incorporation Disk Diffusion Broth Dilution

Fig. 1 Antibiotic susceptibility methods most frequently Varshney et al. (2014), Zhou et al. (2014), Kuwata et al.
used for C. difficile analysis as reported in 46 papers (2015), Shayganmehr et al. (2015), Mackin et al. (2015),
published between 2012 and 2017 Freeman et al. (2015a), Knight et al. (2015), Knight and
Papers: Karlowsky et al. (2012), Liao et al. (2012), Reil Riley (2016), Adler et al. (2015), Eitel et al. (2015),
et al. (2012), Kim et al. (2012), Oka et al. (2012), Tenover Krutova et al. (2015), Lachowicz et al. (2015), Senoh
et al. (2012), Dong et al. (2013), Gouderzi et al. (2013), et al. (2015), Seugendo et al. (2015), Kociolek et al.
Pirs et al. (2013), Eckert et al. (2013), Obuch- (2016), Gao et al. (2016), Kouzegaran et al. (2016),
Woszczatynski et al. (2013, 2014), Rodriguez-Pardo López-Ureña et al. (2016), Santos et al. (2016), Jamal
et al. (2013), Weber et al. (2013), Lee et al. (2014), and Rotimi (2016), Kullin et al. (2016, 2017), Putsahit
Simango and Uladi (2014), Buchler et al. (2014), Norman et al. (2017), Alvarez-Perez et al. (2017), Nyc et al.
et al. (2014), Novak et al. (2014), Terhes et al. (2014), (2017), and Ramı́rez-Vargas et al. (2017)

number of compounds to be tested is high and/or results in MIC decrease towards susceptibility
when only a limited number of strains are to be range (Peláez et al. 2008; Lynch et al. 2013).
analyzed, and the need of skilled and experi- Recent studies suggest the agar incorporation
enced technologists to properly perform it. For (AI) as the method of choice to detect strains
these reasons, most laboratories use the with reduced susceptibility to MTZ compared
epsilometer test, more flexible and simple, for to the AD (Freeman et al. 2005; Moura et al.
routine. Although there were differences in MIC 2013). Differences in the media used (Schaedlers
values between AD and epsilometer test, high broth and Wilkins-Chalgren agar for AI and Bru-
categorical agreement between these methods cella broth/agar for AD) and in the pre-cultured
has been demonstrated (Moura et al. 2013; period (24 h for AD and 48 h for AI) seem to
Baines et al. 2008; Poilane et al. 2000). In addi- affect MIC determination (Baines et al. 2008;
tion, the epsilometer test allows analysis of sus- Moura et al. 2013). The CLSI and the European
ceptibility to multiple antibiotics for numerous Committee on Antimicrobial Susceptibility Test-
strains at the same time. Despite these ing (EUCAST) breakpoints for MTZ are not
advantages, the high cost hinders the extensive equivalent: the first is defined 32 mg/L, the
use of this method in clinical laboratories and second >2 mg/L (Clinical and Laboratory
epidemiological studies. Standards Institute 2015;
Detection of strains with reduced susceptibil- clinical_breakpoints/). Since methodological
ity to MTZ poses problems in choosing the more differences and different interpretation
proper antibiotic susceptibility method to test categories may cause discrepancies in results,
them. In fact, resistance to MTZ is often unstable influencing therapeutic decision and comparison
and laboratory manipulation of strains frequently of data, international committees are currently
146 P. Spigaglia et al.

co-operating with the intention of harmonizing antibiotics has been correlated with resistance
susceptibility testing and breakpoints for this genes and alteration in antibiotic targets, molec-
antibiotic. ular analysis may be considered to investigate
Disk diffusion testing is not recommended by C. difficile resistance beside phenotypic tests.
CLSI for C. difficile but some recent papers sug- The decreased cost of these technologies will
gest that it could be an option for antimicrobial allow their introduction on a large scale as tool
susceptibility testing of this pathogen. A study for infection control in the future, as suggested
carried out in Denmark on 211 isolates, showed by very recent studies that demonstrate the
that an excellent agreement was found between importance of molecular analysis and compara-
MIC results when the epsilometer test and disk tive genomics in the epidemiological surveil-
diffusion were used to test C. difficile strains lance of C. difficile (Ramı́rez-Vargas et al.
susceptibility to VAN, MXF, and MTZ 2017; Cairns et al. 2017).
(Erikstrup et al. 2012). Furthermore, two studies,
performed in Denmark and Brazil, respectively,
successfully used disk diffusion to test C. difficile 5 C. difficile Mechanisms
isolates with reduced susceptibility to MTZ and of Resistance
VAN (Holt et al. 2015; Fraga et al. 2016).
Despite these results, an exact zone diameter Several mechanisms responsible for antibiotic
for breakpoints is still not determined either by resistance have been identified in C. difficile,
CLSI or by EUCAST, therefore the debate about including chromosomal resistance genes,
disk diffusion, as qualified antibiotic susceptibly mobile genetic elements (MGEs), alterations in
testing method for C. difficile, is still open. the antibiotic targets and/or in metabolic
Although in some paper C. difficile MIC pathways, and biofilm formation (Table 4). Fur-
values have been obtained using broth thermore, recent evidences support that
microdilution (Genzel et al. 2014; Lim et al. C. difficile resistance to some antibiotics may
2016), CLSI recommends this method only to be complex and multifactorial.
test Bacteroides species (Clinical Laboratory
Standards Institute 2012). Furthermore, a recent
study of Hastey et al. has demonstrated a nega- 5.1 Antibiotics Associated to CDI
tive bias for the broth microdilution when com-
pared to the AD for C. difficile (Hastey et al. 5.1.1 Cephalosporins
2017). In this study, the MIC values obtained
using the broth microdilution were lower than C. difficile is usually resistant to CFs and several
those obtained with AD. Furthermore, the repro- studies report C. difficile overgrowth after CFs
ducibility with broth microdilution was variable, therapy (Ambrose et al. 1985; de Lalla et al.
probably dependent on the antibiotics tested. 1989; Impallomeni et al. 1995). Although
Therefore, in accordance with the CLSI guide- C. difficile is described as “constitutively resis-
line (Clinical and Laboratory Standards Institute tant” to CFs, the mechanism of resistance to
2012), the results indicate that the broth these antibiotics is still not completely
microdilution method is not equivalent with AD characterized. The variable MICs values
for C. difficile antimicrobial susceptibility observed for the different CFs suggest that resis-
testing. tance may be strain-dependent. Antibiotic-
The phenotypic tests are traditional methods degrading enzymes, β-lactamases, and modifica-
to evaluate antibiotic susceptibility of C. difficile tion of target sites, penicillin-binding proteins
but they need time (almost 1 week to get the (PBPs), are the mainly mechanisms involved in
results) and the isolation of C. difficile from resistance to these antibiotics. A number of cod-
patient stools. Since resistance to several ing DNA sequences (CDSs) potentially involved
Antibiotic Resistances of Clostridium difficile 147

Table 4 C. difficile antibiotic mechanisms of resistance

Mechanism of Genetic
Antibioticsa resistance element Target/protein/gene References
CFs Antibiotic Putative β-lactamases and Spigaglia (2016)
enzymatic PBPs (25 CDSs potentially
destruction; involved identified in
Altered target C. difficile 630)
MLSB 23 S RNA Tn5398 erm B Farrow et al. (2001), Brouwer et al.
methylases; RNA and (2011), Spigaglia et al. (2005, 2011)
methyl Tn5398
transferase -like
Tn6194 erm B Wasels et al. (2013), He et al. (2010,
Tn6215 erm B Goh et al. (2013), Wasels et al.
Tn6218 erm AB/cfr Dingle et al. (2014)
cfr B/cfr C Hansen and Vester (2015), Marin
et al. (2015), Candela et al. (2017)
FQs Altered target gyr A/gyr B Ackermann et al. (2001), Carman
et al. (2009), Dridi et al. (2002),
Walkty et al. (2010) Huang et al.
(2009), Mac Aogáin et al. (2015),
Spigaglia et al. (2008b, 2011),
Kuwata et al. (2015), Liao et al.
MTZ Metabolic Chong et al. (2014), Moura et al.
pathways (2014), Vuotto et al. (2016)
biofilm formation
VAN Altered target; mur G Leeds et al. (2014), Dapa et al. (2013)
biofilm formation
RFs Altered target rpo B Cairns et al. (2017), Carman et al.
(2009), Curry et al. (2009), Pecavar
et al. (2012), O’Connor et al. (2008),
Spigaglia et al. (2011), Huang et al.
(2009), Liao et al. (2012), Miller
et al. (2011b), Walkty et al. (2010)
TET Ribosomal Tn6397 tet M Roberts et al. (2001, 2011)
protection Tn916- tet M Sebaihia et al. (2006), Brouwer et al.
like (2011, 2012), Spigaglia et al. (2005,
Tn6164 tet 44 Corver et al. (2012)
CHL Chloramphenicol Tn4453a cat D Wren et al. (1988, 1989)
acetyltransferase and
CFs cephalosporins, MLSB Macrolide-lincosamide-streptograminB, FQs fluoroquinolones, MTZ metronidazole, VAN
vancomycin, RFs rifamycins, TET tetracycline, CHL chloramphenicol
148 P. Spigaglia et al.

has been identified in C. difficile 630 genome and conjugative transposon Tn6194 identified in
in other C. difficile strains (identity between 73% C. difficile 2007855 (He et al. 2010, 2013; Wasels
and 100%) (Spigaglia 2016). Anyway, further et al. 2013). Tn6194 has a conjugative region
genomic and functional analyses will be neces- related to that of Tn916, a large family
sary to elucidate the role of these potential beta- of conjugative elements widely spread in both
lactam interacting genes. Gram-positive and Gram-negative bacteria, and
an accessory region related to Tn5398, and it is
able to in vitro transfer from C. difficile to Entero-
5.1.2 Macrolide-Lincosamide-
coccus faecalis (Wasels et al. 2014).
StreptograminB (MLSB)
Tn6215 is a peculiar mobilisable transposon
of about 13 kb in length found in C. difficile
In C. difficile, resistance to the macrolide-
CD80 (Goh et al. 2013). Noteworthy,
lincosamide-streptogramin B (MLSB) family is
conjugation-like mechanism or phage ΦC2 trans-
usually conferred by ribosomal methylation.
duction can be involved in the transfer of this
Erythromycin ribosomal methylases (erm) genes
element between C. difficile strain to another.
of class B are the most widespread in C. difficile
Furthermore, it has recently been suggested that
population, even if other erm genes have rarely
a transformation-like mechanism can be respon-
been detected (Roberts et al. 1994; Spigaglia
sible for the transfer of Tn6215 and Tn5398
et al. 2005; Schmidt et al. 2007). In C. difficile,
when C. difficile CD13 is used as recipient strain
ermB is usually located on mobilisable genetic
(Wasels et al. 2015b).
elements and Tn5398, a mobilisable non
Although ermB-containing elements have a
conjugative element of 9.6 kb in length, is the
cost on the C. difficile fitness in vitro (Wasels
best known among these elements (Farrow et al.
et al. 2013), these elements are common in
2001). Tn5398 contains two copies of ermB gene
C. difficile population suggesting that, regardless
and it is able to transfer in vitro from C. difficile to
of the burden on fitness, other factors (i.e. the
Staphylococcus aureus and to Bacillus subtilis
capability of transfer and the intrinsic genetic
(Hächler et al. 1987; Mullany et al. 1995). Integra-
characteristics of strains) are involved in their
tion/excision functions to transfer Tn5398 from
successful spread.
the donor to the recipient strain are provided by
Resistance to both ERY and CLI or only to
other conjugative transposons present in the donor
ERY have been observed also in C. difficile
genome, because Tn5398 does not have genes
strains negative for erm genes (Spigaglia and
encoding a recombinase (Mullany et al. 2015).
Mastrantonio 2004; Pituch et al. 2006;
Integration into the recipient chromosome
Ratnayake et al. 2011; Spigaglia et al. 2011).
occurred either by homologous recombination or
Although alterations in the 23S rDNA or ribo-
by using a site-specific recombinase of the recipi-
somal proteins (L4 or L22) have been found in
ent. It is also possible, as recently suggested, that a
some of these strains, the same changes were also
portion of the donor genome containing Tn5398
observed in susceptible isolates and, therefore,
integrates by homologous recombination into the
their role in resistance has been excluded
recipient (Wasels et al. 2015b).
(Spigaglia et al. 2011). Furthermore, treatment
The majority of C. difficile strains resistant to
of resistant erm-negative strains with two pump
MLSB show ermB-containing elements with a dif-
inhibitors (reserpine and carbonyl cyanide
ferent genetic organizations compared to Tn5398
m-chlorophenyl hydrazone – CCCP), did not
(Farrow et al. 2001; Spigaglia et al. 2005, 2011).
determine any reduction in MICs, suggesting
Seventeen organizations (E1-E17) have been
that resistance is not mediated by efflux-
identified by a PCR-mapping method and the E4
mechanisms (Spigaglia et al. 2011). Recently,
was identified as the most frequent among
other determinants that could have a role in
European C. difficile clinical isolates (Spigaglia
C. difficile resistance to MLSB in the absence of
et al. 2011). Elements E4 are related to the
erm genes have been identified. In particular,
Antibiotic Resistances of Clostridium difficile 149

cfrB or cfrC, which encode a 23S rRNA and Bacteroides fragilis, resistance to MTZ is
methyltransferase and confer resistance to usually conferred by nitroimidazole (nim) genes
PhLOPSA (phenicols, lincosamides, (Gal and Brazier 2004), but these genes have not
oxazolidinones, pleuromutilins, and been identified in C. difficile (Moura et al. 2014).
streptogramin A), have been found in several Although it is not completely understood, data
C. difficile strains resistant to linezolid and obtained in recent studies on strains RT 027 and
other clinically relevant antibiotics (Hansen and RT 010 suggest that C. difficile resistance to
Vester 2015; Marin et al. 2015; Candela et al. MTZ is a multifactorial process that involves
2017). A cfr gene has been identified in a alterations in metabolic pathways, such as activ-
non-conjugative element, denominated Tn6218, ity of nitroreductases, iron uptake and DNA
which is related to Tn916 (Dingle et al. 2014). repair (Chong et al. 2014; Moura et al. 2014).
In addition, biofilm formation seems to play a
5.1.3 Fluoroquinolones role in C. difficile MTZ-resistance (Vuotto et al.
2016). How biofilm growth could contribute to
Alterations in the quinolone-resistance determin- increase C. difficile resistance to MTZ is still
ing region (QRDR) of GyrA and/or GyrB are unclear. However, it can be hypothesized that
responsible for resistance to FQs in C. difficile biofilm matrix can act as a protective barrier,
(Ackermann et al. 2001, 2003; Dridi et al. 2002; inducing, at the same time, an alteration of the
Drudy et al. 2006, 2007). Several amino acidic physiological state of the bacteria within the
substitutions have been identified in the DNA biofilm that determines a higher level of resis-
gyrase subunits (Table 5), but the most common tance to antibiotics.
in C. difficile FQs-resistant strains is the substitu-
tion Thr82Ile in GyrA (Ackermann et al. 2001; 5.2.2 Vancomycin
Dridi et al. 2002; Spigaglia et al. 2008b, 2011;
Kuwata et al. 2015). Interestingly, Thr82Ile in Vancomycin is the first-line antibiotic for mod-
GyrA has not a detectable cost on the fitness of erate to severe CDI (Debast et al. 2014; Jarrad
C. difficile in vitro, suggesting that this substitu- et al. 2015). This antibiotic, which consists of a
tion can be maintained in the bacterial population glycosylated hexapeptide chain and cross linked
even in the absence of antibiotic selective pres- aromatic rings by aryl ether bonds, inhibits the
sure (Wasels et al. 2015a). biosynthesis of peptidoglycan, an essential com-
Resistant mutants to FQs can be obtained with ponent of the bacterial cell wall envelope, and it
high frequency after exposure of C. difficile sus- is poorly absorbed by the gastrointestinal tract
ceptible strains to MXF and levofloxacin (LVX) (Perkins and Nieto 1974; Yu and Sun 2013). The
(Spigaglia et al. 2009). Since the concentration of mechanism of resistance in C. difficile is still
this drug in the human intestine, during the early unclear. Although Tn1549-like elements have
stage of treatment, is not inhibitory, it is possible been found in several strains (Brouwer et al.
for a sub-population of bacteria to acquire 2011, 2012), these elements, differently from
mutations conferring resistance to FQs. the original Tn1549 element described in
E. faecalis, do not have a functional vanB
operon. Interestingly, a vanG-like gene cluster
5.2 Antibiotics for CDI Treatment homologous to that found in E. faecalis have
also been described in C. difficile but it seems
5.2.1 Metronidazole not able to promote resistance to VAN (Ammam
et al. 2012, 2013; Ramı́rez-Vargas et al. 2017).
Metronidazole is a nitro-aromatic pro-drug that Recently, VAN-resistant mutants, showing the
need the reduction of the 5-nitro group of the amino acid change Pro108Leu in the MurG,
imidazole ring to become cytotoxic to bacterial have been obtained in vitro (Leeds et al. 2014).
cells (Goldman 1982). In Helicobacter pylori Since MurG is involved in the membrane-bound
150 P. Spigaglia et al.

Table 5 Amino acid substitutions detected in C. difficile isolates resistant to fluoroquinolones or rifamycins
acid Original Resistance
Antibiotica Target position residue substitution References
FQs GyrA 43 Val Asp Carman et al. (2009)
71 Asp Val Dridi et al. (2002), Walkty et al. (2010), Liao et al.
81 Asp Asn Huang et al. (2009), Liao et al. (2012)
82 Thr Ile or Val Ackermann et al. (2001), Dridi et al. (2002), Spigaglia
et al. (2008b), Kuwata et al. (2015), Liao et al. (2012)
118 Ala Thr Dridi et al. (2002)
384 Ala Asp Mac Aogáin et al. (2015)
GyrB 377 Arg Gly Liao et al. (2012)
416 Ser Ala Liao et al. (2012)
426 Asp Asn or Val Dridi et al. (2002), Spigaglia et al. (2008b), Liao et al.
447 Arg Lys Walkty et al. (2010), Liao et al. (2012)
466 Glu Val Liao et al. (2012)
GyrA/ 82/366 Thr/Ser Ile/Ala Huang et al. (2009), Kuwata et al. (2015)
GyrB 82/366 Thr/Ser Ile/Ala and Walkty et al. (2010), Kuwata et al. (2015)
and 426 and Asp Val
82/366 Thr/Ser Ala/Ala and Kuwata et al. (2015)
and 434 and Gln Lys
82/416 Thr/Ser Ile/Ala Spigaglia et al. (2008b), Liao et al. (2012)
82/426 Thr/Asp Ile/Asn Walkty et al. (2010), Kuwata et al. (2015)
82/426 Thr/Asp Ile/Val Spigaglia et al. (2011)
82/426 Thr/Asp Val/Val Huang et al. (2009), Liao et al. (2012)
82/444 Thr/Leu Ile/Phe Walkty et al. (2010)
RFs RpoB 485 Ser Phe Cairns et al. (2017)
492 Asp Asn or Val Pecavar et al. (2012)
502 His Arg or Asn O’Connor et al. (2008), Pecavar et al. (2012), Miller
or Leu or et al. (2011b)
505 Arg Lys O’Connor et al. (2008), Curry et al. (2009), Miller et al.
(2011b), Spigaglia et al. (2011), Pecavar et al. (2012)
550 Ser Phe or Tyr Pecavar et al. (2012)
448; 505 Ser; Arg Thr; Lys O’Connor et al. (2008), Curry et al. (2009)
487; 502 Leu; His Phe; Tyr Pecavar et al. (2012)
492; 505 Asp; Asn; Lys O’Connor et al. (2008)
498; 505 Ser; Arg Thr; Lys Curry et al. (2009), Miller et al. (2011b)
502; 496 His; Pro Tyr; Ser Carman et al. (2009)
502; 505 His; Arg Asn; Lys O’Connor et al. (2008), Curry et al. (2009), Miller et al.
(2011b), Spigaglia et al. (2011), Pecavar et al. (2012)
505; 548 Arg; Ile Lys; Met O’Connor et al. (2008), Curry et al. (2009), Pecavar
et al. (2012)
FQs Fluoroquinolones, RFs rifamycins
Antibiotic Resistances of Clostridium difficile 151

stage of peptidoglycan biosynthesis, this substi- et al. 2012), with a high local concentration in the
tution may affect VAN activity. In addition, bio- gut and feces (1225.1 μg/g after 10 days of ther-
film formation has been found to probably have a apy) (Goldstein et al. 2012; Sears et al. 2012).
role in VAN-resistance. In fact, C. difficile within Reduced susceptibility to FDX is very rare and
biofilms resulted more resistant to high only one C. difficile clinical isolate with a
concentrations of VAN (20 mg/L) and MIC ¼ 16 mg/L has been described (Goldstein
sub-inhibitory and inhibitory concentrations of et al. 2011). Mutations in rpoB or CD22120,
the antibiotic seems to induce biofilm formation encoding for a homologue to the multidrug
(Dapa et al. 2013). resistance-associated transcriptional regulator
MarR, have been observed in C. difficile mutants
5.2.3 Rifamycins resistant to FDX obtained in vitro (Leeds et al.
2014). Since mutations causing resistance to
Treatment failures and recurrence of infection FDX arise in rpoB gene at distinct loci compared
rates associated with MTZ and VAN treatments to those causing resistance to RFs, FDX retains
have increased in the last years (Vardakas et al. activity against strains resistant to RFs (Anti-
2012) therefore other therapy options for CDI Infective Drugs Advisory Committee Briefing
have been proposed. Document, Optimer Pharmaceuticals, Inc.).
RFs, in particular RFX, have recently been
proposed as “chaser therapy” for treatment of
relapsing CDI (Iv et al. 2014), while fidaxomicin
5.3 Other Antibiotics
(FDX) is a bactericidal new narrow spectrum
macrocyclic antibiotic that is used for the man-
5.3.1 Tetracycline
agement of CDI with high risk for recurrences
(Chaparro-Rojas and Mullane 2013). Both RFs
In C. difficile, resistance to TET is due by tet
and FDX are inhibitors of bacterial transcription
genes (Table 4). The most widespread tet class
but they have different RNA polymerase
is tetM, usually carried by conjugative Tn916-
(RNAP) target sites. FDX binds to the ‘switch
like elements (Spigaglia et al. 2005; Mullany
region’ of RNAP, a target site that is adjacent to
et al. 2012; Dong et al. 2014). This family of
the RIF target but does not overlap (Mullane and
transposon is responsible for the spread of anti-
Gorbach 2011; Srivastava et al. 2011).
biotic resistance (usually referred to TET but also
Different amino acid substitutions have been
to MLSB and other antibiotics) to many impor-
identified within the β-subunit of the RNA poly-
tant pathogens (Roberts and Mullany 2011). The
merase (rpoB) of strains resistant to RFs
best-known C. difficile element of this family is
(Table 5). Among the amino acid substitutions
Tn5397, which is a 21 kb element able to transfer
identified, Arg505Lys is the most common, par-
in vitro between C. difficile and B. subtilis or
ticularly in strains RT027 (Miller et al. 2011b;
E. faecalis (Mullany et al. 1990; Jasni et al.
Spigaglia et al. 2011; Carman et al. 2012;
2010). A group II intron and a different exci-
Pecavar et al. 2012).
sion/insertion module differentiate Tn5397 from
Tn916. In fact, Tn5397 has a tndX gene that
5.2.4 Fidaxomicin
encodes a large serine recombinase, while
Tn916 contains two genes, xisTn and intTn,
This antibiotic provides cure rates not inferior to
encoding an excisionase and a tyrosine integrase
VAN and is associated with a significantly lower
(Roberts et al. 2001). Furthermore, Tn916 inserts
rate of CDI recurrence caused by strains
into multiple regions of the C. difficile genome
non-RT 027 (Louie et al. 2011). Furthermore, it
(Mullany et al. 2012), while Tn5397 inserts DNA
has a minimal impact on the composition of
predicted filamentation processes induced by
indigenous fecal microbiota, in particular on
cAMP (Fic) domain (Wang et al. 2006).
Bacteroides species (Tannock et al. 2010; Louie
152 P. Spigaglia et al.

Different genetic organizations of Tn916-like instead of a catD gene it shows genes predicted
elements and different tetM alleles have been to encode for transcriptional regulator, a two
identified in C. difficile (Spigaglia et al. 2005, component regulatory system, an ABC trans-
2006). In particular, the Tn916-element detected porter, three sigma factors and a putative toxin-
in the clinical isolate CD1911 contains both tetM antitoxin system. The role of these genes is not
and ermB, (Spigaglia et al. 2007). This element is clear and remains to be determined.
non-conjugative and probably originated from
the combination of one or more plasmids and a
Tn916-like element.
6 Conclusions
Albeit more rarely, other tet genes have been
identified in C. difficile. In particular, the
C. difficile infection (CDI) is a growing concern
co-presence of both tetM and tetW have been
for global public health. An increased CDI inci-
described in C. difficile isolates from humans and
dence, morbidity and mortality have been
animals (Spigaglia et al. 2008a; Fry et al. 2012).
reported in the last decades in association with
Interestingly, an element of 106 kb, the
the emergence and spread of C. difficile highly
Tn6164, has been identified in C. difficile strain
virulent types. C. difficile adaptive capability
M120, a RT 078 isolate (Corver et al. 2012). This
and genome plasticity has determined an
transposon is composed by parts of other
increase of strains resistant to multiple
elements from different bacteria, particularly
antibiotics and, currently, most of epidemic
from Thermoanaerobacter sp. and Streptococcus
clinical isolates are MDR. A wide range of
pneumoniae and it contains tet(44) and ant(6)-Ib,
mobile elements and alterations of antibiotic
predicted to confer resistance to TET and strep-
targets mediate resistance to several antibiotics,
tomycin, respectively. Since strain M120 is sus-
including the MLSB family and FQs, which are
ceptible to these antibiotics, Tn6164 does not
significantly associated to CDI. Furthermore, a
seem involved in resistance, but it seems to be
decreased susceptibility to the first-line
associated to higher virulence of strains RT
antibiotics used for CDI therapy, in particular
078, in fact an analysis of data from patients
MTZ and VAN, and to those used for
indicate that mortality was more common in
recurrences, such as RFs, may have a role in
patients infected with strains RT 078 containing
the low rate of response to treatment reported
Tn6164 compared with those infected with
over the last years. Antibiotics resistances
strains without this element.
seem to be maintained in this pathogen regard-
less of the burden imposed by the acquisition of
5.3.2 Chloramphenicol genetic elements/mutations conferring
resistance and the decrease of antibiotics pres-
C. difficile resistance to CHL is usually conferred sure. This feature may explain the persistence
by a CHL acetyltransferase encoded by a catD of “old” resistances and the rapid diffusion of
gene (Wren et al. 1988, 1989) (Table 4). In “new” resistances in C. difficile population. The
C. difficile, the catD gene is located on the multifactorial nature of antibiotic resistances
transposons Tn4453a and Tn4453b, which are and the rapid evolution of C. difficile epidemi-
strictly related to the Clostridium perfringens ology, emphasizing the need for effective anti-
mobilisable element Tn4451 (Lyras et al. 1998). microbial stewardships, implementation of
Recently, a conjugative transposon designed infection control programs, and development
Tn6104, has been described (Brouwer et al. of alternative therapies to prevent and contain
2011). This transposon contains genetic elements the spread of resistant strains and to ensure an
closely related to Tn4453ab and Tn4451 but efficacious therapy for CDI.
Antibiotic Resistances of Clostridium difficile 153

References Brouwer MSM, Roberts AP, Mullany P et al (2012) In

silico analysis of sequenced strains of Clostridium
difficile reveals a related set of conjugative
Ackermann G, Tang YJ, Kueper R et al (2001) Resistance
transposons carrying a variety of accessory genes.
to moxifloxacin in toxigenic Clostridium difficile
Mob Genet Elem 2:8–12
isolates is associated with mutations in gyr
Büchler AC, Rampini SK, Stelling S et al (2014) Antibi-
A. Antimicrob Agents Chemother 45:2348–2353
otic susceptibility of Clostridium difficile is similar
Ackermann G, Tang-Feldman YJ, Schaumann R et al
worldwide over two decades despite widespread use
(2003) Antecedent use of fluoroquinolones is
of broad-spectrum antibiotics: an analysis done at the
associated with resistance to moxifloxacin in Clostrid-
University Hospital of Zurich. BMC Infect Dis 14:607
ium difficile. Clin Microbiol Infect 9:526–530
Burckhardt F, Friedrich A, Beier D et al (2008) Clostrid-
Adler A, Miller-Roll T, Bradenstein R, Block C et al
ium difficile surveillance trends, Saxony, Germany.
(2015) A national survey of the molecular epidemiol-
Emerg Infect Dis 4:691–692
ogy of Clostridium difficile in Israel: the dissemination
Cairns MD, Preston MD, Hall CL et al (2017) Compara-
of the ribotype 027 strain with reduced susceptibility
tive genome analysis and global phylogeny of the
to vancomycin and metronidazole. Diagn Microbiol
toxin variant Clostridium difficile PCR ribotype
Infect Dis 83:21–24
017 reveals the evolution of two independent
Álvarez-Pérez S, Blanco JL, Harmanus C et al (2017)
sublineages. J Clin Microbiol 55:865–876
Subtyping and antimicrobial susceptibility of Clostrid-
Candela T, Marvaud J-C, Nguyen TK et al (2017) A
ium difficile PCR ribotype 078/126 isolates of human
cfr-like Gene cfr(C) conferring linezolid resistance is
and animal origin. Vet Microbiol 199:15–22
common in Clostridium difficile. Int J Antimicrob
Ambrose NS, Johnson M, Burdon D et al (1985) The
influence of single dose intravenous antibiotics on
faecal flora and emergence of Clostridium difficile. J
Carman RJ, Genheimer CW, Rafii F et al (2009) Diversity
Antimicrob Chemother 15:319–326
of moxifloxacin resistance during a nosocomial out-
Ammam F, Marvaud JC, Lambert T (2012) Distribution
break of a predominantly ribotype ARU 027 Clostrid-
of the vanG-like gene cluster in Clostridium difficile
ium difficile diarrhea. Anaerobe 15:244–248
clinical isolates. Can J Microbiol 58:547–551
Carman RJ, Boone JH, Grover H et al (2012) In vivo
Ammam F, Meziane-Cherif D, Mengin-Lecreulx D et al
selection of rifamycin-resistant Clostridium difficile
(2013) The functional vanGCd cluster of Clostridium
during rifaximin therapy. Antimicrob Agents
difficile does not confer vancomycin resistance. Mol
Chemother 56:6019–6020
Microbiol 89:612–625
Chaparro-Rojas F, Mullane KM (2013) Emerging
Baines SD, O’Connor R, Freeman J et al (2008) Emer-
therapies for Clostridium difficile infection – focus
gence of reduced susceptibility to metronidazole in
on fidaxomicin. Infect Drug Resist 6:41–53
Clostridium difficile. J Antimicrob Chemother
Chia JH, Lai HC, Su LH et al (2013) Molecular epidemi-
ology of Clostridium difficile at a medical center in
Baldan R, Trovato A, Bianchini V et al (2015) A success-
Taiwan: persistence of genetically clustering of AB
ful epidemic genotype: Clostridium difficile PCR
+ isolates and increase of A+B+ isolates. PLoS One 8:
ribotype 018. J Clin Microbiol 53:2575–2580
Bartlett JG, Onderdonk AB, Cisneros RL et al (1977)
Chong PM, Lynch T, McCorrister S et al (2014)
Clindamycin-associated colitis due to a toxin-
Proteomic analysis of a NAP1 Clostridium difficile
producing species of Clostridium in hamsters. J Infect
clinical isolate resistant to metronidazole. PLoS One
Dis 136:701–705
Bauer MP, Notermans DW, van Benthem BHB et al
Clements AC, Magalhães RJ, Tatem AJ et al (2010)
(2011) Clostridium difficile infection in Europe: a
Clostridium difficile PCR ribotype 027: assessing the
hospital-based survey. Lancet 377:63–73
risks of further worldwide spread. Lancet Infect Dis
Bignardi GE (1998) Risk factors for Clostridium difficile
infection. J Hosp Infect 40:1–15
Clinical and Laboratory Standards Institute (CLSI) (2012)
Bolton RP, Culshaw MA (1986) Faecal metronidazole
Methods for antimicrobial susceptibility testing of
concentrations during oral and intravenous therapy
anaerobic bacteria. Approved standard-eighth edn.
for antibiotic associated colitis due to Clostridium
CLSI document M11-A8. ISBN 1-56238-789-
difficile. Gut 27:1169–1172
8 (Print); ISBN 1-56238-790-1 (Electronic)
Brazier JS, Fawley W, Freeman J et al (2001) Reduced
Clinical and Laboratory Standards Institute (CLSI) (2015)
susceptibility of Clostridium difficile to metronida-
Performance standards for antimicrobial susceptibility
zole. J Antimicrob Chemother 48:741–742
testing. Twenty-fifth informational supplement. CLSI
Brouwer MSM, Warburton PJ, Roberts AP et al (2011)
document M100-S25. ISBN 1-56238-989-0 (Print);
Genetic organisation, mobility and predicted functions
ISBN 1-56238-990-4 (Electronic)
of genes on integrated, mobile genetic elements in
Corver J, Bakker D, Brouwer MSM et al (2012) Analysis
sequenced strains of Clostridium difficile. PLoS One
of a Clostridium difficile PCR ribotype 078 100
154 P. Spigaglia et al.

kilobase island reveals the presence of a novel trans- Erikstrup LT, Danielsen TK, Hall V et al (2012) Antimi-
poson, Tn6164. BMC Microbiol 12:130 crobial susceptibility testing of Clostridium difficile
Curry SR, Marsh JW, Shutt KA et al (2009) High fre- using EUCAST epidemiological cut-off values and
quency of rifampin resistance identified in an epi- disk diffusion correlates. Clin Microbiol Infect 18:
demic Clostridium difficile clone from a large E266–E272
teaching hospital. Clin Infect Dis 48:425–429 European Centre for Disease Prevention and Control
Dapa T, Leuzzi R, Ng YK et al (2013) Multiple factors (ECDC) (2013) Point prevalence survey of health
modulate biofilm formation by the anaerobic pathogen care associated infections and antimicrobial use in
Clostridium difficile. J Bacteriol 195:545–555 European acute care hospitals
Davies KA, Longshaw CM, Davis GL et al (2014) Under- Falagas ME, Makris GC, Dimopoulos G et al (2008)
diagnosis of Clostridium difficile across Europe: the Heteroresistance: a concern of increasing clinical sig-
European, multicentre, prospective, biannual, point- nificance? Clin Microbiol Infect 14:101–104
prevalence study of Clostridium difficile infection in Farrow KA, Lyras D, Rood JI (2001) Genomic analysis of
hospitalised patients with diarrhoea (EUCLID). Lan- the erythromycin resistance element Tn5398 from
cet Infect Dis 14:1208–1219 Clostridium difficile. Microbiology 147:2717–2728
de Lalla F, Privitera G, Ortisi G et al (1989) Third gener- Fraga EG, Nicodemo AC, Sampaio JL (2016) Antimicro-
ation cephalosporins as a risk factor for Clostridium bial susceptibility of Brazilian Clostridium difficile
difficile-associated disease: a four-year survey in a strains determined by agar dilution and disk. Braz J
general hospital. J Antimicrob Chemother 23:623–631 Infect Dis 20:476–481
Debast SB, Bauer MP, Kuijper EJ (2014) European Soci- Freeman J, Stott J, Baines SD et al (2005) Surveillance for
ety of Clinical Microbiology and Infectious Diseases: resistance to metronidazole and vancomycin in geno-
update of the treatment guidance document for Clos- typically distinct and UK epidemic Clostridium diffi-
tridium difficile infection. Clin Microbiol Infect cile isolates in a large teaching hospital. J Antimicrob
20:1–26 Chemother 56:988–989
Dingle KE, Elliott B, Robinson E et al (2014) Evolution- Freeman J, Vernon J, Morris K et al (2015a)
ary history of the Clostridium difficile pathogenicity Pan-European longitudinal surveillance of antibiotic
locus. Genome Biol Evol 6:36–52 resistance among prevalent Clostridium difficile
Dong D, Zhang L, Chen X et al (2013) Antimicrobial ribotypes. Clin Microbiol Infect 21:248.e9–248.e16
susceptibility and resistance mechanisms of clinical Freeman J, Vernon J, Vickers R et al (2015b) Susceptibil-
Clostridium difficile from a Chinese tertiary hospital. ity of Clostridium difficile isolates of varying antimi-
Int J Antimicrob Agents 41:80–84 crobial resistance phenotypes to SMT19969 and
Dong D, Chen X, Jiang C et al (2014) Genetic analysis of 11 comparators. Antimicrob Agents Chemother
Tn916-like elements conferring tetracycline resistance 60:689–692
in clinical isolates of Clostridium difficile. Int J of Fry PR, Thakur S, Abley M et al (2012) Antimicrobial
Antimicrob Agents 43:73–77 resistance, toxinotype, and genotypic profiling of
Dridi L, Tankovic J, Burghoffer B et al (2002) Gyr A and Clostridium difficile isolates of swine origin. J Clin
gyrB mutations are implicated in cross-resistance to Microbiol 50:2366–2372
ciprofloxacin and moxifloxacin in Clostridium diffi- Gal M, Brazier JS (2004) Metronidazole resistance in
cile. Antimicrob Agents Chemother 46:3418–3421 Bacteroides spp. carrying nim genes and the selection
Drudy D, Quinn T, O’Mahony R et al (2006) High-level of slow-growing metronidazole-resistant mutants. J
resistance to moxifloxacin and gatifloxacin associated Antimicrob Chemother 54:109–116
with a novel mutation in gyrB in toxin-A-negative, Gao Q, Wu S, Huang H, Ni Y et al (2016) Toxin profiles,
toxin-B-positive Clostridium difficile. J Antimicrob PCR ribotypes and resistance patterns of Clostridium
Chemother 58:1264–1267 difficile: a multicentre study in China, 2012–2013. Int
Drudy D, Kyne L, O’Mahony R et al (2007) gyrA J Antimicrob Agents 48:736–739
mutations in fluoroquinolone-resistant Clostridium Genzel GH, Stubbings W, Stingu CS et al (2014) Activity
difficile PCR-027. Emerg Infect Dis 13:504–505 of the investigational fluoroquinolone finafloxacin and
Dubberke ER, Olsen MA (2012) Burden of Clostridium seven other antimicrobial agents against 114 obligately
difficile on the healthcare system. Clin Infect Dis 55 anaerobic bacteria. Int J Antimicrob Agents
(suppl 2):S88–S92 44:420–423
Eckert C, Coignard B, Hebert M et al (2013) Clinical and Goh S, Hussain H, Chang BJ et al (2013) Phage ϕC2
microbiological features of Clostridium difficile mediates transduction of Tn6215, encoding erythro-
infections in France: the ICD-RAISIN 2009 national mycin resistance, between Clostridium difficile
survey. Méd Mal Infect 43:67–74 strains. MBio 4:e00840–e00813
Eitel Z, Terhes G, Sóki J et al (2015) Investigation of the Goldman P (1982) The development of 5-nitroimidazoles
MICs of fidaxomicin and other antibiotics against for the treatment and prophylaxis of anaerobic bacte-
Hungarian Clostridium difficile isolates. Anaerobe rial infections. J Antimicrob Chemother 10(Suppl.
31:47–49 A):23–33
Antibiotic Resistances of Clostridium difficile 155

Goldstein EJ, Citron DM, Sears P et al (2011) Compara- Iv ECO, Iii ECO, Johnson DA (2014) Clinical update for
tive susceptibilities of fidaxomicin (OPT-80) of the diagnosis and treatment of Clostridium difficile
isolates collected at baseline, recurrence, and failure infection. World J Gastrointest Pharmacol Ther
from patients in two fidaxomicin phase III trials of 5:1–26
C. difficile infection. Antimicrob Agents Chemother Jamal WY, Rotimi VO (2016) Surveillance of antibiotic
55:5194–5199 resistance among hospital- and community-acquired
Goldstein EJ, Babakhani F, Citron DM (2012) Antimicro- toxigenic Clostridium difficile isolates over 5-year
bial activities of fidaxomicin. Clin Infect Dis 55 period in Kuwait. PLoS One 11:e0161411
(Suppl. 2):S143–S148 Jarrad AM, Karoli T, Blaskovich MAT et al (2015) Clos-
Goorhuis A, Van der Kooi T, Vaessen N et al (2007) tridium difficile drug pipeline: challenges in discovery
Spread and epidemiology of Clostridium difficile and development of new agents. J Med Chem
polymerase chain reaction ribotype 027/toxinotype 58:5164–5185
III in The Netherlands. Clin Infect Dis 45:695–703 Jasni AS, Mullany P, Hussain H et al (2010) Demonstra-
Goudarzi M, Goudarzi H, Alebouyeh M et al (2013) tion of conjugative transposon (Tn5397)-mediated
Antimicrobial susceptibility of Clostridium difficile horizontal gene transfer between Clostridium difficile
clinical isolates in Iran. Iran Red Crescent Med J and Enterococcus faecalis. Antimicrob Agents
15:704–711 Chemother 54:4924–4926
Gravel D, Miller M, Simor A, Taylor G et al (2009) Johnson S, Schriever C, Patel U et al (2009) Rifaximin
Canadian Nosocomial Infection Surveillance Pro- redux: treatment of recurrent Clostridium difficile
gram. Health care-associated Clostridium difficile infections with rifaximin immediately post-
infection in adults admitted to acute care hospitals in vancomycin treatment. Anaerobe 15:290–291
Canada: a Canadian Nosocomial Infection Surveil- Karlowsky JA, Zhanel GG, Hammond GW et al (2012)
lance Program study. Clin Infect Dis 48:568–576 Multidrug-resistant North American pulsotype 2 Clos-
Hächler H, Berger-Bächi B, Kayser FH (1987) Genetic tridium difficile was the predominant toxigenic
characterization of a Clostridium difficile hospital-acquired strain in the province of Manitoba,
erythromycin-clindamycin resistance determinant Canada, in 2006–2007. J Med Microbiol 61:693–700
that is transferable to Staphylococcus aureus. Khan R, Cheesbrough J (2003) Impact of changes in
Antimicrob Agents Chemother 7:1039–1045 antibiotic policy on Clostridium difficile-associated
Hansen LH, Vester B (2015) A cfr-like gene from Clos- diarrhoea (CDAD) over a five-year period in a district
tridium difficile confers multiple antibiotic resistance general hospital. J Hosp Infect 54:104–108
by the same mechanism as the cfr gene. Antimicrob Kim J, Kang JO, Pai H et al (2012) Association between
Agents Chemother 59:5841–5843 PCR ribotypes and antimicrobial susceptibility among
Hastey CJ, Dale SE, Nary J et al (2017) Comparison of Clostridium difficile isolates from healthcare-
Clostridium difficile minimum inhibitory associated infections in South Korea. Int J Antimicrob
concentrations obtained using agar dilution vs broth Agents 40:24–29
microdilution methods. Anaerobe 44:73–77 Knight DR, Riley TV (2016) Clostridium difficile clade
He M, Sebaihia M, Lawley TD et al (2010) Evolutionary 5 in Australia: antimicrobial susceptibility profiling of
dynamics of Clostridium difficile over short and long PCR ribotypes of human and animal origin. J
time scales. PNAS 107:7527–7532 Antimicrob Chemother 71:2213–2217
He M, Miyajima F, Roberts P, Ellison L et al (2013) Knight DR, Giglio S, Huntington PG et al (2015) Surveil-
Emergence and global spread of epidemic lance for antimicrobial resistance in Australian
healthcare-associated Clostridium difficile. Nat Genet isolates of Clostridium difficile, 2013–2014. J
45:109–113 Antimicrob Chemother 70:2992–2999
Holt HM, Danielsen TK, Justesen US (2015) Routine disc Kociolek LK, Gerding DN, Osmolski JR et al (2016)
diffusion antimicrobial susceptibility testing of Clos- Differences in the molecular epidemiology and antibi-
tridium difficile and association with PCR ribotype otic susceptibility of Clostridium difficile isolates in
027. Eur J Clin Microbiol Infect Dis 34:2243–2246 pediatric and adult patients. Antimicrob Agents
Huang H, Weintraub A, Fang H et al (2009) Antimicro- Chemother 60:4896–4900
bial resistance in Clostridium difficile. Int J Kouzegaran S, Ganjifard M, Tanha AS (2016) Detection,
Antimicrob Agents 34:516–522 ribotyping and antimicrobial resistance properties of
Huang JS, Jiang Z-D, Garey KW et al (2013) Use of Clostridium difficile strains isolated from the cases of
rifamycin drugs and development of infection by diarrhea. Mater Sociomed 28:324–328
rifamycin-resistant strains of Clostridium difficile. Krutova M, Matejkova J, Tkadlec J et al (2015) Antibiotic
Antimicrob Agents Chemother 57:2690–2693 profiling of Clostridium difficile ribotype 176 – a
Impallomeni M, Galletly NP, Wort J et al (1995) multidrug resistant relative to C. difficile ribotype
Increased risk of diarrhoea caused by Clostridium 027. Anaerobe 36:88–90
difficile in elderly patients receiving cefotaxime. Kullin B, Brock T, Rajabally N et al (2016) Character-
BMJ 311:1345–1346 izations of Clostridium difficile strains isolated from
156 P. Spigaglia et al.

Groote Schuur Hospital, Cape Town, South Africa. Lyras D, Cooper MA (2015) Clostridium difficile drug
Eur J Clin Microbiol Infect Dis 35:1709–1718 pipeline: challenges in discovery and development of
Kullin B, Wojno J, Abratt V et al (2017) Toxin new agents. J Med Chem 58:5164–5185
A-negative toxin B-positive ribotype 017 Clostridium Lyras D, Storie C, Huggins AS et al (1998) Chloramphen-
difficile is the dominant strain type in patients with icol resistance in Clostridium difficile is encoded on
diarrhoea attending tuberculosis hospitals in Cape Tn4453 transposons that are closely related to Tn4451
Town, South Africa. Eur J Clin Microbiol Infect Dis from Clostridium perfringens. Antimicrob Agents
36:163–175 Chemother 42:1563–1156
Kuwata Y, Tanimoto S, Sawabe E et al (2015) Molecular Mac Aogáin M, Kilkenny S, Walsh C et al (2015) Identi-
epidemiology and antimicrobial susceptibility of fication of a novel mutation at the primary dimer
Clostridium difficile isolated from a university teach- interface of GyrA conferring fluoroquinolone resis-
ing hospital in Japan. Eur J Clin Microbiol Infect tance in Clostridium difficile. J Glob Antimicrob
34:763–772 Resist 3:295–299
Lachowicz D, Pituch H, Obuch-Woszczatyński P (2015) Mackin KE, Elliott B, Kotsanas D et al (2015) Molecular
Antimicrobial susceptibility patterns of Clostridium characterization and antimicrobial susceptibilities of
difficile strains belonging to different polymerase Clostridium difficile clinical isolates from Victoria,
chain reaction ribotypes isolated in Poland in 2012. Australia. Anaerobe 34:80–83
Anaerobe 31:37–41 Marin M, Martin A, Alcala L et al (2015) Clostridium
Lee J-H, Lee Y, Lee K et al (2014) The changes of PCR difficile isolates with high linezolid MICs harbor the
ribotype and antimicrobial resistance of Clostridium multiresistance gene cfr. Antimicrob Agents
difficile in a tertiary care hospital over 10 years. J Med Chemother 59:586–589
Microbiol 63:819–823 McDonald LC, Killgore GE, Thompson A et al (2005) An
Leeds JA, Sachdeva M, Mullin S et al (2014) In vitro epidemic, toxin gene–variant strain of Clostridium
selection, via serial passage, of Clostridium difficile difficile. N Engl J Med 353:2433–2441
mutants with reduced susceptibility to fidaxomicin or Miller BA, Chen LF, Sexton DJ et al (2011a) Comparison
vancomycin. J Antimicrob Chemother 69:41–44 of the burdens of hospital-onset, healthcare facility-
Lessa FC, Gould CV, McDonald LC (2012) Current status associated Clostridium difficile infection and of
of Clostridium difficile infection epidemiology. Clin healthcare-associated infection due to methicillin-
Infect Dis 55:65–70 resistant Staphylococcus aureus in community
Lessa FC, Mu Y, Bamberg WM et al (2015) Burden of hospitals. Infect Control Hosp Epidemiol 32:387–390
Clostridium difficile infection in the United States. N Miller MA, Blanchette R, Spigaglia P et al (2011b) Diver-
Engl J Med 372:825–834 gent rifamycin susceptibilities of Clostridium difficile
Liao CH, Ko WC, Lu JJ et al (2012) Characterizations of strains in Canada and Italy and predictive accuracy of
clinical isolates of Clostridium difficile by toxin rifampin Etest for rifamycin resistance. J Clin
genotypes and by susceptibility to 12 antimicrobial Microbiol 49:4319–4321
agents, including fidaxomicin (OPT-80) and Miller-Roll T, Na’amnih W, Cohen D et al (2016) Molec-
rifaximin: a multicenter study in Taiwan. Antimicrob ular types and antimicrobial susceptibility patterns of
Agents Chemother 56:3943–3949 Clostridium difficile isolates in different epidemiolog-
Lim SC, Foster NF, Riley TV (2016) Susceptibility of ical settings in a tertiary care center in Israel. Diagn
Clostridium difficile to the food preservatives sodium Microbiol Infect Dis 86:450–454
nitrite, sodium nitrate and sodium metabisulphite. Moura I, Spigaglia P, Barbanti F et al (2013) Analysis of
Anaerobe 37:67–71 metronidazole susceptibility in different Clostridium
López-Ureña D, Quesada-Gómez C, Montoya-Ramı́rez difficile PCR ribotypes. J Antimicrob Chemother
M et al (2016) Predominance and high antibiotic resis- 68:362–365
tance of the emerging Clostridium difficile genotypes Moura I, Monot M, Tani C et al (2014) Multidisciplinary
NAPCR1 and NAP9 in a Costa Rican hospital over a analysis of a nontoxigenic Clostridium difficile strain
2-year period without outbreaks. Emerg Microbes with stable resistance to metronidazole. Antimicrob
Infect 5:e42 Agents Chemother 58:4957–4960
Louie TJ, Miller MA, Mullane KM et al (2011) Mullane KM, Gorbach S (2011) Fidaxomicin: first-in-
Fidaxomicin versus Vancomycin for Clostridium dif- class macrocyclic antibiotic. Expert Rev Anti-Infect
ficile infection. N Engl J Med 364:422–431 Ther 9:767–777
Louie TJ, Cannon K, Byrne B et al (2012) Fidaxomicin Mullany P, Wilks M, Lamb I et al (1990) Genetic analysis
preserves the intestinal microbiome during and after of a tetracycline resistance determinant from Clostrid-
treatment of Clostridium difficile infection (CDI) and ium difficile and its conjugal transfer to and from
reduces both toxin reexpression and recurrence of Bacillus subtilis. J Gen Microbiol 136:1343–1349
CDI. Clin Infect Dis 55(Suppl. 2):S132–S142 Mullany P, Wilks M, Tabaqchali S (1995) Transfer of
Lynch T, Chong P, Zhang J et al (2013) Characterization macrolide-lincosamide-streptogramin B (MLS) resis-
of a stable, metronidazole-resistant Clostridium diffi- tance in Clostridium difficile is linked to a gene
cile clinical isolate. PLoS One 8:e53757 homologous with toxin A and is mediated by a
Antibiotic Resistances of Clostridium difficile 157

conjugative transposon, Tn5398. J Antimicrob Optimer Pharmaceuticals, Inc. (2011) Anti-infective

Chemother 2:305–315 drugs advisory committee briefing document:
Mullany P, Williams R, Langridge GC et al (2012) dificid™ (fidaxomicin tablets) for the treatment of
Behavior and target site selection of conjugative trans- Clostridium difficile infection (CDI), also known as
poson Tn916 in two different strains of toxigenic Clostridium difficile-associated diarrhea (CDAD), and
Clostridium difficile. Appl Environ Microbiol for reducing the risk of recurrence when used for
78:2147–2153 treatment of initial CDI. Available at: http://www.
Mullany P, Allan E, Roberts AP (2015) Mobile genetic
elements in Clostridium difficile and their role in MeetingMaterials/Drugs/Anti-InfectiveDrugsAdviso
genome function. Res Microbiol 166:361–367 ryCommittee/UCM249354.pdf
Musher DM, Aslam S, Logan N et al (2005) Relatively Pecavar V, Blaschitz M, Hufnagl P et al (2012) High-
poor outcome after treatment of Clostridium difficile resolution melting analysis of the single nucleotide
colitis with metronidazole. Clin Infect Dis polymorphism hot-spot region in the rpoB gene as an
40:1586–1590 indicator of reduced susceptibility to rifaximin in
Muto CA, Pokrywka M, Shutt K et al (2005) A large Clostridium difficile. J Med Microbiol 61:780–785
outbreak of Clostridium difficile-associated disease Peláez T, Cercenado E, Alcalá L et al (2008) Metronida-
with an unexpected proportion of deaths and zole resistance in Clostridium difficile is heteroge-
colectomies at a teaching hospital following increased neous. J Clin Microbiol 46:3028–3032
fluoroquinolone use. Infect Control Hosp Epidemiol Pépin JL, Valiquette ME, Alary ME et al (2004) Clostrid-
26:273–280 ium difficile-associated diarrhea in a region of Quebec
Muto CA, Blank MK, Marsh JW et al (2007) Control of from 1991–2003: a changing pattern disease severity.
an outbreak of infection with the hypervirulent Clos- CMAJ 17:466–472
tridium difficile BI strain in a university hospital using Pépin J, Alary ME, Valiquette L et al (2005a) Increasing
a comprehensive “bundle” approach. Clin Infect Dis risk of relapse after treatment of Clostridium difficile
45:1266–1273 colitis in Quebec, Canada. Clin Infect Dis
Norman KN, Scott HM, Harvey RB et al (2014) Compar- 40:1591–1597
ison of antimicrobial susceptibility among Clostrid- Pépin JL, Valiquette ME, Clossette B (2005b) Mortality
ium difficile isolated from an integrated human and attributed to nosocomial Clostridium difficile-
swine population in Texas. Foodborne Pathog Dis associated disease during an epidemic caused by a
11:257–264 hyperviluent strain in Quebec. CMAJ 173:1037–1042
Novak A, Spigaglia P, Barbanti F et al (2014) First clini- Perkins HR, Nieto M (1974) The chemical basis for the
cal and microbiological characterization of Clostrid- action of the vancomycin group of antibiotics. Ann N
ium difficile infection in a Croatian University Y Acad Sci 235:348–363
Hospital. Anaerobe 30:18–23 Pirš T, Avberšek J, Zdovc I et al (2013) Antimicrobial
Nyc O, Tejkalova R, Kriz Z et al (2017) Two clusters of susceptibility of animal and human isolates of Clos-
fluoroquinolone and clindamycin-resistant Clostrid- tridium difficile by broth microdilution. J Med
ium difficile PCR ribotype 001 strain recognized by Microbiol 62:1478–1485
capillary electrophoresis ribotyping and multilocus Pituch H, Brazier JS, Obuch-Woszczatynski P et al (2006)
variable tandem repeat analysis. Microb Drug Resist Prevalence and association of PCR ribotypes of Clos-
23:609–615 tridium difficile isolated from symptomatic patients
O’Connor JR, Galang MA, Sambol SP et al (2008) Rifam- from Warsaw with macrolide-lincosamide-
pin and rifaximin resistance in clinical isolates of streptogramin B (MLSB) type resistance. J Med
Clostridium difficile. Antimicrob Agents Chemother Microbiol 55:207–213
52:2813–2817 Poilane I, Cruaud P, Torlotin JC et al (2000) Comparison
Obuch-Woszczatyński P, Dubiel G, Harmanus C et al of the E test to the reference agar dilution method for
(2013) Emergence of Clostridium difficile infection antibiotic susceptibility testing of Clostridium diffi-
in tuberculosis patients due to a highly rifampicin- cile. Clin Microbiol Infect 6:155–156
resistant PCR ribotype 046 clone in Poland. Eur J Putsathit P, Maneerattanaporn M, Piewngam P et al
Clin Microbiol Infect Dis 32:1027–1103 (2017) Antimicrobial susceptibility of Clostridium dif-
Obuch-Woszczatyński P, Lachowicz D, Schneider A et al ficile isolated in Thailand. Antimicrob Resist Infect
(2014) Occurrence of Clostridium difficile Control 6:58.
PCR-ribotype 027 and it’s closely related 0214-z
PCR-ribotype 176 in hospitals in Poland in Ramı́rez-Vargas G, Quesada-Gómez C, Acuña-Amador L
2008–2010. Anaerobe 28:13–17 et al (2017) A Clostridium difficile lineage endemic to
Oka K, Osaki T, Hanawa T et al (2012) Molecular and Costa Rican hospitals is multidrug resistant by acqui-
microbiological characterization of Clostridium diffi- sition of chromosomal mutations and novel mobile
cile isolates from single, pelapse, and reinfection genetic elements. Antimicrob Agents Chemother 61:
cases. J Clin Microbiol 50:915–921 e02054.
158 P. Spigaglia et al.

Ratnayake L, McEwen J, Henderson N et al (2011) Con- Senoh M, Kato H, Fukuda T et al (2015) Predominance of
trol of an outbreak of diarrhoea in a vascular surgery PCR-ribotypes, 018 (smz) and 369 (trf) of Clostridium
unit caused by a high-level clindamycin-resistant difficile in Japan: a potential relationship with other
Clostridium difficile PCR ribotype 106. J Hosp Infect global circulating strains? J Med Microbiol
79:242–247 64:1226–1236
Redelings MD, Sorvillo F, Mascola L (2007) Increase in Seugendo M, Mshana SE, Hokororo A et al (2015) Clos-
Clostridium difficile-related mortality rates, United tridium difficile infections among adults and children
States, 1999–2004. Emerg Infect Dis 13:1417–1419 in Mwanza/Tanzania: is it an underappreciated patho-
Reil M, Hensgens MPM, Kuijper EJ et al (2012) Season- gen among immunocompromised patients in
ality of Clostridium difficile infections in Southern sub-Saharan Africa? New Microbes New Infect
Germany. Epidemiol Infect 140:1787–1793 8:99–102
Richardson C, Kim P, Lee C et al (2015) Comparison of Shayganmehr F-S, Alebouyeh M, Azimirad M et al
Clostridium difficile isolates from individuals with (2015) Association of tcdA+/tc dB+ Clostridium diffi-
recurrent and single episode of infection. Anaerobe cile genotype with emergence of multidrugresistant
33:105–108 strains conferring metronidazole resistant phenotype.
Roberts AP, Mullany P (2011) Tn916-like genetic Iran Biomed J 19:143–148
elements: a diverse group of modular mobile elements Simango C, Uladi S (2014) Detection of Clostridium
conferring antibiotic resistance. FEMS Microbiol Rev difficile diarrhoea in Harare, Zimbabwe. Trans R Soc
35:856–871 Trop Med Hyg 108:354–357
Roberts MC, McFarland LV, Mullany P et al (1994) Spigaglia P (2016) Recent advances in the understanding
Characterization of the genetic basis of antibiotic of antibiotic resistance in Clostridium difficile infec-
resistance in Clostridium difficile. J Antimicrob tion. Ther Adv Infect Dis 3:23–42
Chemother 33:419–429 Spigaglia P, Mastrantonio P (2004) Comparative analysis
Roberts AP, Johanesen PA, Lyras D et al (2001) Compar- of Clostridium difficile clinical isolates belonging to
ison of Tn5397 from Clostridium difficile, Tn916 from different genetic lineages and time periods. J Med
Enterococcus faecalis and the CW459tet(M) element Microbiol 53:1129–1136
from Clostridium perfringens shows that they have Spigaglia P, Carucci V, Barbanti F et al (2005) ErmB
similar conjugation regions but different insertion determinants and Tn916-like elements in clinical
and excision modules. Microbiology 147:1243–1251 isolates of Clostridium difficile. Antimicrob Agents
Rodrı́guez-Pardo D, Almirante B, Bartolomé RM et al Chemother 49:2550–2553
(2013) Epidemiology of Clostridium difficile infection Spigaglia P, Barbanti F, Mastrantonio P (2006) New
and risk factors for unfavorable clinical outcomes: variants of the tet(M) gene in Clostridium difficile
results of a hospital-based study in Barcelona, Spain. clinical isolates harbouring Tn916-like elements. J
J Clin Microbiol 51:1465–1473 Antimicrob Chemother 57:1205–1209
Salix Pharmaceuticals, Ltd. 10 December 2003, posting Spigaglia P, Barbanti F, Mastrantonio P (2007) Detection
date. Salix receives FDA notification that of a genetic linkage between genes coding for resis-
rifaximin amendment considered a complete response. tance to tetracycline and erythromycin in Clostridium
Salix Pharmaceuticals, Raleigh. http://www. difficile. Microb Drug Resist 13:90–95 Spigaglia P, Barbanti F, Mastrantonio P (2008a) Tetracy-
Salix-Receives-FDA-Notification-Rifamixin-Amend cline resistance gene tet(W) in the pathogenic bacte-
ment-Considered rium Clostridium difficile. Antimicrob Agents
Santos A, Isidro J, Silva C et al (2016) Molecular and Chemother 52:770–773
epidemiologic study of Clostridium difficile reveals Spigaglia P, Barbanti F, Mastrantonio P et al (2008b)
unusual heterogeneity in clinical strains circulating Fluoroquinolone resistance in Clostridium difficile
in different regions in Portugal. Clin Microbiol Infect isolates from a prospective study of C. difficile
22:695–700 infections in Europe. J Med Microbiol 57:784–789
Schmidt C, L€offler B, Ackermann G (2007) Antimicrobial Spigaglia P, Barbanti F, Louie T et al (2009) Molecular
phenotypes and molecular basis in clinical strains of analysis of the gyrA and gyrB quinolone resistance-
Clostridium difficile. Diagn Microbiol Infect Dis determining regions of fluoroquinolone-resistant Clos-
59:1–5 tridium difficile mutants selected in vitro. Antimicrob
Sears P, Crook DW, Louie TJ et al (2012) Fidaxomicin Agents Chemother 53:2463–2468
attains high fecal concentrations with minimal plasma Spigaglia P, Barbanti F, Dionisi AM et al (2010) Clostrid-
concentrations following oral administration in ium difficile isolates resistant to fluoroquinolones in
patients with Clostridium difficile infection. Clin Italy: emergence of PCR ribotype 018. J Clin
Infect Dis 55(Suppl 2):S116–SS12 Microbiol 48:2892–2896
Sebaihia M, Wren BW, Mullany P et al (2006) The Spigaglia P, Barbanti F, Mastrantonio P, European Study
multidrug-resistant human pathogen Clostridium diffi- Group on Clostridium difficile (ESGCD) (2011) Mul-
cile has a highly mobile, mosaic genome. Nat Genet tidrug resistance in European Clostridium difficile
Antibiotic Resistances of Clostridium difficile 159

clinical isolates. J Antimicrob Chemother Wasels F, Spigaglia P, Barbanti F et al (2013) Clostridium

66:2227–2234 difficile erm(B)-containing elements and the burden
Spigaglia P, Barbanti F, Morandi M et al (2015) Diagnos- on the in vitro fitness. J Med Microbiol 62:1461–1467
tic testing for Clostridium difficile in Italian Wasels F, Monot M, Spigaglia P et al (2014) Inter- and
microbiological laboratories. Anaerobe 37:29–33 intraspecies transfer of a Clostridium difficile
Srivastava A, Talaue M, Liu S et al (2011) New target for conjugative transposon conferring resistance to
inhibition of bacterial RNA polymerase: “switch MLSB. Microb Drug Resist 20:555–560
region”. Curr Opin Microbiol Antimicrob/Genomics Wasels F, Kuehne SA, Cartman ST et al (2015a) Fluoro-
14:532–543 quinolone resistance does not impose a cost on the
Tannock GW, Munro K, Taylor C et al (2010) A new fitness of Clostridium difficile in vitro. Antimicrob
macrocyclic antibiotic, fidaxomicin (OPT-80), causes Agents Chemother 59:1794–1796
less alteration to the bowel microbiota of Clostridium Wasels F, Spigaglia P, Barbanti F et al (2015b) Integra-
difficile-infected patients than does vancomycin. tion of erm(B)-containing elements through large
Microbiology 156:3354–3359 chromosome fragment exchange in Clostridium diffi-
Tenover FC, Tickler IA, Persing DH (2012) cile. Mob Genet Elem 1:12–16
Antimicrobial-resistant strains of Clostridium difficile Weber I, Riera E, Déniz C et al (2013) Molecular epide-
from North America. Antimicrob Agents Chemother miology and resistance profiles of Clostridium difficile
56:2929–2932 in a tertiary care hospital in Spain. Int J Med Microbiol
Terhes G, Maruyama A, Latkóczy K et al (2014) In vitro 303:128–133
antibiotic susceptibility profile of Clostridium difficile Wistr€om J, Norrby SR, Myhre EB et al (2001) Frequency
excluding PCR ribotype 027 outbreak strain in of antibiotic-associated diarrhea in 2462 antibiotic-
Hungary. Anaerobe 30:41–44 treated hospitalized patients: a prospective study. J
Vardakas KZ, Polyzos KA, Patouni K et al (2012) Treat- Antimicrob Chemother 47:43–50
ment failure and recurrence of Clostridium difficile Wren BW, Mullany P, Clayton C et al (1988) Molecular
infection following treatment with vancomycin or cloning and genetic analysis of a chloramphenicol
metronidazole: a systematic review of the evidence. acetyltransferase determinant from Clostridium diffi-
Int J Antimicrob Agents 40:1–8 cile. Antimicrob Agents Chemother 32:1213–1121
Varshney JB, Very KJ, Williams JL et al (2014) Charac- Wren BW, Mullany P, Clayton C et al (1989) Nucleotide
terization of Clostridium difficile isolates from human sequence of a chloramphenicol acetyl transferase gene
fecal samples and retail meat from Pennsylvania. from Clostridium difficile. Nucleic Acids Res 17:4877
Foodborne Pathog Dis 11:822–829 Young GP, Ward PB, Bayley N et al (1985) Antibiotic-
Vuotto C, Moura I, Barbanti F et al (2016) Subinhibitory associated colitis due to Clostridium difficile: double-
concentrations of metronidazole increase biofilm for- blind comparison of vancomycin with bacitracin. Gas-
mation in Clostridium difficile strains. Pathog Dis 74: troenterology 89:1038–1045
ftv114. Yu X, Sun D (2013) Macrocyclic drugs and synthetic
Walkty A, Boyd DA, Gravel D et al (2010) Molecular methodologies toward macrocycles. Molecules
characterization of moxifloxacin resistance from 18:6230–6268
Canadian Clostridium difficile clinical isolates. Diagn Zhou Y, Burnham C-AD, Hink T et al (2014) Phenotypic
Microbiol Infect Dis 66:419–424 and genotypic analysis of Clostridium difficile
Wang H, Smith MCM, Mullany P (2006) The conjugative isolates: a single-center study. J Clin Microbiol
transposon Tn5397 has a strong preference for inte- 52:4260–4266
gration into its Clostridium difficile target site. J
Bacteriol 188:4871–4878
Probiotics for Prevention and Treatment
of Clostridium difficile Infection

Lorena Valdés-Varela, Miguel Gueimonde,

and Patricia Ruas-Madiedo

Abstract to evaluate probiotic products, in combina-

Probiotics have been claimed as a valuable tion with antibiotics, in order to select the
tool to restore the balance in the intestinal best candidate for C. difficile infections.
microbiota following a dysbiosis caused by,
among other factors, antibiotic therapy. This Keywords
perturbed environment could favor the over- Probiotic · C. difficile · Clinical study ·
growth of Clostridium difficile and, in fact, Mechanism of action · Antagonism
the occurrence of C. difficile-associated
infections (CDI) is being increasing in
recent years. In spite of the high number of 1 Introduction
probiotics able to in vitro inhibit the growth
and/or toxicity of this pathogen, its applica- The gut microbiota is a complex and diverse micro-
tion for treatment or prevention of CDI is bial community that has co-evolved with humans
still scarce since there are not enough well- in a commensal way (Donaldson et al. 2016). In a
defined clinical studies supporting efficacy. healthy state, this collection of microorganisms
Only a few strains, such as Lactobacillus protects the host by inhibiting colonization and
rhamnosus GG and Saccharomyces growth of pathogens. However, antibiotic exposure
boulardii have been studied in more extent. strongly perturbs the intestinal microbiota, produc-
The increasing knowledge about the probi- ing a decrease in microbial abundance and species
otic mechanisms of action against diversity, as well as a suppression of the innate
C. difficile, some of them reviewed here, immune system disrupting the gut barrier and fre-
makes promising the application of these quently causing antibiotic-associated diarrhea. In
live biotherapeutic agents against CDI. some cases, the intestinal dysbiosis followed after
Nevertheless, more effort must be paid to antibiotic treatment allows the overgrowth of Clos-
standardize the clinical studied conducted tridium difficile given that this perturbed

L. Valdés-Varela · M. Gueimonde · P. Ruas-Madiedo (*)

Department of Microbiology and Biochemistry of Dairy
Products, Instituto de Productos Lácteos de Asturias –
Consejo Superior de Investigaciones Cientı́ficas (IPLA-
CSIC), Villaviciosa, Asturias, Spain

# Springer International Publishing AG 2018 161

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
162 L. Valdés-Varela et al.

environment has a low abundance of short chain the cell and translocate the first component to the
fatty acids, a high abundance of primary bile acids, cytosol (Gerding et al. 2014). In spite of recent
a high carbohydrate availability and a advances in the identification of processes involved
immunosuppressed host in the absence of microbial on receptor binding and entry into mammalian
competitors in the gut (Lawley and Walker 2013). cells, the mode-of-action of clostridial toxins
C. difficile can be found in the gut microbiota remains to be totally elucidated (Orrell et al. 2017).
of both, healthy infants and adults, the occur- The standard treatment for C. difficile infec-
rence being higher in infant (70%) than in the tion is the administration of antibiotics, mainly
adult (17%) population (Ozaki et al. 2004; Jangi metronidazole, vancomycin or fidaxomicin, but
and Lamont 2010). In these healthy carriers the unfortunately the recurrence rate of the disease is
presence of this microorganism does not seem to very high and this treatment becomes less effec-
cause any disease. However, at the same time tive. Indeed, it has been described that some
C. difficile is the main causative agent of C. difficile subpopulations (ribotypes) have a
antibiotic-associated diarrhea in nosocomial reduced susceptibility to metronidazole (Moura
environments (Leffler and Lamont 2015). As previ- et al. 2013). In case of multiple recurrent CDI,
ously indicated the antimicrobial therapy affects the fecal microbiota transplantation (FMT) is being
endogenous gut microbiota diminishing coloniza- more frequently used as the ultimate therapy,
tion resistance, allowing the overgrowth of this although the selection of the appropriate donor
pathogen and causing C. difficile-associated diar- is a critical issue (Woodworth et al. 2017). These
rhea (CDAD). This problem has been traditionally facts have prompted researchers to look for alter-
linked to elderly and institutionalized/hospitalized native therapeutic options (Fig. 1) which have
persons under antibiotic therapy (Rupnik et al. been recently reviewed by different authors
2009); however, the occurrence of C. difficile- (Mathur et al. 2014; Hussack and Tanha 2016;
associated infections (CDI) seems to be increasing Kachrimanidou et al. 2016; Kociolek and
also in traditionally considered low-risk populations Gerding 2016; Martin and Wilcox 2016;
(Carter et al. 2012). This change in the epidemiol- McFarland 2016; Ofosu 2016; Padua and
ogy of CDI has been related to the worldwide Pothoulakis 2016; Unal € and Steinert 2016).
distribution of hyper virulent strains (Yakob et al. Among them, probiotics have been proposed as
2015); besides, foods and animals have been found a potential tool for preventing the dysbiosis of
to act as carriers of this pathogen pointing at microbiota, caused by the administration of
C. difficile as a zoonotic agent and suggesting antibiotics, and for assisting in the microbiota
potential food-borne transmission (Rodriguez et al. restoration after antibiotics or infection (Reid
2016). A range of virulent factors are the cause of et al. 2011); thus, they have also been evaluated
colitis during CDI course, the main ones being for prevention and treatment of CDI (Na and
several toxins, encoded in pathogenicity loci, and Kelly 2011).
the flagella, which are factors allowing mobility and Probiotics were defined in 2001 by a group
adherence of the pathogen (Abt et al. 2016). Patho- of experts joined by FAO/WHO as “live
genesis was initially attributed to the production of microorganisms that, when administered in
toxins A (TcdA) and B (TcdB), belonging to the adequate amounts, confer a health benefit on
large clostridial toxin (LCT) family, which act as the host”; this definition was recently revised,
intracellular glycosyl-transferases that inactivate and accepted after minor grammatical
Rho family GTPases, thus blocking downstream modifications, by members of the International
cellular events (Carter et al. 2012). More recently, Scientific Association for Probiotics and
strains producing a third toxin, the binary toxin Prebiotics (ISAPP) which also propose an
(CDT), have been associated with an increase in overall framework for use of this term,
the CDI severity; this toxin has two components the encompassing diverse end uses (Hill et al.
CDTa, which acts as an ADP-ribosyltransferase 2014). In next sections we will review the
targeting actin, and CDTb that is able to binds to current available data about the efficacy of
Probiotics for Prevention and Treatment of Clostridium difficile Infection 163

Fig. 1 Some therapeutic

options currently under
study for the prevention
and treatment of
Clostridium difficile

probiotics in prevention and therapy for CDI, 2005; Banerjee et al. 2009) or to be able to inhibit
as well as some putative mechanisms involved its growth (Lee et al. 2013; Schoster et al. 2013;
in this anti-C. difficile effect. Valdes-Varela et al. 2016b). Moreover, animal
studies seem to confirm a potential benefit of
probiotics on the inhibition of C. difficile coloni-
zation (Mansour et al. 2017). Nevertheless, to
2 Clinical Studies Evaluating
date most of the clinical studies have focused
Probiotic Efficacy
on prevention and there is a lack of data on the
potential use of probiotics on the treatment of
The ability of probiotics for inhibiting the growth
C. difficile infection.
of C. difficile has been characterized by using
During the last couple of decades several studies
different experimental approaches (Auclair
have evaluated the usefulness of different probiotic
et al. 2015; Forssten et al. 2015; Valdes-Varela
strains in the prevention of CDAD. However, in
et al. 2016b; Fredua-Agyeman et al. 2017). This
spite of the large number of strains screened
use of probiotic microorganisms has long been
in vitro, most of the evidence from clinical trials
considered a potential option to combat CDI.
regards only a few bacterial strains and, most often,
However, despise the large number of in vitro
the studies have focused on the prevention of
studies performed for the selection of probiotic
antibiotic-associated diarrhea, without further con-
strains with activity against C. difficile and for
firmation of C. difficile etiology. Among the
their use for CDI prevention or treatment, the
assessed strains the effect of Lactobacillus
evidence from human clinical trials is still lim-
rhamnosus strain GG (Arvola et al. 1999;
ited. Different probiotic strains have been
Vanderhoof et al. 1999), or the yeast species Sac-
reported to increase the colonization resistance
charomyces boulardii (Kotowska et al. 2005; Can
against C. difficile (Hopkins and Macfarlane
et al. 2006), in the prevention of antibiotic
2003; Kondepudi et al. 2014; Auclair et al.
associated diarrhea has been widely recognized.
2015; Forssten et al. 2015). Certain strains of
Although not so extensively studied, other probiotic
bifidobacteria and lactobacilli have been found
strains and probiotic mixes have also been
to reduce the adhesion of C. difficile to intestinal
evaluated around the world with positive results
epithelial cells or intestinal mucus (Collado et al.
164 L. Valdés-Varela et al.

(Wullt et al. 2003; Maziade et al. 2015). The avail- Dietrich et al. 2014; Maziade et al. 2015). Some
ability of a large number of clinical studies focusing practical examples exist as well, such as that of the
on antibiotic-associated diarrhea has provided “Pierre-Le Gardeur” Hospital in Canada, that after
enough data for carrying out systematic reviews a C. difficile outbreak begun to administer a probi-
and meta-analysis studies, either considering otic mix (BioK+®) together with any antibiotic
probiotics as a group, which shows important prescriptions, achieving a significant reduction on
limitations due to inter-strain and/or inter-product the number of C. difficile disease cases (Maziade
variability, or meta-analyses focused on specific et al. 2015). Recent meta-analyses and systematic
strains. The meta-analysis studies on the general reviews have assessed the effects of probiotic
use of probiotics for the prevention of antibiotic- administration, most of them administering the
associated diarrhea have consistently provided evi- strains together with the antibiotic treatment, on
dence for a beneficial role, especially in children the primary prevention of CDAD in different pop-
(Cremonini et al. 2002; D’Souza et al. 2002; ulation groups (Table 1). In general the data sup-
Sazawal et al. 2006; Johnston et al. 2007; Hempel port a beneficial effect of probiotics on the primary
et al. 2012; Goldenberg et al. 2015). Moreover, prevention of CDAD. However, the high heteroge-
meta-analyses conducted for some specific neity among the available clinical studies makes
probiotics, such as S. boulardii or L. rhamnosus difficult defining the best probiotic to be used, its
GG, have further confirmed the beneficial effect of dose, and the administration regime.
these strains in the prevention of antibiotic- Regarding the prevention of the recurrence of
associated diarrhea (McFarland 2006; Szajewska the disease, the available data are more limited
et al. 2007a, b). This has resulted in than in the case of primary prevention. Some
recommendations issued by the ESPGHAN clinical intervention studies have been conducted
(European Society for Paediatric Gastroenterology with variable results (McFarland et al. 1994;
Hepatology and Nutrition) with regard to the use of Surawicz et al. 2000), with reviews and meta-
probiotics for the prevention of antibiotic-associated analyses indicating that there is only limited evi-
diarrhea in children (Szajewska et al. 2016). dence on the benefit of probiotics in secondary
Furthermore, some studies have specifically prevention of CDI (Allen et al. 2013; O’Horo
focused in confirmed C. difficile-associated diar- et al. 2014; McFarland 2015). The limited data
rhea and these have also provided positive results available on secondary prevention underlines the
for primary prevention (Wullt et al. 2003; Gao et al. need for more clinical intervention trials to be
2010; Sampalis et al. 2010; Allen et al. 2013; conducted in this topic.

Table 1 Recent meta-analyses and systematic reviews on the use of probiotics in primary prevention of C. difficile
Target N eligible N volunteers
population Probiotic RCTsa included Conclusion References
Elderly Any 5 >3400 No significant Vernaya et al. (2017)
Adults Any 19 >6200 Significant Shen et al. (2017)
Adults Lactobacillus 10 >4800 Inconclusive Sinclair et al. (2016)
(any) evidence
Adults and Any 26 >7900 Significant Lau and Chamberlain
children reduction (2016)
Adults and Any (and by 21 >3700 Significant McFarland (2015)
children species) reduction
Adults and Any 31 >4200 Significant Goldenberg et al.
children reduction (2013)
RCT randomized controlled trial
Probiotics for Prevention and Treatment of Clostridium difficile Infection 165

To sum up, the available evidence strongly C. difficile but, as disadvantage, they have the
suggests that probiotics are helpful for primary pre- lack of feedback mechanisms with host and/or
vention with only moderate evidence of a role in host-microbe interactions (Best et al. 2012).
avoiding disease relapse. However, the potential However, these microbial culturing models
role of probiotics in the treatment during the active can be combined with cell culture systems to
phase of the disease remains largely unknown. Per- better mimic the interaction C. difficile- probi-
haps the major criticism that can be done to the otic- host (Venema and van den Abbeele
available data is that there has not been a serious 2013). Co-cultures of toxigenic C. difficile
standardization effort for the probiotic products, strains with probiotic candidates have been
doses, antibiotics and therapeutic protocols to be carried out to determine the potential of the
used. Moreover, analyses of the cost-effectiveness latter for reducing the germination of spores
of probiotic use on the prevention of C. difficile and outgrowth into vegetative toxin-producing
disease have not been performed until recently, cells of the pathogen (Table 2). Models of gut
with variable results, indicating the need for further microbiota have been assayed to in vitro eval-
studies conducted under different healthcare uate the potential of probiotic candidates for
systems (Leal et al. 2016; Starn et al. 2016). decreasing the growth of C. difficile in this
complex microbial ecosystem. These models
range from simple batch fermentations to com-
plex multi-compartmental continuous systems
3 Models to Study Probiotics
(Venema and van den Abbeele 2013). Static
Against C. difficile
batch cultures, containing fecal suspensions,
have been used to observe the influence of
Different experimental models have been
probiotics on the survival of C. difficile
developed in order to study the interaction of
(Tejero-Sariñena et al. 2013). Continuous cul-
C. difficile with the host (recently reviewed by
ture systems (human “colonic” model) allow
Young 2017); additionally, these models can
the study of the pathogen in an environment
be used in the search for new therapeutic
closer to the reality, over considerably longer
alternatives and adjuvant strategies for
periods than in static batch cultures (Best et al.
preventing or treating CDI (Table 2).
2012; Le Lay et al. 2015). Currently, most of
Investigations using in vitro models of bacte-
the colonic simulators consists of four differ-
rial cultures are valuable systems for the
ent units (glass vessels) continuously
screening of potential probiotics against

Table 2 Summary of some in vitro models used to study potential probiotics against Clostridium difficile
In vitro experimental
models References
Microbial vs. Co-cultures of C. difficile Trejo et al. (2010), Best et al. (2012), Kolling et al. (2012),
cultivation probiotic with probiotic candidates Lee et al. (2013), Schoster et al. (2013), Kondepudi et al.
(2014), Yun et al. (2014), Ambalam et al. (2015), Andersen
et al. (2016), Spinler et al. (2016), and Rätsep et al. (2017)
vs. Static-batch system Tejero-Sariñena et al. (2013)
microbiota/ Semi-continuous system Le Lay et al. (2015)
probiotic “Colonic” model Forssten et al. (2015)
Intestinal Adhesion/ HT29-MTX cell Zivkovic et al. (2015)
cell lines exclusion Immobilized intestinal Collado et al. (2005), Banerjee et al. (2009), and Ferreira
mucus et al. (2011)
Cytotoxicity Label-based endpoint Barnerjee et al. (2009), Trejo et al. (2010, 2013), and
methods Valdés-Varela et al. (2016a)
Label-free, RTCA Valdés et al. (2015) and Valdés-Varela et al. (2016a, b)
166 L. Valdés-Varela et al.

connected, having different pH and flow rates, (adapted to methotrexate) thus synthesizing higher
thus representing the ascending, transverse, amounts of mucus (Zivkovic et al. 2015). A study
descending and distal colon (Forssten et al. has suggested that this cell model may be more
2015). suitable for studying cell-pathogen interactions, as
Several in vitro studies investigated the effect well as effectiveness of antimicrobial treatments, as
of probiotic treatment on the interaction of compared to Caco-2 or HT29 models which do not
C. difficile with components of the intestinal have Goblet cells or do not constitutively secrete
mucosa, such as mucus or epithelial cells mucus, respectively (Gagnon et al. 2013).
(Table 2). The cytotoxicity of clostridial cell- In an step forward, several authors have
free supernatants (obtained from co-cultures of evaluated the protective effect of selected probi-
probiotic vs. C. difficile) or of caecum contents otic candidates against CDI in animal models
(collected from animals infected with C. difficile (Best et al. 2012; Kolling et al. 2012; Trejo
and treated with potential probiotics) has been et al. 2013; Kondepudi et al. 2014; Yun et al.
evaluated upon cell lines using classic label- 2014; Andersen et al. 2016; Arruda et al. 2016;
based, endpoint methods (Barnerjee et al. 2009; Spinler et al. 2016; Rätsep et al. 2017). This
Trejo et al. 2010, 2013; Valdés-Varela et al. infection has been studied in different models,
2016a). However, label-free technologies are including mice, hamsters, rats, rabbits, hares,
currently been available and being used in drug guinea pigs, prairie dogs, quails, foals, piglets
development processes, which are non-invasive and monkeys. Moreover, zebrafish embryos
techniques that allow the continuous (real time) have been described as suitable models for iden-
monitoring of the status of live cells (Xi et al. tification of in vivo targets of C. difficile toxins
2008). Indeed the label-free, impedance-based and evaluation of novel candidate therapeutics;
RTCA (real time cell analyzer) technology has zebrafish possess many of the major organs pres-
been applied to develop methods allowing the ent in humans and, due to the transparency of the
clinical diagnosis of toxigenic C. difficile in dif- embryo, damage by toxins can be visualized by
ferent biological samples (Yu et al. 2015). standard light microscopy (Best et al. 2012).
Recently, this RTCA technology was also used Each of the C. difficile animal models has inher-
in our group to develop a model to test the cyto- ent advantages and disadvantages. The hamster
toxicity of C. difficile supernatants upon the model has been widely used to study
intestinal epithelial cell lines HT29 and Caco- pseudomembranous colitis in human because of
2 (Valdés et al. 2015). Moreover, this model extreme sensitivity to infection following antibi-
was used to search for potential probiotic strains otic administration, using clindamycin as agent
able to counteract the toxic effect of C. difficile of choice; however, this model does not represent
supernatants upon HT29 (Valdés-Varela et al. the usual course and spectrum of CDI in humans.
2016a) as well as to evaluate the toxicity of Recently, new mouse and piglet CDI models
C. difficile co-cultured with some of these have been developed which appear to mimic
probiotics (Valdés-Varela et al. 2016b). many of disease symptoms observed in humans
On the other hand, several models have been (Sun et al. 2011; Best et al. 2012; Hutton et al.
used to assess the ability of probiotic candidates to 2014).
modify the adhesion C. difficile to the intestinal
mucosa, such as those using immobilized (human)
intestinal mucus which showed a good correlation 4 Mechanisms of Probiotic
with data obtained with a enterocyte-like (Caco-2) Action
model (Collado et al. 2005; Banerjee et al. 2009;
Ferreira et al. 2011). The ability of potential probi- As pointed in previous sections, probiotics are
otic strains to inhibit the adhesion of C. difficile has gaining more and more interest as preventive and
also been evaluated using intestinal cell lines, such co-adjuvant therapies for treatment of antibiotic-
as HT29-MTX which is a derivative from HT29 associated dysbiosis. However, their modes of
Probiotics for Prevention and Treatment of Clostridium difficile Infection 167

action are poorly understood and vary between context of CDI (Parkes et al. 2009; Ollech et al.
probiotic microorganisms. Indeed the effects of 2016).
any probiotic are strain-specific and, therefore, Some probiotic strains are able to compete
beneficial effects cannot be extrapolated to other with pathogenic bacteria for the adhesion sites,
species or strains (Hickson 2011). It has been i. e. competitive exclusion, thus providing a
described that probiotics could have diverse pos- “physical” barrier that increases the colonization
itive actions on the host by: (i) modulating the resistance (Fig. 2a). In vitro studies showed the
intestinal microbiota and inhibiting pathogenic ability of selected Bifidobacterium and Lactoba-
microorganisms at the intestinal luminal environ- cillus strains to modify the adhesion of
ment, (ii) enhancing of intestinal barrier function C. difficile to intestinal epithelial cells, or intesti-
at the intestinal epithelium, and (iii) modulating nal mucus, the effect being strain-dependent
the immune response, among others (Ng et al. (Collado et al. 2005; Zivkovic et al. 2015). A
2009). Several mechanisms have been proposed reduction from 60% to 3% in the adhesion of
for explaining the potential role of probiotics C. difficile to gingival epithelial cell cultures
against C. difficile. Some of these effects, such (obtained from healthy horses) was reported
as the production of antimicrobial factors (Corr when Lactobacillus reuteri Lr1 was added; addi-
et al. 2007), competitive inhibition of the patho- tionally, it was detected that this strain was able
gen (Collado et al. 2005) or the ability to degrade to co-aggregate with the pathogen (Dicks et al.
and to reduce the toxicity of C. difficile 2015). In this regard, it has been suggested that
(Castagliuolo et al. 1999; Valdes-Varela et al. the aggregation capability between lactobacilli
2016a), could be of help not only in prevention and C. difficile could be a way to reduce the
but also in the treatment of CDI. adhesion of the pathogen to the intestinal mucosa
(Ferreira et al. 2011). S. boulardii is also able to
reduce the adhesion of C. difficile to epithelial
4.1 Microbial Antagonism: cells and the same effect was detected using
Interaction Probiotics extracts obtained from the cell-wall of this yeast
vs. C. difficile (Tasteyre et al. 2002). Similarly, it has been
proved that cell-free supernatants obtained from
The restoration of intestinal microbiota after Lactobacillus delbrueckii ssp. bulgaricus
dysbiosis, caused by any etiological agent, is B-30892 (Banerjee et al. 2009) and different
the main way of action of any treatment against bifidobacterial strains (Trejo et al. 2006) were
intestinal pathogens including C. difficile able to reduce the adhesion of C. difficile to
(Gareau et al. 2010; Reid et al. 2011). This was intestinal epithelial Caco-2 cells. Different
evidenced, for example, in an in vivo study with treatments of the bifidobacterial supernatants
a murine CDI model of antibiotic-induced showed that the factors related to the anti-
dysbiosis, in which the gut microbiota was clostridial adhesion were no heat-resistant,
restored after treatment with a multi-strain probi- non-related with acids (active at neutral pH)
otic supplement (Lactobacillus plantarum F44, and were not affected by proteinases, but its
Lactobacillus paracasei F8, Bifidobacterium nature remains unknown (Trejo et al. 2006).
breve 46, Bifidobacterium animalis subsp. lactis Indirect evidence suggests that exopolysac-
8:8) (Kondepudi et al. 2014). There are several charides covering the surface of some probiotics
mechanisms by which probiotics can help the could be involved in the inhibition of the binding
restoration of the intestinal microbiota, some of capability of some pathogens, including
them being related with typical bacterial antago- C. difficile, by probiotics (Ruas-Madiedo et al.
nism (Ng et al. 2009); however, little is known 2006). Thus, altogether, these studies suggest
about those mechanisms acting specifically in the that different surface molecules and/or secreted
168 L. Valdés-Varela et al.

A: competitive exclusion / co-aggregation B: production of anti-microbial compounds

Organic acids Bacteriocins




C: anti-toxin activity D: reinforcement of the intestinal barrier

Restoraon of barrier funcon

• Anti-inflammatory / regulatory
cytokine synthesis
Pro-inflammatory state • Mucus secretion
• Tight junctions expression…

Neutrophils Mast cells Lymphocytes

TcdA Proteinase EPS

TcdB Unknown S-layer
microbial factor

Fig. 2 Potential mechanisms of action proposed for microbial compounds. (c) anti-toxin activity. (d) rein-
probiotics against Clostridium difficile. (a) competitive forcement of the intestinal barrier
exclusion/co-aggregation. (b) production of anti-

factors might be implicated in the interference of growth in a CDI mouse model, which may be
probiotics against C. difficile adhesion to the related to a reduction in pH as a result of organic
intestinal mucosa. acids produced by the probiotic bacterium (Yun
Another mechanism of probiotic action is the et al. 2014). Several in vitro studies have
inhibition of the pathogen growth through the investigated the activity of probiotics to inhibit
competition for the limiting nutritional sources C. difficile growth; using a fecal, pH-controlled
and/or by the production of antimicrobial factors, (between 6.7 and 6.9), anaerobic batch model it
such as organic acids and bacteriocins (Fig. 2b). was found that Lactobacillus casei
In a study carried out with a CDI animal model it NCIMB30185 and B. breve NCIMB30180 were
was shown that mice treated with Streptococcus able to reduce the numbers of C. difficile in this
thermophilus LMD-9 exhibited less pathology, complex microbial ecosystem (Tejero-Sariñena
and lower detectable toxin levels in cecal et al. 2013). Co-cultivation of C. difficile with
contents, compared with untreated controls; an cell-free supernatants from different commercial
inverse correlation was observed between the probiotics highlighted that the mechanism of
levels of luminal lactate and the abundance of inhibition was pH-dependent; thus, the produc-
C. difficile, suggesting that the anti-clostridial tion of organic acids, mainly lactic and acetic
effect was due to the production of this organic acids, are the inhibition factors controlling the
acid (Kolling et al. 2012). Similarly, the lactic growth of C. difficile (Schoster et al. 2013). In
acid synthesized by Lactobacillus acidophilus another in vitro study, the co-incubation of
GP1B had an inhibitory effect on C. difficile C. difficile with L. rhamnosus LR5, Lactococcus
Probiotics for Prevention and Treatment of Clostridium difficile Infection 169

lactis SL3, B. breve BR3 and B. animalis subsp. 4.2 Probiotics Against C. difficile
lactis BL3 demonstrated their potential to Toxin Activity
decrease C. difficile numbers, mainly mediated
by the organic acid production. However, among The toxins produced by C. difficile are responsi-
those strains, SL3 appeared to have the strongest ble for the clinical profile of the CDI. Therefore,
activity which seems to be pH-independent and therapeutic agents that reduce toxin-induced
likely could be mediated through the action of a damage could be valuable tools to alleviate the
bacteriocin (Lee et al. 2013). Similar severity of symptoms and to improve the course
pH-dependent and pH-independent effects of the disease. Some authors have reported that
against C. difficile were also reported using probiotics are able to reduce the activity of
cell-free supernatants from other commercially C. difficile toxins but, in most cases, the specific
available probiotics (Fredua-Agyeman et al. mechanisms of action by which probiotics exert
2017). With respect to the competition for the protective effect in this infection is unknown
nutrients, some studies have been carried out (Fig. 2c). In a hamster model of enterocolitis
using “synbiotic” combinations, which are induced by C. difficile, Bifidobacterium bifidum
mixtures of probiotics and prebiotic substrates CIDCA5310 protected the animals, and avoided
that (theoretically) will improve the performance mortality, when compared with the control
of probiotics or other beneficial microbes in the (infected) group; besides, the supernatants
gut. In a mice (C57BI/6) model of CDI, the obtained from caecum contents were less toxics
feeding with a synbiotic formulation, consisting upon Vero (cells from monkey’s kidney) cultures
of four strains (L. plantarum F44, L. paracasei in animals fed with the bifidobacteria suggesting
F8, B. breve 46, B. animalis subsp. lactis 8:8) and that this strain is able to in vivo counteract the
three prebiotics (galacto-oligosaccharides, effect of clostridial toxins (Trejo et al. 2013).
isomalto-oligosaccharides and resistant starch), Co-culture of toxigenic strains of C. difficile
conferred protection against this pathogen with different strains of bifidobacteria and
(Kondepudi et al. 2014). Some studies have lactobacilli leads to a reduction of the cytotoxic
suggested that the growth inhibition of effects of spent-culture supernatants on cultured
C. difficile by probiotics is strain but also car- Vero cells, which correlates with a diminution of
bon source specific. Ambalam et al. reported clostridial toxins present in these supernatants
the ability of cell-free supernatants from (Trejo et al. 2010). However, the growth of clos-
L. paracasei F8 and L. plantarum F44 to inhibit tridial strains in BHI medium with different
the growth of C. difficile strains when they concentrations of cell-free supernatants from
grew on glucose, due to the production of bifidobacteria or lactobacilli cultures did not
organic acids and heat-stable antimicrobial decrease the toxic effect of pathogens; taking
proteins, whilst the effect was only into account these results, authors hypothesized
pH-dependent when growing on prebiotics that co-culture of clostridia with lactobacilli or
(Ambalam et al. 2015). Our workgroup bifidobacteria leads to the modification of the
recently analyzed the influence of carbon environment, thus leading to the repression of
sources upon C. difficile growth and toxicity toxin synthesis/secretion pathway. Similarly, a
when co-cultured with Bifidobacterium longum cell extract from L. acidophilus GP1B was able
IPLA20022 or B. breve IPLA20006 in the pres- to decrease the pathogenicity of C. difficile by
ence of short-chain fructo-oligosaccharides inhibiting quorum sensing signaling, probably by
(scFOS) or inulin. The use of scFOS reduced lowering the expression of quorum sensing-
the growth of the pathogen, as well as the regulated toxin genes (Yun et al. 2014).
toxicity of the co-culture supernatants, which On the other hand, it was observed that some
was not observed with inulin (Valdés-Varela microorganisms release metabolites that are able
et al. 2016b). to inhibit the harmful effects of toxins. A
170 L. Valdés-Varela et al.

bacterial cell-free supernatant obtained from reduce the toxic effect of the pathogen; more
L. delbrueckii subsp. bulgaricus LDB B-30892 specifically, the strain B. longum IPLA20022, in
reduced cytotoxic effects of C. difficile a viable state, showed the highest ability to
ATCC9689 upon the human intestinal epithelial reduce the levels of both clostridial toxins and
cell line Caco-2 (Banerjee et al. 2009). These to counteract the cytotoxic effect upon HT29
authors suggested that bioactive components, of (Valdés-Varela et al. 2016a). Furthermore, the
unknown nature, were released by this strain incubation of supernatant from B. longum
which were the probable causative agents of IPLA20022 with the toxigenic C. difficile super-
inhibition of the clostridial toxins. Similarly, natant showed similar effect on the cell line than
bacterial cell-free supernatants obtained from that obtained with the bifidobacterial biomass.
L. lactis CIDCA8221 contained heat-sensitive The treatment of the clostridial supernatant with
metabolites, higher than 10 kDa, that were not this probiotic strain prevented the rounding of
affected by treatment with different proteases or HT29 cells, detected in cells treated only with
proteases-inhibitors, which were able to inhibit C. difficile supernatant, thus keeping a monolayer
cytotoxic effects of C. difficile toxins upon epi- structure resembling that of the control
thelial Vero cells (Bolla et al. 2013). These (non-treated HT29) (Fig. 3). Taking into account
results suggest that the protective effect of these results we hypothesize that the adsorption
L. lactis CIDCA8221 supernatant could be of toxins to the bifidobacterial surface and/or the
owing to a non-covalent interaction between secretion of molecules able to reduce the cyto-
molecules present in the lactococcal supernatant toxic effect by degrading the toxins are both
and toxins. In this regard, surface components of probable mechanisms of action (Valdés-Varela
the bacterial cell envelope, such as exopolysac- et al. 2016a). In this regard, 20 years ago it had
charides which can be released to the environ- been reported that S. boulardii inhibited C. difficile
ment, have been proposed to in vitro inhibit of TcdA effects in the rat ileum by releasing a 54-kDa
adverse effect of pathogenic toxins (Ruas- serine protease which hydrolyzed toxin A and its
Madiedo et al. 2010). A study showed the ability intestinal receptor (Castagliuolo et al. 1996); this
of the outermost (proteinaceous) S-layer from could be the mechanism behind the effectiveness of
Lactobacillus kefir strains to inhibit the damage this yeast in both, the prevention and the treatment
induced by supernatants obtained from of antibiotic-associated colitis in humans
C. difficile upon Vero cells; the protective effect (Castagliuolo et al. 1999). More recently it was
was not affected by inhibitors of proteases or heat observed that a protease secreted by Bacillus
treatment, while pre-incubation with specific clausii O/C is able to inhibit the cytotoxic effect
anti-S-layer antibodies reduced the inhibitory of C. difficile, thus this enzyme could be involved
effect of these proteins (Carasi et al. 2012). in the protective effect of this bacilli in antibiotic-
From this study it was concluded that the capa- associated diarrhea (Ripert et al. 2016). A similar
bility for reducing the toxigenic effect of phenomenon may be taking place with the above
C. difficile could be attributed to an interaction mentioned Bifidobacterium strains (Valdés-Varela
between its toxins and the L. kefir S-layer protein et al. 2016a).
(Carasi et al. 2012). Recently, our workgroup
analyzed the capability of Bifidobacterium and
Lactobacillus strains to reduce the toxic effect of 4.3 Other Mechanisms of Action
supernatants obtained from C. difficile
LMG21717 (TcdA+, TcdB+) culture upon the The intestinal barrier function given, among
human intestinal epithelial cell line HT29. For other factors, by the presence of an intact intesti-
this purpose, the probiotic candidates were nal epithelium enabling the absorption of
incubated together with a toxigenic supernatant nutrients and the exclusion of harmful substances
of C. difficile and the analyzed strains from can be compromised by the activity of enteric
B. longum and B. breve species were able to pathogens including C. difficile (Barreau and
Probiotics for Prevention and Treatment of Clostridium difficile Infection 171

Fig. 3 CSLM (Leica TCSAOBS SP8 X confocal micros- white laser (excited at 578 nm, showing phalloidind-
copy) images obtained, after 20 h incubation, for HT29 alexa-fluor-568 stained F-actin), and the “green” image
cells submitted to different treatments. (a) panel shows resulting from the auto-fluorescence emitted by the intra-
transmission (visible) images and (b) panel shows cellular components of HT29. The 63x/1.4 oil objective
Z-projection snapshots resulting from a combination of was used; bars 10 μm. Individual images of stained
the transmission image with the “blue” image, captured nucleus and/or F-actin were included in the reference
with the violet laser diode (excited at 405 nm, showing Valdes-Varela et al. (2016a)
DAPI stained nucleus), the “red” image, captured with the

Hugot 2014). In fact, internalized clostridial C. difficile ribotypes, thus indicating that this
toxins induce changes in the F-actin cytoskeleton probiotic can prevent intestinal damage and
and a breakdown of the tight junctions, thus inflammation (Koon et al. 2016). In fact, after a
contributing to the disruption of the epithelial literature search conducted by Stier and Bischoff
barrier function; the increase in the permeability (2016) they found that mechanisms of S. boulardii
of this barrier ends with an inflammatory process action involve not only a direct effect on the patho-
due to the infiltration of neutrophils, production gen or its toxins, but also impact on the innate and
of chemokines and pro-inflammatory cytokines, adaptive immune response of the host induced after
and activation of mast cells and lymphocytes CDI. Regarding probiotic bacteria, it has been
among other events (Voth and Ballard 2005; shown that L. rhamnosus L34 and L. casei L39
Rupnik et al. 2009; Abt et al. 2016). Thus some are able to modulate, by different ways, the inflam-
probiotics have been claimed to be able to rein- mation caused by C. difficile, thus making suitable
force the intestinal barrier function, although the use of these vancomycin-resistant lactobacilli
there is not much information in the context of for treating CDI (Boonma et al. 2014). In our
CDI (Fig. 2d). In a hamster model of CDI, the research group we have detected that lactobacilli
oral administration of live S. boulardii 5-days strains are able to increase the synthesis of interleu-
before the infection significantly reduced cecal kin (IL)-8 and mucins by HT29-MTX monolayers
tissue damage, NF-κB phosphorylation and challenged with C. difficile, thus helping to the
TNFα protein expression caused by different reinforcement of the innate immune defense
172 L. Valdés-Varela et al.

(Zivkovic et al. 2015). More recently, a combina- pathogen and/or to degrade the produced toxins.
tion of Lactobacillus helveticus BGRA43, Lactoba- This inhibition of C. difficile toxicity may consti-
cillus fermentum BGHI14 and S. thermophilus tute an interesting strategy for the treatment of
BGVLJ1–44 was in vitro tested against C. difficile CDI by probiotics; first by eliminating the toxins
in a Caco-2 model and results showed an increase in from the intestine and, secondly, by the promo-
the release of transforming growth factor (TGF)-β, tion of the microbiota restoration by the use of
thus resulting in a promising probiotic candidate to selected probiotic strains with both properties.
be further evaluated against CDI (Golic et al. 2017). The existing clinical interest of CDI together
Finally, recombinant lactobacilli, although with the successful application of FMT, allow
they cannot be considered as probiotics, could foreseeing that the interest in the use for probi-
be suitable vehicles for the in situ production and otic therapies, likely using defined combinations
delivery of therapeutic molecules in the intestine. of strains, will continue rising during the next
In a recent study, it was explored the basis for an years. In this regard the development of products,
oral anti-toxin strategy based on engineered Lac- based on the combination of strains with differ-
tobacillus strains expressing TcdB-neutralizing ent properties and anti-C. difficile mechanisms of
antibody fragments in the gastrointestinal tract; action, promises to allow the development of
the results showed that only lactobacilli highly efficacy products for both prevention and
displaying the anti-TcdB variable domain of the treatment of CDI.
heavy chain antibody can inhibit the cytotoxic
effect of TcdB in the gastrointestinal tract of a Acknowledgements The funds supporting this research
hamster model (Andersen et al. 2016). topic in our group are given by the Spanish Ministry of
Economy and Competiveness (current project AGL2015-
64901-R) partially co-funded by FEDER (European
Union) grants.
5 Conclusion and Future Trends

The search for probiotics with anti-C. difficile References

activity has been an active area of research for
more than two decades. However, in spite of the Abt MC, McKenney PT, Pamer EG (2016) Clostridium
abundance of in vitro studies, the in vivo evi- difficile colitis: pathogenesis and host defense. Nat
Rev Microbiol 14:609–620
dence is less conclusive. The role of probiotics in
Allen SJ, Wareham K, Wang D et al (2013) Lactobacilli
preventing antibiotics-associated diarrhea is well and bifidobacteria in the prevention of antibiotic-
established by several clinical intervention associated diarrhoea and Clostridium difficile diar-
studies and meta-analyses. Good evidence is rhoea in older inpatients (PLACIDE): a randomised,
double-blind, placebo-controlled, multicentre trial.
also available regarding the benefit of certain Lancet 382:1249–1257
probiotics in the prevention of specific Ambalam P, Kondepudi KK, Balusupati P et al (2015)
C. difficile diarrhea, being still necessary to Prebiotic preferences of human lactobacilli strains in
define the best conditions for maximizing the co-culture with bifidobacteria and antimicrobial activ-
ity against Clostridium difficile. J Appl Microbiol
efficacy. However, the studies on the use of
probiotics in the treatment of CDI are still scarce; Andersen KK, Strokappe NM, Hultberg A et al (2016)
this is in spite of the several potential Neutralization of Clostridium difficile toxin B
mechanisms of action that would be of interest mediated by engineered lactobacilli that produce
single-domain antibodies. Infect Immun 84:395–406
in the case of C. difficile infection. Among them,
Arruda PHE, Madson DM, Ramirez A et al (2016) Bacte-
the ability of certain strains to inhibit the growth rial probiotics as an aid in the control of Clostridium
of C. difficile, or to promote the restoration of the difficile disease in neonatal pigs. Can Vet J
normal gut microbiota, represent two very direct 57:183–188
Arvola T, Laiho K, Torkkeli S et al (1999) Prophylactic
potentially beneficial mechanisms of action.
Lactobacillus GG reduces antibiotic-associated diar-
Moreover, specific probiotic strains have been rhea in children with respiratory infections: a
found to be able to reduce the toxicity of this randomized study. Pediatrics 104:64
Probiotics for Prevention and Treatment of Clostridium difficile Infection 173

Auclair J, Frappier M, Millette M (2015) Lactobacillus Dicks LMT, Botha M, Loos B et al (2015) Adhesion of
acidophilus CL1285, Lactobacillus casei LBC80R, Lactobacillus reuteri strain Lr1 to equine epithelial
and Lactobacillus rhamnosus CLR2 (Bio-K+): char- cells and competitive exclusion of Clostridium diffi-
acterization, manufacture, mechanisms of action, and cile from the gastro-intestinal tract of horses. Ann
quality control of a specific probiotic combination for Microbiol 65:1087–1096
primary prevention of Clostridium difficile infection. Dietrich CG, Kottmann T, Alavi M (2014) Commercially
Clin Infect Dis 60:S135–S143 available probiotic drinks containing Lactobacillus
Banerjee P, Merkel GJ, Bhunia AK (2009) Lactobacillus casei DN-114001 reduce antibiotic-associated diar-
delbrueckii ssp. bulgaricus B-30892 can inhibit cyto- rhea. World J Gastroenterol 20:15837–15844
toxic effects and adhesion of pathogenic Clostridium Donaldson GP, Lee SM, Mazmanian SK (2016) Gut bio-
difficile to Caco-2 cells. Gut Pathog 1:8 geography of the bacterial microbiota. Nat Rev
Barreau F, Hugot JP (2014) Intestinal barrier dysfunction Microbiol 14:20–32
triggered by invasive bacteria. Curr Opin Microbiol Ferreira CL, Grześkowiak Ł, Collado MC et al (2011) In
17:91–98 vitro evaluation of Lactobacillus gasseri strains of
Best EL, Freeman J, Wilcox MH (2012) Models for the infant origin on adhesion and aggregation of specific
study of Clostridium difficile infection. Gut Microbes pathogens. J Food Prot 74:1482–1487
3:145–167 Forssten SD, R€ oytió H, Hibberd AA et al (2015) The
Bolla PA, Carasi P, Serradell MA et al (2013) Kefir- effect of polydextrose and probiotic lactobacilli in a
isolated Lactococcus lactis subsp. lactis inhibits the Clostridium difficile-infected human colonic model.
cytotoxic effect of Clostridium difficile in vitro. J Microb Ecol Health Dis 26:27988
Dairy Res 80:96–102 Fredua-Agyeman M, Stapleton P, Basit AW et al (2017)
Boonma P, Spinler JK, Venable SF et al (2014) Lactoba- In vitro inhibition of Clostridium difficile by commer-
cillus rhamnosus L34 and Lactobacillus casei L39 cial probiotics: a microcalorimetric study. Int J Pharm
suppress Clostridium difficile-induced IL-8 production 517:96–103
by colonic epithelial cells. BMC Microbiol 14:177 Gagnon M, Zihler Berner A, Chervet N et al (2013)
Can M, Besirbellioglu BA, Avci IY et al (2006) Prophy- Comparison of the Caco-2, HT29 and the mucus-
lactic Saccharomyces boulardii in the prevention of secreting HT29-MTX intestinal cell models to inves-
antibiotic associated diarrhea: a prospective study. tigate Salmonella adhesion and invasion. J Microbiol
Med Sci Monit 12:19–22 Methods 94:274–279
Carasi P, Trejo FM, Pérez PF et al (2012) Surface proteins Gao XW, Mubasher M, Fang CY et al (2010) Dose-
from Lactobacillus kefir antagonize in vitro cytotoxic response efficacy of a proprietary probiotic formula
effect of Clostridium difficile toxins. Anaerobe of Lactobacillus acidophilus CL1285 and Lactobacil-
18:135–142 lus casei LBC80R for antibiotic associated diarrhea
Carter GP, Rood JI, Lyras D (2012) The role of toxin A and Clostridium difficile-associated diarrhea prophy-
and toxin B in the virulence of Clostridium difficile. laxis in adults patients. Am J Gastroenterol
Trends Microbiol 20:21–29 105:1636–1641
Castagliuolo I, Lamont JT, Nikulasson ST (1996) Saccha- Gareau MG, Sherman PM, Walker WA (2010) Probiotics
romyces boulardii protease inhibits Clostridium diffi- and the gut microbiota in intestinal health and disease.
cile Toxin A effects in the rat ileum. Infect Immun Nat Rev Gastroentol Hepatol 7:503–514
64:5225–5232 Gerding DN, JohnsonS RM et al (2014) Clostridium diffi-
Castagliuolo I, Rieger MF, Valenick L et al (1999) Sac- cile binary toxin CDT: mechanism, epidemiology, and
charomyces boulardii protease inhibits the effect of potential clinical importance. Gut Microb 5:15–27
Clostridium difficile toxins A and B in human colonic Goldenberg JZ, Ma SSY, Saxton JD et al (2013)
mucosa. Infect Immun 67:302–307 Probiotics for the prevention of Clostridium difficile-
Collado MC, Gueimonde M, Hernández M et al (2005) associated diarrhea in adults and children. Cochrane
Adhesion of selected Bifidobacterium strains to Database Syst Rev 5:CD006095
human intestinal mucus and its role in enteropathogen Goldenberg JZ, Lytvyn L, Steurich J et al (2015)
exclusion. J Food Protect 68:2672–2678 Probiotics for prevention of pediatric antibiotic-
Corr SC, Li Y, Riedel CU et al (2007) Bacteriocin pro- associated diarrhea. Cochrane Database Syst Rev 12:
duction as a mechanism for the anti-infective activity CD004827
of Lactobacillus salivarius UCC118. Proc Natl Acad Golić N, Veljović K, Popović N et al (2017) In vitro and
Sci U S A 104:7617–7621 in vivo antagonistic activity of new probiotic culture
Cremonini F, Caro S, Nista EC et al (2002) Meta-analysis: against Clostridium difficile and Clostridium
the effects of probiotic administration on antibiotic perfringens. BMC Microbiol 17:108
associated diarrhoea. Aliment Pharmacol Ther Hempel S, Newberry SJ, Maher AR et al (2012)
16:1461 Probiotics for the prevention and treatment of
D’Souza AL, Rajkumar C, Cooke J et al (2002) Probiotics antibiotic-associated diarrhea: a systematic review
in the prevention of antibiotic associated diarrhoea: and meta-analysis. JAMA 307:1959–1969
meta-analysis. Br Med J 324:1361
174 L. Valdés-Varela et al.

Hickson M (2011) Probiotics in the prevention of Leal JR, Heitman SJ, Conly JM et al (2016) Cost-
antibiotic-associated diarrhea and Clostridium difficile effectiveness analysis of the use of probiotics for the
infection. Ther Adv Gastroenterol 4:185–197 prevention of Clostridium difficile-associated diarrhea
Hill C, Guarner F, Reid G et al (2014) The International in a provincial health care system. Infect Cont Hosp
Scientific Association for Probiotics and Prebiotics Epidemiol 37:1079–1086
consensus statement on the scope and appropriate Lee JS, Chung MJ, Seo JG (2013) In vitro evaluation of
use of the term probiotic. Nat Rev Gastroenterol antimicrobial activity of lactic acid bacteria against
Hepatol 11:506–514 Clostridium difficile. Toxicol Res 29:99–106
Hopkins MJ, Macfarlane GT (2003) Nondigestible Leffler DA, Lamont JT (2015) Clostridium difficile infec-
oligosaccharides enhance bacterial colonization resis- tion. N Engl J Med 372:1539–1548
tance against Clostridium difficile in vitro. Appl Envi- Mansour NM, Elkhatib WF, Aboshnad KM et al (2017)
ron Microbiol 69:1920–1927 Inhibition of Clostridium difficile in mice using a
Hussack G, Tanha J (2016) An update on antibody-based mixture of potential probiotic strains Enterococcus
immunotherapies for Clostridium difficile infection. faecalis NM815, E. faecalis NM915, and E. faecium
Clin Exp Gastroenterol 9:209–224 NM1015: novel candidates to control C. difficile infec-
Hutton ML, Mackin KE, Chakravorty A et al (2014) tion (CDI). Probiotics Antimicrob Prot. https://doi.
Small animal models for the study of Clostridium org/10.1007/s12602-017-9285-7. [E-pub ahead of
difficile disease pathogenesis. FEMS Microbiol Lett print]
352:140–149 Martin J, Wilcox M (2016) New and emerging therapies
Jangi S, Lamont JT (2010) Asyntomatic colonization by for Clostridium difficile infection. Curr Opin Infect
Clostridium difficile in infants: implications for dis- Dis 29:546–554
ease in later life. J Pediatr Gastroenterol Nutr 51:2–7 Mathur H, Rea MC, Cotter PD et al (2014) The potential
Johnston B, Supina A, Ospina M et al (2007) Probiotics for emerging therapeutic options for Clostridium diffi-
for the prevention of pediatric antibiotic-associated cile infection. Gut Microb 5:696–710
diarrhea. Cochrane Database Syst Rev 18:CD004827 Maziade PJ, Pereira P, Goldstein EJC (2015) A decade of
Kachrimanidou M, Sarmourli T, Skoura L et al (2016) experience in primary prevention of Clostridium diffi-
Clostridium difficile infection: new insights into ther- cile infection at a community hospital using the probi-
apeutic options. Crit Rev Microbiol 42:773–779 otic combination Lactobacillus acidophilus CL1285,
Kociolek LK, Gerding DN (2016) Breakthroughs in the Lactobacilus casei LBC80R, and Lactobacillus
treatment and prevention of Clostridium difficile rhamnosus CRL2 (Bio-K+). Clin Infect Dis 60(Suppl
infection. Nat Rev Gastroenterol Heptol 13:150–160 2):S144–S147
Kolling GL, Wu M, Warren CA et al (2012) Lactic acid McFarland LV (2006) Meta-analysis of probiotics for the
production by Streptococcus thermophilus alters Clos- prevention of antibiotic associated diarrhea and the
tridium difficile infection and in vitro toxin A produc- treatment of Clostridium difficile disease. Am J
tion. Gut Microb 3:523–529 Gastroenterol 101:812–822
Kondepudi KK, Ambalam P, Karagin PH et al (2014) A McFarland LV (2015) Probiotics for the primary and
novel multi-strain probiotic and synbiotic supplement secondary prevention of C. difficile infections. A
for prevention of Clostridium difficile infection in a meta-analysis and systematic review. Antibiotics
murine model. Microbiol Immunol 58:552–558 4:160–178
Koon HW, Su B, Xu C et al (2016) Probiotic Saccharo- McFarland LV (2016) Therapies on the horizon for Clos-
myces boulardii CNCM I-745 prevents outbreak- tridium difficile infections. Expert Opin Investig
associated Clostridium difficile-associated cecal Drugs 25:541–555
inflammation in hamsters. Am J Physiol Gastroenterol McFarland LV, Surawicz CM, Greenberg RN et al (1994)
Liver Physiol 311:G610–G623 A randomized placebo-controlled trial of Saccharo-
Kotowska M, Albrecht P, Szajewska H (2005) Saccharo- myces boulardii in combination with standard
myces boulardii in the prevention of antibiotic- antibiotics for Clostridium difficile disease. J Am
associated diarrhea in children: a randomized Med Assoc 271:1913
double-blind placebo-controlled trial. Aliment Moura I, Spigaglia P, Barbanti F et al (2013) Analysis of
Pharmacol Ther 21:583–590 metronidazole susceptibility in different Clostridium
Lau CSM, Chamberlaim RS (2016) Probiotics are effec- difficile PCR ribotypes. J Antimicrob Chemother
tive at preventing Clostridium difficile-associated 68:362–365
diarrhea: a systematic review and meta-analysis. Int J Na X, Kelly C (2011) Probiotics in Clostridium difficile
Gen Med 9:27–37 infection. J Clin Gastroenterol 45:S154–S158
Lawley TD, Walker AW (2013) Intestinal colonization Ng SC, Hart AL, Kamm MA et al (2009) Mechanisms of
resistance. Immunology 138:1–11 action of probiotics: recent advances. Inflamm Bowel
Le Lay C, Fernandez B, Hammami R et al (2015) On Dis 15:300–310
Lactococcus lactis UL719 competitivity and nisin O’Horo JC, Jindai K, Kunzer B et al (2014) Treatment of
(Nisaplin®) capacity to inhibit Clostridium difficile recurrent Clostridium difficile infection: a systematic
in a model of human colon. Front Microbiol 6:1020 review. Infection 42:43–59
Probiotics for Prevention and Treatment of Clostridium difficile Infection 175

Ofosu A (2016) Clostridium difficile infection: a review of Schoster A, Kokotovic B, Permin A (2013) In vitro inhibi-
current and emerging therapies. Ann Gastroenterol tion of Clostridium difficile and Clostridium perfringens
29:147–154 by commercial probiotic strains. Anaerobe 20:36–41
Ollech JE, Shen NT, Crawford CV et al (2016) Use of Shen NT, Maw A, Tmanova LL et al (2017) Timely use of
probiotics in prevention and treatment of patients with probiotics in hospitalized adults prevents Clostridium
Clostridium difficile infection. Best Pract Res Clin difficile infection: a systematic review with meta-
Gastroenterol 30:111–118 regression analysis. Gastroenterology 52:1889–1900
Orrell KE, Zhanga Z, Sugiman-Marangosa SN et al Sinclair A, Xie X, Saab L et al (2016) Lactobacillus
(2017) Clostridium difficile toxins A and B: receptors, probiotics in the prevention of diarrhea associated
pores, and translocation into cells. Crit Rev Biochem with Clostridium difficile: a systematic review and
Mol Biol 52:461–473 bayesian hierarchical meta-analysis. CMAJ Open 4:
Ozaki E, Kato H, Kita H et al (2004) Clostridium difficile E706–E718
colonization in healthy adults: transient colonization Spinler JK, Brown A, Ross CL et al (2016) Administra-
and correlation with enterococal colonization. J Med tion of probiotic kefir to mice with Clostridium diffi-
Microbiol 53:167–172 cile infection exacerbates disease. Anaerobe 40:54–57
Padua D, Pothoulakis C (2016) Novel approaches to Starn ES, Hampe H, Cline T (2016) The cost-efficiency
treating Clostridium difficile-associated colitis. Expert and care effectiveness of probiotic administration with
Opin Investig Drugs 10:193–204 antibiotics to prevent hospital-acquired Clostridium
Parkes GC, Sanderson JD, Whelan K (2009) The difficile infection. Qual Manag Health Care
mechanisms and efficacy of probiotics in the preven- 25:238–243
tion of Clostridium difficile-associated diarrhoea. Lan- Stier H, Bischoff SC (2016) Influence of Saccharomyces
cet Infect Dis 9:237–244 boulardii CNCM I-745 on the gut-associated immune
Rätsep M, Kõljalg S, Sepp E et al (2017) A combination system. Clin Exp Gastroenterol 9:269–279
of the probiotic and prebiotic product can prevent the Sun X, Wang H, Zhang Y et al (2011) Mouse relapse
germination of Clostridium difficile spores and infec- model of Clostridium difficile infection. Infect Immun
tion. Anaerobe 47:94–103 79:2856–2864
Reid G, Younes JA, Van der Mei HC et al (2011) Surawicz CM, McFarland LV, Greenberg RN et al (2000)
Microbiota restoration: natural and supplemented The search for a better treatment for recurrent Clos-
recovery of human microbial communities. Nat Rev tridium difficile disease: use of high dose vancomycin
Microbiol 9:27–38 combined with Saccharomyces boulardii. Clin Infect
Ripert G, Racedo SM, Elie AM et al (2016) Secreted Dis 31:1012–1017
compounds of the probiotic Bacillus clausii strain Szajewska H, Skorka A, Dylag M (2007a) Meta-analysis:
O/C inhibit the cytotoxic effects induced by Clostrid- Saccharomyces boulardii for treating acute diarrhoea
ium difficile and Bacillus cereus toxins. Antimicrob in children. AP&T 25:257–264
Agents Chemother 60:3445–3454 Szajewska H, Skorka A, Ruszczynski M et al (2007b)
Rodriguez C, Taminiau B, Van Broeck J et al (2016) Meta-analysis: Lactobacillus GG for treating acute
Clostridium difficile in food and animals: a compre- diarrhoea in children. AP&T 25:871–881
hensive review. Adv Exp Med Biol 4:65–92 Szajewska H, Canani RB, Guarino A et al (2016) Probiotics
Ruas-Madiedo P, Gueimonde M, Margolles A et al (2006) for the prevention of antibiotic-associated diarrhea in
Exopolysaccharides produced by probiotic strains mod- children. J Pediatr Gastroenterol Nutr 62:495–506
ify the adhesion of probiotics and enteropathogens to Tasteyre A, Barc MC, Karjalainen T et al (2002) Inhibition
human intestinal mucus. J Food Prot 69:2011–2015 of in vitro cell adherence of Clostridium difficile by
Ruas-Madiedo P, Medrano M, Salazar N et al (2010) Saccharomyce boulardii. Microb Pathog 32:219–225
Exopolysaccharides produced by Lactobacillus and Tejero-Sariñena S, Barlow J, Costabile A et al (2013)
Bifidobacterium strains abrogate in vitro the cytotoxic Antipathogenic activity of probiotics against Salmo-
effect of bacterial toxins on eukaryotic cells. J Appl nella Typhimurium and Clostridium difficile in anaer-
Microbiol 109:2079–2086 obic batch culture systems: is it due to synergies in
Rupnik M, Wilcox MH, Gerding DN (2009) Clostridium probiotic mixtures or the specificity of single strains?
difficile infection: new developments in epidemiology Anaerobe 24:60–65
and pathogenesis. Nat Rev Microbiol 7:526–536 Trejo FM, Minnaard J, Pereza PF et al (2006) Inhibition
Sampalis J, Psaradellis E, Ranpakakis E (2010) Efficacy of Clostridium difficile growth and adhesion to
of BIO K+ CL1285 in the reduction of antibiotic enterocytes by Bifidobacterium supernatants. Anaer-
associated diarrhea- a placebo controlled double obe 12:186–193
blind randomized, multi-centre study. Arch Med Sci Trejo FM, Pérez PF, De Antoni GL (2010) Co-culture
6:56–64 with potentially probiotic microorganisms antagonises
Sazawal S, Hiremath G, Dhingra U et al (2006) Efficacy virulence factors of Clostridium difficile in vitro.
of probiotics in prevention of acute diarrhoea. A meta- Antonie Van Leeuwenhoek 98:19–29
analysis of masked, randomized, placebo-controlled Trejo FM, De Antoni GL, Pérez PF (2013) Protective
trials. Lancet Infect Dis 6:374–382 effect of bifidobacteria in an experimental model of
176 L. Valdés-Varela et al.

Clostridium difficile associated colitis. J Dairy Res Woodworth MH, Carpentieri C, Sitchenko KL et al
80:263–269 (2017) Challenges in fecal donor selection and screen-

Unal CM, Steinert M (2016) Novel therapeutic strategies ing for fecal microbiota transplantation: a review. Gut
for Clostridium difficile infections. Expert Opin Microb 8:225–237
Investig Drugs 20:269–285 Wullt M, Hagslatt ML, Odenholt I (2003) Lactobacillus
Valdés L, Gueimonde M, Ruas-Madiedo P (2015) Moni- plantarum 299v for the treatment of recurrent Clos-
toring in real time the cytotoxic effect of Clostridium tridium difficile-associated diarrhoea: a double-blind,
difficile upon the intestinal epithelial cell line HT29. J placebo-controlled trial. Scand J Infect Dis
Microbiol Method 119:66–73 35:365–367
Valdés-Varela L, Alonso-Guervos M, Garcı́a-Suárez O Xi B, Yu N, Wang X et al (2008) The application of cell-
et al (2016a) Selection of bifidobacteria and based label-free technology in drug discovery.
lactobacilli able to antagonise the cytotoxic effect of Biotechnol J 3:484–495
Clostridium difficile upon intestinal epithelial HT29 Yakob L, Riley TV, Paterson DL et al (2015) Mechanisms
monolayer. Front Microbiol 7:577 of hypervirulent Clostridium difficile ribotype 027 dis-
Valdés-Varela L, Hernández-Barranco AM, Ruas- placement of endemic strains: an epidemiological
Madiedo P et al (2016b) Effect of Bifidobacterium model. Sci Rep 5:12666
upon Clostridium difficile growth and toxicity when Young VB (2017) Old and new models for studying host-
co-cultured in different prebiotic substrates. Front microbe interactions in health and disease: C. difficile
Microbiol 7:738 as an example. Am J Physiol Gastrointest Liver
Vanderhoof J, Whitney D, Antonson D et al (1999) Lac- Physiol 312:G623–G627
tobacillus GG in the prevention of antibiotic Yu H, Chen K, Wu J et al (2015) Identification of toxemia
associated diarrhea in children. J Pediatr 135:564–568 in patients with Clostridium difficile infection. PLoS
Venema K, van den Abbeele P (2013) Experimental One 10:e0124235
models of the gut microbiome. Best Pract Res Clin Yun B, Oh S, Griffiths MW (2014) Lactobacillus aci-
Gastroenterol 27:115–126 dophilus modulates the virulence of Clostridium diffi-
Vernaya M, McAdam J, Hamptom MD (2017) Effective- cile. J Dairy Sci 97:4745–4758
ness of probiotics in reducing the incidence of Clos- Zivkovic M, Hidalgo-Cantabrana C, Kojic M et al (2015)
tridium difficile-associated diarrhea in elderly Capability of exopolysaccharide-producing Lactoba-
patients: a systematic review. JBI Database Syst Rev cillus paraplantarum BGCG11 and its non-producing
Impl Rep 15:140–164 isogenic strain NB1, to counteract the effect of
Voth DE, Ballard JD (2005) Clostridium difficile toxins: enteropathogens upon the epithelial cell line HT29-
mechanism of action and role in disease. Clin MTX. Food Res Int 74:199–207
Microbiol Rev 18:247–263
Faecal Microbiota Transplantation
as Emerging Treatment in European

Marcello Maida, James Mcilroy, Gianluca Ianiro,

and Giovanni Cammarota

Abstract with promising results. The aim of future

Clostridium difficile infection (CDI) is one of research is therefore to standardize protocols
the most common healthcare-associated and develop FMT as a therapeutic option for
infections in the world and is a leading cause these patients.
of morbidity and mortality in hospitalized This review summarizes data on the use of
patients. FMT as a treatment for CDI and IBD, with
Although several antibiotics effectively special attention given to studies conducted in
treat CDI, some individuals do not respond European countries.
to these drugs and may be cured by
transplanting stool from healthy donors. This Keywords
procedure, termed Faecal Microbiota Trans- Clostridium difficile · European · Faecal
plantation (FMT), has demonstrated remark- microbiota transplantation · Fecal ·
able efficacy as a treatment for recurrent CDI. Inflammatory bowel disease
FMT has also been investigated in other
diseases and disorders where perturbations to
the gut microbiota have been theorized to play
a causative role in pathogenesis and severity, 1 Introduction
such as inflammatory bowel disease (IBD).
Although FMT is currently not recommended Gut microbiota is critical to health and functions
to cure IBD patients in clinical practice, sev- and therefore emerges as a “virtual” organ with a
eral studies have recently been carried out level of complexity comparable to that of any other
organ system. Fecal microbiota transplantation
(FMT) is a medical treatment that aims to restore
the normal gut microbiota in diseases or infections
All authors contributed to writing the paper and had full associated with bacterial imbalances. FMT has the
control over preparation of manuscript; all authors
approved the final draft manuscript.
M. Maida
Section of Gastroenterology, S.Elia – Raimondi Hospital,
G. Ianiro · G. Cammarota (*)
Caltanissetta, Italy
Gastroenterological Area, Fondazione Policlinico
J. Mcilroy Universitario Gemelli, Università Cattolica del Sacro
School of Medicine, Medical Sciences and Nutrition, Cuore, Rome, Italy
University of Aberdeen, Aberdeen, UK e-mail:

# Springer International Publishing AG 2018 177

P. Mastrantonio, M. Rupnik (eds.), Updates on Clostridium difficile in Europe, Advances
in Experimental Medicine and Biology 1050,
178 M. Maida et al.

potential to compete with powerful antibiotics as a incidences of CDI in 2011 (Lessa et al. 2015)
treatment strategy in several gastrointestinal CDIs, 83,000 cases of first recurrences and an
disorders. Clostridium difficile infection (CDI) is estimated number of deaths of 29,300 only in
one of the most common healthcare-associated 2011 (Lessa et al. 2015). In Europe the extent
infections in the world and is a leading cause of of CDI is less clear. The burden of healthcare-
morbidity and mortality in hospitalized patients. associated CDIs in acute care hospitals has been
Although several antibiotics effectively treat CDI, estimated at 123,997 cases annually with a mor-
some individuals do not respond to these drugs and tality of 3700 per year (European Surveillance of
may be cured by FMT, which has demonstrated CDI 2015). A prospective study conducted in
extraordinary efficacy for the cure of recurrent 2005 in 38 hospitals in 14 different European
CDI (rCDI). FMT has also been investigated in countries reported a mean incidence of nosoco-
other diseases and disorders where perturbations to mial CDI of 2.45 per 10,000 patient-days (range
the gut microbiota have been theorized to play a 0.1–7.1) (Barbut et al. 2007). Beside this, a more
causative role in pathogenesis and severity, such as recent and larger hospital-based survey
inflammatory bowel disease (IBD) (Ianiro et al. performed through a network of 97 hospitals
2014; Cammarota et al. 2015a). The current thera- from 34 European countries, reported a higher
peutic options for IBD have limitations with regards CDI incidence of 4.1 per 10,000 patient days
to cost, safety profile and the onset of drug resis- (Bauer et al. 2011).
tance and dependence. There is therefore a need to Similar epidemiological data are observed in the
develop novel therapeutic avenues that are both safe eastern countries. A meta-analysis of 51 studies,
and effective to control the disease. Although FMT showed similar rates of CDI in Asia compared to
is currently not recommended to cure IBD patients Europe and North America (Borren et al. 2017).
in clinical practice (Cammarota et al. 2017), several Beside this, epidemiological trends show that
studies have recently been carried out with the incidence of CD has increased over recent
promising results. The aim of future research is decades. In the United States, reported cases of
therefore to standardize protocols and develop CDI doubled from 2000 to 2010 and are expected
FMT as a therapeutic option for these patients. to increase further (Lessa et al. 2015). A recent
This review summarizes data on the use of FMT retrospective cohort study that analysed more
for the treatment of both CDI and IBD, with special than 38 billion commercially insured patients in
attention given to studies carried out in European the United States showed that between the years
countries. of 2001 and 2012, the annual incidence of CDI
and multiply recurrent CDI (mrCDI) per 1000
person-years increased by 42.7% (from 0.4408
to 0.6289 case) and 188.8% (from 0.0107 to
2 Faecal Microbiota
0.0309 case) respectively (Ma et al. 2017). How-
Transplantation for Clostridium
ever, it should be noted that these results may be
difficile Infection
biased by the selection of the only insured
2.1 The Burden of C. difficile
This raising in incidence and virulence of CD
can been explained, at least in part, by inappro-
CDI is the most common cause of hospital
priate antibiotic usage, outbreaks of CDI in
associated diarrhoea in the western world and is
healthcare facilities, and the diffusion of
one of the leading causes of morbidity and mor-
fluoroquinolone-resistant strains belonging to
tality in hospitalized patients globally
the PCR-ribotype 027 (Warny et al. 2005;
(Bagdasarian et al. 2015). CDI is highly
McDonald et al. 2005).
prevalent in North America and Europe. A
CDI infection is also palaces a significant
population-based study performed in the United
economic burden on the health services. A recent
States reported that there were 453,000
Faecal Microbiota Transplantation as Emerging Treatment in European Countries 179

analysis of health-care associated infections in 2.2 Faecal Microbiota

the United States ranked CDI fourth in terms of Transplantation and C. difficile
attributable costs and length of hospital stay
(Zimlichman et al. 2013). In recent decades, FMT has been trialed as a
The bacterium Clostridium difficile (CD) is treatment for rCDI and, over the years, a consid-
spread via the faecal-oral route. CDI generally erable body of evidence has emerged in support
requires two things: the presence (endogenous of its effectiveness. Consequently, FMT is
infection) or acquisition (exogenous infection) recommended as a treatment option for rCDI in
of CD and an altered composition of gut guidelines produced by the European Society for
microbiota. Risk factors facilitating infection Microbiology and Infectious Disease and the
are older age, hospitalization, recent use of American College of Gastroenterology
antibiotics, long-term therapy with proton pump (Surawicz et al. 2013; Debast et al. 2014). Fur-
inhibitors and chronic kidney disease (Asha et al. thermore, a recent European consensus confer-
2006; Mullane et al. 2013; Stevens et al. 2011). ence on FMT was held with the aim of
Once the bacterium is present in the large standardizing FMT guidance across Europe.
intestine it proliferates, taking advantage of an According to the statements of the conference,
impaired gut microbiota. The production of FMT is recommended as treatment option for
toxins create its main virulence factors. Toxin both mild and severe rCDI (Cammarota et al.
A (TcdA) and B (TcdB) induce mucosal inflam- 2017).
mation, disruption of colonic epithelium with Three randomized controlled trials (RCTs)
pseudomembrane formation resulting in lower have been performed, to date, with the aim to
abdominal pain, fever and diarrhea. Clinical assess the effectiveness of FMT compared to
pictures of CDI are variable and range widely conventional therapy, two of the RCT’s were
from mild colitis to fulminant disease with conducted in European countries and one in
associated toxic megacolon and death. Canada (Table 1). The first RCT was conducted