V O L U M E 1 1 0 • J A N U A RY 2 0 2 1 • N O .

ACTA PÆDIATRICA
N U R T U R I N G T H E C H I L D

Tablets, toddlers and tantrums

Sleep
Reduced parental awareness of SIDS?

www.actapaediatrica.com

apa_15355_issueinfo.indd 1 16-12-2020 11:33:33

 ACTA PÆDIATRICA
N U R T U R I N G T H E C H I L D

EDITOR-IN-CHIEF ACTA PAEDIATRICA IS OWNED BY THE COPYRIGHT AND COPYING

Hugo Lagercrantz MD, PhD ACTA PAEDIATRICA FOUNDATION
Stockholm, Sweden
T +46 73 2024447 BOARD OF TRUSTEES
E hugo.lagercrantz@actapaediatrica.se Göran Wennergren, Chairman
Jan Gustafsson, Treasurer
DEPUTY EDITOR-IN-CHIEF Martin Ritzén
Annika Janson Lena Hellström-Westas
Johan Hirschfeldt
SENIOR EDITORS Gunilla Hedlin
Ingemar Engström Örebro, Sweden Mats Eriksson
Josef Milerad Stockholm, Sweden
Martin Ritzén Stockholm, Sweden EDITORIAL COMMITTEE
Kjell Tullus London, UK C. Agostoni Milan, Italy
P. Anderson Melbourne, Australia
ASSOCIATE EDITORS
R. Andersson Gothenburg, Sweden
Peter Born Copenhagen, Denmark
A. G. Bechensteen Oslo, Norway
Björn Fischler Stockholm, Sweden
C. Casper Toulouse, France
Liisa Lehtonen Turku, Finland
S. Chemtob Montreal, Canada
Alf Meberg Tønsberg, Norway
M. Hermanussen Altenhof, Germany
Michael Weindling Liverpool, UK
A. Hjern Stockholm, Sweden
I. A. Kelmanson St. Petersburg, Russia
ADJUNCT EDITORS
B. Koletzko Munich, Germany
Hans Christian Erichsen
M. Korppi Tampere, Finland
Emma Goksör
P Kuhn Strasbourg, France
Rafael Krmar
S Kutzsche Kuala Lumpur, Malaysia
Olli Lohi
R. Ljung Malmö, Sweden
Eirik Nestaas
H. P. Lukkarinen Turku, Finland
Nelly Padilla
K. M. Main Copenhagen, Denmark
Sven-Arne Silfverdal
A. Malamitsi-Puchner Athens, Greece
Terhi Tapiainen
R. J. Martin Cleveland, USA
MANAGING EDITOR L. Å. Persson Addis Abeba, Ethiopia
Anna Käll H. Rabe Brighton, UK
M. Ranke Tübingen, Germany
ASSISTANT EDITORS O Saugstad Oslo, Norway
Elin Alvehag E. Shinwell Jerusalem, Israel
Margareta Munkert Karnros B. Sun Shanghai, China
B. Trollfors Gothenburg, Sweden
A DIFFERENT VIEW EDITOR R. D. White South Bend, USA
Annie Janvier Montreal, Canada A. Whitelaw Bristol, UK
EBNEO COMMENTARIES
J. Zeman Prague, Czech Republic
Clyde Wright U. Ådén Stockholm, Sweden
LANGUAGE EDITOR
Annette Whibley
©2021 Foundation Acta Pædiatrica.
Published by John Wiley & Sons Ltd.
All rights reserved.
No part of this publication may be
reproduced, stored or transmitted in any form
or by any means without the prior permission
in writing from the copyright holder. Authori-
zation to copy items for internal and personal
use is granted by the copyright holder for
libraries and other users registered with their
local Reproduction Rights Organisation
(RRO), e.g. Copyright Clearance Center
(CCC), 222 Rosewood Drive, Danvers,
MA 01923, USA (www.copyright.com),
provided the appropriate fee is paid directly
to the RRO. This consent does not
extend to other kinds of copying such
as copying for general distribution, for
advertising or promotional purposes, for re-
publication, for creating new collective works
or for resale. Permissions for such reuse can
be obtained using the RightsLink “Request
Permissions” link on Wiley Online Library.
Special requests should be addressed to:
permissions@wiley.com.
DISCLAIMER
The Publisher, Foundation Acta Pædiatrica
and Editors cannot be held responsible for
errors or any consequences arising from the
use of information contained in this journal;
the views and opinions expressed do
not necessarily reflect those of the Publisher,
Foundation Acta Pædiatrica and Editors,
neither does the publication of advertisements
constitute any endorsement by the Publisher,
Foundation Acta Pædiatrica and Editors of
the products advertised.
For submission instructions, subscription
and all other information visit:
http://wileyonlinelibrary.com/journal/apa.

EDITORIAL OFFICE
Acta Pædiatrica
Tomtebodavägen 9
171 65 Solna, Sweden
T +46 73 2024447
E mail@actapaediatrica.se

apa_15355_issueinfo.indd 2 16-12-2020 11:33:39

 ACTA PÆDIATRICA
N U R T U R I N G T H E C H I L D

VO LUM E 1 1 0 • JAN UARY 2 0 2 1 • NO. 1

C ON TE N TS
HIGHLIGHTS IN THIS ISSUE

8 Highlights in this issue

Anna Käll, Hugo Lagercrantz

EDITORIALS

10 Daunting but not impossible—Addressing child maltreatment in Bangladesh

Steven Lucas
12 Equal oral health for young children—A new approach?
Karin Ridell
14 Global investment is needed so that countries can reduce neonatal mortality to below 12 deaths per 1000 live births
by 2030
Lars Åke Persson
17 Bronchopulmonary dysplasia: A problem of prediction or a problem of diagnosis?
Luigi Gagliardi

PERSPECTIVE

19 The burden of Europe’s immigration crisis on mother-child healthcare services and opportunities for culturally
sensitive family-centred care
Livio Provenzi

A DIFFERENT VIEW

22 Do mothers of extremely preterm babies have a duty to express breastmilk?

Fiona Woollard
25 Exploring respiratory syncytial virus prophylaxis for children with all grades of bronchopulmonary dysplasia
Bosco Paes, Eugenio Baraldi, Brigitte Fauroux & Xavier Carbonell-Estrany

PAEDIATRIC ESSAY

28 On the diagnosis of childhood coeliac disease: Past and present

Lars Stenhammar, Anna Myléus, Olof Sandström & Lotta Högberg

CLINICAL OVERVIEW

30 Identical twins affected by congenital cytomegalovirus infections showed different audio-vestibular profiles
Andra Lazar, Ulrika Löfkvist, Luca Verrecchia & Eva Karltorp

REVIEW ARTICLES

36 Current ideas about the roles of rapid eye movement and non–rapid eye movement sleep in brain development
Marit S. Knoop, Eline R. de Groot & Jeroen Dudink
45 Systematic review and meta-analysis suggest that the duration of Kangaroo mother care has a direct impact on
neonatal growth
Nathalie Charpak, Adriana Montealegre-Pomar & Adriana Bohorquez

apa_15355_issueinfo.indd 3 16-12-2020 11:33:39

 Contents continued

60 The impact of Caesarean section on the infant gut microbiome

Delphine M. Hoang, Elvira I. Levy & Yvan Vandenplas

REGULAR ARTICLES & BRIEF REPORTS

Neonatology
68 Limited agreement between clinical assessment of infant colour at birth and oxygen saturation in a hospital in
Ethiopia
Francesco Cavallin, Maria Sofia Cori, Senait Negash, Gaetano Azzimonti, Fabio Manenti, Giovanni Putoto &
Daniele Trevisanuto
72 Accurate neonatal heart rate monitoring using a new wireless, cap mounted device
Caroline Henry, Lara Shipley, Carole Ward, Siavash Mirahmadi, Chong Liu, Steve Morgan, John Crowe,
James Carpenter, Barrie Hayes-Gill & Don Sharkey
79 Serum (1 ൺ 3)-β-D-glucan could be useful to rule out invasive candidiasis in neonates with an adapted cut-off
Pauline Cliquennois, Pauline Scherdel, Rose-Anne Lavergne, Cyril Flamant, Florent Morio, Jeremie F. Cohen,
Elise Launay & Christele Gras Le Guen
85 Hypothermic treatment for neonatal asphyxia in low-resource settings using phase-changing material—An easy
to use and low-cost method
Hang T. T. Tran, Ha T. T. Le, Hanh T. P. Tran, Dung T. K. Khu, Hugo Lagercrantz, Dien M. Tran,
Birger Winbladh, Lena Hellström-Westas, Tobias Alfvén & Linus Olson
94 How doctors communicated with parents in a neonatal intensive care: Communication and ethical issues
Gaelle Sorin, Lionel Dany, Renaud Vialet, Laurent Thomachot, Sophie Hassid, Fabrice Michel & Barthélémy Tosello
101 Early parenting intervention promotes 24-month psychomotor development in preterm children
Camilla Pisoni, Livio Provenzi, Michela Moncecchi, Camilla Caporali, Cecilia Naboni, Mauro Stronati,
Rosario Montirosso, Renato Borgatti & Simona Orcesi
109 Faecal calprotectin and gut microbiota do not predict enteropathy in very preterm infants
Florence Campeotto, Caroline Elie, Clotilde Rousseau, Agnès Giuseppi, Taymme Hachem, Ponny Gobalakichenane,
Mathilde Le Touzey, Marie de Stefano, Marie-José Butel & Nathalie Kapel
117 Suctioning at birth showed low adherence to official recommendations in a low-resource setting
Francesco Cavallin, Shaimaa Abuelnoor Ahmed Abdelghany, Serena Calgaro, Amir Hussein Abubacar Seni,
Bonifacio Rodriguez Cebola, Giovanni Putoto & Daniele Trevisanuto
119 The effects of Xenon gas inhalation on neuropathology in a placental-induced brain injury model in neonates: A
pilot study
Thomas Phillips, David A. Menassa, Simon Grant, Nicki Cohen & Marianne Thoresen
123 Managing mother’s own milk for very preterm infants in neonatal units in 11 European countries
Carina Rodrigues, Jennifer Zeitlin, Emilija Wilson, Liis Toome, Marina Cuttini, Rolf F. Maier, Véronique Pierrat,
Henrique Barros & the EPICE Research Group
Outcome after preterm birth
127 Visual-motor functions are affected in young adults who were born premature and screened for retinopathy of
prematurity
Dýrleif Pétursdóttir, Gerd Holmström, Eva Larsson & Birgitta Böhm
134 Preeclampsia was a risk factor for pulmonary interstitial emphysema in preterm infants born 32 weeks of
gestational age
Judith Behnke, Anita Windhorst, Frank Oehmke, Lars D. Berthold, Klaus-Peter Zimmer, Markus Waitz &
Harald Ehrhardt
141 Cholestasis after very preterm birth was associated with adverse neonatal outcomes but no significant long-term
liver disease: A population-based study
Jonas Teng, Kajsa Bohlin, Antal Nemeth & Björn Fischler
149 Controlled case series demonstrates how parents can be trained to treat paediatric feeding disorders at home
Tessa Taylor, Neville Blampied & Nikolas Roglić
158 Influence of maternal region of birth on neonatal outcomes of babies born small
Nicole E. Young, Miranda Davies-Tuck & Atul Malhotra
166 Bronchopulmonary dysplasia outcome estimator in current neonatal practice
Elizabeth K. Baker & Peter G. Davis

apa_15355_issueinfo.indd 4 16-12-2020 11:33:39

 Contents continued

168 Post-haemorrhagic hydrocephalus management: Delayed neonatal transport negatively affects outcome
Alessandro Parodi, Ilaria Giordano, Laura De Angelis, Mariya Malova, Maria Grazia Calevo, Deborah Preiti,
Marcello Ravegnani, Armando Cama, Carlo Bellini & Luca A. Ramenghi
Head circumference
171 Breastfeeding practices and weight gain predicted head circumference in young Amazonian children
Isabel Giacomini, Lalucha Mazzucchetti, Thainara A. B. Lima, Maíra B. Malta, Bárbara H. Lourenço
Marly A. Cardoso & the MINA-Brazil Study Group
Sudden unexpected infant death
174 Factors associated with age of death in sudden unexpected infant death
Kelty Allen, Tatiana M. Anderson, Urszula Chajewska, Jan-Marino Ramirez & Edwin A. Mitchell
184 Maternal intentions towards infant sleeping practices in Ireland
Niamh O’ Brien, Cliona McGarvey, Karina Hamilton & Breda Hayes
COVID-19
194 Italian parents welcomed a telehealth family-centred rehabilitation programme for children with disability during
COVID-19 lockdown
Livio Provenzi, Serena Grumi, Alice Gardani, Valentina Aramini, Erika Dargenio, Cecilia Naboni, Valeria Vacchini,
Renato Borgatti & Engaging with Families through On-line Rehabilitation for Children during the Emergency
(EnFORCE) Group
Infections
197 Retrospectively diagnosing congenital cytomegalovirus infections in symptomatic infants is challenging
Natalia Lüsebrink, Matthias Kieslich, Holger F. Rabenau, Rolf L. Schlößer & Horst Buxmann
203 Device-associated multidrug-resistant bacteria surveillance in critically ill children: 10 years of experience
Mònica Girona-Alarcón, Elena Fresán, Ana Garcia-Garcia, Sara Bobillo-Perez, Monica Balaguer, Aida Felipe,
Maria Esther Esteban & Iolanda Jordan
210 The epidemiology of Staphylococcus aureus bacteraemia in Israeli children: Community- vs hospital-acquired or
healthcare related infections
Halima Dabaja-Younis, Wakar Garra, Yael Shachor-Meyouhas, Tanya Mashiach, Yuval Geffen & Imad Kassis
219 Interleukin 17F polymorphisms showed no association with lung function at school age after infant bronchiolitis
Annukka Holster, Riikka Riikonen, Johanna Teräsjärvi, Matti Korppi, Kirsi Nuolivirta, Sari Törmänen, Qiushui He
& Eero Lauhkonen
Lungs
222 IL17F rs763780 single nucleotide polymorphism is associated with asthma after bronchiolitis in infancy
Annukka Holster, Johanna Teräsjärvi, Alex-Mikael Barkoff, Eero Lauhkonen, Sari Törmänen, Merja Helminen,
Matti Korppi, Qiushui He & Kirsi Nuolivirta
228 Cystic fibrosis drug approved for patients aged 6-11 years worked well in clinical practice
Ioanna Loukou, Argyri Petrocheilou, Maria Moustaki, Christina N. Katsagoni & Konstantinos Douros
Oral health
230 Impact of an extended postnatal home visiting programme on oral health among children in a disadvantaged
area of Stockholm, Sweden
Ida Brännemo, Göran Dahllöf, Fernanda Cunha Soares & Georgios Tsilingaridis
Paediatric advanced care
237 Evaluation showed that stakeholders valued the support provided by the Implementing Pediatric Advance Care
Planning Toolkit
Jurrianne Fahner, Judith Rietjens, Agnes van der Heide, Megan Milota, Johannes van Delden & Marijke Kars
Social health
247 Early childhood social determinants and family relationships predict parental separation and living arrangements
thereafter
Anders Hjern, Malin Bergström, Stine Kjaer Urhoj & Anne-Marie Nybo Andersen
255 Tablets, toddlers and tantrums: The immediate effects of tablet device play
Tiffany G. Munzer, Alison L. Miller, Yujie Wang, Niko Kaciroti & Jenny S. Radesky
257 Children’s exposure to psychological abuse and neglect: A population-based study in rural Bangladesh
Md Atiqul Haque, Syed Moniruzzaman, Staffan Janson, AKM Fazlur Rahman, Saidur Rahman Mashreky &
Ulla-Britt Eriksson

apa_15355_issueinfo.indd 5 16-12-2020 11:33:41

 Contents continued

265 Identifying language disorder in bilingual children aged 2.5 years requires screening in both languages
Laleh Nayeb, Dagmar Lagerberg, Anna Sarkadi, Eva-Kristina Salameh & Mårten Eriksson
273 Internationally adopted children with and without a cleft lip and palate showed no differences in language ability
at school-age
AnnaKarin Larsson, Christina Persson, Kristina Klintö & Carmela Miniscalco
Arthritis
280 Muscle function and architecture in children with juvenile idiopathic arthritis
Pierre Bourdier, Anthony Birat, Emmanuelle Rochette, Éric Doré, Daniel Courteix, Frédéric Dutheil, Bruno Pereira,
Sébastien Ratel, Etienne Merlin & Pascale Duché
Obesity
288 A strength and neuromuscular exercise programme did not improve body composition, nutrition and
psychological status in children with obesity
Alexandra Thajer, Katharina Truschner, Anselm Jorda, Gabriele Skacel, Brian Horsak & Susanne Greber-Platzer
Neurology
290 Temporal and spatial localisation of general movement complexity and variation—Why Gestalt assessment
requires experience
Ying-Chin Wu, Ilse M. van Rijssen, Maria T. Buurman, Linze-Jaap Dijkstra, Elisa G. Hamer &
Mijna Hadders-Algra
301 Pain is frequent in children with cerebral palsy and negatively affects physical activity and participation
Cecilie Schmidt Østergaard, Nanna Sofie Astrup Pedersen, Anne Thomasen, Inger Mechlenburg &
Kirsten Nordbye-Nielsen
307 Structured Observation of Motor Performance in Infants: Level and quality associated with later motor
development
Cecilia Montgomery, Ylva F. Kaul, Katarina Strand Brodd, Kristina Persson & Lena Hellström-Westas
Gastroenterology
314 A nationwide questionnaire survey on accidental magnet ingestion in children in Japan
Ryosuke Miyamoto, Masumi Okuda, Shogo Kikuchi, Hideyuki Iwayama, Hiroshi Hataya & Akihisa Okumura
326 Growth failure is rare in a contemporary cohort of paediatric inflammatory bowel disease patients
James J. Ashton, Zachary Green, Aneurin Young, Florina Borca, Tracy Coelho, Akshay Batra, Nadeem A. Afzal,
Sarah Ennis, Mark J. Johnson & R. Mark Beattie
335 Involvement of clinical characteristics and HLA-DQ2 genotypes in paediatric patients with coeliac disease
suffering from allergies
Carmen M. Cabrera, Lorenzo Sánchez-Godoy, Víctor M. Navas-López & José M. Urra
337 Point-of-Care test screening versus Case finding for paediatric coeliac disease: A pragmatic study in primary care
Giuseppe Primavera, Andrea Aiello, Caterina Grosso, Gianluca Trifirò, Stefano Costa, Anne-Marie Grima, Socrate
Pallio, Mondher Toumi, Giuseppe Magazzu’, Salvatore Pellegrino & on behalf of the Sicilian Celiac Disease Study
Group
Urology
340 Urolithiasis in second-generation immigrant children younger than 18 years of age in Sweden
Per Wändell, Axel C. Carlsson, Xinjun Li, Jan Sundquist & Kristina Sundquist
347 A scoring system for predicting downgrading and resolution of high-grade infant vesicoureteral reflux
Sofia Sjöström, Aldina Pivodic, Kate Abrahamsson, Rune Sixt, Eira Stokland & Sverker Hansson

EBNEO COMMENTARY

357 Predicting poor physical activity outcomes in children born extremely preterm or extremely low birthweight
Kate L. Cameron, Jeanie L. Y. Cheong & Alicia J. Spittle
359 Early neurodevelopmental screening: Parent perspectives from the neonatal intensive care unit
Amanda K. L. Kwong, Abbey L. Eeles & Alicia J. Spittle

apa_15355_issueinfo.indd 6 16-12-2020 11:33:41

 Contents continued

READER’S FORUM

361 Letter in response to identifying language disorder in bilingual children aged 2.5 years requires screening in both
languages
Daisy I. Perry & Gurdas V. Singh
362 Answer to the letter concerning our published paper about identifying language disorder in bilingual children
Laleh Nayeb, Dagmar Lagerberg, Anna Sarkadi, Eva-Kristina Salameh & Mårten Eriksson
364 SIDS or suffocation—The problem continues
Roger W. Byard
365 Response to: SIDS or suffocation—The problem continues
Niamh O’ Brien, Cliona McGarvey, Karina Hamilton & Breda Hayes
366 ACKNOWLEDGEMENTS

CORRIGENDUM

372 Corrigendum

Front Cover image: Parent taking out touch pad from child. / istockphoto.com, Nadezhda1906

apa_15355_issueinfo.indd 7 16-12-2020 11:33:41

 | |
Received: 13 May 2020    Revised: 24 June 2020    Accepted: 1 July 2020

DOI: 10.1111/apa.15463

REGULAR ARTICLE

Limited agreement between clinical assessment of infant

colour at birth and oxygen saturation in a hospital in Ethiopia

Francesco Cavallin1 | Maria Sofia Cori2 | Senait Negash3 | Gaetano Azzimonti4 |

Fabio Manenti4 | Giovanni Putoto5 | Daniele Trevisanuto6

1
Independent statistician, Solagna, Italy
2 Abstract
Department of Pediatrics, Università
Cattolica del Sacro Cuore, Rome, Italy Aim: To evaluate the relationship between clinical assessment of infant colour and
3
St. Luke Catholic Hospital in Wolisso, oxygen saturation at birth in a low-resource setting.
Wolisso, Ethiopia
4
Methods: Classification of infant colour (cyanotic, pink or unclear) by midwives
Doctors with Africa CUAMM, Wolisso,
Ethiopia was compared to pulse-oximeter data at 60-90-120-300 seconds after birth in 60
5
Doctors with Africa CUAMM, Padua, Italy neonates.
6
Department of Woman's and Child's Health, Results: Overall, oxygen saturation increased over time (P < .0001) and was different
University of Padua, Padua, Italy
according to infant colour (P < .0001). Median oxygen saturation in pink infants was
Correspondence 87% at 60 seconds (n = 1), 90% (IQR 83-91) at 90 seconds (n = 5), 86% (IQR 81-94) at
Daniele Trevisanuto, Department of
Woman's and Child's Health, University of 120 seconds (n = 11) and 93% (IQR 90-96) at 300 seconds (n = 20). Median oxygen
Padua, Via Giustiniani 3, 35128 Padua, Italy. saturation in cyanotic infants was 60% (IQR 45-70) at 60 seconds (n = 52), 64% (IQR
Email: daniele.trevisanuto@unipd.it
52-69) at 90 seconds (n = 42), 63% (IQR 56-68) at 120  seconds (n  = 35) and 66%
Funding information (IQR 62-74) at 300 seconds (n = 22). Median oxygen saturation in unclear-coloured
This study was supported by a grant of the
Italian Episcopal Conference. infants was 57% (IQR 56-60) at 60 seconds (n = 7), 78% (IQR 71-81) at 90 seconds
(n = 13), 81% (IQR 79-88) at 120 seconds (n = 14) and 80% (IQR 76-84) at 300 sec-
onds (n = 18). The proportion of infants with unclear colour ranged from 12% to 30%.
Conclusion: The variability of oxygen saturation among pink and cyanotic infants,
and the substantial proportion of unclear infant colour, suggest the possible benefit
of the availability of pulse oximetry in low-resource settings.

KEYWORDS

birth, colour, infant newborn, low-resource-setting, oxygen saturation

However, the evaluation of infant colour remains the most used

1 | I NTRO D U C TI O N
The International Guidelines for Neonatal Resuscitation recommend the
assessment of heart rate, breathing and oxygenation to guide the resus-
citation intervention.1,2 Since 2010, pulse oximetry has replaced infant
colour as measure of oxygenation in resuscitation guidelines because the
1-3
latter has been shown to be an unreliable indicator of oxygen saturation.
Abbreviation: IQR Interquartile range
indicator of oxygenation in middle- and low-resource settings where
pulse oximetry is often unavailable.4,5 In such settings, oxygen admin-
istration is exclusively driven by clinical observation, but there is lack of
information on whether assessing infant colour may be useful to health-
care providers in determining when to give supplemental oxygen.
Therefore, this study aimed to evaluate the relationship between
clinical assessment of infant colour and oxygen saturation at birth in
a low-resource setting.

©2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

|
68     
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:68–71.
CAVALLIN et al.
2 |  M E TH O DS
2.1 | Study design
This study includes the secondary analysis of infant colour and oxygen
saturation data from a main study (ClinicalTrials.gov NCT03854435)
on heart rate assessment in a low-resource setting. This analysis was
included in the study protocol and was approved by the institutional
review board of the St. Luke Catholic Hospital. All parents gave writ-
ten informed consent to participate in the study.
2.2 | Settings
The study was conducted at the St. Luke Catholic Hospital in Wolisso
(Ethiopia), which is a tertiary hospital with around 3600 deliveries per year.
2.3 | Patients
Patients enrolled in the main study were inborn infants at risk of as-
phyxia (ie suspected foetal distress assessed by intermittent foetal
auscultation, shoulder dystocia, suspected placenta abruption, cord pro-
lapse, foetal malpresentation, vacuum- or forceps-assisted vaginal deliv-
ery and meconium-stained amniotic fluid). Exclusion criteria were major
congenital malformations or parental refusal to participate to the study.
2.4 | Procedure
During the main study, an external observer, who was not involved in
the care of the newborn, positioned a pulse-oximeter sensor (LNCS®
SENSORS, Masimo Radical, Masimo) on the right hand of the baby. The
maximum sensitivity setting was chosen. Time of birth was the ‘time 0’,
and oxygen saturation was evaluated at 60, 90, 120 and 300 seconds by
the external observer with a stopwatch. The midwife involved in the care
of the baby was asked to evaluate infant colour at the same time points.
Infant colour was evaluated by looking at perioral area including lips and
tongue, and was classified as pink, cyanotic or unclear. Colour was defined
as unclear when the midwife defined the baby as neither pink nor cyanotic.
2.5 | Blinding
The midwife assessing infant colour was masked to the pulse oxime-
try data. The statistician who analysed the data was masked to the
classification of infant colour.
2.6 | Outcome measure
The outcome measure was the association between infant colour
and oxygen saturation (%) at 60, 90, 120 and 300 seconds after birth.
|
      69
Key notes
• In low-resource settings, oxygen administration at birth
is exclusively driven by clinical observation, but it is un-
known whether assessing infant colour may determine
when to give supplemental oxygen.
• There was an association between oxygen saturation
and infant colour after birth, but a considerable pro-
portion of infants were difficult to classify as pink or
cyanotic.
• In low-resource settings, the availability of a pulse oxi-
meter may improve clinical evaluation and guide oxy-
genation administration.
2.7 | Data collection
An external observer, who was not involved in the care of the new-
born, was responsible for data collection. Maternal and infant char-
acteristics included maternal age, attendance to antenatal clinic,
antenatal steroids, HIV, intrauterine growth restriction (IUGR),
meconium-stained amniotic fluid, foetal bradycardia, mode of deliv-
ery, gestational age, sex, birthweight and Apgar score at 1 and 5 min-
utes. Infant colour and oxygen saturation were recorded at 60, 90,
120 and 300 seconds after birth.
2.8 | Statistical analysis
This is a secondary analysis of data from the main study; thus, no formal
sample size calculation was performed. Continuous variables were ex-
pressed as median and interquartile range (IQR), while categorical data
variables were expressed as frequency and percentage. Oxygen satura-
tion during the first 5 minutes of life was evaluated with a mixed re-
gression model including infant colour (cyanotic, unclear, pink), time and
interaction colour*time. All tests were 2-sided, and a P-value <.05 was
considered statistically significant. Statistical analysis was performed
using R 3.5 (R Foundation for Statistical Computing, Vienna, Austria).6
3 | R E S U LT S
The analysis included 60 neonates who were enrolled in the main
study from February to April 2019. Participant characteristics are
summarised in Table 1. Face-mask ventilation was administered to
27/60 neonates (45%), chest compressions to 2/60 (3%) and medica-
tions to 2/60 (3%).
Clinical assessment of infant colour is shown in Figure 1. Infants
who looked cyanotic were 52/60 (86%) at 60 seconds, 42/60 (70%) at
90 seconds, 35/60 (59%) at 120 seconds and 22/60 (37%) at 300 sec-
onds after birth. Infant colour was judged unclear in 7/60 (12%) neonates
70      | CAVALLIN et al.

at 60 seconds, 13/60 (22%) at 90 seconds, 14/60 (23%) at 120 seconds
and 18/60 (30%) at 300 seconds after birth. Infants who looked pink
were 1/60 (2%) at 60 seconds, 5/60 (8%) at 90 seconds, 11/60 (18%) at
120 seconds and 20/60 (33%) at 300 seconds after birth.
Oxygen saturation according to infant colour at 60, 90, 120 and
300 seconds after birth is summarised in Figure 2. Oxygen saturation
was not recorded in 6/60 (10%) infants at 60 seconds and 4/60 (7%)
infants at 90-120 seconds. Median oxygen saturation in infants who
looked cyanotic was 60% (IQR 45-70) at 60 seconds, 64% (IQR 52-69)
at 90 seconds, 63% (IQR 56-68) at 120 seconds and 66% (IQR 62-74) at
300 seconds. Median oxygen saturation in infants with unclear colour
was 57% (IQR 56-60) at 60 seconds, 78% (IQR 71-81) at 90 seconds,
81% (IQR 79-88) at 120 seconds and 80% (IQR 76-84) at 300 seconds.
In infants who looked pink, oxygen saturation was 87% (only one in-
fant) at 60 seconds, and median of oxygen saturation was 90% (IQR
83-91) at 90 seconds, 86% (IQR 81-94) at 120 seconds and 93% (IQR
90-96) at 300 seconds. Overall, oxygen saturation increased over time
(P < .0001) and was different according to infant colour (P < .0001).
F I G U R E 1   Clinical assessment of infant colour at 60, 90, 120
and 300 s after birth
TA B L E 1   Participant characteristics
N 60
Maternal age, ya  28 (24-30)
At least one antenatal visit 55 (92)
Antenatal steroids 12 (20)
The differences of oxygen saturation among infant colours (cyanotic,
unclear, pink) persisted over time (interaction colour*time P = .28).
4 | D I S CU S S I O N
Our findings showed an association between oxygen saturation and
infant colour after birth, with higher saturation levels in pink infants
compared to cyanotic infants. A considerable proportion of infants
were difficult to classify as pink or cyanotic.
To our knowledge, this is the first study investigating the rela-
tionship between clinical assessment of infant colour and oxygen-
ation in a low-resource setting. According to guidelines on neonatal
resuscitation, assessment of oxygenation should be performed by
pulse oximetry in high-resource settings,1,2 while clinical evaluation
remains the only option in low-resource settings due to guidelines
recommendations and equipment constraints.7 Two previous studies
evaluated oxygen saturation and infant colour at birth in a high-re-
source setting.3,8 O'Donnell et al reviewed the videos of 20 infants
and reported disagreement among clinicians assessing infant colour,
with wide variation in oxygen saturation in infants considered to be
pink.3 Dawson et al assessed tongue colour in 68 infants and found
that tongue not pink had good specificity but low sensitivity as proxy
of oxygen saturation below 70%.8
Our study adds data on the association between oxygen satura-
tion and infant colour assessed by midwives in a low-resource setting.
We involved local midwives because they were usually in charge of
assessing clinical conditions in newborn infants without the avail-
ability of a pulse oximeter. Therefore, their experience may enhance
the perception of infant colour compared to healthcare providers in
high-resource settings, where oxygen supplementation is based on
pulse oximetry. In our study, higher saturation levels were found in
pink infants compared to cyanotic infants, thus suggesting the clin-
ical importance of assessing infant colour in low-resource settings.
However, there was some variability in oxygen saturation levels
within colour groups, in agreement with data from a high-resource
setting.3 While previous studies classified infant colour as pink or
not pink,3,8 we added a third class (‘unclear’) to give assessors the

Intrauterine growth restriction (IUGR) 2 (3) chance to express their uncertainty when infant colour could not be
clearly perceived. This situation happened in a substantial proportion
Meconium-stained amniotic fluid 26 (43)
of infants who were therefore classified as ‘unclear’, while they would
Foetal bradycardia 10 (17)
have likely converged in the ‘not pink’ class in previous study classi-
Mode of delivery
fication. Of note, infants with ‘unclear’ colour had oxygen saturation
Vaginal 35 (58)
levels between those found in cyanotic and pink groups, and their
Cesarean 25 (42)
proportion increased from 12% to 30% during the first 5 minutes of
Gestational age, wka  38 (35-40) life. Interestingly, oxygen saturation in infants with ‘unclear’ colour
Male:female 36:14 increased after 90 seconds to levels that supported the use of pulse
Birthweight, gramsa  2600 oximetry in order to not start oxygen too early. These findings, along
(2100-3125)
with the variability of oxygen saturation among pink and cyanotic
Apgar score at 1 min 5 (3-6) infants, are in agreement with guidelines on neonatal resuscitation
Apgar score at 5 min 7 (6-8) recommended pulse oximetry to measure oxygenation at birth.1,2
Note: Data expressed as n (%). This study has some limitations. First, it is a secondary analysis
a
Median (IQR). of data from the main study; thus, no formal sample size calculation
CAVALLIN et al.
F I G U R E 2   Oxygen saturation by pulse
oximeter in relation to the colour of the
infant as judged by midwives blinded to
the equipment at 60, 90, 120, and 300 s
after birth
was performed. Second, oxygen saturation and infant colour at
10 minutes after birth could provide further information on adapta-
tion of oxygenation status, but such data were not recorded.
Further studies may replicate this investigation outside of the
delivery room (ie for sick babies in neonatal units) and with different
healthcare providers (ie traditional birth attendants in homebirths).
5 |  CO N C LU S I O N S
Our findings suggest that clinical evaluation of infant colour remains
an important approach for healthcare providers during neonatal
resuscitation. However, the variability of oxygen saturation among
pink and cyanotic infants, and the substantial proportion of unclear
infant colour, suggest the possible benefit of the availability of pulse
oximetry in low-resource settings.
AC K N OW L E D G E M E N T S
We are very grateful to the parents of the enrolled newborns, to the
midwives and to all the local CUAMM staff of the St. Luke Catholic
Hospital in Wolisso (Ethiopia) for their support during this study.
C O N FL I C T O F I N T E R E S T
All the authors have no conflicts of interest to disclose.
ORCID
Daniele Trevisanuto  https://orcid.org/0000-0002-6462-0079
REFERENCES
1. Wyckoff MH, Aziz K, Escobedo MB, et al. Part 13: neonatal resus-
citation: 2015 American Heart Association Guidelines Update for
|
      71
cardiopulmonary resuscitation and emergency cardiovascular care.
Circulation. 2015;132(18 Suppl 2):S543-S560.
2. Wyllie J, Bruinenberg J, Roehr CC, Rüdiger M, Trevisanuto D,
Urlesberger B. European Resuscitation Council Guidelines for
Resuscitation 2015: Section 7. Resuscitation and support of transi-
tion of babies at birth. Resuscitation. 2015;95:249-263.
3. O'Donnell CP, Kamlin CO, Davis PG, Carlin JB, Morley CJ. Clinical
assessment of infant colour at delivery. Arch Dis Child Fetal Neonatal
Ed. 2007;92:F465-F467.
4. Trevisanuto D, Cavallin F, Arnolda G, et al. Equipment for neonatal re-
suscitation in a middle-income country: a national survey in Vietnam.
BMC Pediatr. 2016;16:139.
5. Umphrey L, Breindahl M, Brown A, et al. When Helping Babies
Breathe is not enough: designing a novel, mid-level neonatal resusci-
tation algorithm for Médecins Sans Frontières Field Teams working in
low-resource hospital settings. Neonatology. 2018;114:112-123.
6. R Core Team. R: A language and environment for statistical comput-
ing. Vienna, Austria: R Foundation for Statistical Computing; 2018.
https://www.R-proje​c t.org/.
7. http://www.helpi​ngbab​iesbr​eathe.org. Accessed on February 26,
2020.
8. Dawson A, Ekström A, Frisk C, et al. Assessing the tongue colour of
newly born infants may help to predict the need for supplemental
oxygen in the delivery room. Acta Paediatr. 2015;104:356-359.
How to cite this article: Cavallin F, Cori MS, Negash S, et al.
Limited agreement between clinical assessment of infant
colour at birth and oxygen saturation in a hospital in Ethiopia.
Acta Paediatr. 2021;110:68–71. https://doi.org/10.1111/
apa.15463
 | |
Received: 30 January 2020    Revised: 5 April 2020    Accepted: 6 April 2020

DOI: 10.1111/apa.15303

REGULAR ARTICLE

Accurate neonatal heart rate monitoring using a new wireless,

cap mounted device

Caroline Henry1  | Lara Shipley1  | Carole Ward1 | Siavash Mirahmadi2 |

Chong Liu2  | Steve Morgan2 | John Crowe2  | James Carpenter3 |
Barrie Hayes-Gill2  | Don Sharkey1

1
Division of Child Health, Obstetrics &
Gynaecology, University of Nottingham, Abstract
Nottingham, UK Aim: A device for newborn heart rate (HR) monitoring at birth that is compatible
2
Faculty of Engineering, University of
with delayed cord clamping and minimises hypothermia risk could have advantages
Nottingham, Nottingham, UK
3
SurePulse Medical Limited, Nottingham, UK
over current approaches. We evaluated a wireless, cap mounted device (fhPPG) for
monitoring neonatal HR.
Correspondence
Don Sharkey, Division of Child Health,
Methods: A total of 52 infants on the neonatal intensive care unit (NICU) and im-
Obstetrics & Gynaecology, University mediately following birth by elective caesarean section (ECS) were recruited. HR was
of Nottingham, Queen’s Medical Centre
Campus, Derby Road, Nottingham, UK.
monitored by electrocardiogram (ECG), pulse oximetry (PO) and the fhPPG device.
Email: don.sharkey@nottingham.ac.uk Success rate, accuracy and time to output HR were compared with ECG as the gold

Funding information
standard. Standardised simulated data assessed the fhPPG algorithm accuracy.
Innovate UK Biomedical Catalyst (MRC Results: Compared to ECG HR, the median bias (and 95% limits of agreement) for the
MC_PC_15012); Engineering and Physical
Sciences Research Council (EP/K 503101/1
NICU was fhPPG −0.6 (−5.6, 4.9) vs PO −0.3 (−6.3, 6.2) bpm, and ECS phase fhPPG
via Tioga Ltd PhD CASE award). Additional −0.5 (−8.7, 7.7) vs PO −0.1 (−7.6, 7.1) bpm. In both settings, fhPPG and PO correlated
support by the Faculty of Engineering,
University of Nottingham's Hermes
with paired ECG HRs (both R 2 = 0.89). The fhPPG HR algorithm during simulations
Fellowship scheme (A1H361), and from the demonstrated a near-linear correlation (n = 1266, R 2 = 0.99).
University of Nottingham EPSRC Impact
Acceleration Award (EP/R511730/1 and
Conclusion: Monitoring infants in the NICU and following ECS using a wireless, cap
RA45VJ). mounted device provides accurate HR measurements. This alternative approach
could confer advantages compared with current methods of HR assessment and war-
rants further evaluation at birth.

KEYWORDS

heart rate, infant, newborn, photoplethysmography, resuscitation, technologies, wireless

1 |  I NTRO D U C TI O N or resuscitation.1 Preterm infants often require more advanced

stabilisation or resuscitation whilst avoiding hypothermia, known
Approximately 10% of newborns require assistance at birth to to increase mortality. 2,3 Heart rate (HR) is the most sensitive
establish breathing with 3% needing more sustained stabilisation predictor of the infant's clinical status and the efficacy of any

Abbreviations: BT, Bluetooth; DCC, Delayed cord clamping; ECG, Electrocardiogram; ECS, Elective caesarean section; fhPPG, Forehead photoplethysmography; HR, Heart rate; IQR,
Interquartile range; JIS, Japanese Industrial Standard; LOA, Limits of agreement; NICU, Neonatal Intensive Care Unit; PO, Pulse oximetry; PPG, Photoplethysmography; RMSE, Root
mean square error; SD, Standard deviation.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2020 The Authors. Acta Paediatrica published by John Wiley & Sons Ltd on behalf of Foundation Acta Paediatrica

|
72     
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:72–78.
HENRY et al.
interventions, and international guidelines suggest the electro-
cardiogram (ECG) may be used in the delivery room to monitor
it.4 However, electrode application can be difficult, ECG monitors
are not always available and electrical cardiac activity may not
always equate to effective cardiac output. 5,6 Pulse oximetry (PO)
is widely used on neonatal intensive care units (NICUs), but up-
7 preterm infants.
take in the delivery room is not universal potentially because of
the delay of a few minutes to obtain a reliable HR after birth due
to poor peripheral perfusion and motion. 8 This could explain why
PO underestimates HR when compared to ECG during the first
minutes of life.9,10
Delayed cord clamping (DCC) for term infants is widely practised
and could become more common for preterm infants, but there is a
need to better monitor these babies during this transition to maxi-
mise the benefits of DCC whilst avoiding any adverse effects of de-
layed resuscitation.11,12 Maximising placental transfusion requires
the attending team to be confident of the baby's condition including
an acceptable HR and avoidance of hypothermia. A non-intrusive,
wireless, easy to apply HR monitoring system, which avoids the risk
of hypothermia, could be advantageous in this setting.
We have previously described a prototype wired, fore-
head-mounted sensor studied in NICU patients that utilises reflec-
tance-mode green light photoplethysmography (PPG) to measure
HR.13 This has the advantages that this wavelength of light (525 nm) is
optimised for reflection-mode detection of blood flow and captures
this at the forehead where brain perfusion shares arterial pathways
via the carotid artery and so is less susceptible to poor peripheral
perfusion.14,15 This original device was held in place on the forehead
by a spun bond laminate headband and successfully demonstrated
that HR measurement by this method was feasible but had a Bland-
Altman limit of agreement (LOA) of up to ± 12bpm. In addition, it was
not compatible with current neonatal care due to its inability to ac-
commodate endotracheal tube and CPAP attachments. Furthermore,
the relatively large sensor (22 mm diameter), large electronic circuit
boards and the presence of cumbersome wiring were less suited for
bedside stabilisation of preterm infants during DCC.
This device has now been substantially modified (Video S1,
Figure S1 and Figure S2) in terms of its practicality, and several major
modifications have taken place including the following:
(i) A T-shape cap compatible with respiratory equipment fittings.
(ii) A miniaturised and ruggedised sensor (10 mm diameter).
(iii) Bluetooth (BT) wireless connection removing the need for trail-
ing wires.
The primary aim of this study was to assess, via a clinical trial, the
accuracy and reliability of this new cap mounted newborn HR device
the SurePulse VS (fhPPG) (SurePulse Medical Ltd). The study was
designed to satisfy the regulatory requirements of the CE Medical
Devices Directive 93/42 and its 2007 amendment for measuring HR
compared to PO, with ECG as a gold standard. A secondary aim was
to investigate acquisition time.
      73|
Key notes
• Wireless, continuous heart rate (HR) monitoring of
newborns using a cap mounted device (fhPPG) could be
beneficial during delayed cord clamping, especially in
• Using electrocardiogram HR as a gold standard, a novel
fhPPG performed similarly to pulse oximetry in the neo-
natal intensive care and immediately after caesarean
section birth.
• The fhPPG is an accurate, wireless alternative to current
methods of HR monitoring in newborn infants.
2 | M ATE R I A L S A N D M E TH O DS
2.1 | Study population
This prospective observational study was conducted at the
Nottingham University Hospitals NHS Trust, UK following ethi-
cal approval (NHS Health Research Authority Yorkshire & The
Humber—Sheffield Research Ethics Committee 15/YH/0522).
Informed parental consent was obtained prior to infants being en-
rolled in the study, and evaluations were performed in the NICU
and on babies born by elective caesarean section (ECS). For the
NICU phase, infants were included if they were >22 weeks gesta-
tion, required HR monitoring and were deemed clinically stable
by the attending team. Inclusion criteria for the ECS phase were
infants ≥37 weeks gestation with no obvious requirement for re-
suscitation, based on antenatal history, and delivered by planned
caesarean section.
2.2 | Study design
There were two phases to this study, the NICU and ECS. On the
NICU, an appropriately sized fhPPG cap was fitted to the infant's
head and connected wirelessly to a bespoke synchronised data-
logging system via the inbuilt wireless BT module. As part of
their routine care, infants had 3 neonatal ECG monitoring elec-
trodes (PD50-F4C, SKINTACT, Leonhard Lang GmbH) attached
and connected to a CARESCAPE Monitor B450 (General Electric
Healthcare) with Masimo SET® for SpO2 (Masimo). A transmission
mode PO sensor (LNCS Neo, Masimo) was attached to the wrist
of the right hand and connected to the same monitor. An in-house
designed system using LabVIEW 2014 (National Instruments) si-
multaneously collected synchronised data from all devices and
displayed waveforms from each device in real time on a laptop
(Lenovo Y50 20378, Lenovo Group Ltd). In keeping with normal
practice, if any device presented a poor signal output the elec-
trodes or optical sensors (head or wrist) were repositioned. Up
 |
74       HENRY et al.

to 30  minutes of raw physiological data were collected with the TA B L E 1   Patient demographics and baseline variables during
infant in their cot or incubator. NICU and elective caesarean section (ECS). Values are absolute
numbers or median and interquartile range (IQR)
During the ECS phase, the same monitoring system described
above was deployed. Following birth, the baby was shown to par- Demographic or
ents before being placed on the resuscitaire. All monitoring equip- Variable NICU n = 34 ECS n = 18

ment was attached in the same sequence for each patient by the Gestational age (wk) 31 + 2 (28 + 5-35 + 5) 39 + 0 (38 + 6-39 + 3)
research team. The fhPPG cap was fitted, ECG leads attached and a Age (d) 23 (10-39) Birth
PO sensor attached to the wrist of the right hand16-18 and raw data Birthweight (g) 1460 (1019-2288) 3335 (2833-3561)
collected for up to 20 minutes. Male sex, n (%) 16 (47) 7 (39)
The B450 monitor generates HR data which provides ECG and
Ethnicity, n (%)
PO values every 5 seconds. All fhPPG HR data extracted were aver-
White 32 (94) 14 (77)
aged over the same 5 seconds window for direct comparison.
Mixed white/ 1 (3) –
Slow and rapidly changing HRs can be unpredictable in the NICU
Middle Eastern
and ECS settings. To ensure the fhPPG HR algorithm can effectively
Mixed white/ – 1 (6)
extract a wide range of rapidly changing HRs, as experienced during Afro-Caribbean
newborn care, we subjected the algorithm to simulated HR data Asian – 1 (6)
covering 25-250 bpm using the Japanese Industrial Standard (JIS T
Black 1 (3) 2 (11)
1303:2005, revised 2018).19 Although the JIS is designed for test-
Resuscitation during – 2 (positive pressure
ing foetal heart rate monitors, it appropriately models newborns as study mask ventilation)
the HR baseline and variation are similar to those of the foetus. 20
A simulated PPG signal was input across the JIS range of HRs, the
algorithm's performance tracked and linear regression applied to median (range) or median (IQR) where appropriate. Non-parametric
evaluate the overall performance. data were compared using the Wilcoxon signed-rank test, and be-
tween-groups were compared using Friedman's test with Dunn's
correction for multiple comparisons. The study was registered at
2.3 | Data analysis Clinicaltrials.gov NCT 02701920.

HR data from either PO or fhPPG were excluded if there was no ECG

(the gold standard) HR comparator, a protocol violation occurred or
a device was detached from the patient. To ensure good quality gold
standard data, erroneous ECG HR data, as previously observed,13
were excluded if there was no discernible QRS complexes or there
had been an algorithm drop out. If the ECG HR value deviated by
>20% from the other test devices (PO and/or fhPPG), the RR interval
on the raw ECG trace corresponding to that time point was manually
checked and excluded if the manual ECG HR check deviated from
the reported ECG HR by >10%.
The success rate was calculated as the percentage of time during
which the device in question was attached to the patient and re-
porting a HR once all three devices were simultaneously outputting
a valid HR.
Comparison of the HR accuracy (fhPPG and PO) with the
gold standard (ECG) utilised three methods: (a) Positive Percent
Agreement (PPA) which is the percentage of time the test device
generated a valid HR within 10% of the paired ECG HR, 21 (b) mod-
ified Bland-Altman plots to correct for multiple measurements and
its limits of agreement (LOA), 22,23 and (c) Root Mean Square Error
(RMSE) calculated between paired HRs (fhPPG vs ECG, and PO vs
ECG). HR output time for each device is the time taken from comple-
tion of its attachment to the display of a HR.
Analyses were performed using GraphPad Prism 7.01
(GraphPad). Continuous variables were tested for normality using
the Kolmogorov-Smirnov test. Data were presented as mean (SD),
3 | R E S U LT S
A convenience sample of 60 infants was recruited (NICU  =  40,
ECS  =  20) when researchers were available. Six infants were ex-
cluded from the NICU phase and two from the ECS phase due to
ECG data not being stored or corrupted leading to data loss and mis-
alignment. The patient demographics can be seen in Table 1.
3.1 | Success rate
During the NICU phase, the median ECG success rate was 100% (IQR,
100-100, n = 12 967), PO was 100% (IQR, 99.1-100, n = 12 688) and
fhPPG was 98.7% (IQR, 93.9-100, n = 11 501). During the ECS phase,
the median ECG success rate was 96.2% (IQR, 91.9-100, n = 2039),
PO was 98.1% (IQR, 89.5-100, n = 1903) and fhPPG was 94.1% (IQR,
84.1-98.4, n = 1805).
3.2 | Accuracy of devices
During the NICU phase, the median PPAs were similar for fhPPG
(99.6%; IQR 99.0-100, n  =  11  353) and PO (99.4%; IQR 99.0-100,
n = 12 249). During the ECS phase, the median PPAs were again simi-
lar for fhPPG (98.6%; IQR 96.4-100, n = 1584) and PO (98.7%; IQR
HENRY et al. |
      75

97.2-100, n = 1684). Device comparison with the gold standard was All devices provided similar success rates when outputting a HR.
also determined using both Bland-Altman plots (Table 2) and RMSE During the NICU phase, the fhPPG device had fewer data points
(Table 2 and Figure 1). because of the requirement to remove and replace the device for
Linear regression of all pooled paired HRs demonstrated PO safety checks as required by the ethics panel. The high ECG success
and fhPPG had similar goodness-of-fit with both having an R 2 = .89 rate reflects the use of electrodes that were already in place for nor-
(Figure  2). In five babies, the ECG measured a HR <100  bpm. For mal neonatal care and, unlike the fhPPG device, were not removed
these, there were 23 paired values with the PO of which two re- and re-sited for safety and acquisition time purposes.
ported a HR >100 bpm and for the fhPPG there were 13 paired val- Accuracy, as measured by PPA, Bland-Altman and RMSE
ues with two reporting a HR >100 bpm, and on both occasions the demonstrated the fhPPG device was similar to the PO in both set-
fhPPG output was identified as a poor signal triggering a warning to tings. The measures of variance for each of these measures were
the user. within clinically acceptable limits, both optical devices providing
accurate data when compared to the ECG that detects electrical
activity. This requires the algorithms to respond in a timely fash-
3.3 | Algorithm simulation ion to rapid changes in HR, a key requirement of any device in
these settings. For PO in the NICU, Singh et al (2008) had demon-
Simulated PPG data were applied as described across a range of HRs strated the Bland-Altman LOA of ±12  bpm. 24 In our study, the
and with varying rates of change as defined by JIS T (Figure 3). The PO performed much better in the NICU and ECS groups. For the
fhPPG algorithm demonstrated excellent correlation with the simu- new fhPPG device, accuracy in the NICU has improved to approx-
lated HR (Figure 4). imately ±5  bpm from the previous iteration with LOAs of up to
±12 bpm.13 This improved accuracy remained for the fhPPG device
during the ECS phase with LOAs of approximately ±8  bpm. PO
3.4 | Heart rate output time has been studied in the delivery room across a range of gesta-
tional ages. Kamlin et al’s 8study was mostly mature infants, with
Median time for HR output was similar for all three devices during a mean gestational age of 35  weeks, and the majority (65%) did
the ECS phase. For ECG, this was 24.5s (IQR 14.2-31.2), fhPPG 35.5s not require any intervention. The Bland-Altman demonstrated a
(IQR 16.7-53.7) and PO 24.5s (IQR 13.5-53.5) with no statistical sig- mean difference of −2 bpm and LOA of ±26 bpm. The results for
nificance between any of the devices. both PO and fhPPG in the present study were significantly bet-
ter although the groups were not entirely comparable with only
infants ≥37  weeks gestation recruited at birth in the ECS group.
4 |  D I S CU S S I O N The additional measures of HR accuracy also add confidence that
the fhPPG device performs as well as current devices with a PPA
We aimed to establish the accuracy and reliability of a new wireless, of approximately 99%, in the NICU and ECS phase, and an RMSE
cap mounted HR monitoring device (fhPPG) in the neonatal setting 3-4 bpm. Furthermore, both the PO and fhPPG devices produced
and compare this to the HR obtained from PO. The fhPPG device high correlations with an R 2 = .89 for goodness-of-fit compared to
provided accurate and reliable HR data compared to current predi- ECG HR. For the ECS phase, these data are reassuring given the
cate devices generating a median RMSE of 2.8  bpm compared to vigorous nature of the babies at birth and the significant motion
3.3 bpm for PO in the NICU. In the ECS phase at birth, median RMSE artefact introduced.
values of 4.1 bpm and 3.7 bpm were obtained for the fhPPG and PO It is essential that any newborn HR monitoring device is able to
devices, respectively. Time to output a HR after placement in the detect low rates, particularly those below 100 bpm. The majority
delivery room was similar across all three devices. of babies in this study were stable with limited HR variability so
limiting the number of HRs <100  bpm. Use of the JIS methodol-
TA B L E 2   Bland-Altman limits of agreement (LOA), bias (fhPPG ogy allowed us to test the software algorithm to rapidly chang-
HR-ECG HR) and Root Mean Square of the Error (RMSE) expressed ing simulated HRs across the typical newborn range. The fhPPG
as medians from patient level data algorithm was able to track the HRs quickly and accurately with
an R 2 = .994. This strong correlation supports the accuracy of the
NICU (n = 34) ECS (n = 18)
device's software based on simulation data, but the whole system
Measure fhPPG PO fhPPG PO requires real-world evaluation with infants undergoing resuscita-
LOA (bpm) −5.6, 4.9 −6.3, 6.2 −8.7, 7.7 −7.6, 7.1 tion or stabilisation.
Bias (bpm) −0.6 −0.3 −0.5 −0.1 The secondary aim of this study was to explore the acquisition
RMSE (bpm) 2.8 3.3 4.1 3.7 time of the new device, as it is essential for fast delivery of infor-
mation to the attending team to allow correct management of the
Abbreviations: ECS, elective caesarean section; fhPPG, forehead
wireless cap PPG device; NICU, neonatal intensive care unit; PO, pulse baby. Output of the HR at birth for all devices was similar although
oximeter. more babies in the study would be required to confirm this persisted
 76       | HENRY et al.

NICU ECS F I G U R E 1   Violin plots of the Root

50 Mean Square of the Error (RMSE, median
14
and interquartile range) for the wireless
45
12 cap (fhPPG) device and the pulse oximeter
40
(PO) in the neonatal intensive care unit
10 35
RMSE (bpm)

RMSE (bpm)
(NICU, n = 34) and at birth in the elective
30 caesarean section (ECS) phase (n = 18)
8
25
6 20
4 15
10
2
5
0 0
fhPPG PO fhPPG PO

PO versus ECG fhPPG versus ECG

200 200

150 150
PO HR (bpm)

100 fhPPG HR (bpm) 100

50 50

R2 = .89 R2 = .89
0 0
0 50 100 150 200 0 50 100 150 200
ECG HR (bpm) ECG HR (bpm)

F I G U R E 2   Linear regression of heart rate between pulse oximetry (PO) vs ECG (n = 13 935) and wireless cap (fhPPG) vs ECG (n = 12 939).
Each data plot represents paired HR values between the devices

275 F I G U R E 3   Comparison of the

Japanese Industrial Standard (JIS) input
250 JIS Input (bpm)
signal with the wireless cap (fhPPG)
225 fhPPG Output (bpm) algorithm output rate over simulated heart
200
rate signals ranging from 25 to 250 bpm.
Fluctuations are designed to simulate a
Heart rate (bpm)

175 number of heart rate changes observed in

150 the newborn

125

100

75

50

25

0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500
Time (s)

or identify any differences. ECG measures electrical activity and increasing in the first few minutes of life. The fhPPG device uses
has been shown to acquire a HR quicker than pulse oximeters, pos- green light to measure HR, but also has red and infrared light that are
sibly reflecting delayed acquisition from early cutaneous perfusion optimised for oxygen saturation measurement and used by POs to
HENRY et al.
250
200
fhPPG output (bpm)
150
100
50
2
R = .99
0 Further development of the HR algorithm, addition of oxygen satu-
0 50 100 150 200 250
JIS input (bpm)
F I G U R E 4   Linear regression plot of Japanese Industrial
Standard (JIS) input signal (n = 1266) with the wireless cap (fhPPG)
algorithm output (n = 1264) rate. Each data plot represents paired
HR values between the device and input signal
measure HR. Green light is better suited to detecting pulsatile blood
13
flow so potentially optimising HR estimation particularly during
low perfusion states. The sequence of device attachment after birth
was fhPPG, ECG and PO, respectively, potentially disadvantaging the
fhPPG device as it was placed earlier during the transition period.
Whilst fast detection of electrical activity of the heart is useful, when
cardiac output is poor or non-existent5,6 there could be advantages
to having a parallel optical sensor that is able to detect blood flow and
hence confirm the electrical activity is associated with cardiac output.
Any potential benefits of a centrally placed, green light optical sensor,
that is fhPPG which detects blood flow from the forehead sharing a
common central arterial blood supply to the brain, compared to a pe-
ripherally sited transmission mode device, require further evaluation
in larger number of patients including preterm infants.
The main limitation of the present study is the low number of
HRs below 100, reflecting the nature of the population who were
stable NICU patients or ECS term babies born following uncompli-
cated pregnancies. There were fewer paired ECG and fhPPG values
<100 bpm, compared to the PO, due to sensor repositioning as the
signal had been identified as poor. We mitigated against low HRs by
using the JIS methodology fully exercising the software algorithm to
create HR values over the range 25-250 bpm confirming successful
algorithm functionality. However, additional studies are required to
establish this in compromised newborns, especially those with HRs
<100 bpm, where the HR variability changes frequently and a larger
number of preterm infants.
5 |  CO N C LU S I O N
With the increasing use of DCC, the development of a wireless,
continuous HR monitor that aligns with the normal care pathway,
|
      77
including placement of a cap with attachment for respiratory equip-
ment, could better support transitioning of newborn babies and
audit interventions. 25 Confidence in such a device could allow moni-
toring of the newborn and aid decisions around when to clamp the
cord and instigate any stabilisation or additional resuscitation steps.
fhPPG operates in reflectance mode allowing detection of a PPG
signal on many parts of the body. The forehead allows inclusion in a
cap, normally applied to keep the baby warm, and could potentially
improve reliability during low perfusion states with a shared carotid
artery blood supply to the brain.13 The wireless design avoids mul-
tiple trailing wires that could become accidentally detached during
bedside care with DCC. This device provides healthcare profession-
als with an accurate alternative to current neonatal HR monitoring.
ration monitoring and evaluation during low perfusion states are
underway.
AC K N OW L E D G E M E N T S
Dr Andrew Prayle helped to implement the modified Bland-Altman
for multiple measurements. Thank you to all the families who kindly
agreed to participate in this study.
C O N FL I C T O F I N T E R E S T
Following are shareholders in SurePulse Medical Ltd—Don Sharkey,
Barrie Hayes-Gill, John Crowe, James Carpenter and Steve Morgan.
James Carpenter is the chief executive officer for SurePulse Medical
Ltd. Don Sharkey and Barrie Hayes-Gill are non-executive directors
of SurePulse Medical Ltd.
ORCID
Caroline Henry  https://orcid.org/0000-0003-1192-9985
Lara Shipley  https://orcid.org/0000-0002-8538-9531
Chong Liu  https://orcid.org/0000-0002-2340-4329
John Crowe  https://orcid.org/0000-0002-5438-2169
Barrie Hayes-Gill  https://orcid.org/0000-0003-4237-7001
Don Sharkey  https://orcid.org/0000-0002-4989-8697
REFERENCES
1. Finer N, Rich W. Neonatal resuscitation for the preterm infant: evi-
dence versus practice. J Perinatol. 2010;30(Suppl):S57-66.
2. Laptook AR, Salhab W, Bhaskar B. Admission temperature of
low birth weight infants: predictors and associated morbidities.
Pediatrics. 2007;119(3):e643-e649.
3. de Almeida MF, Guinsburg R, Sancho GA, et al. Hypothermia
and early neonatal mortality in preterm infants. J Pediatr.
2014;164(2):271-275.
4. Wyllie J, Perlman JM, Kattwinkel J, et al. Part 7: neonatal resusci-
tation: 2015 International Consensus on cardiopulmonary resusci-
tation and emergency cardiovascular care science with treatment
recommendations. Resuscitation. 2015;95:e169-201.
5. Luong D, Cheung PY, Barrington KJ, et al. Cardiac arrest with pulse-
less electrical activity rhythm in newborn infants: a case series.
Arch Dis Child Fetal Neonatal Ed. 2019;104:F572-F574.
6. Patel S, Cheung PY, Solevag AL, et al. Pulseless electrical activity:
a misdiagnosed entity during asphyxia in newborn infants? Arch Dis
Child Fetal Neonatal Ed. 2019;104(2):F215-F217.
 |
78       HENRY et al.
7. Mann C, Ward C, Grubb M, et al. Marked variation in newborn
resuscitation practice: a national survey in the UK. Resuscitation.
2012;83(5):607-611.
8. Kamlin CO, Dawson JA, O'Donnell CP, et al. Accuracy of pulse oxim-
etry measurement of heart rate of newborn infants in the delivery
room. J Pediatr. 2008;152(6):756-760.
9. van Vonderen JJ, Hooper SB, Kroese JK, et al. Pulse oximetry mea-
sures a lower heart rate at birth compared with electrocardiogra-
phy. J Pediatr. 2015;166(1):49-53.
10. Murphy MC, De Angelis L, McCarthy LK, O'Donnell CPF.
Randomised study comparing heart rate measurement in newly
born infants using a monitor incorporating electrocardiogram and
pulse oximeter versus pulse oximeter alone. Arch Dis Child Fetal
Neonatal Ed. 2019;104:F547-F550.
11. Tarnow-Mordi W, Morris J, Kirby A, et al. Delayed versus im-
mediate cord clamping in preterm infants. N Engl J Med.
2017;377(25):2445-2455.
12. Manley BJ, Owen LS, Hooper SB, et al. Towards evi-
dence-based resuscitation of the newborn infant. Lancet.
2017;389(10079):1639-1648.
13. Grubb MR, Carpenter J, Crowe JA, et al. Forehead reflectance pho-
toplethysmography to monitor heart rate: preliminary results from
neonatal patients. Physiol Meas. 2014;35(5):881-893.
14. Fernandez M, Burns K, Calhoun B, George S, Martin B, Weaver
C. Evaluation of a new pulse oximeter sensor. Am J Crit Care.
2007;16(2):146-152.
15. Berkenbosch JW, Tobias JD. Comparison of a new forehead reflec-
tance pulse oximeter sensor with a conventional digit sensor in pe-
diatric patients. Respir Care. 2006;51(7):726-731.
16. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Feasibility of and
delay in obtaining pulse oximetry during neonatal resuscitation.
J Pediatr. 2005;147(5):698-699.
17. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Obtaining pulse
oximetry data in neonates: a randomised crossover study of
sensor application techniques. Arch Dis Child Fetal Neonatal Ed.
2005;90(1):F84-F85.
18. Louis D, Sundaram V, Kumar P. Pulse oximeter sensor applica-
tion during neonatal resuscitation: a randomized controlled trial.
Pediatrics. 2014;133(3):476-482.
19. Japanese Industrial Standards JT. 分べん（娩）監視装置 Fetal moni-
tors. Tokyo: Japanese Standards Association; 2005 (revised 2018).
Japanese Industrial Standards T1303.
20. Bustos R, Bejar R, Arroyave H, et al. Heart rate in fetuses and neo-
nates in normal conditions and with mild depression. J Perinat Med.
1975;3(3):172-179.
21. Cohen WR, Ommani S, Hassan S, et al. Accuracy and reliability of
fetal heart rate monitoring using maternal abdominal surface elec-
trodes. Acta Obstet Gynecol Scand. 2012;91(11):1306-1313.
22. Bland JM, Altman DG. Agreement between methods of measure-
ment with multiple observations per individual. J Biopharm Stat.
2007;17(4):571-582.
23. Altman DG, Bland JM. Measurement in medicine: the analysis of
method comparison studies. Statistician. 1983;307-317.
24. Singh JK, Kamlin CO, Morley CJ, O'Donnell CP, Donath SM, Davis
PG. Accuracy of pulse oximetry in assessing heart rate of in-
fants in the neonatal intensive care unit. J Paediatr Child Health.
2008;44(5):273-275.
25. Pickup L, Lang A, Shipley L, et al. Development of a clinical interface
for a novel newborn resuscitation device: human factors approach
to understanding cognitive user requirements. JMIR Hum Factors.
2019;6(2):e12055.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Henry C, Shipley L, Ward C, et al.
Accurate neonatal heart rate monitoring using a new
wireless, cap mounted device. Acta Paediatr. 2021;110:72–78.
https://doi.org/10.1111/apa.15303
 | |
Received: 13 November 2019    Revised: 7 April 2020    Accepted: 16 April 2020

DOI: 10.1111/apa.15321

REGULAR ARTICLE

Serum (1 → 3)-β-D-glucan could be useful to rule out invasive

candidiasis in neonates with an adapted cut-off

Pauline Cliquennois1  | Pauline Scherdel2,3  | Rose-Anne Lavergne4  |

Cyril Flamant1  | Florent Morio4  | Jeremie F. Cohen3,5  | Elise Launay2,6  |
Christele Gras Le Guen2,6

1
Neonatal Intensive Care Unit, University
Hospital of Nantes, Nantes, France Abstract
2
Clinical Investigation Center (CIC004), Aim: We assessed the diagnostic accuracy of serum (1  →  3)-β-D-glucan (BDG) for
University Hospital of Nantes, Nantes,
neonatal invasive candidiasis (NIC) using the recommended cut-off usually used in
France
3
Inserm UMR 1153, Obstetrical, Perinatal
adults for detecting invasive candidiasis and searched for an optimal cut-off for rul-
and Pediatric Epidemiology Research Team ing out NIC.
(Epopé), Center of Research in Epidemiology
and Statistics (CRESS), University of Paris,
Methods: We conducted a prospective cross-sectional study at Nantes University
Paris, France medical centre from January 2017 to July 2018. All consecutive newborn infants of
4
Parasitology and Mycology Laboratory, less than 28 days of corrected age, with clinically suspected NIC, who underwent
Institute of Biology, University Hospital of
Nantes, Nantes, France BDG assay, were included. Sensitivity and specificity were calculated by using the
5
Department of General Pediatrics and recommended cut-off of 80 pg/mL. Receiver operating characteristic curve analysis
Pediatric Infectious Diseases, Necker
was used to identify an optimal cut-off value.
Hospital for Sick Children, APHP, University
of Paris, Paris, France Results: We included 55 newborn infants with 61 episodes of suspected NIC. Their
6
Pediatric and Emergency Department, median gestational and chronological ages were 28.0 weeks (interquartile range
University Hospital of Nantes, Nantes,
France [IQR] 26.4-34.1) and 10.0 days (IQR 6.0-22.0), respectively. Of 61 episodes, seven
revealed NIC. Sensitivity and specificity were 85.7% (95% confidence interval [CI]
Correspondence
Pauline Cliquennois, Service de 42.1%-99.6%) and 51.9% (37.8%-65.7%) with the recommended cut-off, respectively.
néonatologie, Hôpital mère-enfant, Centre An optimal cut-off of 174 pg/mL offered the same sensitivity but higher specificity
Hospitalier Universitaire de Nantes, 38
boulevard Jean Monnet, 44093 Nantes 77.8% (64.4%-88.0%).
Cedex 1, France. Conclusion: The recommended cut-off of 80 pg/mL was probably too low for ruling
Email: pauline.cliquennois@chu-nantes.fr
out NIC. A higher cut-off might have been more appropriate.

KEYWORDS

β-D-glucan, biomarker, cut-off, diagnostic accuracy, neonatal invasive candidiasis

1 |  I NTRO D U C TI O N
The incidence of proven neonatal invasive candidiasis (NIC) has
ranged from 2.6% to 13.2% in children with birth weight of <1500 g
and from 6.6% to 26.0% with birth weight of <1000 g.1,2 Candida
Abbreviations: BDG, (1 → 3)-β-D-glucan; CI, confidence interval; IQR, interquartile
range; NIC, neonatal invasive candidiasis; ROC, receiver operating characteristic.
albicans and Candida parapsilosis are the most common fungal spe-
cies in this population. 2 The incidence of NIC has decreased since
the 2000s with the control of identified risk factors, improved infec-
tion control and the use of fluconazole prophylaxis.1-5 However, NIC
is still associated with high morbidity rates in neonates and mortality
rates of about 30%. 2,6 Because of the delay in fungal culture results,
with a risk of false-negative findings, the diagnosis is difficult. That's

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

Acta Paediatrica. 2021;110:79–84.  |

wileyonlinelibrary.com/journal/apa     79
 |
80       CLIQUENNOIS et al.
why with severe NIC, clinicians tend to start empirical antifungal
treatments in case of clinical deterioration. 2 Biomarkers with high
sensitivity could allow for stopping empirical antifungal treatments
earlier in newborn infants with negative results.
(1 → 3)-β-D-glucan (BDG), a fungal cell wall polysaccharide, is a
biomarker notably recommended for the early diagnosis of invasive
fungal infections such as pneumocystis in high-risk adults.7,8 Some
studies reported the diagnostic accuracy of serum BDG in children
resulting in sensitivities from 63% to 100% and specificities from
64% to 100%.9-16 However, these studies were heterogeneous: they
used several cut-off values and different laboratory tests for BDG.
Also, diagnostic confirmation was not the same in all studies. Several
studies reported that extrinsic factors such as bacterial infection,
specific antibiotics, surgery or blood products could modify the di-
agnostic accuracy of the BDG assay.8,17,18
Because of heterogeneous data about the diagnostic use of
serum BDG in newborn infants, we assessed the diagnostic accuracy
of the BDG assay according to the cut-off value recommended by
the manufacturer for adults and currently used for newborn infants
with suspected NIC. We also searched for an optimal cut-off value
for the BDG assay for detecting and for ruling out NIC in order to
stop empirical antifungal treatments earlier.
2 |  PATI E NT S A N D M E TH O DS
2.1 | Study design and population
We conducted a prospective cross-sectional study at Nantes
University Medical Centre, France. The protocol was approved
by the local ethics committee. Parents gave their approval before
their child was included in the study. We followed the Standards
for Reporting Diagnostic Accuracy for reporting the study
19
(Table S1).
All consecutive newborn infants admitted to the neonatal inten-
sive care unit from January 14, 2017 to July 14, 2018 and with one or
multiple episodes of suspected NIC were eligible for inclusion. Those
who had a serum BDG assay were included. The routine protocol
in our neonatal intensive care unit was to perform a BDG assay for
all newborn infants with NIC suspected by clinicians before starting
antifungal treatments, when possible. Clinically suspected NIC was
identified by the physician by using clinical signs of secondary de-
terioration, such as respiratory, hemodynamic, digestive instability
and fever, associated with risk factors of NIC or lack of improvement
despite broad-spectrum antibiotic treatment. Newborn infants with
indeterminate or unavailable fungal cultures were excluded from the
study. We excluded episodes of suspected NIC when the newborn
infant was more than 28 days of corrected age or received curative
antifungal therapy for more than 72 hours. The result of BDG assay
was checked twice for confirmation. In case of disagreement, the
episode of suspected NIC was excluded. Antifungal prophylaxis with
fluconazole was administered twice a week for six weeks to new-
born infants with birth weight of <1000 g or with an endotracheal
Key notes
• This study assessed the diagnostic accuracy of serum
(1 → 3)-β-D-glucan (BDG) in newborn infants with clini-
cally suspected neonatal invasive candidiasis (NIC).
• Serum BDG showed high sensitivity (85.7%) for the di-
agnosis of NIC with the recommended cut-off of 80 pg/
mL, but low specificity (51.9%).
• The recommended cut-off was probably too low for
neonates and a higher cut-off provided better diagnos-
tic accuracy (sensitivity 85.7% and specificity 77.8%) for
ruling out NIC.
tube or a central venous catheter, with gastric fluid positive for
Candida species plurimae.
2.2 | BDG assay
For each infant, the BDG value was determined on a serum sample
by using the Fungitell assay kit (Associates of Cape Cod Inc). The
BDG assay was performed and interpreted as a blinded test with-
out all clinical information and blood culture results. A BDG value of
at least 80 pg/mL was considered a positive result according to the
manufacturer's recommendation.
2.3 | Reference standard used to define NIC
As part of routine diagnostic procedures, fungal and bacterial blood
cultures along with samples from other non-sterile body sites such as
urine without catheterisation, stool or trachea were collected for all
newborn infants with clinically suspected NIC. The diagnosis of NIC
was classified as proven, probable or possible based on criteria of the
2008 European Organisation for Research and Treatment of Cancer/
Mycoses study group adapted to neonates.10,20 NIC was classified
according to the presence of signs and symptoms of neonatal sepsis,
risk factors for invasive candidiasis, evidence of Candida colonisation
and Candida cultured from a normally sterile site (Table S2).10 On the
basis of the BDG results with blinding, newborn infants were clas-
sified as positive for NIC, including proven and probable NIC com-
bined or negative for NIC, including possible NIC.
2.4 | Data collection
Data prospectively collected from medical records included sex,
chronological age, birth weight, gestational age, mode of delivery,
comorbidities, duration of hospitalisation, patient outcome at the
end of hospitalisation and fluconazole prophylaxis. We also col-
lected the main treatments used and risk factors for NIC. Finally,
CLIQUENNOIS et al.
the results of fungal and bacterial cultures were collected. The time
interval between obtaining the BDG assay, blood culture and anti-
fungal therapy was assessed.
2.5 | Statistical analysis
First, we described the characteristics of newborn infants included
and the suspected NIC episodes according to NIC classification.
Quantitative data were analysed with median and interquartile range
(IQR) and categorical variables with numbers and percentages. We
compared newborn infants included with eligible newborn infants
not included with the Mann-Whitney test. Second, we described and
compared BDG values by NIC classification with the Mann-Whitney
test. Next, we assessed the diagnostic accuracy of serum BDG with
the recommended cut-off of 80 pg/mL in terms of sensitivity, speci-
ficity and negative and positive likelihood ratios with corresponding
95% confidence intervals (CIs) by using the exact binomial method.
Then, we used receiver operating characteristic (ROC) curve analy-
sis to study the diagnostic accuracy of serum BDG with different
cut-off values. The optimal cut-off was identified by maximising the
Youden's index. We compared the diagnostic accuracy of serum
BDG for the recommended and optimal cut-off values in terms of
sensitivity and specificity by using the McNemar test for paired data
F I G U R E 1   Flow chart of study participants
|
      81
with continuity correction, if needed. We graphically represented
both cut-off values in terms of pre-test and post-test probabilities
plotted in a Fagan nomogram. The P value <.05 was considered sta-
tistically significant. All analyses involved use of R version 3.4.2 (R
Foundation for Statistical Computing).
3 | R E S U LT S
3.1 | Characteristics of newborn infants and
suspected NIC episodes
Among 112 eligible episodes of suspected NIC in 98 newborn infants,
73 (65%) episodes in 67 (68%) newborn infants underwent BDG assay
(Figure 1). Overall, 61 suspected NIC episodes in 55 newborn infants
met the inclusion criteria, with a median chronological age of 10 days
(IQR 6.0-22.0), a median gestational age of 28.0 weeks (IQR 26.4-34.1)
and a median birth weight of 1100 g (IQR 847.5-1820.0) (Table 1). The
31 eligible newborn infants with 39 suspected NIC episodes who did
not have the BDG assay did not differ from the 55 included newborn
infants in terms of gestational age (P = .33) or birth weight (P = .17).
Of the 61 episodes of suspected NIC, two (3%) were classified
as proven NIC, five (8%) probable NIC and 54 (89%) possible NIC,
corresponding to seven (11%) and 54 (89%) episodes of positive and
82      | CLIQUENNOIS et al.

TA B L E 1   Characteristics of newborn infants by neonatal invasive candidiasis (NIC) classification

NIC classification (n = 61 episodesa )

Positive Negative
Newborn infant
  included (n = 55) Proven (n = 2) Probable (n = 5) Possible (n = 54)

Sex (female), n (%) 24 (43.6) 1 (50.0) 0 (0.0) 25 (46.3)

Chronologic age (d), median [IQR] 10.0 (6.0-22.0) 27.5 (16.3-38.8) 10.0 (10.0-13.0) 11.5 (6.3-23.0)
Birth weight (g), median [IQR] 1100.0 (847.5-1820.0) 3058.0 (2424.0-3691.0) 1200.0 (1120.0-2010.0) 1015.0 (800.0-1596.0)
Gestational age (wk), median [IQR] 28.0 (26.4-34.1) 36.8 (35.5-38.1) 29.0 (27.3-36.0) 27.5 (25.9-32.0)
Mode of delivery (caesarean), n (%) 35 (63.6) 1 (50.0) 2 (40.0) 36 (66.7)
Comorbidities, n (%) 24 (43.6) 2 (100) 2 (40.0) 21 (38.9)
Duration of hospitalisation (d), 11.0 (6.0-23.5) 27.5 (16.3-38.8) 10.0 (10.0-13.0) 11.0 (6.3-23.0)
median [IQR]
Death at the end of hospitalisation, 4 (7.3) 1 (50.0) 0 (0.0) 3 (5.6)
n (%)
Fluconazole prophylaxis (before or 21 (38.2) 0 (0.0) 0 (0.0) 26 (48.1)
at inclusion), n (%)
a
Three newborn infants with suspected NIC at different times of the study were classified differently according to the 2008 European Organisation
for Research and Treatment of Cancer/Mycoses study group criteria.

negative NIC, respectively. We did not observe any NIC in the ex-
cluded infants (no BDG available). Treatments used and some NIC
risk factors are described in Table  S2. The most frequent first-line
empirical antifungal treatment used for NIC was fluconazole. The
first episode classified as proven NIC was in an infant born at term
and who died with a C albicans maternofetal infection. This newborn
infant had positive cultures from blood, urine, stool and trachea, and
the BDG value was 25 000 pg/mL. The second one was in a preterm
infant born at 34 weeks of gestational age in a context of cardiac
surgery (transposition of the great arteries). Blood and urine cultures
were positive for C parapsilosis, and the BDG value was 7840 pg/mL.
Of the five episodes classified as probable NIC, no bacteria were
found in blood culture, no alternative diagnosis was suspected and
confirmed, and C albicans was the only species identified from urine,
stool or trachea culture. Among the 54 NIC-negative episodes, 37
(69%) had an alternative diagnosis: 19 (35%) had a positive blood cul-
ture for bacteria, six (11%) were associated with necrotising entero-
colitis, two (4%) were associated with pulmonary haemorrhage, two
(4%) were associated with virus and eight (15%) had other explana-
tions for clinical deterioration.
3.2 | Diagnostic accuracy of serum BDG
to the recommended cut-off of 80 pg/mL were 85.7% (95% CI 42.1%-
99.6%) and 51.9% (37.8%-65.7%), respectively. The positive and
negative likelihood ratios were 1.78 (1.18-2.68) and 0.28 (0.04-1.72),
respectively.
3.3 | ROC analysis
The area under the ROC curve for the BDG assay for detecting NIC
was 0.88 (0.75-1.00; Figure S1). The optimal cut-off was 174 pg/mL,
for sensitivity 85.7% (42.1%-99.6%) and specificity 77.8% (64.4%-
88.0%). The positive and negative likelihood ratios were 3.86 (2.15-
6.91) and 0.18 (0.03-1.13), respectively. The sensitivity was similar
(85.7%) but the specificity was significantly higher with the opti-
mal cut-off of 174 pg/mL than the recommended cut-off of 80 pg/
mL (77.8% vs 51.9%; P < .01). The pre-test probability of NIC was
seven (12%) in this population of newborn infants at risk of NIC.
With the optimal cut-off of 80 pg/mL, the negative and positive like-
lihood ratios yielded post-test probabilities of 18.8% (13.3%-25.9%)
and 3.5% (0.6%-18.3%), respectively (Figure  S2). With the optimal
cut-off of 174  pg/mL, the negative and positive likelihood ratios
yielded post-test probabilities of 33.4% (21.9%-47.3%) and 2.3%
(0.4%-12.8%), respectively (Figure S2).

The BDG value was assessed for the 54/61 (89%) of the suspected
NIC episodes in the first 48 hours after the clinical suspicion and
blood culture sample collection. The median BDG value was sig-
nificantly higher in NIC-positive episodes (proven and probable NIC
combined: 605 pg/mL, IQR 338-4448; proven NIC: 16 420 pg/mL,
IQR 12  130-20  710; and probable NIC: 500  pg/mL, IQR 176-605)
than in NIC-negative episodes (75 pg/mL, IQR 31-165) (P < .01). The
sensitivity and specificity of serum BDG for detecting NIC according
4 | D I S CU S S I O N
In a high-risk population of newborn infants under fluconazole
prophylaxis, we founded that the BDG assay showed good sensitiv-
ity for the diagnosis of NIC with the recommended cut-off of 80 pg/
mL (85.7%) but poor specificity (51.9%). Even with a good area under
the ROC curve value (0.88), the specificity was improved with the
CLIQUENNOIS et al.
cut-off of 174 pg/mL but remained poor (77.8%) in terms of our ob-
jective to rule out the diagnosis of NIC in order to stop empirical
antifungal treatments.
In comparison with previous findings, the median values of BDG
increased with the clinical probability of NIC, such as 16 420 pg/mL
with proven NIC, 500  pg/mL with probable NIC and 75  pg/mL with
possible NIC, which agrees with previous studies.10-13,16 A system-
atic review of newborn infants reported a sensitivity from 75% to
100% and a specificity from 47% to 100% using the same BDG assay
16
(Fungitell) with the recommended cut-off of 80  pg/mL. Moreover,
some studies suggested that the cut-off value used in adults was too
low for children 21,22 because many factors in neonates could increase
the BDG values. Indeed, the different cut-off values suggested in neo-
11,12,23
nate infants in previous studies ranged from 106 to 305 pg/mL.
Otherwise, the diagnostic performance may vary depending on the
BDG assay kit used such as Fungitell, Dynamiker Fungus or GKT-5M
and the studied population.13,24 Indeed, some studied populations in
previous publications did not reflect neither current clinical practices
nor the characteristics of newborn infants at high risk of NIC admitted
to neonatal intensive care units.
Despite some strengths, especially the choice of a targeted
population with clinically suspected NIC reflecting the current
clinical practices in neonatal intensive care units, our study has
limitations. First, it was performed in a single centre. Second, sev-
eral eligible newborn infants with clinically suspected NIC during
the study period could not be included because the BDG assay
was not performed. Finally, the scarcity of proven and probable
episodes of NIC in this population limits the statistical power and
the external validity of the study. The proposed optimal cut-off
of 174  pg/mL should be validated in other populations and set-
tings before any recommendation can be made. We did not include
enough children to analyse environmental factors associated with
false-positive BDG results as observed in previous studies.7,8,25
Such factors included the use of cotton gauze during surgery,
blood product transfusions, broad-spectrum antibiotics or antifun-
8,17,18
gal therapy and systemic bacterial infections. For example,
Goudjil et al reported that blood product transfusion were associ-
ated with increased BDG value in neonates.17
5 | CO N C LU S I O N
The BDG assay seems potentially useful for optimising antifungal
treatments by decreasing the treatment duration after ruling out
the diagnosis of NIC in case of negative results. The recommended
cut-off of 80  pg/mL for the BDG assay is probably too low for a
population of newborn infants at high risk of NIC. Higher cut-off
values may be more appropriate to help the clinician in evaluating
the risk of NIC and deciding to stop empirical antifungal treatments.
AC K N OW L E D G E M E N T S
We thank all the personal of neonatal intensive care unit who con-
tributed to the study. We also thank Dr Frédéric Dalle, university
|
      83
hospital of Dijon, France, for his biological expertise. We thank
Laura Smales for correcting the English language several times in the
manuscript.
C O N FL I C T O F I N T E R E S T
The authors declare no conflicts of interest.
ORCID
Pauline Cliquennois  https://orcid.org/0000-0001-5525-9931
Pauline Scherdel  https://orcid.org/0000-0003-4517-9430
Rose-Anne Lavergne  https://orcid.org/0000-0001-9339-1562
Cyril Flamant  https://orcid.org/0000-0002-2665-8244
Florent Morio  https://orcid.org/0000-0002-5859-6105
Jeremie F. Cohen  https://orcid.org/0000-0003-3572-8985
Elise Launay  https://orcid.org/0000-0002-4081-4653
Christele Gras Le Guen  https://orcid.org/0000-0001-9060-2344
REFERENCES
1. Aliaga S, Clark RH, Laughon M, et al. Changes in the inci-
dence of candidiasis in neonatal intensive care units. Pediatrics.
2014;133:236-242.
2. Benjamin DK, Stoll BJ, Gantz MG, et al. Neonatal candidia-
sis: epidemiology, risk factors, and clinical judgment. Pediatrics.
2010;126:e865-e873.
3. Fisher BT, Ross RK, Localio AR, Prasad PA, Zaoutis TE. Decreasing
rates of invasive candidiasis in pediatric hospitals across the United
States. Clin Infect Dis. 2014;58:74-77.
4. Cleminson J, Austin N, McGuire W. Prophylactic systemic anti-
fungal agents to prevent mortality and morbidity in very low birth
weight infants. Cochrane Database Syst Rev. 2015;(10):CD003478.
5. Kaufman DA, Morris A, Gurka MJ, Kapik B, Hetherington S.
Fluconazole prophylaxis in preterm infants: a multicenter case-con-
trolled analysis of efficacy and safety. Early Hum Dev. 2014;90(Suppl
1):S87-S90.
6. Stoll BJ, Hansen NI, Adams-Chapman I, et al. Neurodevelopmental
and growth impairment among extremely low-birth-weight infants
with neonatal infection. JAMA. 2004;292:2357-2365.
7. Cuenca-Estrella M, Verweij PE, Arendrup MC, et al. ESCMID* guide-
line for the diagnosis and management of Candida diseases 2012: di-
agnostic procedures. Clin Microbiol Infect. 2012;18(Suppl 7):9-18.
8. Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guideline
for the management of candidiasis: 2016 update by the Infectious
Diseases Society of America. Clin Infect Dis. 2016;62:e1-e50.
9. Huppler AR, Fisher BT, Lehrnbecher T, Walsh TJ, Steinbach WJ.
Role of molecular biomarkers in the diagnosis of invasive fungal dis-
eases in children. J Pediatric Infect Dis Soc. 2017;6:S32-S44.
10. Cohen JF, Ouziel A, Matczak S, et al. Diagnostic accuracy of serum
(1,3)-beta-d-glucan for neonatal invasive candidiasis: systematic re-
view and meta-analysis. Clin Microbiol Infect. 2020;26(3):291-298.
11. Cornu M, Goudjil S, Kongolo G, et al. Evaluation of the (1,3)-be-
ta-D-glucan assay for the diagnosis of neonatal invasive yeast in-
fections. Med Mycol. 2018;56:78-87.
12. Goudjil S, Kongolo G, Dusol L, et al. (1–3)-beta-D-glucan levels in
candidiasis infections in the critically ill neonate. J Matern Fetal
Neonatal Med. 2013;26:44-48.
13. Liu Y, Chen F, Zhu X, Shen L, Zhang SX. Evaluation of a novel plasma
(1,3)-beta-d-glucan detection assay for diagnosis of candidemia in
pediatric patients. J Clin Microbiol. 2015;53:3017-3020.
14. Mackay CA, Ballot DE, Perovic O. Serum 1,3-beta-D-Glucan assay
in the diagnosis of invasive fungal disease in neonates. Pediatr Rep.
2011;3:e14.
 |
84       CLIQUENNOIS et al.
15. Montagna MT, Coretti C, Lovero G, et al. Diagnostic performance
of 1→3-beta-d-glucan in neonatal and pediatric patients with can-
didemia. Int J Mol Sci. 2011;12:5871-5877.
16. Ramos JT, Villar S, Bouza E, et al. Performance of a quantitative
PCR-based assay and beta-d-glucan detection for diagnosis of inva-
sive candidiasis in very-low-birth-weight preterm neonatal patients
(CANDINEO Study). J Clin Microbiol. 2017;55:2752-2764.
17. Goudjil S, Chazal C, Moreau F, Leke A, Kongolo G, Chouaki T. Blood
product transfusions are associated with an increase in serum
(1–3)-beta-d-glucan in infants during the initial hospitalization in
neonatal intensive care unit (NICU). J Matern Fetal Neonatal Med.
2017;30:933-937.
18. Liss B, Cornely OA, Hoffmann D, Dimitriou V, Wisplinghoff H.
1,3-beta-D-Glucan contamination of common antimicrobials. J
Antimicrob Chemother. 2016;71:913-915.
19. Cohen JF, Korevaar DA, Altman DG, et al. STARD 2015 guidelines
for reporting diagnostic accuracy studies: explanation and elabora-
tion. BMJ Open. 2016;6:e012799.
20. De Pauw B, Walsh TJ, Donnelly JP, et al. Revised definitions of in-
vasive fungal disease from the European Organization for Research
and Treatment of Cancer/Invasive Fungal Infections Cooperative
Group and the National Institute of Allergy and Infectious Diseases
Mycoses Study Group (EORTC/MSG) Consensus Group. Clin Infect
Dis. 2008;46:1813-1821.
21. Calley JL, Warris A. Recognition and diagnosis of invasive fungal
infections in neonates. J Infect. 2017;74(Suppl 1):S108-S113.
22. Smith PB, Benjamin DK, Alexander BD, Johnson MD, Finkelman
MA, Steinbach WJ. Quantification of 1,3-beta-D-glucan levels in
children: preliminary data for diagnostic use of the beta-glucan
assay in a pediatric setting. Clin Vaccine Immunol. 2007;14:924-925.
23. Calitri C, Caviglia I, Cangemi G, et al. Performance of 1,3-be-
ta-D-glucan for diagnosing invasive fungal diseases in children.
Mycoses. 2017;60:789-795.
24. Shabaan AE, Elbaz LM, El-Emshaty WM, Shouman B. Role of
serum (1,3)-beta-d-glucan assay in early diagnosis of invasive fun-
gal infections in a neonatal intensive care unit. J Pediatr (Rio J).
2018;94:559-565.
25. Koo S, Bryar JM, Page JH, Baden LR, Marty FM. Diagnostic per-
formance of the (1→3)-beta-D-glucan assay for invasive fungal dis-
ease. Clin Infect Dis. 2009;49:1650-1659.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Cliquennois P, Scherdel P, Lavergne
R-A, et al. Serum (1 → 3)-β-D-glucan could be useful to rule out
invasive candidiasis in neonates with an adapted cut-off. Acta
Paediatr. 2021;110:79–84. https://doi.org/10.1111/apa.15321
 | |
Received: 23 December 2019    Revised: 22 April 2020    Accepted: 24 April 2020

DOI: 10.1111/apa.15331

REGULAR ARTICLE

Hypothermic treatment for neonatal asphyxia in low-resource

settings using phase-changing material—An easy to use and
low-cost method

Hang T. T. Tran1,2,3  | Ha T. T. Le1 | Hanh T. P. Tran1 | Dung T. K. Khu1,4 |

Hugo Lagercrantz5 | Dien M. Tran2 | Birger Winbladh5 | Lena Hellström-Westas6 |
Tobias Alfvén3,7  | Linus Olson1,4,5,6,8

1
Neonatal Intensive Care Unit, Vietnam
National Children’s Hospital (VNCH), Hanoi, Abstract
Vietnam Aim: To evaluate whether phase-changing material can be used for therapeutic hypo-
2
Research Institute for Child Health, Hanoi,
thermia of asphyxiated newborns in low-resource settings.
Vietnam
3
Department of Global Public Health,
Methods: Prospective interventional study of asphyxiated term infants fulfilling
Karolinska Institutet, Stockholm, Sweden criteria for hypothermia treatment at Vietnam National Children's Hospital from
4
Training and Research Academic September 2014 to September 2016. Hypothermia was induced within 6 hours after
Collaboration, Vietnam-Sweden
5 birth and maintained for 72 hours by a phase-changing material mattress with melt-
Department of Women’s and Children’s
Health, Karolinska Institutet, Stockholm, ing point of 32°C. Rectal temperature was continuously measured, and deviations
Sweden
6
from target temperature range 33.5-34.5°C were recorded.
Department of Womens and Children’s
Health, Uppsala University, Uppsala, Sweden Results: In total 52 infants (mean gestational age 39.3  ±  1.1  weeks) included and
7
Sachs’ Children and Youth Hospital, cooled, the median temperature at initiation of cooling was 35.3 (IQR 34.5-35.9)°C.
Stockholm, Sweden
The median time to reach target temperature was 2.5 (IQR 2-3) hours. The mean tem-
8
Department of Clinical Research and
Education, Södersjukhuset, Karolinska
perature during the cooling phase was 33.95 ± 0.2°C. Throughout the cooling phase,
Institutet, Stockholm, Sweden the target temperature range (33.5-34.5°C) was maintained more than 80% of the

Correspondence
time. Rate of rewarming was 0.5 ± 0.14°C/hour.
Hang T. T. Tran, Neonatal Intensive Care Conclusion: Phase-changing material can be used as an effective cooling method.
Unit, Vietnam National Children’s Hospital
(VNCH), Hanoi, Vietnam.
Though not a servo-controlled system, it is easy to induce hypothermia, maintain
Email: hang231@gmail.com target temperature and rewarm infants in a slow and controlled manner without need

Funding information
for frequent changes and minimum risk of skin injury.
Swedish Research Council
KEYWORDS

asphyxia, cooling, encephalopathy, low-income setting, phase-changing materials

Abbreviations: aEEG, amplitude-integrated electroencephalography; CFM, cerebral function monitoring; HIE, hypoxic-ischemic encephalopathy; IQR, Interquartile range; MRI, magnetic
resonance imaging; NICU, neonatal intensive care unit; OBGY, obstetrics and gynaecology; PCM, phase-changing material; PLIC, posterior limb of the internal capsule; SD, standard
deviation; TOBY trial, Total Body Hypothermia; TPN, total parenteral nutrition; VNCH, Vietnam national children's hospital.
Tobias Alfvén and Linus Olson equal contributed to this study.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in
any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2020 The Authors. Acta Paediatrica published by John Wiley & Sons Ltd on behalf of Foundation Acta Paediatrica

Acta Paediatrica. 2021;110:85–93.  wileyonlinelibrary.com/journal/apa     85|

 |
86       TRAN et al.
1 |  I NTRO D U C TI O N
Neonatal hypoxic-ischemic encephalopathy (HIE) in term infants
constitutes a serious public health problem which often has life-
long neurological consequences. An estimated 3-5 out of every
1000 live term births are affected,1 a quarter of which have severe
symptoms; 10%-30% of affected children do not survive; 25-30%
of HIE survivors will have long-term neurodevelopmental dis-
abilities that include cerebral palsy, seizure disorder and mental
retardation. 2 The incidence may be 10-fold higher in low-income
countries. 3
Clinical trials in term newborns in high-income countries have
shown that therapeutic hypothermia is effective and safe as a treat-
ment for neonatal encephalopathy caused by birth asphyxia.4,5
Three systematic reviews6–8 concluded that therapeutic hypother-
mia can significantly reduce death and medium-term disability after
neonatal encephalopathy and that it is safe in an intensive care set-
ting. A prerequisite is that the treatment is started within 6  hours
after birth and, ideally, as soon as possible.9,10
There is still limited experience with hypothermia treatment
in low- and middle-income countries, and a great heterogeneity in
study protocols limits the evaluation.11 Current methods of cooling
are expensive, need continuous power supply and a range of expen-
sive consumables, and are very difficult to use during transporta-
tion,12 a major issue because it is hard to reach a neonatal unit with
adequate resources for cooling treatment in time. Phase-changing
materials (PCMs) appear to be ideal for such purposes. PCMs do
not require electricity; they are biologically safe for humans, are
low-cost and can be reused almost infinitely, and they are likely to
provide a more stable cooling temperature than other ‘low tech’
methods such as ice.13
A PCM is a substance (usually a salt hydride, fatty acid, and ester
or paraffin, such as octadecane) with a high heat of fusion, which
is capable of storing or releasing large amounts of energy per unit
weight (or volume) while maintaining a given temperature.14,15 Like
water, a PCM has three phases: solid, liquid and steam, although only
the solid-liquid change is used. A PCM substance with a set ‘melting
point’ of, for example 33.5°C will be solid at room temperatures of
~25°C.
When the human body (37°C) is in contact with a PCM, the PCM
will absorb body heat, and the temperature of the PCM itself will
rise. When the PCM reaches its melting point, a great deal of addi-
tional energy is absorbed from the body in contact with the PCM,
which remains at its specific set temperature until the phase change
eventually occurs. It is this natural property of PCMs that effectively
cools contacting bodies for extended periods of time. A PCM works
as a heat sink and thus stabilises the temperature of whatever it is
in contact with.
So far, there have been several animal and human trials world-
wide 14,16,17
that have used PCMs as a mode of cooling neonates.
Results from the multicentre randomised controlled trial in India
concluded that therapeutic hypothermia of neonates with HIE
Keynotes
• Cooling is standard treatment for neonatal hypoxic-
ischemic encephalopathy; however, there are limited
resources for cooling equipment in low-middle income
countries.
• We show that phase-changing material (PCM) can in-
duce hypothermia effectively and maintain temperature
at target range during 80% of the treatment period.
• Future studies are needed to determine the feasibility of
PCM during transportation.
using a PCM-based cooling device was feasible and safe when
practiced in level 3 NICUs. A PCM-based device was comparable
to standard servo-controlled equipment in maintaining the target
temperature.16,17
Based on results from previous studies, we proposed PCMs as a
cooling technique in order to achieve hypothermia without the need
for electricity or water, and without risk of overcooling in order to
treat more patients who otherwise would be at risk for HIE damages
of the brain, and lower the amount of injuries due to asphyxia or
HIE. The aim of this study was to determine the feasibility of whole
body cooling to 33.5-34.5°C for 72 hours by use of a PCM mattress
in a low-resource setting, such as at Vietnam National Children's
Hospital (VNCH).
2 | M E TH O DS
2.1 | Study participants
This non-randomised intervention study took place in Hanoi be-
tween September 2014 and September 2016. Asphyxiated term
newborn infants from 8 provincial hospitals with an average 2-hour
transport distance from Vietnam National Children's Hospital
(Hanoi Obstetrics and Gynecology Hospital, National Hospital
of Obstetrics and Gynecology, Ha Dong General Hospital, Thai
Binh Provincial Hospital of Obstetrics and Pediatrics, Hai Duong
Provincial Hospital of Pediatrics, Bac Giang Provincial Hospital
of Obstetrics and Pediatrics, Vinh Phuc Provincial Hospital of
Obstetrics and Pediatrics, Hung Yen Provincial Hospital) who ful-
filled criteria for hypothermia treatment after birth asphyxia and
referred to VNCH were included in the study, as there was no in-
born delivery at VNCH.
2.2 | Study procedures
The study process is summarised in Figure 1.18,19
TRAN et al.
F I G U R E 1   Summary of study procedures
2.3 | Patient recruitment and consent
2.3.1 | Referring hospitals
Doctors at the obstetric hospitals identified patients potentially
20
eligible for cooling using eligibility criteria, as in the TOBY study
(Table 1): upon being granted verbal consent (yes/no) in which the
parents were explained about the newborn's condition and thera-
peutic hypothermia treatment, and once contact was made with the
on-call doctor at VNCH NICU for a transfer. During transportation,
passive cooling was used (removing clothes, no heater). 21 Basic data
regarding the mother (pregnancy complications, mother having any
sign of infection, type of delivery, time of ruptures of membranes,
amniotic fluid check and contact information) and the child (place of
birth, gestational age, weight, resuscitation Y/N, Apgar score, first
gasp in min, whether the infant was heated) was collected.
2.3.2 | At VNCH
On arrival at VNCH, all patients were evaluated again for eligibility
by the NICU doctor. Written consent form was acquired from any
one of the parents who was transferred with the infant; in all cases,
it was the father because the mother was still at the hospital. Once
TA B L E 1   Criteria for cooling
Infants ≥ 36 wk gestational age and ≤6 h after birth, and at least one
A criterion and one B criterion
Criteria A (perinatal factors)
Apgar score ≤ 5 at 10 min
Continued need for resuscitation, 10 min
pH < 7.0 and/or base deficit >16 mmol/L (Blood gas within 60 min)
Criteria B (encephalopathy)
Altered consciousness
Abnormal tone (hypertonia or hypotonia)
Abnormal primitive reflexes
Exclusion
>6 h after birth at time of referral/evaluation; coagulopathy with
active bleeding; prenatally diagnosed syndromes, malformations or
metabolic disorders not compatible with survival.
|
      87
a patient was enrolled, the complete obstetric history was obtained,
and the degree of HIE was determined using Sarnat's clinical stag-
ing system (Appendix 1). Basic patient data were recorded, includ-
ing respiratory symptoms, oxygen dependence, oxygen saturation,
capillary refill time, convulsions (clinical and/or seizures on aEEG),
vitality (scored as spontaneous activity, alertness, reaction to stimu-
lation and muscular tonus), need for immediate shock therapy and
time from birth until hypothermia treatment was started, seizures
before hypothermia treatment and highest registered rectal tem-
perature before treatment.
The infant was put on the PCM mattress (Medical Cooling
Sweden AB; by TST AB) and nursed in a cot naked or lightly covered
with a cotton sheet. The PCM mattress that was used contained 8
PCM packs with a specific melting point of 32°C, built into two lay-
ers of 4 × 2 PCM sheets (illustrated in Figure 2). The mattress was
covered with a fabric with possibilities for providing temperature
changes between the PCM and the patient through thermal conduc-
tance. A rectal temperature probe was inserted 2-3 cm into the rec-
tum and was connected to a multiparameter monitor to monitor core
temperature. The target rectal temperature was 33.5°C-34.5°C. The
temperature was continuously monitored and recorded every hour
for 72 hours. If the temperature decreased below 33.5°C, a folded
cloth sheet was inserted between the infant and the PCM mattress,
and the infant was covered with a sheet. If the temperature reached
<33.0°C (lower alarm limit), the steps taken to increase the core
temperature were to place a sheet between the PCM bed and the
infant, to cover the infant with another sheet or to switch on the
radiant warmer. The radiant warmer was used in manual mode with
an output of 10% to begin, and was then adjusted in increments or
decrements of 5%, depending on the infant's temperature.
After 72 hours of cooling, the infant was rewarmed by removing
the mattress and then naturally rewarming in room temperature to
a normal temperature by no more than 0.5°C/h. The radiant warmer
was added for rewarming if the targeted temperature of 36°C was
not reached.
2.3.3 | Treatment and monitoring
During the hypothermia treatment, standard medical care and treat-
ment of the NICU were given to all patients. During cooling, infants
 |
88       TRAN et al.
were sedated with an intravenous infusion of low-dose morphine
10 μg/kg/hour as soon as cooling started. Phenobarbital was admin-
istered if clinical and/or an aEEG seizure were detected (loading dose
20 mg/kg and maintenance dose 10 mg/kg/day). aEEG was used to
monitor patient during active cooling and for 24 hours after rewarm-
ing was completed. All patients were given total parenteral nutrition
(TPN) 40 mL/kg/day and nil by mouth for 72 hours. Adjustments for
TPN on successive days were made individually based on patients’
kidney functions, and serum sodium and fluid balances. Kidney and
liver function, serum sodium, potassium, calcium and magnesium
were tested daily during treatment. The severity of the infant's en-
cephalopathy was assessed and scored daily during hospitalisation
and at discharge using encephalopathy scores.
2.4 | Outcomes
2.4.1 | Primary outcome
Stability of rectal temperature during treatment as determined by
time out of target temperature range 33.5-34.5 C (%).
2.4.2 | Secondary outcomes
Number of temperature corrections needed, time to reach target
temperature (hour), rate of rise in temperature during rewarming
(°C/h), seizure episodes during cooling and during rewarming, brain
MRI (normal/abnormal), mortality during and after hypothermia
treatment (%).
2.5 | Data analysis
Demographic factors and clinical characteristics were summarised
with counts (percentages) for categorical variables, mean (standard
deviation [SD]) for normally distributed continuous variables or me-
dian (interquartile or entire range) for other continuous variables.
F I G U R E 2   A phase-changing mattress
setting. The left part of the figure shows
the mattress with the eight PCM packs
and foam to isolate from below, as well to
make it more comfortable for the infant;
the photograph is from VNCH during
cooling (photograph with permission of
the parents)
Demographic and clinical variables were compared using unpaired
comparisons obtained from chi-square for categorical variables and
t tests for continuous variables.
2.6 | Ethics
This study followed its procedures in accordance with the ethical
standards of the responsible committee on human experimenta-
tion and with the Helsinki Declaration of 1975, as revised in 1983,
and was approved by the Ethical Review Board of National Hospital
of Pediatrics Research Institute for Child Health (RICH) (NHP
- RICH- 13-002).
Verbal consent was elicited at the local hospital by the referring
doctor regarding the transfer for possible cooling therapy. Official
written consent was obtained from the parents when they came to
the neonatal ward of VNCH by the admitting doctor.
3 | R E S U LT S
In this study, a total of 52 newborns were cooled using PCMs be-
tween September 2014 and September 2016. The mean gestational
age was 39.3 ± 1.1 weeks. Of the 52 infants, 12 cases (23%) were
delivered in a emergency cesarean section due to complications of
labour such as foetal distress, cord prolapse or shoulder dystocia.
Regarding Apgar scores, only 53.8% infants had scores registered. In
many other cases, the referring hospitals did not give Apgar score so
we have to use other A criterion to assess whether the infant fulfilled
cooling criteria. The baseline characteristics of the mothers and the
infants prior to therapeutic hypothermia treatment are, respectively,
shown in Tables 2 and 3. We only have information on complications
at delivery for 14 mothers; others were missing from the referring
hospitals.
Regarding infants’ levels of encephalopathy according to Sarnat
scoring, 3(6%) infants had mild, 38 (73%) had moderate and 11 (21%)
had severe encephalopathy; 30 (57.7%) infants had convulsions (clin-
ical and/or aEEG) on admission. Of those 30 infants with convulsion,
TRAN et al. |
      89

TA B L E 2   Maternal baseline characteristics of enrolled infants
(n = 52)
Baseline characteristic
Age – y
Intrapartum complications
Foetal heart rate deceleration
Cord prolapse
Shoulder dystocia
Emergency ± cesarean delivery
TA B L E 3   Baseline characteristics of enrolled infants (n = 52)
Baseline characteristic
was 5  ±  1.23  hours. The temperature profiles of all 52 infants are
shown in Figures 3 and 4.
No % Among the 30 infants that presented with seizures on admis-
sion, seizure was controlled with phenobarbital in 25 cases and 5
27 ± 6  
remained with seizures until rewarming phase. Cooling had to be dis-
continued for 4 (7.7%) infants because parents decided to withdraw
10 19
treatment due to poor neurological prognosis based on cerebral
2 4
function monitoring (aEEG) and seizure control. Forty-eight infants
2 4
who completed cooling therapy received brain MRI at 7 days of age.
12 23 Twenty-seven (56.2%) had normal MRI, 8 had abnormal signal inten-
sity within the posterior limb of the internal capsule (PLIC), and 13
had diffuse abnormalities.
The median hospitalisation time was 12 days (IQR 8-18).
No %

Age when cooling started – h 3.25 ± 1.1  

Male sex – No 35 67 4 | D I S CU S S I O N
Birth weight – grams 3360 ± 620  
a In this study, we have shown that infants with HIE can be cooled ef-
Apgar score ≤ 5 at 10 min   28 54
fectively using a PCM mattress.
Lowest blood gas pHa  7.12 (6.8-7.25)  
All patients were born at local hospitals with travel distances of
Status on admission to VNCH
approximately 2-3  hours or more during rush hours by ambulance
Seizure (clinical and/or aEEG) 30 58
without proper transport incubators, and temperature and vitals
Use of anticonvulsant agent 24 46
monitoring during transportation. Therefore, most of them presented
Hypotension 8 15 with low temperatures at initiation of cooling, median 35.3 (IQR
Sepsis 5 10 34.5-35.9)°C. This is lower than in the NICHD trial (36.6 ± 1.0°C),5
Coagulation disorder 2 6 where all patients were born at the hospital where cooling was per-
Mechanical ventilation 50 96 formed. However, in comparison with Thayyil S et al trial in India
Total duration of hospital care (d) (35.2  ±  1.3°C),17 our mean temperature was slightly higher. In the
Median 12   Indian trial, no information was provided on whether patients were

IQR 8-18   inborn or transported. The study also pointed out about that PCM

a
Even though in the protocol Apgar score and blood gas at 60 min were
asked for, some referring hospitals had not been capable of sending this
information so we had to use other A criterion to assess whether the
infant fulfilled cooling criteria. Blood gas was only available for 32/52
infants, in 3 out of 6 hospitals possibility for analysing blood gas not
available.
24 (80%) were treated with an anticonvulsant agent, either pheno-
barbital or midazolam, or both, at referring hospitals.
A total of 3744 temperature readings were recorded for 52
patients during 72  hours of therapeutic hypothermia treatment.
Median temperature at initiation of cooling was 35.3 (IQR 34.5-
35.9)°C, which was also the temperature on arrival at VNCH as the
patients were immediately placed on PCM. Median time for reaching
the target temperature of 33.5°C was 2.5 (IQR 2-3) hours.
The mean  ±  SD temperature during the cooling phase was
33.95 ± 0.2°C. Throughout the cooling phase, the target tempera-
ture range (33.5-34.5°C) was maintained 80% of the time. The oc-
currence of temperatures reading below 33.5°C, above 34.5° and
below 32°C was 10.21%, 9.23% and 0.58%, respectively. No reading
above 35.5° was recorded within 72 hours for all infants.
As for the rewarming phase, the rate of rewarming was
0.5 ± 0.14°C/hour, and the average time for infants to reach 36.5°C
cooling was effective only when the ambient temperature was
<28°C; in this study, we always set the room temperature inside the
nursery at 26-27°C and continuously measured and regulated during
treatment. In the current study, hypothermia treatment was initiated
in all infants within 6 hours, although it would have been even better
if we could have started within 3 hours, due to reasons mentioned
by Thoresen et al.10
Though not a servo-controlled system, it was possible to in-
duce hypothermia to reach the target temperature using the
PCM mattress within 2.5 (IQR 2-3) hours. Compared to Thayyil
et al trial,17 the time taken to reach the target temperature in
our study was longer, but their temperature at initiation was also
lower than ours, and the PCM used in their study had lower melt-
ing points compared to ours (29°C vs 32°C),22 which could have
caused the differences in effectivity. Both studies required more
time to reach the target temperature compared to the trial using
servo-controlled equipment, as in that conducted by the NICHD
(within 90 minutes). 5 On the other hand, we feel that our way of
cooling offers more stability and less of fluctuation than some
servo-controlled systems. This is because our PCM mattress is in
very close contact with the skin of the patient and PCMs do not
have a swing in phase as a servo-controlled system does, where
the temperature is fed back to the processor of the servo-control
 |
90       TRAN et al.
F I G U R E 3   Temperature during hypothermia treatment of 52 patients
F I G U R E 4   Temperature range during treatment of 52 patients
system, which adjusts the temperature of the cooling system to
reach the intended temperature of the patient, who will then try
to compensate while getting a lower temperature then that which
is fed back to the servo-control system. Normally, the adjustment
at the servo-control system is performed every minute.12
In terms of the ability to maintain the target temperature of
33.5°C-34.5°C, our results show that the PCM mattress provided
stable body temperatures during cooling and that infants had tem-
peratures below 32°C only 1% of the time. However, since the PCM
mattress is a low-technology cooling device, it requires trained
nurses and pre-specified interventions in order to maintain the
target temperature. Interventions mentioned in the study proce-
dures such as inserting a folded cloth sheet between the infant and
the PCM mattress, covering the infant with another sheet or switch-
ing on the radiant warmer, were required in all cases. In the study by
Thomas et al, 22 different types of PCM blocks with different melt-
ing points (29°C and 21°C) were used for interventions similar to
those of our study, but with PCM blocks added or removed to adjust
temperatures. Many other ‘low-tech’ cooling methods, such as ice or
frozen gel packs, are also labour intensive, and may result in marked
temperature fluctuations and shivering, with a potential loss of neu-
roprotective efficacy. 23–25
TRAN et al.
The rewarm of cooling infants was initiated in a slow and con-
trolled manner (0.5°C/h) using PCMs. The advantage of PCMs over
cool gel packs is that there is no need for frequent changes, there
are fewer fluctuations in temperature and the risk of skin injury is
minimised, which in this study resulted in no patients developing
any skin ailment during treatment. An even more stable tempera-
ture might be possible using a blanket with PCMs integrated into the
fabric; some non-medical fabrics are now on the market for other
temperature spans; however, their use in a study such as this would
26
require a medically approved version with our target temperature
Working in a low- to middle-income setting in a country on a
fast track towards modern health care in the most advanced hospi-
tals is challenging. Vietnam is one of the fastest growing economies,
and technology and knowledge are rapidly changing. The staff size
at hospitals is always limited, and the staff members need to learn
how to operate new equipment and to increase general medical
knowledge at the same time as they are busy handling the ward.
Also, what was needed and recorded in medical journals at the be-
ginning of the study is not the same as in the later phases. This is
even more noticeable in referral hospitals. The theoretic knowledge
at the referral hospitals and the receiving hospital, and the amount
of patient history obtained at admission, improved during the study.
Examples of other challenges are that the ward where the study
was performed had a severe shortage of doctors and nurses who
were caring for 150-200 neonatal patients in a small and crowded
space designed for 80 patients. The ward also had various types
of medical equipment from different donors that were often being
used without appropriate training, which created handling prob-
lems, a problem frequently observed in NICUs around the globe.
The situation improved when a part of one room was reserved for
hypothermia that could be isolated from air conditioning units and
fans. The standard practice of care in Vietnam is that severely ill
neonates are isolated from their parents due to practical reasons,
mainly limited space and crowded wards. The goal is to have care-
takers more involved and closer to their newborns, as many studies
have shown that the first connection between mother and child is
of critical importance 27,28
The study has some limitations. The nursing interventions during
which it was discovered that the temperature readings were out of
target range were not objectively measured. Our study was not de-
signed to incorporate variations in endogenous thermogenesis or
environmental temperatures. A different ambient temperature, es-
pecially in cases of diurnal temperature swing, may also affect the
efficacy, stability and life expectancy of PCMs. These factors have
to be carefully considered when interpreting the findings. Further,
the effective cooling duration of the PCM mattress and its optimal
melting point for newborn infants needs to be addressed more thor-
oughly; shielding the patients with a transparent cover that keeps
the infants on the mattress and also reflects outside temperature
swings better will be done during follow-up studies.
We could not use the eEEG/CFM as an inclusion criteria, since it
was not available at the local hospitals and could not be used during
transportation. The evidence of eEEG/CFM from other studies shows
|
      91
it is valuable if used correctly with the right timing. 29,30 However, it
should be used during the whole treatment to be comparable with
other studies.
5 | CO N C LU S I O N
Use of a PCM mattress is an effective and low-cost method of cool-
ing infants in low- to middle-income countries, and probably during
transportation anywhere. However, as this is not a servo-controlled
system, careful monitoring and good nursing care are needed, es-
pecially during the induction and rewarming phases. However, this
caveat applies to conventional servo-controlled systems as well.
Further studies are needed to look at the influence of the environ-
mental temperature on the performance of PCM mattresses and to
determine their feasibility during transportation. Larger studies in
other clinical settings are also needed to detect more unusual com-
plications and handling problems with the method. Further, the op-
timal usage of PCM mattresses (if used for other brief terms, such
as for short-distance transportation, which is a potential follow-up
study to this one) and the optimal melting point for newborn infants
need to be addressed in depth.
AC K N OW L E D G E M E N T S
This work was undertaken at VNCH, which received a proportion
of funding from the Swedish Research Council (VR), the Swedish
Foundation for International Cooperation in Research and Higher
Education, STINT (through the TRAC Collaboration Sweden-
Vietnam). The authors would like to thank Medical Cooling Sweden
for supplying the PCM mattress used in this study and all the clini-
cal staff, from the director of the hospital to the nurses and clean-
ing staff of the Neonatal and Neurology departments at VNCH, for
all their hard work. We would also like to acknowledge the support
of Karolinska Institutet, the Research Institute for Child Health and
Vietnam National Children's Hospital for giving us time to do im-
portant research that might save lives or help newborns with HIE to
enjoy a better long-term outcome.
C O N FL I C T O F I N T E R E S T
None of the authors have any conflict of interest to present.
ORCID
Hang T. T. Tran  https://orcid.org/0000-0003-0254-2535
Tobias Alfvén  https://orcid.org/0000-0002-2328-3512
Linus Olson  https://orcid.org/0000-0003-0046-6348
REFERENCES
1. Kurinczuk JJ, White-Koning M, Badawi N. Epidemiology of neona-
tal encephalopathy and hypoxic-ischaemic encephalopathy. Early
Hum Dev. 2010;86(6):329-338.
2. Vannucci RC, Perlman JM. Interventions for perinatal hypoxic-isch-
emic encephalopathy. Pediatrics. 1997;100(6):1004-1014.
3. Lawn JE, Cousens S, Zupan J. 4 million neonatal deaths: when?
Where? Why? Lancet. 2005;365(9462):891-900.
 |
92       TRAN et al.
4. Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head cooling
with mild systemic hypothermia after neonatal encephalopathy:
multicentre randomised trial. Lancet. 2005;365(9460):663-670.
5. Shankaran S, Laptook AR, Ehrenkranz RA, et al. Whole-body hypo-
thermia for neonates with hypoxic-ischemic encephalopathy. N Engl
J Med. 2005;353(15):1574-1584.
6. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder TE, Davis
PG. Cooling for newborns with hypoxic ischaemic encephalopathy.
Cochrane Database Syst Rev. 2013;(1):Cd003311.
7. Schulzke SM, Rao S, Patole SK. A systematic review of cooling for
neuroprotection in neonates with hypoxic ischemic encephalopa-
thy - are we there yet? BMC Pediatr. 2007;7:30.
8. Shah PS, Ohlsson A, Perlman M. Hypothermia to treat neonatal
hypoxic ischemic encephalopathy: systematic review. Arch Pediatr
Adolesc Med. 2007;161(10):951-958.
9. Thoresen M. Cooling the asphyxiated brain - ready for clinical tri-
als? Eur J Pediatr. 1999;158(Suppl 1):S5-S8.
10. Thoresen M, Tooley J, Liu X, et al. Time is brain: starting therapeu-
tic hypothermia within three hours after birth improves motor out-
come in asphyxiated newborns. Neonatology. 2013;104(3):228-233.
11. Montaldo P, Pauliah SS, Lally PJ, Olson L, Thayyil S. Cooling in a
low-resource environment: lost in translation. Semin Fetal Neonatal
Med. 2015;20(2):72-79.
12. Robertson NJ, Kendall GS, Thayyil S. Techniques for therapeutic
hypothermia during transport and in hospital for perinatal asphyxial
encephalopathy. Semin Fetal Neonatal Med. 2010;15(5):276-286.
13. Prashantha YN, Suman Rao PN, Nesargi S, Chandrakala BS, Balla
KC, Shashidhar A. Therapeutic hypothermia for moderate and se-
vere hypoxic ischaemic encephalopathy in newborns using low-cost
devices – ice packs and phase changing material. Paediatr Int Child
Health. 2019;39(4):234-239.
14. Iwata S, Iwata O, Olson L, et al. Therapeutic hypothermia can be
induced and maintained using either commercial water bottles or a
"phase changing material" mattress in a newborn piglet model. Arch
Dis Child. 2009;94(5):387-391.
15. Yang G, Yim YJ, Lee JW, Heo YJ, Park SJ. Carbon-filled organic
phase-change materials for thermal energy storage: a review.
Molecules. 2019;24(11):E2055.
16. Higgins RD, Raju T, Edwards AD, et al. Hypothermia and other
treatment options for neonatal encephalopathy: an executive sum-
mary of the Eunice Kennedy Shriver NICHD workshop. J Pediatr.
2011;159(5):851.e1-858.e1.
17. Thayyil S, Shankaran S, Wade A, et al. Whole-body cooling in neo-
natal encephalopathy using phase changing material. Arch Dis Child.
2013;98(3):F280-F281.
18. Age & Stages questionnaires. https://agesa​ndsta​ges.com. Accessed
October 9, 2018.
19. Haataja L, Mercuri E, Regev R, et al. Optimality score for the neuro-
logic examination of the infant at 12 and 18 months of age. J Pediatr.
1999;135(2 Pt 1):153-161.
20. Azzopardi D, Brocklehurst P, Edwards D, et al. The TOBY study.
Whole body hypothermia for the treatment of perinatal asphyx-
ial encephalopathy: a randomised controlled trial. BMC Pediatr.
2008;8:17.
21. Hallberg B, Olson L, Bartocci M, Edqvist I, Blennow M. Passive in-
duction of hypothermia during transport of asphyxiated infants: a
risk of excessive cooling. Acta Paediatr. 2009;98:942-946.
22. Thomas N, Abiramalatha T, Bhat V, et al. Phase changing ma-
terial for therapeutic hypothermia in neonates with hypoxic
ischemic encephalopathy - a multi-centric study. Indian Pediatr.
2018;55(3):201-205.
23. Bharadwaj SK, Bhat BV. Therapeutic hypothermia using gel packs
for term neonates with hypoxic ischaemic encephalopathy in re-
source-limited settings: a randomized controlled trial. J Trop Pediatr.
2012;58(5):382-388.
24. Horn AR, Woods DL, Thompson C, Eis I, Kroon M. Selective ce-
rebral hypothermia for post-hypoxic neuroprotection in neonates
using a solid ice cap. S Afr Med J. 2006;96(9 Pt 2):976-981.
25. Shabeer MP, Abiramalatha T, Smith A, et al. Comparison of two low-
cost methods of cooling neonates with hypoxic ischemic encepha-
lopathy. J Trop Pediatr. 2017;63(3):174-181.
26. Shanghai Cloud Chemical Co. L. [cited 2019. PCM fabric]. https://
cloud​chem.en.made-in-china.com/produ​c t/Njzxv​nEhEK​k Z/China​
-Ice-Cold-Insta​n t-Cooli​n g-Towel​- Runni​n g-Sport​s-Phase​- Chang​
e-Mater​ial.html. Accessed May 13, 2018.
27. Angelhoff C, Blomqvist YT, Sahlén Helmer C, et al. Effect of
skin-to-skin contact on parents' sleep quality, mood, parent-in-
fant interaction and cortisol concentrations in neonatal care
units: study protocol of a randomised controlled trial. BMJ Open.
2018;8(7):e021606.
28. Csaszar-Nagy N, Bokkon I. Mother-newborn separation at birth
in hospitals: a possible risk for neurodevelopmental disorders?
Neurosci Biobehav Rev. 2018;84:337-351.
29. Bruns N, Blumenthal S, Meyer I, Klose-Verschuur S, Felderhoff-
Muser U, Muller H. Application of an amplitude-integrated
EEG monitor (Cerebral function monitor) to neonates. J Vis Exp.
2017;127.
3 0. de Vries LS, Hellstrom-Westas L. Role of cerebral function monitor-
ing in the newborn. Arch Dis Child Fetal Neonatal Ed. 2005;90(3):F2
01-F207.
How to cite this article: Tran HTT, Le HTT, Tran HTP, et al.
Hypothermic treatment for neonatal asphyxia in low-resource
settings using phase-changing material—An easy to use and
low-cost method. Acta Paediatr. 2021;110:85–93. https://doi.
org/10.1111/apa.15331
TRAN et al. |
      93

APPENDIX 1

Level of encephalopathy (Sarnat staging)

Severity Stage 1 (mild) Stage 2 (moderate) Stage 3 (severe)

Level of consciousness Hyperalert Lethargic or obtunded Stupor or coma

Activity Normal Decreased Absent
Neuromuscular control
Muscle tone Normal Mild hypotonia Flaccid
Posture Mild distal flexion Strong distal flexion Intermittent decerebration
Stretch reflexes Overactive Overactive Decreased or absent
Complex or primitive reflexes
Suck Weak Weak or absent Absent
Moro (startle) Strong Weak Absent
Tonic neck Slight Strong Absent
Autonomic function
Pupils Mydriasis Miosis Variable
Heart rate Tachycardia Bradycardia Variable
Seizures None Common Uncommon
 | |
Received: 10 January 2020    Revised: 28 April 2020    Accepted: 29 April 2020

DOI: 10.1111/apa.15339

REGULAR ARTICLE

How doctors communicated with parents in a neonatal

intensive care: Communication and ethical issues

Gaelle Sorin1,2  | Lionel Dany3,4 | Renaud Vialet2 | Laurent Thomachot2 |

Sophie Hassid2 | Fabrice Michel1,5 | Barthélémy Tosello1,2

1
Hospital La Timone, Aix-Marseille
University/EFS/CNRS, UMR 7268 ADÉS, Abstract
Espace Éthique Méditerranéen, Marseille, Aim: Doctors have a moral and legal obligation to keep patients and their families
France
2 informed, and this is an integral part of care. We explored the communication strate-
Department of Neonatal Medicine, North
Hospital, Assistance Publique-Hôpitaux de gies used by doctors when they spoke to parents in a French neonatal intensive care
Marseille, Marseille, France
unit (NICU).
3
Aix-Marseille University, Aix-en-Provence,
France
Methods: This was a single-centre qualitative pilot study carried out from October
4
Department of Oncology, Hospital La 2015 to January 2016. We asked five doctors (three female) to audiotape their dis-
Timone, Assistance Publique-Hôpitaux de cussions with the parents of newborn infants during their NICU stay. The doctors’
Marseille, Marseille, France
5 mean age was 43 years, and they had a mean of 14 years of NICU experience. Each
Department of Pediatric Intensive Care
Unit, Hospital La Timone, Assistance- discussion was subjected to thematic content analysis.
Publique des Hôpitaux de Marseille,
Results: We analysed 40 discussions carried out between doctors on 26 newborn
Marseille, France
infants. Five communication strategy themes emerged: building understanding, how
Correspondence
the communication was constructed, the role of the doctor, and of the parents, in
Gaelle Sorin, Department of Anesthesia
and Intensive Care, Neonatal and Pediatric the overall care of the newborn infant and how the information given to the parents
Intensive Care Unit, Hospital Nord, AP-HM,
developed over time.
13015 Marseille, France.
Email: gaelle.sorin@ap-hm.fr Conclusion: Analysing the content of the information discussed with parents pro-
vided us with the opportunity to understand the communication and ethical issues
surrounding the delivery of information in a NICU. This could be used to improve
future discussions between doctors and parents.

KEYWORDS

audio recording, content analysis, decision-making, medical information, neonatal intensive

care unit

1 | I NTRO D U C TI O N
Communication between doctors and their patients and families
1
is one of the most important elements of medical care for three
key reasons. Firstly, it enables shared decision-making.1,2 Secondly,
it creates good inter-personal relationships, which is important be-
cause the perceived quality of these relationships can be regarded
as a prerequisite for medical care. Thirdly, it promotes interaction
Abbreviations: NICU, neonatal intensive care unit.
between the relevant parties so that they can exchange information,
ask questions and share views.
The environment of the neonatal intensive care (NICU) presents
healthcare staff and parents with particular challenges with regard
to communication and information. It is important that parents are
kept informed while their baby is in the NICU, because they are con-
fronted with feelings of shock and stress and need to cope with fear
and to deal with anticipation and uncertainty.3-5 The parents’ emo-
tions have consequences with regard to their ability to concentrate,

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

|
94     
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:94–100.
GAELLE et al.
communicate and make decisions.6 In addition, if the parents have
unrealistically optimistic expectations, this can lead to a discrepancy
between the information given by healthcare staff and the parents’
understanding of that information.7
In most studies, parents have expressed specific concerns about
their need for information.8 Complete information helped them to
trust healthcare professionals and provided them with a sense of
control. They also wanted to control how they participated in the de-
cision-making process.9,10 These needs were a challenge for health-
care staff in the NICU, especially when it came to enhancing the
psychosocial support they offered parents.6
Medical communication in the NICU has been shown to pres-
ent healthcare professionals with specific professional, ethical and
moral dilemmas. For example, neonatologists provided information,
but made parents aware that there was some uncertainty.11,12 The
trajectory of the sick newborn infant guided the information that
was given to infants. This trajectory could be considered as a whole
or broken down by certain phases or important events, such as
pre-admission, acute, crisis, downward or death phases.13 Possible
deterioration of the infants’ clinical states and his emotional ups and
downs experienced by the parents had an impact on the possibility
4
and the capabilities to inform parents. Some studies reported that
information and medical decision gradually built up over time.14,15,16
A literature review on the information needs of parents of chil-
13
dren admitted to a NICU found that only three of the 78 studies
used audio or videotaped recording of conferences between parents
and healthcare professionals.
The aim of this pilot study, which was based on the natural set-
ting of a NICU, was to improve knowledge of the communication
strategies used by doctors when they talked to parents.
2 |  PATI E NT S A N D M E TH O DS
2.1 | Design
This was a qualitative exploratory study, conducted with doctors
practising in one French NICU. Each physician used a digital dicta-
phone to record their usual discussions with parents of newborn
infants hospitalised in the unit over a period of four months. The
reasons for the discussions were left to the discretion of the doc-
tors, who did not use a standardised discussion grid. The aim was to
capture the doctors’ natural communication.
2.2 | Data analysis
The recorded discussions were fully transcribed verbatim and an-
onymised. Several items were collected: the place of the discussion,
the parents’ socio-professional category, which parents partici-
pated in the discussion, the gestational age of the newborn infant,
the infant's clinical status at the time of any discussion and their
clinical status at the end of their NICU stay. Each discussion was
|
      95
Key Notes
• This qualitative study focused on five doctors who au-
diotaped 40 discussions with the parents of 26 newborn
infants in a French neonatal intensive care unit.
• Five communication strategy themes emerged that
could help inform future discussions.
• These were as follows: building understanding, how the
communication was constructed, the role of the doctor,
and of the parents, in the overall care of the newborn in-
fant and how the information provided developed over
time.
subjected to thematic content analysis.17 This analysis consisted of
building a thematic structure based on analysing the unit of sense
of the discussions. A unit of sense is the part of the data that gives
meaning to the speech and it can range from a number of words
to paragraphs. The units of sense were grouped by themes and
sub-themes in order to build a global thematic structure. This was
carried out by the first two authors (GS, LD) on behalf of all the
researchers. All the discussions were used to build the thematic
structure. These two investigators each read the first discussion
carefully in order to identify the units of sense. They then modified
their initial thematic structure by using content analysis on the fol-
lowing discussions. These modifications consisted of adding to, and
reorganising, the themes and sub-themes. During working group
meetings, all the authors shared the coding for the units of sense,
agreed on the analysis of any ambiguous units of sense and devel-
oped a common thematic structure. This common thematic struc-
ture was finalised by the tenth discussion.
3 | R E S U LT S
A total of five doctors (three women) with an average age of 43 years
old agreed to participate: three had more than 10 years of experience
in the NICU, and the mean NICU experience of the five doctors was
14  years. A total of 40 recordings were made between October 1,
2015, and January 31, 2016. There was an average of 1.6 ± 1.0 re-
cordings for each of the 26 children, according to the duration of their
stay in the unit. Doctors recorded an average of 8.0 ± 5.4 discussions.
Of the 40 discussions, 32 (80%) were in the presence of both parents,
6 (15%) only with the mother and 2 (5%) only with the father. Twenty
newborns (76.9%) were released alive, and 6 died (23.1%) from the
unit. The characteristics of the 26 infants are reported in Table 1 and
the characteristics of interviews in Table 2.
The content analysis (Table 3) shows that five main themes
emerged from the analysis. These were as follows: building under-
standing, how the communication was constructed, the role of the
doctor, and of the parents, in the overall care of the newborn infant
and how the information given to the parents developed over time.
96      | GAELLE et al.

TA B L E 1   Characteristics of the 26 newborn infants A good understanding of the information transmitted is essential
Birth, n (%) for the involvement of parents and the good doctor-patient relation-
Premature 20 (76.9) ship. ‘Did she explain to you a little bit about what happened? What

Term
Gestational age, mean (SD) (weeks)
Male gender, n (%)
Singletons, n (%)
Maternal age, mean (SD) (years)
Parents in low socio-professional category, n (%)
Length of stay (days), median (IQR)
Clinical status at the end of the stay, n (%)
Alive
Died
Number of meeting per child, mean (SD)
Duration of meeting (minimum), median, (IQR)
a 3.2 | Construction of communication
In five cases, death occurred after treatment was limited or withdrawn;
SD, standard deviation; IQR, interquartile range.
TA B L E 2   Characteristics of interviews
 
Place, n (%)
6 (23.1)
did you understand about the situation?’
The doctors use the information given upstream and sum-
31 (± 4.6)
marise the essentials. ‘What we said with your husband was the
13 (50)
context of the birth and the fact that she was resuscitated at birth
15 (57.6)
with the need to give her a cardiac massage, which is very bad
30.5 (± 5.2)
actually for babies as small. I think that's the information you had
17/22 (77.2)
from Dr…’.
13 (8-23) Throughout the interview, the doctor pauses for a good under-
standing of the information given. It will then be able to validate or
20 (76.9) correct the parent's understanding and continue to deliver the infor-
a
6   (23.1) mation (feedback). ‘Exactly, and now we've seen a brain bleed’.
1.6 (± 1.02)
11 (7.5-18)
This theme was of concern to doctors throughout the discussions.
Several sub-themes were found to be a necessary part of this element.
It was important that the doctor was clear about the information
Interviews that the parents had received during the initial discussions and to
(n = 40) make sure that they had understood what they had been told. As one
doctor said: ‘You had questions already about what was explained

Child's room 17 (42.5) to you?’

The doctors tried to use simple vocabulary and explained the
Room dedicated to interviews 18 (45)
medical information the parents had received in a way that they
Mother's room 5 (12.5)
could understand. As one doctor said: ‘So it's not an illness in the
Parents' prior meeting, n (%)
sense that it's something that your baby has caught. It's something
Yes 18 (45)
her has had from the beginning. The medical term for it is not be easy
No 14 (35)
to understand or explain. It is basically the state of his lungs’.
Not specified 8 (20) A period of reflection or silence was necessary throughout this
Presence at the interview, n (%) communication to give parents the chance to take in the information
Two parents 32 (80) they had received and ask questions. At the end of the discussion,
Father only 2 (5) the doctor made it clear they were available to answer questions or
Mother only 6 (15) to meet for another discussion. As one doctor said:’I’m sure it's not
Stable state of newborn infant at the time of the interview, n (%) easy for you (long silence). We are thinking about what's best for the
Unstable 28 (70) baby … I'm here tomorrow and if you're here do not hesitate to ask

Stable 12 (30)
more. Okay.’

The first two themes were included in all of the discussions, the role
of the doctor was discussed in three-quarters and the role of the
parents appeared in about half of the discussions. The notion of pro-
gressivity of discourse figured in all the discussions.
3.1 | Building understanding
This theme was of concern to doctors throughout the discussions.
Several sub-themes were found to be a necessary part of this
element.
3.3 | Role of the doctor in the global care of the
newborn infant
Several rules or principles seemed to guide the doctor's speech, such
as using the phrase ‘do no harm’ to the child and other important
phrases. One doctor sought to reassure a baby girl's parents by say-
ing: ‘We are very careful about how she is, how comfortable she is,
that she does not feel pain, that she does not suffer.’
The doctor role also appeared to place the medical knowl-
edge and experience of the child and his family in similar situa-
tions. ‘Because in these terms, we know, that's what I tell you,
GAELLE et al. |
      97

TA B L E 3   Content analysis

Themes Sub-themes Examples

Build understanding
Construction of
communication
Role of the doctor in the
global care of the newborn
infant
Parents' place in the project
of care of their child
Progressivity of information
Upstream knowledge
Summary of information already
given
Make sure you understand
Validation/ correction of
understanding
Question search
Clarity
Reflection time
Maintaining the link/availability
Do not harm
Medical knowledge/experience
Collegiality/equip
Compassion/reassurance/hope/
empathy
Loyalty
Limit of medicine/humility
Role/involvement of parents
Consent/notice wanted
(Figure 1)
‘Did she explain to you a little bit about what happened? What did you
understand about the situation?’
‘What we said with your husband was the context of the birth and the
fact that she was resuscitated at birth with the need to give her a cardiac
massage, which is very bad actually for babies as small. I think that's the
information you had from Dr…’
You understand? ‘Well, I was clear?’
‘Exactly. There is risk already because he is born at 25 weeks. Okay?’
‘You had questions already about what was explained to you?’
‘So it's not an illness in the sense that it's something that your baby has
caught. It's something her has had from the beginning. The medical term for
it is not be easy to understand or explain. It is basically the state of his lungs’.
‘It's sure it's not easy for you… it's… it's (long silence) we are thinking about
what's best for the baby’
‘After me I'm here tomorrow, it's still me who takes care of (name of the child).
Okay? So anyway I'll see you tomorrow, if you have questions in the morning.
Okay?’
‘We are very careful about how she is, how comfortable she is, that she does
not hurt, that she does not suffer. We are very careful’.
‘Our role is to think about what's best for the baby. Because in these terms,
we know, that's what I tell you, on this kind of very serious hemorrhage, we
know there's a bad prognosis’ (…)’
‘With this kind of very serious haemorrhage, we involve several doctors,
because complicated decisions have to be made. These are decisions that
can have serious consequences’.
‘It's hard too but it's good… we had you explained a little at first… but it's sure
it's never easy even if you're prepared, it's always… it's never easy anyway’
‘It's sure it's not easy for you…’
‘What I said to you, the decision, you expressed your will, but it's a decision
we had… that will be medical.’
‘I sincerely think that the situation… the situation escapes us completely. It
does not make sense to go on…’
‘For you it's really about being patient and wise and coming to see your child
every day. It's important.’
‘What I told you, the decision, you made us by your will but… it was a decision
we had… medical’.

on this kind of very serious haemorrhage, we know there's a bad
prognosis’.
The doctors stressed the fact that the baby was being cared for
by a multi-disciplinary NICU team to support what they were say-
ing and the decision they had to make, especially in the most diffi-
cult situations. One doctor said: ‘Our role is to think about what's
best for the baby… With this kind of very serious haemorrhage,
we involve several doctors, because complicated decisions have to
be made. These are decisions that can have serious consequences.’
The doctors made it clear that they wanted to provide the most
accurate and regular information for the children in their care. As
one doctor said: ‘Whatever happens, whatever you see on the exams
we do, we will keep you informed about everything that's happening
to your baby …’.
3.4 | Role of the parent in the global care of the
newborn infant
This theme was the least represented in the discussions and, when it
was discussed, it seemed to focus on the role the parents played in
being there for their child and being patient. As one doctor explained
‘For you it's really about being patient and wise and coming to see
your child every day. It's important.’
 |
98       GAELLE et al.

Steps Abstracts Areas of improvement

1st step: basic uncertainty
2nd step: current situation
rd
3 step: medical knowledge
4th step: means implemented
against
5th step: diagnostic
6th step: project change
7th step: persistent uncertainty
"We are in a situation that is quite disturbing for your preamble: create a safe space
daughter, uh for now it's a bit early to say"
"So, um, it's a really, very, very small baby that's born a lot,
well, at the end of the day, it's a little bit eh 25 weeks, uh,
and it's a baby that needed a lot of assistance at birth.” Percepon: idenfy
"In general, we know that it's children as they are very level of
understanding
small and that they need very heavy care at birth ... that's
what I told you, there is a major risk of
complication...bleeding in the brain ". Summuary: Reflect Elicit Hopes, Fears
Parents’ views and Values
"After that we're never sure of anything, especially when we
did not do the exams. Tomorrow we will do an ultrasound
of the brain to see what it is, and we will monitor after its
evolution "
"We did a brain scan this afternoon which confirmed that
there was bleeding ... which is quite important"
Knowledge: assess
"It's something for which we have no way to improve things Emoon:
their ability to hear
Achnowledge
... So what we do in these situations is that we meet, we Emoons
what you have to
say
have a meeting with several doctors, (...) for talk to us about
what's best for your baby. " Invitaon:
"If it's a child who survives, if it was to survive this determine what the
parent wants to
bleeding, what we know is that the after-effects, the know

handicap after, is a handicap that is likely to be extremely

heavy. We do not know if she will be able to walk, we do not
know if she will get to have a normal intelligence, right? Is
she going to eat all by herself, is that ... all that "

F I G U R E 1   Example of the construction of information progressivity (Background: birth at 25SA + 4days, 625 g, premature rupture of
membranes, diagnosis of intraventricular and parenchymal haemorrhage stage 4)

On the other hand, the doctors made it clear that sometimes the
parents needed to be there in just an advisory capacity, but the final
decisions had to be made by the clinical staff on medical grounds.
‘What I said to you, the decision, you expressed your will, but it's a
decision we had… that will be medical.’
3.5 | Progression of the information
The progression of how the medical information was delivered to the
parents, which was identified during the content analysis process,
is demonstrated in Figure 1. This has led us to propose a number of
approaches to communication. Although these approaches do not
cover all eventualities, the basic communication principals are the
same.
4 |  D I S CU S S I O N
To our knowledge, this was one of the first studies to examine the
discussion between doctors to parents in the natural setting of a
NICU. The analysis allowed us to highlight existing communication
practices or strategies by exploring what information was provided
and how it was delivered.
Our analysis showed that how the discussions were constructed
by the doctors and understood by the parents was very important as
it shaped how any information was provided. The goals were to build
trusting relationships, kindness and empathy, and help to make the
family feel more in control, especially when it came to making deci-
sions. Delivering information is a key element of providing neonatal
counselling. Indeed, providing medical information and follow-up
allows parents to be involved in their child's decision-making and
care. It also promotes a sense of control over the situation, which is
beneficial to the relationship with the healthcare professionals, the
parents’ well-being and that of their child.18,19 Furthermore, it has
been reported that the perception of shared decision-making has
been associated with lower grief scores in the long term, compared
to paternalistic decision-making. 20 However, the results of our study
indicated that the doctors tended to place the greatest emphasis on
the information that explained the current medical situation. They
focussed less on delivering this information by exploring the history
of the child and parents or their belief systems or cultural or religious
references.
The role the doctor played in caring for the child, which was
the third theme of our analysis, was based on non-maleficence,
which is providing a person who is not able to express his or her
will. 21,22 This notion corresponds to the primum non nocere de-
scribed by Hippocrate, which translates as first do no harm. But,
unlike adult patients, newborn infants cannot express their pref-
erences. Its legal representatives are its parents or guardians who
have parental authority. Thus, the ways that doctors communicate
with parents and what they tell them are key to establishing a part-
nership of care for the child. The language that doctors use when
discussing medical conditions differs from that of parent and some
doctors protect their emotions by using very technical, rational-
ising language. Actively listening to parents is essential. Parents
GAELLE et al.
may be more receptive to a doctor's explanations if they feel that
their emotions are understood and their patenting role is rec-
ognised. 23 Trust is built by paying attention to others. It is not just
about the team's medical skills, but about their intentions towards
the child and its parents and their ability to predict the child's fu-
ture and support the child. It would have been very interesting to
explore the parents' views on how this relationship of trust was
built. These situations have a strong emotional impact on medical
staff. In addition to providing medical care, they have to take into
account the values and emotions of the parents when they make
decisions. There is a close relationship between emotions and val-
ues, because the emotions that are felt in a situation reveal the
value of a person's ethical principles. 24 Compassion or empathy
reveals the principles of doing good to others, doing no harm and
autonomy, together with fear and respect. In difficult situations,
several emotions can be felt at the same time, such as compassion
and fear. These can lead to conflicting values between doing good
to others and doing no harm and the best interests of the child
often move between these two principles. Possible divergent de-
cisions can lead to argumentative discussions, and doctors need
to listen to the views of others. That is why doctors are advised to
involve other colleagues when difficult decisions need to be made,
especially when cases raise ethical issues or they need to make
decisions about limiting or withdrawing active treatment.
Our fourth theme, on the role parents played in the care of
their child, was less developed in medical discussions, contrary
to current policies. Until recently, doctors saw their primary role
as providing parents with detailed, objective information about
treatment choices and likely outcomes. We now have a newer
approach, which recommends that doctors help parents discover
their own values and ethical commitments as they face an unan-
ticipated situation and a series of life-altering decisions. 25 Parents
find that intimate contact with their newborn infant can be diffi-
cult, painful and stressful for a number of reasons. These include
the often unstable and precarious state of their child and their in-
tense emotions following the shock of a birth that did not happen
as expected. In addition, they are suddenly immersed in their new
role as parents, while most of the time the life of their child de-
pends on complex interventions and technologies,. 6,26-28 Although
shared decision-making is the ethically preferred approach, as op-
posed to pure parental autonomy or pure medical paternalism in
29
the grey zone of ethical ambiguity, most neonatologists do not
share decision-making well. 30
The shared decision-making model may seem ideal, but rigid
schemes are not adapted to all situations, even when it comes to
20
reaching this type of decision together. Therefore, it is important
that doctors give parents the opportunity to express themselves at
different times and after reflection. The parents also need to decide
what degree of involvement they want. It is important to recognise
that parents are in unfamiliar territory when their baby is in the NICU
and their thoughts and feelings may change as current and future
situations unfold. Their positions can also be shaped by ongoing
|
      99
discussions and information and their autonomy will be weakened,
or even destroyed, if doctors do not make allowances for their
changing views. The way that they react is partly shaped by what
medical professionals say to them on an ongoing basis.
The progression of how information was provided was a key el-
ement of our analysis. In some ways, it was a bit like a movie script,
which tells a story, but also contains technical and organisational de-
tails. This progression did not seem to be a matter of chance and the
doctors in our study used it as a real communication strategy. It was
used as a way to fight against the uncertainty inherent in the situa-
tion by implementing approaches such as paraclinical examinations.
It was also a way of creating a bond with the family and organising
future meetings in the hope that the clinical picture would become
clearer with fewer possible scenarios. Lastly, delivering information
at various stages seemed to be used as a tool for mitigating informa-
tion, especially when it was considered difficult. The phenomenon
of psychologisation, which corresponds to attributing psychologi-
cal states to others, can provide the basis of this communicational
strategy. Progressivity, by essence, implies that the final information
cannot be delivered straight away, because it could have a negative
emotional or psychological effect on the recipient. The information
needs to be watered down or diluted over the course of several dis-
cussions to make it more bearable.
This study had some limitations. The constraints of the medical
activity in the NICU meant that the five doctors who took part over
the four-month period spoke to the parents of the same 26 children
more than once during the 40 discussions. In addition, the reason
for the discussions was at the discretion of the individual doctors
and we cannot neglect the fact that they may have modified how
the information was delivered because they were recording the con-
versations. Finally, the parent's responses were a vital part of the
discussions, but we were unable to use these data.
5 | CO N C LU S I O N
Our study focused on the information that doctors provided parents
during their child's NICU stay and also explored how it was provided.
This revealed five key themes that can help to inform future commu-
nication strategies. These were as follows: building understanding,
how the communication was constructed, the role of the doctor, and
of the parents, in the overall care of the newborn infant and how
the information given to the parents developed over time. The latter
theme, and the model of shared decision-making, is of particular in-
terest for future communication strategies. The findings also under-
lined the importance of making healthcare professionals feel more
comfortable in these sometimes difficult situations. A new study is
currently underway, concerning the medical information delivered in
critical care situations.
C O N FL I C T S O F I N T E R E S T
The authors have no conflicts of interest to declare.
 |
100       GAELLE et al.
E T H I C A L A P P R OVA L
Both physicians have given their consent to participate in this study.
An information leaflet was given to the parents of the newborn
infants by the doctors. They were they could withdraw from the
research at any time and assured that the identities would remain
confidential. Ethical approval for this research was provided by the
university ethics committee of Aix-Marseille University, France
(number 2015-06-03-001).
ORCID
Gaelle Sorin  https://orcid.org/0000-0002-5283-9030
REFERENCES
1. Ong LM, De Haes JC, Hoos AM, Lammes FB. Doctor-patient commu-
nication: a review of the literature. Soc Sci Med. 1995;40:903-918.
2. Beisecker AE, Beisecker TD. Patient information-seeking behaviors
when communicating with doctors. Med Care. 1990;28:19-28.
3. Brazy JE, Anderson BM, Becker PT, Becker M. How parents of
premature infants gather information and obtain support. Neonat
Netw. 2001;20:41-48.
4. Charchuk M, Simpson C. Hope, disclosure, and control in the neo-
natal intensive care unit. Health Commun. 2005;17:191-203.
5. Ward K. Perceived needs of parents of critically ill infants in a neo-
natal intensive care unit. Pediatr Nurs. 2001;27:281-286.
6. Hall SL, Cross J, Selix NW, et al. Recommendations for enhancing
psychosocial support of NICU parents through staff education and
support. J Perinatol. 2015;35:529-536.
7. De Wit S, Donohue P, Shepard J, Boss RD. Mother–clinician dis-
cussions in the neonatal intensive care unit: agree to disagree? J
Perinatol. 2013;33:278-281.
8. Bass LS. What do parents need when their infant is a patient in the
NICU? Neonatal Netw. 1991;10:25-33.
9. Brinchman BS, Forde R, Nortvedt P. Parents' experiences with life-
and-death decisions concerning their premature infants. Tidsskr Nor
Laegeforen. 2002;122:2098-2101.
10. Wocial LD. Life support decisions involving imperiled infants. J
Perinat Neonatal Nurs. 2000;14:73-86.
11. Alderson P, Hawthorne J, Killen M. Parents’ experiences of sharing
neonatal information and decisions: consent, cost and risk. Soc Sci
Med. 2006;62:1319-1329.
12. Vermeulen E. Dealing with doubt: making decisions in a neonatal
ward in The Netherlands. Soc Sci Med. 2004;59:2071-2085.
13. De Rouck S, Leys M. Information needs of parents of children ad-
mitted to a neonatal intensive care unit: a review of the literature
(1990–2008). Patient Educ Couns. 2009;76:159-173.
14. Tosello B, Dany L, Le Coz P, Gire C, Vriet-Ndour ME, Einaudi MA.
Perceptions of lethal fetal abnormality among perinatal profes-
sionals and the challenges of neonatal palliative care. J Palliat Med.
2014;17:924-930.
15. Orzalesi M, Aite L. Communication with parents in neonatal inten-
sive care. J Matern Fetal Neonatal Med. 2011;24:135-137.
16. Koh TH, Butow PN, Coory M, et al. Provision of taped conversa-
tions with neonatologists to mothers of babies in intensive care:
randomised controlled trial. BMJ. 2007;334:28.
17. Flick U. Thematic coding and content analysis. In: Metzler K, ed.
An Introduction to Qualitative Research, 5th edn. London: Sage
Publications; 2014:420–438.
18. Arockiasamy V, Holsti L, Albersheim S. Fathers’ experiences in
the neonatal intensive care unit: a search for control. Pediatrics.
2008;121:215-222.
19. Macdonell K, Omrin D, Pytlik K, Pezzullo S, Bracht M, Diambomba
Y. An effective communication initiative: using parents’ experiences
to improve the delivery of difficult news in the NICU. Journal of
Neonatal Nursing. 2015;21:142-149.
20. Caeymaex L, Jousselme C, Vasilescu C, et al. Perceived role in end-
of-life decision making in the NICU affects long-term parental grief
response. Arch Dis Child Fetal Neonatal Ed. 2013;98:26-31.
21. Dehan M, Gold F, Grassin M, et al. Ethical dilemmas in the peri-
natal period: guidelines for end-of-life decisions. Arch Pediatr.
2001;8:407-419.
22. Moriette G, Rameix S, Azria E, et al. Very premature births:
Dilemmas and management. Second part: Ethical aspects and rec-
ommendations. Arch Pediatr. 2010;17:527-533.
23. Boss RD, Hutton N, Sulpar LJ, West AM, Donohue PK. Values par-
ents apply to decision-making regarding delivery room resuscita-
tion for high-risk newborns. Pediatrics. 2008;122:583-589.
24. Beauchamp TL, Childress JF. Principles of Biomedical Ethics, 6th edn.
New York, NY: Oxford University Press; 2009.
25. Gaucher N, Nadeau S, Barbier A, Janvier A, Payot A. Personalized
antenatal consultations for preterm labor: responding to mothers’
expectations. J Pediatr. 2016;178:130-134.
26. Haines DKJ, Remedois L, Berney SC, Knott DC, Denehy L. The con-
ceptualisation of health-related quality of life in decision-making
by intensive care physicians: a qualitative inquiry. Aust Crit Care.
2017;30:152-159.
27. Ainbinder JG, Blanchard LW, Singer GH, et al. A qualitative study
of Parent to Parent support for parents of children with special
needs. Consortium to evaluate Parent to Parent. J Pediatr Psychol.
1998;23:99-109.
28. Bourque CJ, Dahan S, Mantha G, Robson K, Reichherzer M, Janvier
A. Improving neonatal care with the help of veteran resource par-
ents: an overview of current practices. Semin Fetal Neonatal Med.
2018;23:44-51.
29. Rosenbaum L. The paternalism preference choosing unshared deci-
sion making. N Engl J Med. 2015;373:589-592.
3 0. Lantos J. Ethical problems in decision making in the neonatal ICU.
Engl J Med. 2018;379:1851-1860.
How to cite this article: Sorin G, Dany L, Vialet R, et al. How
doctors communicated with parents in a neonatal intensive
care: Communication and ethical issues. Acta Paediatr.
2021;110:94–100. https://doi.org/10.1111/apa.15339
 | |
Received: 11 December 2019    Revised: 30 April 2020    Accepted: 5 May 2020

DOI: 10.1111/apa.15345

REGULAR ARTICLE

Early parenting intervention promotes 24-month psychomotor

development in preterm children

Camilla Pisoni1  | Livio Provenzi2  | Michela Moncecchi2 | Camilla Caporali2 |

Cecilia Naboni2 | Mauro Stronati1 | Rosario Montirosso3  | Renato Borgatti2,4  |
Simona Orcesi2,4

1
NICU, Fondazione IRCCS Policlinico San
Matteo, Pavia, Italy
2
Child Neurology and Psychiatry Unit, IRCCS
Mondino Foundation, Pavia, Italy
3
0-3 Center for the at-Risk Infant, Scientific
Institute IRCCS E. Medea, Lecco, Italy
4
Department of Brain and Behavioral
Sciences, University of Pavia, Pavia, Italy
Correspondence
Livio Provenzi, Child Neurology and
Psychiatry Unit, IRCCS Mondino
Foundation, via Mondino 2, Pavia 27100,
Italy.
Email: livio.provenzi@mondino.it
Funding information
This study was supported by funds from the
Italian Ministry of Health, Ricerca Corrente
2016, 2017, 2018, 2019, 2020, and Cinque
per Mille, 2017.
Abstract
Aim: Although parenting is key to promoting healthy development of at-risk preterm
infants, parents have often restricted access to neonatal intensive care units (NICUs).
This study aimed to assess the effect of an early parenting intervention on the psy-
chomotor outcome in preterm children at 24 months of corrected age.
Methods: Forty-two preterm children and their parents were consecutively recruited
at a level III NICU in Northern Italy and randomly allocated to early intervention
(two educational peer-group sessions and four individual infant observation ses-
sions) or care as usual (no educational or infant observation sessions). During NICU
stay, parents provided information on daily holding and skin-to-skin. Psychomotor
development was measured at 24 months of corrected age using the Griffith Mental
Development Scales.
Results: There were no significant differences in socio-demographic and clinical vari-
ables between early intervention (n = 21; 13 females) and care as usual (n = 21; 12
females) groups. At 24  months of corrected age, children in the early intervention
arm had greater scores for global psychomotor development as well as for Hearing-
Speech and Personal-Social sub-scales, compared to those in the care as usual group.
Conclusion: The present NICU parenting intervention was found to be associated
with better psychomotor outcomes in preterm children at 24-month age. The effects
were especially evident for domains related to language and socio-emotional func-
tioning. Results are promising and should be retested with more heterogeneous and
representative preterm sample.

KEYWORDS

developmental quotient, early intervention, neonatal intensive care unit, parenting, preterm
birth

Abbreviations: DQ, developmental quotient; GMDS, Griffith Mental Development Scales; NICU, neonatal intensive care unit.
The members of the Early Intervention Group are Giada Ariaudo, Linda Gasparini, Ivana Olivieri, Silvia Spairani and Giovanna Tritto (Child Neurology and Psychiatry Unit, IRCCS
Mondino Foundation, Pavia, Italy) and Lina Bollani and Chryssoula Tzialla (NICU, Fondazione IRCCS Policlinico San Matteo, Pavia, Italy)

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

Acta Paediatrica. 2021;110:101–108.  wileyonlinelibrary.com/journal/apa     101 |

 |
102       PISONI et al.
1 |  I NTRO D U C TI O N
Preterm birth is one of the major challenges for the healthcare sys-
tems worldwide, representing the leading cause of long-lasting mor-
bidities in infants and children.1 Preterm infants are especially at risk
for detrimental developmental consequences, mainly due to their
immature central nervous system and to higher probability perina-
2 dren at 24 months of corrected age.
tal injuries. The infants’ long-term hospitalisation in the neonatal
intensive care unit (NICU) often results in early separation from the
parents, thus making it difficult to establish immediate bonding and
attachment.3 During the NICU stay, preterm infants are exposed to
several sources of stress, such as high-intensity lights and sounds,
invasive procedures and painful stimulations. Overall, NICU-related
stress may further increase the risk of suboptimal and comorbid de-
velopmental outcomes.4
The NICU environment can be especially stressful for the par-
ent-child system. On the one hand, the neurobehavioural and com-
municative signals of the preterm infants may be less pronounced
and difficult to be interpret by parents.5,6 On the other hand, the
early separation from the infant and the need of specialised intensive
medical and nursing actions may lead the mothers and the fathers to
feel less confident in caregiving and more alienated in their parental
role.7-9 Thus, it is not surprising that skin-to-skin and holding care
approaches are promoted to increase the physical proximity and
emotional closeness between parents and infants, and to facilitate
10-12
attachment processes that are crucial for healthy development.
Specifically, the parental caregiving ability to read infants’ signals,
respond contingently and provide support to emotional regulation is
early interactive behaviours that may be challenged in the NICU.13-15
Moreover, it should be highlighted that the needs of both mothers
and fathers have be addressed in NICU to support their engagement
in the early caregiving and to promote an adaptive post-discharge
transition to home.16-18
As such, early interventions that focus on providing information
and psychoeducational guidelines for parents should be prioritised
in the NICU environment.19 Mother-focused parenting programmes
have been found to be successful in improving the cognitive and
behavioural outcomes of preterm-born children up to five years
of age. 20 When maternal involvement in caregiving is greatly sup-
ported, this can lead to increased scores in the psychomotor profile
of preterm children at 18  months of age. 21 A meta-analytic study
suggests that early parenting interventions in the NICU are suc-
cessful in promoting the psychomotor developmental outcomes of
preterm children with a peak between 24 and 36 months. 22 Notably,
the effects were much more pronounced for language and person-
al-social domains, rather than physical performance. While these re-
sults are promising, previous research in this field has mainly focused
on parenting interventions delivered to mothers rather than to both
the parents of preterm infants. 23,24
In the present study, we assessed the effects of an early and brief
parenting intervention that includes both psychoeducational support
on the psychomotor developmental outcomes of preterm children at
24 months of corrected age. We hypothesised that, compared to a
Key notes
• Preterm-born children are at high risk for suboptimal
psychomotor development in the first years of life.
• An early parenting intervention was successful in pro-
moting better psychomotor outcomes in preterm chil-
• Parenting interventions are needed to be implemented
precociously in order to improve the psychomotor out-
comes of preterm children.
control group exposed to usual NICU care without any dedicated
parenting support, the children of parents who received the early
intervention would show better psychomotor development.
2 | M E TH O DS
2.1 | Participants
Between January 2015 and December 2016, 42 preterm children
from the NICU of the Fondazione IRCCS Policlinico San Matteo,
Pavia, Italy, and their parents were enrolled in this study. All preterm
infants with gestational age ≤32 weeks or birthweight ≤1500 grams
were eligible for the study. Subjects were not considered eligible to
the study if at least one parent was unable to converse in Italian,
had psychiatric morbidities or reported alcohol/drug abuse, was less
than 18  years old, and if the infants had any genetic syndrome. A
complete flowchart of enrolment is reported in Figure  1. Twenty-
one children and their parents were allocated to the early interven-
tion arm, whereas twenty-one were allocated to the care as usual
arm. In order to minimise the risk of biases related to the interactions
among parents belonging to different groups, allocation occurred
in two blocks: care as usual allocation occurred between January
2015 and May 2016, whereas early intervention allocation occurred
between July and December 2016, with a wash-out period of one
month during June 2016. Enrolment occurred as soon as the infant
was not considered at risk for survival. The study was approved
by the Ethical Committee of the Fondazione IRCCS Policlinico San
Matteo and the IRCCS Mondino Foundation in Pavia, Italy. All the
parents signed an informed consent.
2.2 | Study design and procedures
Figure 2 reports an overview of the study design. The parents of chil-
dren allocated to the early intervention group participated in group
and individual sessions (Table 1). A detailed description of the interven-
tion is reported in Table S2, according to the Template for Intervention
Description and Replication (TIDieR) checklist. Two psychoeducational
group sessions of 60-90 minutes of duration were led by a paediatric
PISONI et al. |
      103

F I G U R E 1   Flow-chart of enrolment

F I G U R E 2   Timeline and procedures of the study. CA, corrected age; CU, Care as Usual; EI, Early Intervention; GMDS, Griffith Mental
Development Scales; GS, Group Session; IS, Individual Session; NE, neurological examination; NICU, Neonatal Intensive Care Unit; PCA,
post-conceptional age; SCENE, Separation and Closeness Experiences in the Neonatal Environment

neuropsychiatrist and a neurodevelopmental therapist. These sessions
provided information on typical characteristics of preterm infants, spe-
cial needs in caregiving and the procedures used by the NICU staff.
The group sessions were interspersed with up to four individual ses-
sions that were led by a trained neurodevelopmental therapist. These
sessions lasted approximately 45  minutes and consisted of a joint
observation of the infant by focusing on behaviours, temperament,
communicative signals, signs of distress and regulation processes. The
joint observation sessions were informed by the methodological ap-
proach of Dr Boukydis, namely the collaborative consultation with
Parents and Infants in the Perinatal Period,25 which integrates princi-
ples from evidence-based infant research and well-validated neurobe-
havioural assessments. These principles included both guidelines for
the joint observation of the infant and communicative skills to support
 |
104       PISONI et al.

TA B L E 1   Description of the early intervention

Session Description Participants Main themes

Group Session #1 Psychoeducational Parents Dealing with unexpected birth

intervention
Group Session #2 Psychoeducational
intervention
Individual Session #1 Conjoint collaborative
consultation
Individual Session #2 Conjoint collaborative
consultation
Individual Session #3 Conjoint collaborative
consultation
Individual Session #4 Conjoint collaborative
consultation
Child neuropsychiatrist
Neurodevelopmental therapist
Parents
Child neuropsychiatrist
Neurodevelopmental therapist
Parents
Neurodevelopmental therapist
Parents
Neurodevelopmental therapist
Parents
Neurodevelopmental therapist
Parents
Neurodevelopmental therapist
Preterm newborn skills
Physiological needs
Neurobehavioural states
Stability needs
Communicative cues
Caregiving actions
Sensory profile and stimulations
Reading the baby signals and co-regulation
principles
Holding, handling, wrapping
Sensitive stimulations
Tailored hints for neurobehavioural stability
Transition to home

parenting. On the one hand, the therapists were trained to focus on
maternal regulatory behaviours, infant communicative signals, mutual
dyadic processes, sensitive caregiving and intrusive behaviours. On
the other hand, specific relational skills included asking examples from
the parents, elicit their own view of the infant and themselves in the
interaction, active listening by the therapist, and stimulating parents’
curiosity about infants’ signals and behaviour. The parents allocated to
the care as usual group received no structured parenting support, and
there were no differences in how the NICU staff cared for the infants.
Any parents of the care as usual group who reported high levels of psy-
chological distress could be referral for a psychotherapist consultation.
two groups. Mothers and fathers completed the diary separately.
The parental closeness diary allows the collection of quantitative
data on the presence of parents, time spent holding the infant and
time spent in skin-to-skin contact. This instrument has been used in
previous research with NICU parents. 26 The holding time score was
computed as the percentage of the total time spent by parents hold-
ing the infant with respect to the time they were present in the unit.
The skin-to-skin time score was computed as the percentage of the
total time spent by parents in skin-to-skin contact with respect to
the time they were present in the unit. The absolute values (in min-
utes) for presence, holding and skin-to-skin are reported in Table S1.

2.3 | Socio-demographic variables and 2.5 | Psychomotor outcome

clinical variables
The following socio-demographic variables were collected: infants’
sex, gestational age, birthweight, Apgar score at the first and fifth
minute, parental age and educational level measured as years of
study. The clinical characterisation of the children included the
following variables: vaginal or Caesarean delivery, being small for
gestational age, intrauterine growth restriction, bronchopulmonary
dysplasia, retinopathy of prematurity, periventricular leukomalacia,
ultrasound scan results and clinical outcomes of neurological exami-
nations at discharge, term-equivalent age and 3 months of corrected
age.
The Griffith Mental Development Scales (GMDS) is a well-validated
and standardised tool for assessing psychomotor development of
children up to 8  years. It is based on five subscales that provide
scores for the following domains: locomotor, personal-social, hear-
ing-speech, hand-eye coordination and performance. A standard-
ised developmental quotient (DQ) score (range: 50-150) is obtained
for each domain (mean = 100, standard deviation = 16) and for the
global score (mean = 100, standard deviation = 12). The GMDS has
been extensively used in previous research with preterm infants and
children.13,22 The clinician administering the GMDS was blind to the
random allocation.

2.4 | Parental presence, holding and skin-to- 2.6 | Plan of analysis

skin care
Parents were asked to complete the parental closeness diary26 start-
ing from the day of enrolment up to discharge, to control for differ-
ences related to the amount of skin-to-skin and holding between the
Both groups were compared for socio-demographic and clini-
cal variables as well as for holding and skin-to-skin percentage
scores by means of pair-wise mean comparisons (continuous vari-
ables), non-parametric Mann-Whitney test (discrete variables) and
PISONI et al.
chi-square test (dummy variables). General linear models were
used to assess the presence of significant group differences in
the total and domain DQ scores of the GMDS. Statistical analyses
were carried out using SPSS Statistics for Windows, version 25
(SPSS Inc.) setting P < .05.
3 | R E S U LT S
Descriptive statistics for the socio-demographic and clinical charac-
teristics as well as for holding and skin-to-skin percentage scores are
reported in Table 2. There were no significant differences between
the two groups for any of these variables. Apgar score at the first
(W = 381.00, P > .05) and fifth minute (W = 432.50, P > .05) did not
significantly differ between early intervention.
A significant difference emerged for the total GMDS score
(t(40) = 2.40, P = .02) (Figure 3). Children from parents in the early
intervention arm had higher global DQ compared to those in the
care as usual arm. Moreover, a significant multivariate effect was de-
tected (F (5,36) = 4.49, P = .003, η2p = 0.38). Significant univariate dif-
ferences emerged for two domains: personal-social (F (1,40) = 14.93,
2
P < .001, η p = 0.27) and hearing-speech (F (1,40) = 5.36, P = .026,
2
η p  =  0.12). For both the domains, children of parents in the early
intervention group had higher DQ scores compared to those in the
care as usual group. These effects persisted even when the analyses
were restricted to children without cerebral palsy.
4 | D I S CU S S I O N
In the present study, we report on the effects of an early NICU
psychoeducational parenting intervention including both moth-
ers and fathers on the psychomotor outcome of preterm infants at
24  months of CA. Children of parents who took part in the early
intervention programme had greater global DQ and higher scores
in the personal-social and hearing-speech domains of the GMDS
assessment.
Previous research has showed that early interventions that ac-
tively involve the parents in informative and/or psychoeducational
programmes have the greatest long-term efficacy in reducing det-
rimental psychomotor and cognitive outcomes. 27 Moreover, the
effect of maternal caregiving support may be especially observed
between 24 and 36 months, with the greatest benefits reported for
socio-emotional and behavioural domains. 22 Our findings further
extend previous evidence and suggest that engaging both parents
in an early and brief parenting intervention by providing informa-
tive support and psychoeducational guidance in understanding the
infant's communicative and behavioural signals may be significantly
beneficial for the psychomotor development of preterm children.
The finding that parental support may be more effective in im-
proving social-related domains, rather than motor or physical com-
28
petences, resonates with previous reports in older children. This
specific effect may suggest pathways of neuroprotection. First, one
|
      105
may speculate that promoting parental knowledge through a psy-
choeducational training aimed at improving caregiving skills may
benefit the quality of early parent-infant relationship so that the
infant receives adequate stimulations and contingent responses.
From this perspective, the evidence that social-related domains are
the ones most impacted by early parenting interventions is not sur-
prising. Second, the lack of significant effects on motor and physical
performance domains further highlights the need to develop more
integrated programmes capable of acting on multiple levels of the
developmental trajectories of the preterm child. Consistently, future
research should explore the role of parental engagement in physical
therapies for preterm infants and children.
Notably, this intervention contributes to the existing literature
on the importance of early family-centred interventions in the
NICU.12,29,31 The intervention reported here moves from theoreti-
cal and methodological principles of the collaborative consultation
approach and adopted both psychoeducational group sessions and
tailored joint infant observation sessions with parents. The main goal
of the intervention was to increase parental knowledge of factors
related to infant's development and to improve both mothers and
fathers’ skills in reading the baby's signals, responding contingently
and supporting emotional regulation. As previously suggested,33,33
investing resources in improving parental involvement in NICU is a
critical protective factor for the development of preterm infants and
children. Although these findings promisingly support the relevant
of early parenting interventions in the NICU, in the light of the lim-
ited sample size and the dropout rate reported for the present study,
they should be retested with more heterogeneous and representa-
tive preterm samples in future research.
This study had limitations. First, the sample size was relatively
small. As such, we could not include more complex models in our
statistical plan to control for potential confounders such as life
events, parental educational style and socio-economic conditions.
Nonetheless, the two groups were not significantly different in
terms of socio-demographic and clinical characterisation, which
indirectly indicates the robustness of the study and the reliability
of findings. Second, parenting a preterm infant is often associated
with increased levels of stress, depression and anxiety.30,34,35 We
could not exclude that affective problems may have had a role in
moderating the effects of the intervention on children outcomes.
For example, parents with significant depressive and/or anxious
symptomatology may be less able to engage in psychoeducational
programmes focused on parent-child interaction. Third, the selected
sample included low-risk families, as single parents and those with
psychiatric morbidities were not eligible for the study. As such, our
findings only have partial generalisability to the broader population
of families of preterm infants hospitalised in the NICU. Finally, the
domain specificity of these findings indirectly suggests that a po-
tential process underlining the effectiveness of the intervention may
be ascribed to the improved daily quality of parent-child interaction
or at least to sensitive and protective parenting behaviour. 22 As
such, we suggest that future studies assessing the psychomotor out-
comes of early parenting interventions in the NICU should include
 |
106       PISONI et al.

TA B L E 2   Descriptive statistics for socio-demographic, clinical and NICU care activities variables

Intervention arms

Early Intervention (EI) group (N = 21) Care as Usual (CU) group (N = 21) Group comparison

Mean SD Min Max Mean SD Min Max t test sig

Socio-demographic characteristics
Infants' gestational age 29.6 3.0 23.0 34.0 30.4 2.8 24.0 35.0 0.91 0.367
(wk)
Infants' birthweight 1244 345 630 1700 1309 307 654 1695 0.65 0.521
(grams)
Maternal age (y) 34 5 25 39 35 5 28 43 1.76 0.087
Maternal education (years 15 4 5 22 16 4 5 22 0.23 0.818
of study)
Paternal age (y) 35 4 26 43 38 5 27 47 2.12 0.041
Paternal education (years 14 4 5 22 15 4 5 22 0.69 0.497
of study)
N % N % Χ2 sig
Infants' sex (females) 13 61.9 — — 12 57.1 — — 0.10 0.753
Couple marital status 21 100 — — 20 95.2 — — 1.02 0.311
(cohabitant)
N % N % t test sig
Clinical characteristics
Small for gestational age 2 9.5 — — 1 4.8 — — 0.36 0.549
(SGA)
Intrauterine growth 2 9.5 — — 2 9.5 — — 0.00 1.000
restriction (IUGR)
Bronchopulmonary 7 33.3 — — 5 23.8 — — 0.47 0.495
dysplasia (BDP)
Retinopathy of 5 23.8 — — 3 14.3 — — 0.62 0.432
prematurity (ROP)
Periventricular 0 0.0 — — 1 4.8 — — 1.02 0.311
leukomalacia (PVL)
Severe anomalies at 0 0.0 — — 1 4.8 — — 1.20 0.549
ultrasound scan
Cerebral palsy 1 4.8 — — 1 4.8 — — 0.00 1.000
Abnormal neurological 7 33.3 — — 8 38.1 — — 0.10 0.747
examination (discharge)
Abnormal neurological 9 42.9 — — 7 33.3 — — 0.89 0.346
examination (term-
equivalent age)
Abnormal neurological 6 28.6 — — 2 9.5 — — 2.47 0.120
examination (3 mo)
Mean SD Min Max Mean SD Min Max t test sig
NICU care activities (%)
Maternal holding 45 28 13 79 31 24 1 79 1.59 0.120
Maternal skin-to-skin 25 27 3 99 9 13 2 46 1.87 0.074
Paternal holding 37 22 2 83 28 27 1 85 0.96 0.344
Paternal skin-to-skin 14 9 4 24 6 6 1 19 1.77 0.101
Parental holding 39 23 10 82 32 25 5 82 0.84 0.411
Parental skin-to-skin 15 8 4 22 8 10 2 35 1.04 0.321
PISONI et al.
F I G U R E 3   Griffith Developmental Mental Scales (GDMS) Developmental Quotient (DQ) scores for children of parents in the Care as
Usual (CU) or Early Intervention (EI) groups. Error bars represent Standard Errors. *P < .05; ***P < .001
video-recording of parent-child dyadic and triadic interactions to
better highlight potential relational mechanisms of action.
5 | CO N C LU S I O N
We found that a brief and early psychoeducational parenting inter-
vention during the NICU stay was significantly associated with im-
proved psychomotor outcome in children at 24 months of corrected
age. Social-related domains were significantly improved, suggesting
that the potential pathways of protective action may be related to
improvements in the quality of parent-child relationships in the first
months of life. As parenting is advocated to be the major preventive
factor for the child health and well-being,36 these findings suggest
that investing in early parenting engagement and caregiving skills
may be crucial to further promote better developmental outcomes
in preterm children.
C O N FL I C T S O F I N T E R E S T
The authors have no conflicts of interest to declare.
ORCID
Camilla Pisoni  https://orcid.org/0000-0001-9292-9682
Livio Provenzi  https://orcid.org/0000-0001-7424-8744
Rosario Montirosso  https://orcid.org/0000-0002-2175-3082
Renato Borgatti  https://orcid.org/0000-0001-8165-4994
Simona Orcesi  https://orcid.org/0000-0001-8947-9342
REFERENCES
1. Blencowe H, Cousens S, Chou D, et al. Too Soon: The global epide-
miology of 15 million preterm births. Reprod Health. 2013;10(Suppl
1). Available from http://www.scopus.com/inwar​d/record.
|
      107
url?eid=2-s2.0-84888​14881​1&partn​e rID=40&md5=893a0​7b4ba​
d2d61​7c8ac​831e5​dd62cd1
2. Lester BM, Miller RJ, Hawes K, et al. Infant neurobehavioral devel-
opment. Semin Perinatol. 2011;35:8-19.
3. Nyström K, Axelsson K. Mothers’ experience of being separated
from their newborns. J Obstet Gynecol Neonatal Nurs. 2002;31:275-
282. Available from http://www.scopus.com/inwar​d/record.
url?eid=2-s2.0-00365​81688​& partn​e rID=40&md5=5d70e​8 ec34​
daa6f​5552e​4a94a​a6a11cd
4. Montirosso R, Provenzi L. Implications of epigenetics and stress
regulation on research and developmental care of preterm infants.
JOGNN - J Obstet Gynecol Neonatal Nurs. 2015;44:174-182.
5. Muller-Nix C, Forcada-Guex M, Pierrehumbert B, Jaunin L, Borghini
A, Ansermet F. Prematurity, maternal stress and mother–child
interactions. Early Hum Dev. 2004;79:145-158. Available from
ht tps://linki​n ghub.elsev​i er.com/retri​e ve/pii/S 0378​3 7820​
4000775
6. Pisoni C, Spairani S, Fauci F, Ariaudo G, Tzialla C, Tinelli C, et al.
Effect of maternal psychopathology on neurodevelopmental out-
come and quality of the dyadic relationship in preterm infants: an
explorative study. J Matern Neonatal Med. 2020;33(1):103-112.
7. Montirosso R, Provenzi L, Calciolari G, Borgatti R. Measuring
maternal stress and perceived support in 25 Italian NICUs. Acta
Paediatr Int J Paediatr. 2012;101:136-142. Available from http://
www.scopus.com/inwar​d /record.url?eid=2-s2.0-84855​7 3344​
2&partn​erID=MN8TOARS
8. Aftyka A, Rozalska I, Rybojad B, Samardakiewicz M. Polish version
of the Parental Stressor Scale: neonatal intensive care unit. Ann
Agric Environ Med. 2019;26:67-72. Available from http://www.ncbi.
nlm.nih.gov/pubme​d/30922032
9. Franck LS, Cox S, Allen A, Winter I. Measuring neonatal intensive
care unit-related parental stress. J Adv Nurs. 2005;49:608-615.
10. Hoffenkamp HN, Tooten A, Hall RAS, et al. The impact of prema-
ture childbirth on parental bonding. Evol Psychol. 2012;10:542-561.
11. Provenzi L, Fumagalli M, Bernasconi F, et al. Very preterm and full-
term infants’ response to socio-emotional stress: the role of post-
natal maternal bonding. Infancy. 2016;1-18. Available from http://
doi.wiley.com/10.1111/infa.12175
12. Welch MG, Hofer MA, Brunelli SA, et al. Family nurture interven-
tion (FNI): methods and treatment protocol of a randomized con-
trolled trial in the NICU. BMC Pediatr. 2012;12:14 http://www.
 |
108       PISONI et al.
pubme​dcent​r al.nih.gov/artic​leren​der.fcgi?artid​=33940​87&tool=p-
mcen​trez&rende​r type​=abstract
13. Fumagalli M, Provenzi L, De Carli P, et al. From early stress to
12-month development in very preterm infants: preliminary find-
ings on epigenetic mechanisms and brain growth. PLoS ONE.
2018;13:e0190602. Available from http://www.ncbi.nlm.nih.gov/
pubme​d/29304146
14. Feldman R. Infant-mother and infant-father synchrony: the co-
regulation of positive arousal. Infant Ment Health J. 2003;24:1-
23. Available from http://www.scopus.com/inwar​d/record.
url?eid=2-s2.0-35430​8 3158​& partn​e rID=40&md5=55b1a​d 3a63​
086f0​71486​188b9​fc312e5
15. Tronick E, Beeghly M. Infants' meaning-making and the develop-
ment of mental health problems. Am Psychol. 2011;66:107-119.
Available from http://psycn​et.apa.org/journ​als/amp/66/2/107/
16. Provenzi L, Santoro E. The lived experience of fathers of preterm
infants in the neonatal intensive care unit: a systematic review
of qualitative studies. J Clin Nurs. 2015;24(13-14):1784-1794.
Available from http://doi.wiley.com/10.1111/jocn.12828
17. Guillaume S, Michelin N, Amrani E, et al. Parents' expectations of
staff in the early bonding process with their premature babies in the
intensive care setting: a qualitative multicenter study with 60 par-
ents. BMC Pediatr. 2013;13:18. Available from http://www.biome​
dcent​ral.com/1471-2431/13/18
18. Obeidat HM, Bond EA, Callister LC. The parental experience of
having an infant in the newborn intensive care unit. J Perinat Educ.
2009;18:23-29.
19. Hall SL, Cross J, Selix NW, et al. Recommendations for enhanc-
ing psychosocial support of NICU parents through staff educa-
tion and support. J Perinatol. 2015;35(Suppl 1):S29-S36. Available
from http://www.pubme​dcent​ral.nih.gov/artic​leren​der.fcgi?artid​
=46600​46&tool=pmcen​trez&rende​r type​=abstract
20. Nordhov SM, Rønning JA, Ulvund SE, Dahl LB, Kaaresen PI. Early
intervention improves behavioral outcomes for preterm infants:
randomized controlled trial. Pediatrics. 2012;129(1):e9-e16.
21. Lester BM, Hawes K, Abar B, et al. Single-family room care and neu-
robehavioral and medical outcomes in preterm infants. Pediatrics.
2014;134:754-760. Available from http://pedia​trics.aappu​blica​
tions.org/cgi/doi/10.1542/peds.2013-4252
22. Vanderveen J a, Bassler D, Robertson C m t, Kirpalani H. Early
interventions involving parents to improve neurodevelopmen-
tal outcomes of premature infants: a meta-analysis. J Perinatol.
2009;29(5):343-351. Available from 10.1038/jp.2008.229%5Cnht​
tp://www.nature.com/doifi​nder/10.1038/jp.2008.229
23. Fisher D, Khashu M, Adama EA, et al. Fathers in neonatal units:
improving infant health by supporting the baby-father bond and
mother-father coparenting. J Neonatal Nurs. 2018;24:306-312.
Available from https://linki​nghub.elsev​ier.com/retri​eve/pii/S1355​
18411​8300930
24. Feeley N, Waitzer E, Sherrard K, Boisvert L, Zelkowitz P. Fathers'
perceptions of the barriers and facilitators to their involvement
with their newborn hospitalised in the neonatal intensive care unit.
J Clin Nurs. 2013;22:521-530.
25. Boukydis Z. Collaborative consultation with parents and infants in
the perinatal period. Infant Ment Health J. 2015;36:240-241.
26. Raiskila S, Axelin A, Toome L, et al. Parents' presence and parent–
infant closeness in 11 neonatal intensive care units in six European
countries vary between and within the countries. Acta Paediatr Int J
Paediatr. 2017;106:878-888.
27. Spittle A, Orton J, Anderson P, Boyd R, Doyle LW. Early develop-
mental intervention programmes post-hospital discharge to pre-
vent motor and cognitive impairments in preterm infants. Spittle
A, ed. Cochrane database Syst Rev. 2012;12:CD005495 http://doi.
wiley.com/10.1002/14651​858.CD005​495.pub3
28. Johnson S, Ring W, Anderson P, Marlow N. Randomised trial of pa-
rental support for families with very preterm children: outcome at 5
years. Arch Dis Child. 2005;90:909-915.
29. Als H, McAnulty GB. The Newborn Individualized Developmental
Care and Assessment Program (NIDCAP) with Kangaroo Mother
Care (KMC): comprehensive care for preterm infants. Curr Womens
Health Rev. 2011;7:288-301. http://www.ncbi.nlm.nih.gov/pubme​
d/25473384
3 0. Ahlqvist-Björkroth S, Axelin A, Korja R, Lehtonen L. An educational
intervention for NICU staff decreased maternal postpartum de-
pression. Pediatr Res. 2019;85(7):982-986.
31. Provenzi L, Barello S. Behavioral Epigenetics of Family-Centered
Care in the Neonatal Intensive Care Unit. JAMA Pediatr.
2015;169(7):697-698. https://jaman​etwork.com/journ​als/jamap​
ediat​rics/artic​le-abstr​act/2337233
32. Montirosso R, Casini E, Del Prete A, et al. Neonatal developmental
care in infant pain management and internalizing behaviours at 18
months in prematurely born children. Eur J Pain (United Kingdom).
2016;20:1010-1021.
33. Montirosso R, Del Prete A, Bellu R, et al. Level of NICU quality
of developmental care and neurobehavioral performance in very
preterm infants. Pediatrics. 2012;129:e1129-e1137 http://www.
pubme​dcent​r al.nih.gov/artic​leren​der.fcgi?artid​= 40746​10&tool=p-
mcen​trez&rende​r type​=abstract
3 4. Shields-Poë D, Pinelli J. Variables associated with parental stress
in neonatal intensive care units. Neonatal Netw. 1997;16:29-
37. Available from http://www.scopus.com/inwar​d/record.
url?eid=2-s2.0-00310​62943​& partn​e rID=40&md5=6be93​49e67​
1dde3​75f2d​854cd​b6c95aa
35. Pisoni C, Spairani S, Manzoni F, Ariaudo G, Naboni C, Moncecchi
M, et al. Depressive symptoms and maternal psychological distress
during early infancy: A pilot study in preterm as compared with
term mother–infant dyads. J Affect Disord. 2019;257:470-476.
36. Perrin EC, Leslie LK, Boat T. Parenting as Primary Prevention. JAMA
Pediatr. 2016;170:637. Available from http://www.ncbi.nlm.nih.
gov/pubme​d/27182902
37. ProvenziLivio, Barello Serena. Behavioral Epigenetics of Family-
Centered Care in the Neonatal Intensive Care Unit. JAMA Pediatrics.
2015;169(7):697–698. https://jaman​etwork.com/journ​als/jamap​
ediat​rics/artic​le-abstr​act/2337233
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Pisoni C, Provenzi L, Moncecchi M,
et al. Early parenting intervention promotes 24-month
psychomotor development in preterm children. Acta Paediatr.
2021;110:101–108. https://doi.org/10.1111/apa.15345
 | |
Received: 5 December 2019    Revised: 30 April 2020    Accepted: 12 May 2020

DOI: 10.1111/apa.15354

REGULAR ARTICLE

Faecal calprotectin and gut microbiota do not predict

enteropathy in very preterm infants

Florence Campeotto1,2,3 | Caroline Elie4 | Clotilde Rousseau2,3,5 | Agnès Giuseppi6 |

Taymme Hachem6 | Ponny Gobalakichenane7 | Mathilde Le Touzey8 |
Marie de Stefano2,3 | Marie-José Butel2,3  | Nathalie Kapel2,3,9

1
Department of Pediatric Gastroenterology,
Necker-Enfants Malades Hospital, AP-HP, Abstract
Paris, France Aim: Very preterm birth is associated with a high risk of enteropathies. Diagnosis is
2
Faculty of Pharmacy, INSERM, U1139,
challenging, especially in mild forms, leading to unnecessary periods of cessation of
Paris, France
3
UMR-S U1139, Hospital-University
enteral feeding. This study aimed at establishing a prognosis score of enteropathy
Department Risks In Pregnancy, Paris combining clinical parameters and faecal calprotectin concentration.
Descartes University, Paris University, Paris,
France
Methods: This prospective multicentric study included preterm neonates born at a
4
Clinical Research Unit, Clinical Investigation gestational age of 33 weeks or less. Stools were collected weekly until hospital dis-
Center, Necker-Enfants Malades Hospital, charge, and daily in case of digestive events for calprotectin measurement (ELISA
AP-HP, Paris, France
5 and immunochromatography) and microbiota analyses (16S rRNA gene sequencing).
Microbiology Department, St-Louis
Hospital, APHP, Paris, France Results: Among the 121 neonates included, 21 experienced at least one episode of
6
Neonatology Department, Necker-Enfants enteropathy, mainly mild forms. By ELISA testing, median faecal calprotectin was
Malades Hospital, AP-HP, Paris, France
7 88 (8-798) µg/g faeces. No statistically significant association was found between
Neonatology Department, Armand
Trousseau Hospital, AP-HP, Paris, France the outset of enteropathy and maternal and neonatal characteristics, and calprotec-
8
Neonatology Department, Poissy-Saint tin levels. The agreement between ELISA and immunochromatography assay was
Germain Intercommunal Hospital Center,
Poissy, France
moderate (intra-class correlation coefficient 0.58, 95%CI [0.47-0.66]). Comparison
9
Coprology Department, Pitié-Salpétrière of species diversity and relative bacterial abundance profiles between infants with
Hospital, APHP, Paris, France or without enteropathy revealed no specific alterations associated with enteropathy.
Correspondence Conclusion: The study failed to propose a prognostic score of enteropathy, prob-
Florence Campeotto, Hôpital Necker- ably due the large inter- and intra-individual variability of faecal calprotectin in very
Enfants Malades, Gastroentérologie
pédiatrique, hépatologie et nutrition preterm neonates.
pédiatriques, 149 rue de Sèvres, 75015
Paris, France. KEYWORDS
Email: florence.campeotto@aphp.fr
enteropathy, faecal calprotectin, gut microbiota, necrotising enterocolitis, preterm neonates
Funding information
The study was funded by a grant from
Programme Hospitalier de Recherche
Clinique—PHRC 2012 (AOR12088, Ministry
of Health).

Abbreviations: FC, faecal calprotectin; NEC, necrotising enterocolitis; IC, immunochromatography; NICU, neonatal intensive care nit; POCT, point-of-care testing; GA, gestational age;
OTU, operational taxonomic unit.
Marie-José Butel and Nathalie Kapel are contributed equally to this work.

Clinical trial registration number NCT02010268

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction
in any medium, provided the original work is properly cited and is not used for commercial purposes.
© 2020 The Authors. Acta Paediatrica published by John Wiley & Sons Ltd on behalf of Foundation Acta Paediatrica

Acta Paediatrica. 2021;110:109–116.  wileyonlinelibrary.com/journal/apa     109|

 |
110       CAMPEOTTO et al.
1 |  I NTRO D U C TI O N
Very premature birth is associated with a high risk of serious com-
plications, including gastrointestinal diseases, that is enteropathies,
which the most severe form is necrotising enterocolitis (NEC), one
of the leading cause of morbidity and mortality in neonatal intensive
care units.1 Enteropathies are classified according to Bell's stages
2 • Faecal calprotectin, a reliable marker of inflammation
from mild-to-severe forms. The prevalence of the milder form of
enteropathy (Bell's stages Ia and Ib, suspected NEC) can be as high
as 44%.3 The severe form, that is definite NEC (Bell's stages IIb to
IIIb), has an incidence depending on the country, from 2% to about
10% in extremely preterm neonates, associated with a mortality up
to 50%.1,4
Diagnosis is mainly supported by a body of presumptions based
on clinical and radiological signs according to Bell's stage. However,
it can be challenging in numerous suspected NEC cases, especially
in the early course of the disease or in mild forms. This diagnostic
uncertainty leads to frequent unnecessary and extended periods of
cessation of enteral feeding that could impair the intestinal matu-
ration, and prolonged courses of parenteral nutrition and antibiot-
ics. New tools are thus required to facilitate an early and precise
diagnosis and to optimise the management of preterm neonates with
intestinal symptoms.
Biomarkers are molecular indicators of disease process, diag-
nosis, and/or prognosis. There is currently a paucity of reliable and
5
robust biomarkers for neonatal enteropathies, including NEC.
Noninvasive biomarkers have obvious advantages in preterm in-
fants. Among the noninvasive faecal biomarkers, the most studied is
faecal calprotectin (FC), a protein mainly expressed in the cytoplasm
of neutrophils and highly resistant to proteolysis, which is now rou-
tinely measured to evaluate mucosal inflammation in patients with
inflammatory bowel diseases.6-8 Physiological values of neonatal FC
are higher than adult ones with a large inter-individual variability,
reflecting the permeability of neonatal digestive tract, the immune
stimulation induced by the microbiota settlement and the first con-
tacts with feeding allergens.9 Several studies have shown increased
FC levels in infants with enterocolitis. However, there is still no con-
sensus about cut-off levels in the neonatal population and regarding
its use in current clinical practice.10
In a retrospective study, we previously proposed a FC thresh-
old at 350 µg/g faeces (sensitivity 0.65, specificity 0.82) using en-
zyme-linked immunosorbent assay (ELISA) for early diagnosis of
neonatal enteropathy requiring strict enteral feeding interruption in
preterm neonates.11 The current prospective cohort study was de-
signed to establish a prognosis and diagnosis score of enteropathy
based on combined criteria including neonatal characteristics and
FC concentration. Furthermore, the study aimed at establishing the
agreement between the ELISA method considered as the reference
assay performed in laboratories and a point of care testing (POCT),
that is an immunochromatographic (IC) assay, more convenient for
rapid diagnosis. In addition, faecal microbiota composition during
the digestive event was analysed.
Key Notes
• Early diagnosis of enteropathies in very preterm neo-
nates is challenging, leading to unnecessary enteral
feeding discontinuation and prolonged courses of
antibiotics.
in children and adults, showed highly variable levels
in very preterm neonates, limiting its use to predict
enteropathy.
• By contrast with necrotising enterocolitis, the most se-
vere form of enteropathies, gut microbiota did not show
a specific profile linked to milder forms.
2 | PATI E NT S A N D M E TH O DS
2.1 | Patients
This prospective multicentre study was conducted in three French
Neonatal Intensive Care Units (NICUs). Eligible preterm neonates
had a gestational age (GA) of 33 weeks or less, and did not present
with any relevant malformation. They were enrolled at birth, and var-
ious perinatal parameters were recorded: sex, GA, mode of delivery,
Apgar score, intra-uterine growth retardation, acute foetal distress,
respiratory distress, antenatal and per partum antibiotic therapy, and
suspicion of maternofoetal infection. Neonates underwent a weekly
clinical follow-up until hospital discharge, including weight, height,
head circumference, abdominal circumference, collateral venous ab-
dominal circulation and mode of feeding. Digestive tolerance was
evaluated through stool frequency and characteristics according to
‘Amsterdam’ stool form scale,12 stool bleeding, bloating, type and
volume of feeding, and residual gastric volume. Enteropathy was
suspected when occurred an intestinal distress requiring the inter-
ruption of enteral feeding for more than 48  hours, the commonly
used treatment for a suspected enterocolitis (Bell's stage ≥ Ia). When
neonates experienced such events, a daily monitoring was set-up in-
cluding all relevant previously described parameters, and biological
parameters (C-reactive protein (CRP), procalcitonin). Medical files
were further checked by both a gastropaediatrician and a public
health specialist to ensure that the event had a digestive origin and
was not secondary to another infectious disease. Cases, that is defi-
nite enteropathies, were defined by the concurrent presence of an
abdominal distension (either clinically diagnosed by visual observa-
tion and/or measurement of the abdominal circumference, or radio-
logically diagnosed, or both) and either rectal bleeding or elevated
gastric retention (the threshold was higher than one third of the vol-
ume of milk ingested per 24 hours) or both. All cases were classified
according to the Bell's scale. 2
Stools were collected at enrolment, then weekly until discharge,
for FC assays and gut microbiota analysis, and immediately stored at
CAMPEOTTO et al.
−20°C and −80°C, respectively. During the event, complementary
stool samples were collected daily from the first digestive symp-
toms. The expected quantity of stools was over 2g to make it possi-
ble high-quality grade assays.
This study was approved by an Ethic Committee (CPP Ile-de-
France I, number 2012-Janv.-12813) and registered on https://clini​
caltr​ials.gov/ (NCT02010268). Parental written consent was signed
prior to any study procedure.
2.2 | Calprotectin assays
Stool samples were assayed for FC by Calprest ELISA method
(Eurospital, Italy) and Calfast IC assay (Eurospital, Italy)—as the
POCT—according to manufacture recommendations. Assays were
performed at the end of the sampling period and simultaneously
for both methods to avoid any bias in the measurements. Hence,
samples were stored at −20°C before the assay no more than
12 months, which is known to prevent from significant alteration
13
of FC levels.
2.3 | Gut microbiota analysis
Neonates who had a stool sample collected at the time of an en-
teropathy event (from −2 to  +  5  days) were enrolled in the gut
microbiota analysis. Two control neonates were selected per case
and matched on the following criteria: absence of enteropathy,
stool specimen collected in the same postnatal week, birth GA
(± 1.5 weeks), birthweight (± 100g), mode of delivery and neonatal
intensive care unit.
Total DNA was extracted from 0.2g of faecal samples as pre-
viously described.14 Bacterial diversity and composition were
assessed by 16S rRNA gene sequencing–based method. Variable
regions V3-V4 were amplified by polymerase chain reaction (PCR)
and sequenced on an Illumina MiSeq platform. Next-generation
sequencing data set was analysed using the open-source bio-
informatics pipeline Find, Rapidly, OTUs with Galaxy Solution
(FROGS).15 After sequence processing, operational taxonomic
units (OTUs) were assigned to different taxonomic levels using
the NCBI reference databases and Ribosomal database Project
16,17
Classifier.
2.4 | Statistical analysis
Cases were defined as infants with intestinal distress leading to
interruption of enteral feeding for more than 48  hours, associated
with an abdominal distension and either rectal bleeding or elevated
gastric retention or both. The sample size was calculated to show
that a FC threshold (ELISA) higher than 350 µg/g faeces is a risk fac-
tor of feeding interruption in preterm neonates as previously dem-
onstrated.11,18 We hypothesised that the prevalence of enteropathy
|
      111
was 9% in neonates with FC < 350 µg/g. For a RR equal to 4 with
alpha risk 5% and power 90%, it was needed to include forty-seven
infants per class group (FC ≥ 350 µg/g, FC < 350 µg/g). Ninety-four
infants were required, and the sample size was extended to 125 in-
fants overall to account for other possible risk factors.
All statistical analyses were undertaken using R 2.3.3 software
(https://cran.rproj​ect.org/). Statistical tests were two-sided with
statistical significance of P < .05.
Characteristics of the neonates were described in the overall pop-
ulation. Mean ± SD or median [range] were reported for quantitative
variables, and frequencies (%) for qualitative variables. Univariate anal-
yses were carried out to determine the association between enterop-
athy and FC levels using ELISA, clinical parameters and the occurrence
of an enteropathy. Since a prognosis score was seeking, this analysis
was performed on neonates with at least one sample available before
day 7 of life (D7) and before the event when it occurred before D7. The
agreement between faecal calprotectin concentrations determined by
ELISA and IC assay was evaluated by intra-class correlation coefficient
(ICC) and Bland and Altman plot.19 Gut microbiota was analysed using
Wilcoxon test and permutational multivariate analysis of variance
(PERMANOVA).
3 | R E S U LT S
3.1 | Population
Between September 2013 and November 2014, 122 preterm neo-
nates were included in the study with a sex ratio of 1.1. One neo-
nate died before any follow-up. One patient with a malformation
(clubfoot) was maintained in the analysis population. One-hundred
and twenty-one patients were included in the analysis popula-
tion. Forty infants (33%) were prematurely transferred before the
planned NICU discharge. Neonates’ characteristics, and pregnancy
and delivery conditions are detailed in Table 1. The median GA was
30.9  weeks ranging from 24.9 to 32.9  weeks. The majority of the
neonates were very premature, with only 20 extremely prema-
ture infants, that is born at a GA under 28 weeks, and among them
only 4 under 27 weeks. All infants were included from 0 to 7 days
old (median age: 2 days). All documented neonates had an enteral
feeding, and for 95% (n = 114), it was exclusive. Exclusive banked
breastfeeding was the main diet type (n = 90, 75%), and 11 neonates
(9.2%) had exclusively preterm milk formula.
3.2 | Enteropathies
During their study follow-up (median duration 41 days [1-83]), 21
neonates (17%) experienced at least one episode of enteropathy
(one event for 19 neonates and two events for 2 neonates. The
first enteropathy occurred at a median age of 15 days [1-38], and
the second at 21 and 71 days, respectively. The enteropathy was
diagnosed within 48h after the first digestive symptoms. The 23
 |
112       CAMPEOTTO et al.

TA B L E 1   Characteristics of the 121 infants included TA B L E 2   Characteristics of the enteropathies’ events (n = 23) in
21 patients (two patients had two events separated by more than
Pregnancy's characteristics
2 wk)
Multiple birth (n, %) 68 56.2
Age at the event time (d) 20 [1-71]
Intrauterine retardation (n, %) 15 12.4
Bell's stage (n, %)
Maternal corticosteroid (n, %) 112 92.6
Ia 12 (52.2)
Maternal smoking (n, %)a  18 17.8
Ib 6 (26.1)
Maternal antibiotics (n, %)a 
IIa 4 (17.4)
Antenatal (one week before delivery) 31 26.5
Intestinal distension (n, %) 23 (100.0)
Intrapartum 53 44.5
Abdominal circumference (cm, mean ± SD)a  25.2 ± 2.2
Caesarean section (n, %) 94 77.7
Collateral venous circulation (n, %)a  4 (18.2)
Neonatal characteristics
Gastric residual volumea 
Gestational age (weeks, median, range) 30.9 24.9-32.9
Gastric residual volume over 24 h (mL, median 13 [0-49]
APGAR score (median, range) 1 min 8 0-10
[range])
5 min 9 2-10
Ratio of gastric residual volume per ingested 7.5
10 min 10 3-10 milk volume over 24 h (%,median [range]) [1.6-118.8]
Acute foetal distress (n, %)a,b  4 3.4 Pneumatosis intestinalis (n, %)a  3 (13.6)
Maternofoetal infection (n, %) 79 65.3 Stools within the 48 h prior to the event 21 (91.3)
Antibiotic therapy (n, %) 77 97.5 Number of stools within 24h prior to the event a  5 [1-11]
Confirmed infection (n, %) 4 5 Blood in the stool (n, %)a  7 (31.8)
Birthweight (g, mean, SD) 1330 329 Biological markers
Age at inclusion (day, median, range) 2 0-7 CRP (μg/L, median, range)a  6 [0-157]
Enteral feeding (n, %)a  120 99.1 PCT (μg/L, median, range)   a
0.4 [0.0-9.7]
Exclusive bank milk feeding 90 74.4 Abbreviations: CRP, C-reactive protein; PCT, procalcitonin.
Exclusive premature milk feeding 11 9.2 a
Missing values: abdominal circumference, 10; collateral venous
Exclusive breastfeeding 1 0.8 circulation, 1; gastric residual volume over 24 h, 2; gastric residual
volume ratio, 8; number of stools within 24 h prior to the event, 2;
Breast + bank milk feeding 12 10.0
pneumatosis intestinalis, 1; blood in the stools, 1; CRP, 2; PCT, 2.
Other mixed feeding 6 5.0
a
Missing values: maternal smoking during pregnancy, 20; antenatal
and intrapartum antibiotic therapy, 4 and 2, respectively; acute foetal The agreement between ELISA and IC results was moderate
distress, 3; enteral feeding, 1. (ICC = 0.58, 95% CI [0.47-0.66] (Figure S1, the Bland and Altman plots
b
At least 2 critical criteria among at-risk Apgar score (<4 at 1 min or < 7 corroborate the moderate agreement between FC measured using
at 5 min), foetal cardiac rhythm abnormality (type II dips,) umbilical ELISA and IC). However, FC levels present a similar evolution with both
artery pH < 7, meconium-stained amniotic fluid.
methods, as shown by the individual plots overlaying the results of neo-
nates who experienced enteropathy (Figure 1) or not (Figure S2, which
digestive events were mainly mild enteropathies [Bell's stage: Ia- shows FC dynamics using ELISA and IC in the 99 without enteropathy),
-Ib, 18 patients (78.3%); IIa, four patients (17.4%); IIb, one patient with a high inter- and intra-individual variability.
(4.3%)] (Table 2).

3.4 | Prognosis factors of enteropathy based on

3.3 | Faecal calprotectin faecal calprotectin and clinical parameters

Overall, 681 samples were assayed for FC levels. One sample was
excluded for an aberrant value (5250 µg/g faeces), and 634 samples
(93%) were studied by both ELISA and IC assay. The median FC-ELISA
was 88 µg/g faeces (range [8-798], inter-quartile range [44.3-140]).
FC-ELISA ≥ 350 µg/g faeces were found only in 12 patients (10%)
from the cohort. Among them, 7 had no enteropathies during hospi-
tal stay, 4 had a high FC-ELISA during or after the event, and 1 before
the event. Thus, the threshold for enteropathy of 350 µg/g faeces
was not confirmed in the current cohort.
Prognostic value of FC-ELISA was evaluated on the 84 neonates for
whom we collected at least one sample before D7 and before the
event (when it occurred before D7). Among these neonates, 14 (17%)
presented with an enteropathy. Univariate analysis did not show any
significant association between the onset of enteropathy, FC-ELISA
at inclusion and/or before the event, and the clinical parameters re-
corded prior inclusion, precluding the establishment of a prognostic
score (supplementary file, Table  S1, which demonstrates the ab-
sence of significant association between FC and clinical parameters).
CAMPEOTTO et al. |
      113

F I G U R E 1   Individual evolution of faecal calprotectin in the 21 neonates suffering from enteropathy event(s). Events are identified by the
first day (D) of enteral feeding discontinuation (vertical dotted line): 18 patients with mild enteropathy (suspected NEC, Bell stage Ia and Bell
stage Ib), and 5 with severe enteropathy (definite NEC, patients 80 (D71), 86, 120 and 121: Bell stage IIA; patient 104: Bell stage IIb). Blue
line: faecal calprotectin (FC) by ELISA; pink line: FC by immunochromatographic (IC) assay

Diagnostic score could not be evaluated due to the low occur-
rence of NEC cases in our cohort.
3.5 | Gut microbiota characteristics
To identify microbiota signatures of enteropathy, microbiota com-
position was analysed at various phylogenetic levels. Two main
profiles with either Proteobacteria or Firmicutes predominance
were observed in both enteropathy and control groups (Figure 2A).
Comparison of species diversity and relative bacterial abundance
profiles between infants with or without enteropathy revealed no
specific alterations associated with the onset of enteropathy, except
a trend towards a higher diversity in infants without enteropathy
(Figure 2B,C). No correlations with perinatal factors were identified.
4 | D I S CU S S I O N
This multicentre prospective study on 121 very preterm neonates
aimed at defining a combined prognostic and diagnostic score of
enteropathy including perinatal parameters and FC level. Univariate
analysis results did not make possible to propose neither a prognos-
tic score nor a diagnostic score as FC-ELISA did not differentiate
neonates that will or will not have enteropathy.
Literature reported wide variations in FC levels in premature in-
fants.10 In our cohort, FC levels were of the same order than those
described recently by MacQueen et al20 and Ho et al21 but lower
than those of Van Zoonen et al22 or Nakayuenyongsuk et al. 23 These
discrepancies could be related to either neonates’ characteristics as
a recent study indicated a relationship between FC and corrected
GA, 24 or to the use of different methods of assay, which are known
to give numerical values which differ substantially. 25 This out-
come points out that results between cohorts are not interchange-
able. Moreover, it should necessitate the use of adapted clinical
cut-off values in the purpose of help on clinical decision-making.
Interestingly, very recently, a calprotectin upper reference interval
incorporating corrected GA has been shown to best predict NEC. 20
We failed to determine a prognosis score of enteropathy. Of
note, the prevalence of enteropathies in our cohort (23 enteropa-
thies in 21 patients, including only 5 with Bell's stage II) was lower
than expected although our population included patients with the
 |
114       CAMPEOTTO et al.

F I G U R E 2   Microbiota profiles in neonates suffering from enteropathy and matching controls. A, Relative abundance of sequence reads
at the phylum level assigned to different bacterial taxa in each neonate. B, Bacterial alpha-diversity using the Chao-1, Shannon and Inverse
Simpson Diversity Indexes in control and neonates suffering from enteropathy. C, Beta-diversity was analysed using multidimensional
scaling (MDS) plot of samples according to disease status (enteropathy vs control) on the weighted-UniFrac distance metrics

highest risk of NEC (low GA and low birthweight). This can be due
to improvements in the care of very preterm neonates that may
have led to a decrease in the risk of enteropathy and NEC, with
an incidence of NEC in France of 3.2% in very preterm infants in
26
a recent nationwide cohort. Moreover, only 14 cases out of the
23 with enteropathy could be included in the prognosis score cal-
culation. The hypothesised threshold of 350  µg/g faeces to char-
acterise enteropathy requiring feeding interruption was reached in
very few neonates with and without enteropathy. Several cut-off
values have been proposed, especially for prediction of severe form
of NEC.10 However, these cut-off levels remain within the range of
FC levels found in infants without digestive symptom enteropathies.
Recent studies also failed to establish such relationship. Van Zoonen
et al did not find any differences between control infants and the ten
ones who further developed NEC (Bell’ stages 2 or 3). 22 Interestingly,
MacQueen et al20 found more frequently lower calprotectin levels in
preterm infants with bloody stools (Bell's stage Ib), but who did not
progress to NEC, suggesting that calprotectin level may be a marker
of the severity of the enteropathy. However, this relationship was
imperfect, with neonates with elevated FC levels and without NEC,
and neonates with levels within the reference range and with NEC.
The usefulness of point-of-care test to determine FC in preterm
infants is obvious, providing results rapidly available to the paedia-
tricians. POCT assay was carried out centrally and compared with
CAMPEOTTO et al.
FC-ELISA as the reference assay performed on the same samples.
We have previously shown a correlation between the two tech-
niques in adult patients with IBD. 27 The expected agreement was
not observed in our very preterm neonate population, even if the
individual plots overlaying the results of both techniques presented
slightly similar evolution whatever the digestive status. A recent
study using a different POCT confirmed the high calprotectin lev-
els with major fluctuations, and nonspecific increase in some infants
that developed or not NEC. 20
There are few information on microbiota in mild-to-moderate
enteropathies by contrast to NEC. We found no specific microbi-
ota profiles by comparison with infants without digestive events,
by contrast with that we26 and others previously reported for
NEC. 28 However, the high heterogeneity of microbiota of very
preterm infants that we and others described, and the low num-
ber of digestive events in this study may have limited the analysis
power.
4.1 | Strengths and limitations
This prospective study on preterm neonates born at less than
33  weeks of GA gave detailed information on their characteristics
that may help to define population of further studies. Neonates with
enteropathy leading to interruption of enteral feeding were fully de-
scribed. The strength of our study also lies in the repeated meas-
urements of the faecal calprotectin. However, in our cohort, the
occurrence of enteropathy was lower than expected and reached
only half of the one observed in our previous study used to sup-
port the hypothesis for sample size calculation,9 thus precluding the
power of the study to define a prognosis score.
5 | CO N C LU D I N G R E M A R K S
This study did not achieve its goal to propose a prognostic score of en-
teropathy mainly due to the large inter- and intra-individual variability
of faecal calprotectin levels that limit its use as a biomarker of enter-
opathy in very preterm infants. However, the interest of a prognostic
score in this population at high risk of mild-to-severe enteropathy that
will help to take appropriate management of the disease as early as
possible deserves to design a larger multicentric study which would
take into account the difficulties faced in this study.
AC K N OW L E D G E M E N T S
The authors would like to thank the support of all the staff of the
neonatal units participating in this project and the parents for their
agreement in the study. They would also like to thank the staff
Clinical Research Unit of Necker-Enfants Malades for the monitor-
ing of the clinical trial, Robert Campos Oriola from AUXESIA for
the manuscript preparation, Julie Salomon, gastropaediatrician at
Necker-Enfants Malades hospital and Maria-Elisabetta Baldassarre,
neonatologist at Aldo Moro University, Bari, Italy, for their helpful
|
      115
advice. The sponsor was Assistance Publique—Hôpitaux de Paris
(Clinical Research and Innovation Department).
C O N FL I C T O F I N T E R E S T
None declared.
ORCID
Marie-José Butel  https://orcid.org/0000-0003-4249-2130
REFERENCES
1. Battersby C, Santhalingam T, Costeloe K, Modi N. Incidence
of neonatal necrotising enterocolitis in high-income coun-
tries: a systematic review. Arch Dis Child Fetal Neonatal Ed.
2018;103:F182-F189.
2. Kliegman RM, Walsh MC. Neonatal necrotizing enterocolitis:
pathogenesis, classification, and spectrum of illness. Curr Probl
Pediatr. 1987;17:219-288.
3. Campeotto F, Kalach N, Lapillonne A, Butel MJ, Dupont C, Kapel N.
Time course of faecal calprotectin in preterm newborns during the
first month of life. Acta Paediatr. 2007;96:1531-1533.
4. Rich BS, Dolgin SE. Necrotizing Enterocolitis. Pediatr Rev.
2017;38:552-559.
5. Gephart SM, Gordon PV, Penn AH, et al. Changing the paradigm
of defining, detecting, and diagnosing NEC: Perspectives on Bell's
stages and biomarkers for NEC. Semin Pediatr Surg. 2018;27:3-10.
6. Fagerhol MK. Calprotectin, a faecal marker of organic gastrointesti-
nal abnormality. Lancet. 2000;356:1783-1784.
7. van Rheenen PF, Van de Vijver E, Fidler V. Faecal calprotectin for
screening of patients with suspected inflammatory bowel disease:
diagnostic meta-analysis. BMJ. 2010;341:c3369.
8. Heida A, Van de Vijver E, van Ravenzwaaij D, et al. Predicting in-
flammatory bowel disease in children with abdominal pain and diar-
rhoea: calgranulin-C versus calprotectin stool tests. Arch Dis Child.
2018;103:565-571.
9. Kapel N, Campeotto F, Kalach N, Baldassare M, Butel MJ, Dupont
C. Faecal calprotectin in term and preterm neonates. J Pediatr
Gastroenterol Nutr. 2010;51:542-547.
10. Pergialiotis V, Konstantopoulos P, Karampetsou N, et al.
Calprotectin levels in necrotizing enterocolitis: a systematic review
of the literature. Inflamm Res. 2016;65:847-852.
11. Campeotto F, Baldassarre M, Butel MJ, et al. Fecal calprotectin:
cutoff values for identifying intestinal distress in preterm infants. J
Pediatr Gastroenterol Nutr. 2009;48:507-510.
12. Bekkali N, Hamers SL, Reitsma JB, Van Toledo L, Benninga
MA. Infant stool form scale: development and results. J Pediatr.
2009;154(521–6):e1.
13. Ton H, Brandsnes DS, Holtlund J, et al. Improved assay for fecal
calprotectin. Clin Chim Acta. 2000;292:41-54.
14. Rousseau C, Levenez F, Fouqueray C, Dore J, Collignon A, Lepage
P Clostridium difficile colonization in early infancy is accompanied
by changes in intestinal microbiota composition. J Clin Microbiol.
2011;49:858-865.
15. Escudie F, Auer L, Bernard M, et al. FROGS: Find, Rapidly,
OTUs with Galaxy Solution. Bioinformatics (Oxford, England).
2018;34:1287-1294.
16. Federhen S. The NCBI Taxonomy database. Nucleic Acids Res.
2012;40:D136-D143.
17. Cole JR, Wang Q, Fish JA, et al. Ribosomal Database Project: data
and tools for high throughput rRNA analysis. Nucleic Acids Res.
2014;42:D633-D642.
18. Yang Q, Smith PB, Goldberg RN, Cotten CM. Dynamic change of
fecal calprotectin in very low birth weight infants during the first
month of life. Neonatology. 2008;94:267-271.
 |
116       CAMPEOTTO et al.

19. Bland JM, Altman DG. Statistical methods for assessing agree-
ment between two methods of clinical measurement. Lancet.
1986;1:307-310.
20. MacQueen BC, Christensen RD, Yost CC, et al. Reference intervals
for stool calprotectin in preterm neonates and their utility for the di-
agnosis of necrotizing enterocolitis. J Perinatol. 2018;38:1379-1385.
21. Ho TTB, Groer MW, Kane B, et al. Enteric dysbiosis and fecal
calprotectin expression in premature infants. Pediatr Res.
2019;85:361-368.
22. van Zoonen A, Hulzebos CV, Muller Kobold AC, Kooi EMW, Bos
AF, Hulscher JBF. Serial fecal calprotectin in the prediction of
necrotizing enterocolitis in preterm neonates. J Pediatr Surg.
2019;54:455-459.
23. Nakayuenyongsuk W, Christofferson M, Stevenson DK, Sylvester
K, Lee HC, Park KT. Point-of-Care Fecal Calprotectin Monitoring
in Preterm Infants at Risk for Necrotizing Enterocolitis. J Pediatr.
2018;196(98–103):e1.
24. Yoon JM, Park JY, Ko KO, Lim JW, Cheon EJ, Kim HJ. Fecal calpro-
tectin concentration in neonatal necrotizing enterocolitis. Korean J
Pediatr. 2014;57:351-356.
25. Labaere D, Smismans A, Van Olmen A, et al. Comparison of six
different calprotectin assays for the assessment of inflammatory
bowel disease. United European Gastroenterol J. 2014;2:30-37.
26. Roze JC, Ancel PY, Lepage P, et al. Nutritional strategies and gut
microbiota composition as risk factors for necrotizing enterocolitis
in very-preterm infants. Am J Clin Nutr. 2017;106:821-830.
27. Benahmed NA, Manene D, Barbot-Trystram L, Kapel N. Evaluation
of Calfast(R) immunochromatographic quantitative assay for
the measurement of calprotectin in faeces. Clin Chem Lab Med.
2014;52:e143-e145.
28. Baldassarre ME, Di Mauro A, Capozza M, et al. Dysbiosis and pre-
maturity: is there a role for probiotics? Nutrients. 2019;11(6):1273.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Campeotto F, Elie C, Rousseau C,
et al. Faecal calprotectin and gut microbiota do not predict
enteropathy in very preterm infants. Acta Paediatr.
2021;110:109–116. https://doi.org/10.1111/apa.15354
DOI: 10.1111/apa.15448
BRIEF REPORT
Suctioning at birth showed low adherence to official
recommendations in a low-resource setting
Education on neonatal resuscitation is most urgent in settings with
poor access to intrapartum obstetric care, where immediate postna-
tal mortality can be reduced by 30% with basic training in neonatal
1
resuscitation. Management of newborns at birth includes different
interventions based on progressive steps (initial steps, ventilation,
chest compressions and medications). Initial steps include oropha-
ryngeal suctioning, which is recommended under vision in infants
who have obvious obstruction to spontaneous breathing or who re-
quire positive pressure ventilation (PPV). 2 The Manual of Neonatal
Resuscitation describes when and how to perform oropharyngeal
suctioning, 2 but information on how the procedure is actually per-
formed in low-resource settings is lacking.
The aim of this study was to evaluate the occurrence and the
characteristics of suctioning at birth in newly born infants in a
low-resource setting.
The study was performed at Beira Central Hospital (Beira,
Mozambique) where about 4500 deliveries occur every year. Beira
Central Hospital is the referral hospital for a geographical area that
covers about 7 million people, with large referral services for mater-
nal and neonatal care in the province.
This study presents a secondary analysis of data from a pro-
spective study on education in neonatal resuscitation using video
recording.3 The National Committee of Bioethics and the Minister of
Health of the Republic of Mozambique approved the research pro-
tocol.3 Parental consent to record neonatal delivery room manage-
ment and to use the data was obtained before every delivery.
Among 150 neonates from the original study, who required suc-
tioning were included in this secondary analysis. Lack of parental
consent was the only exclusion criterion.
Neonatal resuscitation was performed routinely under ra-
diant warmers in the delivery room or in the operating room.
Suctioning was performed with bulb syringe or suction catheter
by the attending caregiver in infants who have obvious obstruc-
tion to spontaneous breathing or who require PPV. 2,3 A camera
installed above the radiant warmers video-recorded all interven-
3
tions as previously described. Two researchers (SA and SC) re-
viewed and evaluated all 150 videos of neonatal resuscitation,
with a third researcher (DT) resolving any conflicts. Time of birth
was defined as the time the Apgar clock was started or birth was
announced.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
Acta Paediatrica. 2021;110:117–118. 
The outcome measures were the proportion of infants who re-
ceived suctioning, the characteristics of suctioning and the propor-
tion of infants who needed PPV.
Since this study presents a secondary analysis of data collected
during a prospective, a convenience sample comprising all 150 vid-
eos of the original study was evaluated. Continuous data were sum-
marised using median and interquartile range (IQR), and categorical
data as number and percentage.
Of the 150 video recordings, suctioning was performed in 117
neonates (78%). There were 77 males (66%) and 40 females (34%).
Median gestational age was 38 weeks (IQR 36-40), and median birth
weight was 2900 g (IQR 2350-3200). Median Apgar score at 5 min-
utes was 6 (IQR 4-7). Suctioning resolved the airway obstruction in
40 neonates (34%), while ventilation was needed after suctioning in
70 neonates (60%) who displayed persisting bradycardia and/or ap-
noea. In seven neonates (6%), suctioning was performed due to fail-
ure of ventilation for suspected airway block. Information on suction
is shown in Table 1. Suction started at median 70 seconds (IQR 40-
120) and lasted median 100 seconds (IQR 50-140). A catheter was
used in 99 neonates (85%) and a bulb syringe in 18 (15%). The correct
sequence (suctioning the mouth before the nose) was performed in
106 neonates (91%), but suctioning the stomach was observed in 78
neonates (67%).
Overall, our findings showed that suctioning was performed in
three out of four neonates receiving resuscitation, with further PPV
needed in the majority of them. The video review underlined late
initiation and long duration of the procedure, with preference for
deep suctioning and the use of a catheter. This study has the merit of
describing magnitude and quality of suctioning in neonates receiving
resuscitation in a low-resource setting by using video recording, but
could not provide information on the impact of suctioning on clinical
outcomes.
While oro/nasopharyngeal suctioning is no longer recommended
as routine care immediately after birth, this procedure can be useful
in clearing the airway or when PPV is required. 2 Our data suggested
that, after receiving drying and stimulation, suctioning was suffi-
cient to avoid further resuscitation procedures in one out of three
neonates. Although suctioning may activate vagal response induc-
ing bradycardia and apnoea, we can speculate that deep suctioning
acted as some form of stimulation in our neonates.
|
wileyonlinelibrary.com/journal/apa     117
 |
118       BRIEF REPORT
TA B L E 1   Characteristics of suctioning
No. of neonates receiving suction 117
Time of suctioning initiation, seconds a
70 (40-120)
Number of suctions a
1 (1-2)
Duration of first suction, secondsa 65 (43-105)
Total duration of suction, secondsa 100 (50-140)
Bulb syringe 18 (15)
Catheter 99 (85)
Suctioning the mouth before the nose 106 (91)
Suctioning the stomach 78 (67)
Note: Data expressed as No. (%) or amedian (IQR).
Our data disclosed low adherence to international guidelines on
neonatal resuscitation in terms of start and duration of the proce-
dure. When indicated, suctioning should be performed within the
first minute of life and should last few seconds. 2 However, at least
half of our procedures began beyond the first minute of life and
lasted almost two minutes. Previous research suggested that many
non-breathing infants may be in primary apnoea with a heart rate
and may initiate spontaneous breathing in response to the initial
steps of resuscitation (including drying, stimulation and suctioning)
only when timely implemented.4 Delayed start and prolonged dura-
tion of suctioning represent an additional hazard for neonates, be-
cause of the delay of PPV initiation. The correct order of suctioning
(mouth before the nose) was followed in the majority of cases, in
order to reduce the risk of aspiration. 2 However, the high prevalence
of deep suctioning had the potential risk of opening the oesophagus
and leading to more air leaks into the stomach.
In conclusion, suctioning was largely performed in neonates re-
ceiving resuscitation, but adherence to official recommendations
was limited in terms of late initiation, prolonged duration and deep
suctioning. Appropriate training is warranted to improve the quality
and the timing of this procedure. On the other hand, further evi-
dence is required to shed lights on benefits and/or harms of suction-
ing in neonates needing resuscitation at birth.5
AC K N OW L E D G E M E N T S
We are very grateful to the midwifery and medical staff of the Beira
Central Hospital, Beira, Mozambique, for their participation and in-
valuable cooperation in this study.
C O N FL I C T O F I N T E R E S T
The authors declare that they have no conflict of interest.
F U N D I N G I N F O R M AT I O N
Associazione Pulcino, Italy, supported the study providing funds
for video cameras. The funders had no role in the study design or
conduct, the management, analysis or interpretation of the data, the
preparation, review or approval of the report, or the decision to sub-
mit the manuscript for publication.
Francesco Cavallin1
Shaimaa Abuelnoor Ahmed Abdelghany2
Serena Calgaro2,3
Amir Hussein Abubacar Seni4
Bonifacio Rodriguez Cebola4
Giovanni Putoto3
Daniele Trevisanuto2
1
Independent Statistician, Solagna, Italy
2
Department of Woman's and Child's Health, University of
Padua, Padua, Italy
3
Doctors with Africa CUAMM, Padua, Italy
4
Beira Central Hospital, Beira, Mozambique
Correspondence
Daniele Trevisanuto, Department of Woman's and Child's
Health, University of Padua, Via Giustiniani 3, 35128, Padua,
Italy.
Email: daniele.trevisanuto@unipd.it
ORCID
Daniele Trevisanuto  https://orcid.org/0000-0002-6462-0079
REFERENCES
1. Lawn JE, Blencowe H, Oza S, et al. Lancet Every Newborn Study
Group. Progress, priorities, and potential beyond survival. Lancet.
2014;384:189-205.
2. American Academy of Pediatrics and American Heart Association.
NRP Neonatal Resuscitation Textbook, 7th edn. Chicago, IL: American
Academy of Pediatrics; 2016.
3. Cavicchiolo ME, Cavallin F, Bertuola F, et al. Effect of a low-dose/
high-frequency training on real-life neonatal resuscitation in a
low-resource setting. Neonatology. 2018;114:294-302.
4. Ersdal HL, Mduma E, Svensen E, Perlman JM. Early initiation of
basic resuscitation interventions including face mask ventilation
may reduce birth asphyxia related mortality in low-income coun-
tries: a prospective descriptive observational study. Resuscitation.
2012;83:869-873.
5. Foster JP, Dawson JA, Davis PG, Dahlen HG. Routine oro/nasopha-
ryngeal suction versus no suction at birth. Cochrane Database Syst
Rev. 2017;4:CD010332.
 | |
Received: 3 June 2020    Revised: 10 July 2020    Accepted: 14 July 2020
DOI: 10.1111/apa.15486
BRIEF REPORT
The effects of Xenon gas inhalation on neuropathology in a
placental-induced brain injury model in neonates: A pilot study
Improved obstetric and neonatal care have reduced the preva-
lence of severe hypoxic-ischaemic-encephalopathy (HIE). However,
1-3/1000 newborns in the developed world1 suffer death or neuro-
developmental disability from HIE. The normal development of the
brain during gestation can also be altered by placental reprogram-
ming under oxidative stress. Under these conditions, the placenta
releases DNA-damaging molecules, bone morphogenic proteins,
2
microRNAs and glutamate. At present, one is unable to diagnose or
treat these factors.
We have previously applied Xenon, a rare noble gas used in an-
aesthesia, at 50% using a closed-circuit system with and without hy-
3,4
pothermia in the newborn piglet and rodent models of HIE.
other inhalational anaesthetics, Xenon did not induce neuroapopto-
sis in the immature brain5 and improved cardiovascular control after
hypoxia-ischaemia (HI). A clinical feasibility and ongoing randomised
phase-two study are testing the effects of breathing Xenon
term infants undergoing therapeutic hypothermia (TH) against those
undergoing TH alone. Experimentally, inhaling Xenon50% improves
motor function and cognition after long-term survival in rats post-in-
4
jury. In rat models of HI brain injury, Xenon is neuroprotective by
upregulating neurotrophic factors and anti-apoptotic proteins,6 by
inducing hypoxia-inducible factor (HIF-1α) pathways allowing for
pre-conditioning,7 by suppressing the astroglial response to injury
and limiting glutamate release to counter excitotoxicity thereby im-
proving neuronal survival.4
Xenon's neuroprotective properties may be extended to treat
brain injury arising from placental reprogramming under oxida-
tive stress. To test this hypothesis, we used a rat injection model
whereby media obtained directly from human placentae under oxi-
dative stress were injected into postnatal day 4(P4) rat brain (human
gestational age 29-31  weeks equivalent). In brief, media were col-
lected from human first trimester placentae cultured under 21% O2
(CM + 21%) and 2%-8% O2 (CM + 2%-8%). An additional group was
injected with saline (Sal) and constituted the sham condition. The
pups were then allowed to survive into the juvenile age, brains were
culled, and neuropathology was examined (see Supplementary files
for more information).
We have tested (a) the effects of hypoxic injury modelled by the
injection of hypoxia-derived conditioned media from the placenta
into P4 rat pup brains and (b) whether breathing 50% Xenon for
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction
in any medium, provided the original work is properly cited and is not used for commercial purposes.
© 2020 The Authors. Acta Paediatrica published by John Wiley & Sons Ltd on behalf of Foundation Acta Paediatrica
Acta Paediatrica. 2021;110:119–122. 
4 hours after this injury could reduce neuropathology in those pups
surviving into juvenile age.
We report here that the injection of hypoxia-derived conditioned
media to healthy pups causes a modest loss of parvalbumin neurons
in the thalamic reticular nucleus (TRN), the hippocampus and the
cortex at P30 of survival (Figure 1). The glial response to neuronal
loss was assessed by GFAP immunofluorescence and showed a
marked increase in astrocytes in addition to an activated morphol-
ogy. The injury also affected dendrite lengths, a proxy measure for
degree of arborisation/connectivity, and this was consistent with
previous in vitro findings.8 There were no changes in overall neu-
Unlike ronal counts, but dopaminergic neurons process lengths decreased
in some areas. Importantly, Xenon treatment conferred some resis-
tance to the increase in glial numbers (P  <  0.05) in the cortex and
hippocampus (Figure 1). Most strikingly, we observed Xenon treat-
50%
in ment to be protective against the loss of parvalbumin cells in the
TRN caused by the injury (P < 0.05). Interestingly, Xenon treatment
did not protect dendritic arborisations/complexity but did greatly in-
crease the lengths of dopaminergic (tyrosine hydroxylase) processes
and overall neuronal numbers post-injury.
These promising results albeit limited in scope support that Xenon
treatment after mild injury due to maternal hypoxia does offer some
protection. Our findings are consistent with previously reported effects
of Xenon in toning down gliosis in neonatal rat cortex, hippocampus
and thalamus in a classical HI rodent model.9,10 Most neurons including
parvalbumin neurons are sensitive to hypoxia and are lost in the 2%-8%
condition in most areas. Xenon acts as an anti-apoptotic agent, and as
neurogenesis is still very active in the early developing postnatal rodent
brain, Xenon may be promoting a compensatory neuroblast differen-
tiation response in vulnerable areas explaining the higher number of
densities we observed compared with 21%. This is speculative, and we
do not know the underlying mechanisms. We also cannot speculate on
the behavioural significance of an overall increased number of neurons
without further study. Xenon is thought to provide partial protection
of dopaminergic cells by acting as a trophic factor in conditions of ex-
citotoxicity and by suppressing the astroglial response.11 This improves
cell survival but may also result in outgrowth around the area of dam-
age. Fibre outgrowth is linked with altered connectivity in the brain and,
therefore, may not necessarily beneficial. Behavioural work is needed
to test how motor skills have been affected with/ without Xenon after
|
wileyonlinelibrary.com/journal/apa     119
 |
120       PHILLIPS et al.
PHILLIPS et al. |
      121

F I G U R E 1   Methodology and effect of placental hypoxic secretions with/without Xenon treatment on neuropathology. (A) Diagram
of experimental set-up (see Supplementary files). Placental explants were incubated for 24 hours at either 21% or 2% oxygen conditions
in trophoblast media and maintained in neurotrophic medium at 21% or at 2%-8%. Media were collected according to two conditions:
CM + 21% (control) and CM + 2%-8% (hypoxia/reoxygenation). Media were injected into the brain of P4 rats (n=3/condition) and an
n=3 was injected with saline as the sham condition).  3 pups from the CM + 2%-8% condition were subsequently placed in a closed-
loop system at Xenon50% for 4 hours. Pups were kept in normal conditions until P30. The P30 brains were collected, fixed and assessed
for neuropathology. All identifiers were hidden for assessment and de-blinded for statistical analyses. (B-G) Results from cell density
quantifications in specific anatomical areas such as the thalamic reticular nucleus, cortex and hippocampus at P30: (B) Parvalbumin neuron
densities; (C) MAP2 neuron densities; (D) GFAP astrocyte densities. We also report process lengths for dopaminergic cells (E), MAP2
neurons (F) and astrocytes (G) process lengths in the thalamic reticular nucleus, hippocampus and cortex. Testing was done using a two-
way ANOVA with Tukey's post hoc test (*P < 0.05, **P < 0.01, ***P < 0.001; n=3/group; data shown as means ± SD). (H-J) Representative
photomicrographs of MAP2 (H) and GFAP (I) cells in the cortex, parvalbumin cells (J) in the reticular nucleus of pups injected with
conditioned media from placental explants cultured at 21% and 2%-8% oxygen with/without Xenon treatment. Images were converted to
grey scale and inverted for clarity using ImagePro Premier. Scale bar = 40 µm. CM = conditioned media

injury. Although we did not perform cognitive tests on the rodents,
no significant neuropathology in the tissues was detected. Previously,
when Xenon was administered shortly after a carotid ligation + hypoxic
insult on P7 rat pups (near-term equivalent), which survived into adult-
hood, both neuropathology and behaviural testing were improved by
Xenon50%.4 In this 10-week survival to adulthood model, we defined
that the time-window for neuroprotection with Xenon was 5  hours,
which is the basis for our clinical trial of administering Xenon within
5 hours after birth since delayed Xenon significantly improved outcome.
Another clinical trial delivered Xenon30% starting at 10 hours of age and
short-term outcome (using Magnetic Resonance Spectroscopy) after
10 days did not show any difference between TH and TH + Xenon.12
It is likely that any effect of Xenon on cognition needs long-term fol-
low-up. We are currently undertaking full IQ testing at 3-5 years in the
CoolXenon trial (ending 010321).
Several clinical conditions present with hypoxic insults to the
placenta including pre-eclampsia, maternal gestational diabetes13
and stress.14 Placental reprogramming can alter fetal neurodevelop-
ment due to maternal hypoxia and Xenon inhalation may yet offer a
promising therapeutic strategy.
K E Y WO R D S
brain injury, cooling, hypoxic-ischaemic encephalopathy,
neurodevelopment, placental reprogramming, Xenon
AC K N OW L E D G E M E N T S
We acknowledge the financial support of the JP Moulton founda-
tion, the Perivoli Trust and SPARKS-The Children's Medical Research
Charity. We also acknowledge Dr Charles Patrick Case for his input
in this study and Dr John Dingley for providing Xenon-delivery
expertise.
C O N FL I C T O F I N T E R E S T
The authors declare no conflict of interest.
Thomas Phillips1,2
David A. Menassa1,3,4
Simon Grant5
Nicki Cohen6
Marianne Thoresen7,8
1
Translational Health Sciences, Bristol Medical School,
University of Bristol, Bristol, UK
2
UK Dementia Research Institute, Cardiff University, Cardiff, UK
3
Biological Sciences, Faculty of Environmental and Life Sciences,
University of Southampton, Southampton, UK
4
Nuffield Department of Clinical Neurosciences, University of
Oxford, Oxford, UK
5
Department of Obstetrics and Gynaecology, Southmead
Hospital, Bristol, UK
6
Department of Medical Education, King's College London,
London, UK
7
Neonatal Neuroscience, Translational Health Sciences, Bristol
Medical School, University of Bristol, Bristol, UK
8
Institute of Basic Medical Sciences, Section for Physiology,
University of Oslo, Oslo, Norway
Correspondence
Marianne Thoresen, Neonatal Neuroscience, Translational
Health Sciences, Bristol Medical School, University of
Bristol, Bristol, UK.
Email: marianne.thoresen@bristol.ac.uk
Thomas Phillips and David A. Menassa equal contribution.
ORCID
David A. Menassa  https://orcid.org/0000-0002-5984-8407
Marianne Thoresen  https://orcid.org/0000-0002-9615-9109
REFERENCES
1. Lai M-C, Yang S-N. Perinatal hypoxic-ischemic encephalopathy. J
Biomed Biotechnol. 2011;2011:609813.
2. Phillips TJ, Scott H, Menassa DA, et al. Treating the placenta to pre-
vent adverse effects of gestational hypoxia on fetal brain develop-
ment. Sci Rep. 2017;7(1):9079.
3. Dingley J, Tooley J, Liu X, et al. Xenon ventilation during therapeu-
tic hypothermia in neonatal encephalopathy: a feasibility study.
Pediatrics. 2014;133(5):809-818.
4. Thoresen M, Hobbs CE, Wood T, Chakkarapani E, Dingley J. Cooling
combined with immediate or delayed xenon inhalation provides
equivalent long-term neuroprotection after neonatal hypoxia-isch-
emia. J Cereb Blood Flow Metab. 2009;29(4):707-714.
 |
122       PHILLIPS et al.
5. Sabir H, Walløe L, Dingley J, Smit E, Liu X, Thoresen M. Combined
treatment of xenon and hypothermia in newborn rats–additive or
synergistic effect? PLoS One. 2014;9(10):e109845.
6. Fan X, Kavelaars A, Heijnen CJ, Groenendaal F, van Bel F.
Pharmacological neuroprotection after perinatal hypoxic-ischemic
brain injury. Curr Neuropharmacol. 2010;8(4):324-334.
7. Ma D, Lim T, Xu J, et al. Xenon preconditioning protects against
renal ischemic-reperfusion injury via HIF-1alpha activation. J Am
Soc Nephrol. 2009;20(4):713-720.
8. Curtis DJ, Sood A, Phillips TJ, et al. Secretions from placenta, after
hypoxia/reoxygenation, can damage developing neurones of brain
under experimental conditions. Exp Neurol. 2014;261:386-395.
9. Dingley J, Tooley J, Porter H, Thoresen M. Xenon provides short-
term neuroprotection in neonatal rats when administered after hy-
poxia-ischemia. Stroke. 2006;37(2):501-506.
10. Campos-Pires R, Hirnet T, Valeo F, et al. Xenon improves long-term
cognitive function, reduces neuronal loss and chronic neuroinflam-
mation, and improves survival after traumatic brain injury in mice.
Br J Anaesth. 2019;123(1):60-73.
11. Lavaur J, Le Nogue D, Lemaire M, et al. The noble gas xenon pro-
vides protection and trophic stimulation to midbrain dopamine neu-
rons. J Neurochem. 2017;142(1):14-28.
12. Azzopardi D, Robertson NJ, Bainbridge A, et al. Moderate hypo-
thermia within 6 h of birth plus inhaled xenon versus moderate
hypothermia alone after birth asphyxia (TOBY-Xe): a proof-of-
concept, open-label, randomised controlled trial. Lancet Neurol.
2016;15(2):145-153.
13. Gauster M, Desoye G, Tötsch M, Hiden U. The placenta and gesta-
tional diabetes mellitus. Curr Diab Rep. 2012;12(1):16-23.
14. Gheorghe CP, Goyal R, Mittal A, Longo LD. Gene expression in the
placenta: maternal stress and epigenetic responses. Int J Dev Biol.
2010;54(2–3):507-523.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
 | |
Received: 28 February 2020    Revised: 27 April 2020    Accepted: 25 May 2020

DOI: 10.1111/apa.15378

REGULAR ARTICLE

Visual-motor functions are affected in young adults who were

born premature and screened for retinopathy of prematurity

Dýrleif Pétursdóttir1  | Gerd Holmström1  | Eva Larsson1  | Birgitta Böhm2

1
Institution of Neuroscience/
Ophthalmology, Uppsala University, Abstract
Uppsala, Sweden Aim: To assess visual-motor integration in young adults previously included in a pro-
2
Department of Women’s and Children’s
spective study on the incidence of retinopathy of prematurity (ROP).
Health, Karolinska Institute, Stockholm,
Sweden Methods: The study encompassed 59 preterm individuals, born 1988-1990, with a
birth weight ≤1500 g, and 44 full-term controls, aged 25-29 years. Ophthalmological
Correspondence
Eva Larsson, Institution of Neuroscience/ examination, including visual acuity and contrast sensitivity, and the Beery Visual-
Ophthalmology, Uppsala University, S-75185
Motor Integration (VMI) with supplemental tests of visual perception and motor co-
Uppsala, Sweden.
Email: eva.larsson@neuro.uu.se ordination, were performed. A short questionnaire was filled in.
Results: The preterm individuals had significantly lower scores than the controls in
Funding information
Crown Princess Margaretha Foundation; all VMI tests, median values and interquartile ranges: Beery VMI 87 (21) vs 103 (11),
The Carmen and Bertil Regnér Foundation;
visual perception 97 (15) vs 101 (8) and motor coordination 97 (21) vs 102 (15), re-
Synskadades Vänner i Uppsala län;
Ögonfonden spectively. Within the preterm group, no correlations were found between the VMI
tests and ROP, gestational age, birth weight or visual acuity. Contrast sensitivity was
correlated to visual perception. Neurological complication at 2.5 years was a risk fac-
tor for lower scores on Beery VMI. The preterm subjects reported six times as many
health problems as compared to the controls.
Conclusion: Being born preterm seemed to have life-long effects. This study shows
that visual-motor integration was affected in young adults born preterm.

KEYWORDS

preterm birth, retinopathy of prematurity, visual-motor integration, young adults

1 |  I NTRO D U C TI O N
Preterm birth can affect a child´s development in many ways. The
immature central nervous system is vulnerable, not only to injury
but also to disruptions of normal development. With improve-
ments in neonatal intensive care and survival of more immature
infants, the rate of major disabilities associated with preterm
births has remained relatively constant, while milder neurocogni-
tive deficits have become more prominent.1-3 A higher prevalence
of low intelligence quotient (IQ) among very low birth weight
preterm individuals than in term controls has been reported.4
Some studies have shown that children born preterm have more
difficulties with nonverbal reasoning and visuospatial tasks than
verbal ability. 5,6 This pattern seems to persist in adolescence and

Abbreviations: BW, birth weight; CI, confidence interval; g, grams; GA, gestational age; IQ, intelligence quotient; IQR, interquartile range; ROP, retinopathy of prematurity; VA, visual
acuity; VLBW, very low birth weight; VMI, visual-motor integration.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in
any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2020 The Authors. Acta Paediatrica published by John Wiley & Sons Ltd on behalf of Foundation Acta Paediatrica

Acta Paediatrica. 2021;110:127–133.  wileyonlinelibrary.com/journal/apa     127 |

 |
128       PÉTURSDÓTTIR et al.
young adulthood.7,8 Visual-motor problems can have an effect on
academic performance such as reading, writing and mathematics
and become more problematic with age.9,10 Very and extremely
preterm individuals also have an increased risk of attention-def-
icit/hyperactivity disorder, anxiety, depression and autism.11,12 In
adulthood, this group has a higher risk of medical and social dis-
abilities, with the risk increasing with lower gestational age (GA)
12
at birth.
Retinopathy of prematurity (ROP) is one of the complications
of preterm birth. It can have a negative impact on the future visual
function8,13 of preterm individuals and is related to other impair-
ments, although this association could be mediated by other neo-
natal risk factors.8
Between 1988 and 1990, a prospective population-based
study was conducted on the incidence of ROP in preterm infants in
Stockholm County, Sweden.14 At 10 years, the same group partici-
pated in a follow-up study of ophthalmological outcome, together
15,16
with a control group. At young adult age, a larger study on the
ophthalmological outcome within the same groups was performed.
The visual function has been reported.17
The aim of the present study was to report on the visual-motor
functions and the educational and health aspects in the two groups.
Furthermore, the study explored whether the prematurity, previous
ROP and neurological complications, and current visual acuity (VA)
and contrast sensitivity had any impact on visual-motor integration
at this age.
2 |  M ATE R I A L A N D M E TH O DS
2.1 | Participants
The present study group included 59 preterm individuals who were
born between 1 November 1988 and 31 October 1990, with a BW
of ≤1500 grams (g), in Stockholm county, Sweden. All were screened
for ROP in the neonatal period and treated with cryotherapy when
indicated. The criterion for treatment at that time was ROP 3 in four
contiguous clock hours in zone II, even in the absence of plus dis-
ease.14 The children were followed for 3.5 years.14,18 At 10 years, the
same group of children were comprehensively examined ophthal-
mologically again, together with a control group of children born in
15
Stockholm County, during the same time period but born at term.
The control group of the present study included 34 individuals from
the 10-year study as well as 10 new individuals randomly chosen
from the Swedish National Board of Health and Social Welfare
Register fulfilling the same inclusion criteria. The examinations were
performed at 25-29 years of age for all participants. Figure 1 shows a
flow diagram of the very low birth weight (VLBW) cohort over time,
including some demographic data of those who attended and did not
attend at 25 years of age compared with 10 years of age.
Regarding ROP, the preterm group was divided into subgroups
classified as: no ROP, untreated ROP and treated ROP, according to
the most affected eye.
Key Notes
• Visual-motor integration was assessed in a follow-up of
young adults born prematurely and a control group born
at term to evaluate the influence from preterm birth and
especially retinopathy of prematurity.
• The preterm individuals had lower scores on all parts of
the visual-motor integration tests than the controls.
• The preterm subjects reported more frequent health
and social disabilities than the controls.
At 2.5 years of age, neurological complications were defined as
intraventricular haemorrhage in the neonatal period and or obvious
neurologic sequelae, that is epilepsy, cerebral palsy or mental retar-
dation.18 Further neurological examinations were not carried out.
The study was approved by the regional Ethical Review Board of
Uppsala, Sweden (Dnr 2014/584) and was performed in accordance
with the Declaration of Helsinki. All participants signed a written
consent form.
2.2 | Methods
Best-corrected distance VA was measured monocularly with the
logarithmic Early Treatment Diabetic Retinopathy Study chart and
the best-corrected near VA was estimated binocularly with the loga-
rithmic Near Visual Acuity Early Treatment Diabetic Retinopathy
Study Chart. Contrast sensitivity was assessed monocularly with
the Vistech Contrast Sensitivity Test System (VCTS 6500) (Vistech
Consultants), and the logarithmic values were used to calculate the
area under the curve.17 Furthermore, a comprehensive ophthalmo-
logical examination was performed including cover test, motility, an-
terior and posterior segment and refraction during cycloplegia.
Before dilating the pupils, the participants answered a short
questionnaire about their vision, education, work and actual health,
and the Developmental test of Visual-Motor Integration, the Beery
VMI test was given together with the two optional tests: Beery visual
perception and motor coordination.19 The participants were given a
booklet with a sequence of 30 increasingly complicated geometrical
forms, from a vertical line to a three-dimensional star to copy. In the
visual perception test, the subject is asked to identify the exact copy
among several variants of each of 27 geometric forms. The motor
coordination test is a proof of tracing different forms with a pencil
without going outside the double-lined paths. Each test was evalu-
ated individually and anonymously by an experienced psychologist
(BB), blinded to the group members. The raw scores were converted
into standardised scores. The results of the three tests range from
45 to 107 points, with a mean of 100 points and standard deviation
of 15. All preterm and control subjects could accomplish the three
tests.
PÉTURSDÓTTIR et al. |
      129

260 included 1988-1990

7 died
1 emigrated
4 excluded

248 included 3.5 years

1 died
6 emigrated
1 immigrated
1 protected identy

241 located 10 years

7 did not respond

15 declined
3 did not aend
216 included 10 years

2 could not be located

214 located
25-29 years

133 did not respond

5 leers of invitaon returned
7 declined
10 did not aend

59 included 157 not parcipang

BW 1217 g (405)* BW 1200 g (356)*

GA 29 w (3)* GA 29 w (4)*
Females 37 (63%) Females 74 (47%)
Previous ROP 25 (42%) Previous ROP 61 (39%)
Treated ROP 13 (22%) Treated ROP 12 (8%)
Neurological complicaon 8 (14%) Neurological complicaon 24 (15%)

F I G U R E 1   Flow diagram of the very low birth weight cohort from infancy, when included in a prospective population-based study of the
incidence of retinopathy of prematurity, up to young adulthood

2.3 | Statistical methods
Descriptive data were analysed using IBM SPSS Statistics version 25
(IBM Corp). Mann-Whitney and Kruskal-Wallis tests were used to com-
pare the results of the two study groups on the three-part Beery VMI
test. Post hoc analyses were carried out using Dunn´s test. Since there
was a difference in proportions of females and males in the preterm
and control groups, linear regression models adjusting for gender, also
were used to estimate the differences between the groups on Beery
VMI. Simple regression analyses were performed within the preterm
group, and the variables were ROP, yes or no, in the most affected eye,
treated ROP, yes or no, GA in weeks, BW, a diagnosis of a neurologi-
cal disorder at 2.5 years of age, current distance VA in the better eye,
binocular near VA and contrast sensitivity in the better eye. Better eye
was defined as the eye with the best VA. Thereafter, multivariable lin-
ear regression analyses were performed to identify predictors for the
three parts of the Beery VMI. The multivariable analyses included ROP,
yes or no, treated ROP, yes or no, neurology at 2.5 years of age, VA
and contrast sensitivity, together with GA in weeks in one and BW in
one, since GA and BW were highly correlated. The analyses were per-
formed using R version 3.6.1 (R Foundation for Statistical Computing,
Vienna, Austria). A P value of < 0.05 was considered significant.
 |
130       PÉTURSDÓTTIR et al.

3 |   R E S U LT S female participants in either the study group or the control group,
see Table 3.
The demographic data are shown in Table 1. The results of the three Within the VLBW group, no correlations between the Beery VMI
Beery tests for the VLBW and control groups are given in Figure 2. test and GA or BW were found, and no significant differences ac-
Table 2 shows the results for the total study groups and for the pre- cording to level of ROP in any of the three tests, although those with
term group divided according to level of ROP. There was a significant previously treated ROP had the lowest scores.
difference between the two study groups in all three tests: Beery Further, among the preterm individuals, no correlations between
VMI (P < 0.01), visual perception (P < 0.01) and motor coordination distance and near VA and Beery VMI were detected. However, there
(P < 0.01). When comparing the three ROP subgroups with the con- was a correlation between the results of the visual perception test
trol group, the preterm individuals without previous ROP had signifi- and contrast sensitivity, as the area under the curve in the better
cantly lower values than the control group on Beery VMI (P < 0.01), eye (coefficient 20.1, 95% CI 2.9-37.3, P < 0.05), in the multivariable
but not on the visual perception or the motor coordination tests. analysis, but not regarding the other tests.
The results of those with untreated ROP did not differ enough to At 2.5 years corrected age, eight individuals of the pres-
reach statistical significance compared to the control group on any ent preterm group had a neurological complication, see Table 1.
test, but the cryotherapy treated individuals had significantly lower Regarding Beery VMI, there was a significant difference between
values than the controls on all three tests, Beery VMI P < 0.01, visual the subjects with a neurological complication at 2.5 years, median
perception P < 0.05, motor coordination P < 0.05. value 74.5 (IQR 37), and those without 92.0 (IQR 15) (P < 0.05), but
When comparing the results of the three Beery tests accord- not with regard to visual perception, 92.0 (IQR 15) vs 101.0 (IQR 15)
ing to sex, no significant difference was found between male and or motor coordination 97.0 (IQR 27) vs 97.0 (IQR 16). In the multivari-
able regression analysis, neurological complication was a risk factor
TA B L E 1   Demographics of 59 prematurely born individuals and for reduced scores regarding Beery VMI (coefficient −15.2, 95% CI
44 term born individuals −26.6 to −3.9, P < 0.05).
The questionnaire revealed that 36/59 preterm individuals and
VLBW group Controls
(59) (44) 33/44 controls had completed or were pursuing a university edu-
cation. Furthermore, 36 of the preterm individuals and 35 of the
Birth weight (grams)
controls had a driving licence. Regarding daily activities, 43 of the
Median (IQR) 1217 (405) n/a
preterm individuals and 38 of the controls were working or studying
Range 700-1490 3000-4000
full time.
Gestational age (wk)
The results of the questionnaire regarding self-reported health
Median (IQR) 29 (3) n/a revealed that the preterm individuals reported medical and mental
Range 24-34 39-41 health disorders more often than the controls. Eight preterm indi-
BCVA logMAR viduals and one of the controls reported depression. Five preterm
Median (IQR) −0.08 (0.12) −0.16 (0.08) individuals but no full-term subjects had a diagnosis of autism, and
Range −0.26 to 0.42 −0.30 to 0.02 neurological morbidity was only reported by the preterm individuals,
Binocular near-VA logMAR except for epilepsy, which was started from one participant of each

Median (IQR) −0.08 (0.10) −0.12 (0.09) group.

Range −0.26 to 0.48 −0.30 to 0.00

Contrast sensitivity (AUC)
4 | D I S CU S S I O N
Median (IQR) 2.06 (0.23) 2.18 (0.17)
Range 1.37-2.31 1.60-2.32
The current population-based follow-up study showed that visual-
Sex m/f 22/37 26/18
motor integration was affected in prematurely born adults compared
No ROP m/f (N = 34) 16/18 n/a with controls. They had a lower standardised score on all three parts
Untreated ROP m/f (N = 12) 4/8 n/a of the Beery VMI, which is in line with previous studies involving
Treated ROP m/f (N = 13) 2/11 n/a children and adolescents.7,20 Those with previous ROP had the low-
Neurology (at 2.5 y of age) 4/4 n/a est scores. Within the preterm group, no correlations were found
m/fa  with GA, BW, level of ROP or VA. However, better visual contrast
Abbreviations: AUC, area under the curve; BCVA, best corrected visual sensitivity was correlated to better performance on the visual per-
acuity; f, female; IQR, interquartile range; m, male; N,number; n/a, not ception part of the Beery VMI, and neurological complication was a
applicable; ROP, retinopathy of prematurity; VA, visual acuity; VLBW,
risk factor for lower scores on Beery VMI.
very low birth weight.
a There are very few population-based studies on the visual func-
Defined as an intraventricular haemorrhage in the neonatal period
and/or obvious neurological sequelae (epilepsy, cerebral palsy, or tion, visual-motor integration and social outcome of preterm individu-
mental retardation). als, who were prospectively screened for ROP in infancy and examined
PÉTURSDÓTTIR et al.
F I G U R E 2   A-C, Boxplots with results of the Beery VMI, visual perception and motor coordination in very low birth weight and control
adults, with the median and IQR inside the boxes. IQR, interquartile range, VMI, visual-motor integration
throughout childhood and in young adulthood. Darlow et al have re-
ported on the visual function13 and social outcome21 in a large cohort
of preterm young adults from New Zealand, born before the introduc-
tion of treatment for ROP. Molloy et al reported on the visual function
and visual perception in a cohort from Australia of individuals aged 14
to 20 born extremely preterm22 and found significantly poorer perfor-
mance in visual perceptual tasks in the preterm adolescents compared
with the controls. In addition, the authors showed that those with
previous severe ROP performed more poorly regarding visual-motor
integration,8 which was in accordance with the present study. Molloy
et al concluded that this association was largely mediated by other
neonatal risk factors. It may reflect the fact that children who develop
treatment-requiring ROP are the most immature and vulnerable.23
In a cohort from Stockholm, Sweden, Lundequist et al found that
extremely preterm (GA < 28 weeks) individuals had significantly
lower scores than controls on the Beery VMI, but the results of the
very premature group (GA 28-31 weeks) did not differ significantly
from those of the controls.7 In the current study, we found differ-
ences in all parts of the VMI test compared with controls (Figure 2),
but did not further divide the VLBW group into subgroups of imma-
turity. However, within the preterm group we found no correlations
with BW or GA.
|
      131
In the present study, an extensive ophthalmological examination
was performed,17 which made it possible to analyse the association
between the results of the VMI test and visual function, assessed as
VA and contrast sensitivity. Best-corrected distance VA and near VA
did not correlate with the results of the Beery VMI in the preterm
group. One can hypothesise that the VA was generally sufficiently
good to accomplish the Beery VMI, even if VA was significantly
lower among those born preterm compared with the controls.17 This
emphasises that VA is only one of the many factors involved in the
integration of visual perception and hand movements. Moreover, it
relates to the finding that the preterm individuals with neurological
complications in the present study had significantly lower scores on
the main test, the Beery VMI, a test that is not only visually demand-
ing, but also cognitively.19
Contrast sensitivity, which is an important part of visual func-
tion, may be affected in persons with brain lesions. 24 In the current
study, a correlation between contrast sensitivity and performance
on the visual perception test was found in the preterm group, con-
firmed in the multivariable analysis. It cannot be concluded if this
finding was due to minor changes in the brain.
A lower proportion of preterm individuals than controls had or was
pursuing university education at the age of 25-29. Relatively fewer
 |
132       PÉTURSDÓTTIR et al.

TA B L E 2   Results of Beery VMI, Visual Perception and Motor TA B L E 3   Results of Beery VMI, Visual Perception and Motor
Coordination for preterm and control groups, median, IQR and Coordination, median, IQR and range, for control and preterm
range for different degrees of ROP in the eye with the higher groups, males and females
degree of ROP
Beery Visual Motor
Beery Visual Motor VMI Perception Coordination
VMI Perception Coordination
Median Median Median
Median Median Median (IQR) (IQR) (IQR)
(IQR) (IQR) (IQR)
Range Range Range
Range Range Range
Prematures (59)
Prematures (59) 87.0 (21) 97.0 (15) 97.0 (21) Males (22) 87.0 (26) 97.0 (26) 92.0 (21)
45-107 45-107 45-107 45-107 45-107 45-102
No ROP (34) 89.5 (17) 99.0 (11) 97.0 (17) Females (37) 87.0 (21) 101.0 (15) 97.0 (19)
45-103 45-107 45-107 50-107 45-107 56-107
Untreated ROP (12) 89.5 (27) 96.5 (26) 94.5 (16) Controls (44)
57-107 45-107 71-107 Males (26) 103.0 (11) 101.0 (7) 97.0 (20)
Treated ROP (13) 83.0 (18) 92.0 (18) 92.0 (16) 72-107 97-107 71-107
45-103 81-107 64-102 Females (18) 100.5 (15) 101.0 (9) 104.5 (11)
Controls (44) 103.0 (11) 101.0 (8) 102.0 (15) 77-107 86-107 72-107
72-107 86-107 71-107
Abbreviations: IQR, interquartile range; VMI, visual-motor integration.
Abbreviations: IQR, interquartile range; ROP, retinopathy of
prematurity; VMI, visual-motor integration.
at 10 years and prevalence of neurological complications from those
who did not attend compared with the 10-year follow-up.17 Another
were also working or studying full-time. This was in accordance with limitation was that no neurological examinations or magnetic reso-
4,12,21
the findings of other studies. In the study of social outcome, in nance imaging studies were performed on the premature group,
which 71% of their original cohort participated, Darlow et al found which could have given valuable information on structural changes
lower rates of higher education and more welfare dependence.21 in the brain.
However, the authors also concluded that in many aspects of health
and social functioning the VLBW cohort had equivalent scores as the
controls. 5 | CO N C LU S I O N
Driving a car is a task that requires amongst other skills, good
visual-motor integration; in this study, significantly fewer preterm The present study on prematurely born young adults shows that
individuals had a driving licence compared with controls. The ques- visual-motor integration was affected. The subjects born preterm,
tionnaire also revealed an increased occurrence of health problems, and especially those with severe treated ROP had the lowest test
notably depression and autism. This is in accordance with several results. However, even those without a diagnosis of ROP fell below
previous reports on general health of prematurely born individu- the controls, an indication of the cognitive dysfunction caused by
11,12
als. Moreover, the risk seems to be proportional to the degree of the prematurity itself. The affected visual-motor integration found
prematurity at birth, 25 but this does not seem to affect the quality of in the present study might potentiate visual problems of the young
life to the same extent. 21 prematurely born adults.

4.1 | Strengths and limitations
The strength of this project is its long-term prospective ophthalmo-
logical follow-up from birth to young adult age of a population-based
group of preterm subjects, screened for ROP in the neonatal period.
The ophthalmological examination in this study was extensive and
performed on all participants by the same ophthalmologist and the
Beery VMI by the same psychologist.
A limitation of the study was the low follow-up attendance rate,
as can be the case with longitudinal studies. However, those who at-
tended at the 25-year follow-up did not differ regarding GA, BW, VA
AC K N OW L E D G E M E N T S
We thank Eva Nuija, study nurse, for her skilful help with the study,
and Fabian Söderdahl MSc, who assisted with the statistical analysis.
C O N FL I C T O F I N T E R E S T
The authors have no conflicts of interest to declare.
ORCID
Dýrleif Pétursdóttir  https://orcid.org/0000-0002-9757-1373
Gerd Holmström  https://orcid.org/0000-0002-5600-7186
Eva Larsson  https://orcid.org/0000-0001-9674-0094
Birgitta Böhm  https://orcid.org/0000-0002-9406-0930
PÉTURSDÓTTIR et al. |
      133

REFERENCES prematurely born children in the Stockholm area of Sweden. Br J

Ophthalmol. 1993;77(7):417-423.
1. Aylward GP. Neurodevelopmental outcomes of infants born prema-
15. Larsson EK, Rydberg AC, Holmstrom GE. A population-based study
turely. J Dev Behav Pediatr. 2005;26(6):427-440.
on the visual outcome in 10-year-old preterm and full-term chil-
2. Robertson CM, Watt MJ, Dinu IA. Outcomes for the extremely
dren. Arch Ophthalmol. 2005;123(6):825-832.
premature infant: what is new? And where are we going? Pediatr
16. Larsson E, Rydberg A, Holmstrom G. Contrast sensitivity in 10 year
Neurol. 2009;40(3):189-196.
old preterm and full term children: a population based study. Br J
3. Brostrom L, Vollmer B, Bolk J, Eklof E, Aden U. Minor neurological
Ophthalmol. 2006;90(1):87-90.
dysfunction and associations with motor function, general cogni-
17. Petursdottir D, Holmstrom G, Larsson E. Visual function is reduced
tive abilities, and behaviour in children born extremely preterm.
in young adults formerly born prematurely: a population-based
Dev Med Child Neurol. 2018;60(8):826-832.
study. Br J Ophthalmol. 2020;104(4):541-546.
4. Hack M, Flannery DJ, Schluchter M, Cartar L, Borawski E, Klein N.
18. Holmstrom G, el Azazi M, Kugelberg U. Ophthalmological follow up
Outcomes in young adulthood for very-low-birth-weight infants. N
of preterm infants: a population based, prospective study of visual
Engl J Med. 2002;346(3):149-157.
acuity and strabismus. Br J Ophthalmol. 1999;83(2):143-150.
5. Johnson S. Cognitive and behavioural outcomes following very
19. Beery K, Beery N. The Beery-Buktenica Developmental Test of Visual-
preterm birth. Semin Fetal Neonatal Med. 2007;12(5):363-373.
Motor Integration (Beery VMI). Administration, Scoring, and Teaching
6. van Veen S, van Wassenaer-Leemhuis AG, Oosterlaan J, van Kaam
Manual. 6th ed. USA: Psych Corp; 2010.
AH, Aarnoudse-Moens CSH. Eight-year-old very and extremely
20. Goyen TA, Lui K, Woods R. Visual-motor, visual-perceptual, and
preterm children showed more difficulties in performance intelli-
fine motor outcomes in very-low-birthweight children at 5 years.
gence than verbal intelligence. Acta Paediatr. 2019;109(6):1175–
Dev Med Child Neurol. 1998;40(2):76-81.
1183. https://doi.org/10.1111/apa.15095. [Epub ahead of print].
21. Darlow BA, Horwood LJ, Pere-Bracken HM, Woodward LJ.
7. Lundequist A, Bohm B, Lagercrantz H, Forssberg H, Smedler AC.
Psychosocial outcomes of young adults born very low birth weight.
Cognitive outcome varies in adolescents born preterm, depending
Pediatrics. 2013;132(6):e1521-e1528.
on gestational age, intrauterine growth and neonatal complications.
22. Molloy CS, Wilson-Ching M, Anderson VA, et al. Visual processing
Acta Paediatr. 2015;104(3):292-299.
in adolescents born extremely low birth weight and/or extremely
8. Molloy CS, Anderson PJ, Anderson VA, Doyle LW. The long-term
preterm. Pediatrics. 2013;132(3):e704-e712.
outcome of extremely preterm (<28 weeks' gestational age) infants
23. Austeng D, Kallen K, Ewald U, Jakobsson P, Holmstrom G. Incidence
with and without severe retinopathy of prematurity. J Neuropsychol.
of retinopathy of prematurity in infants born before 27 weeks' ges-
2016;10(2):276-294.
tation in Sweden. Arch Ophthalmol. 2009;127(10):1315-1319.
9. Aylward GP. Cognitive and neuropsychological outcomes: more
24. Good WV, Hou C, Norcia AM. Spatial contrast sensitivity vision loss
than IQ scores. Ment Retard Dev Disabil Res Rev. 2002;8(4):234-240.
in children with cortical visual impairment. Invest Ophthalmol Vis Sci.
10. Aarnoudse-Moens CS, Weisglas-Kuperus N, van Goudoever JB,
2012;53(12):7730-7734.
Oosterlaan J. Meta-analysis of neurobehavioral outcomes in
25. Raju TNK, Buist AS, Blaisdell CJ, Moxey-Mims M, Saigal S. Adults
very preterm and/or very low birth weight children. Pediatrics.
born preterm: a review of general health and system-specific out-
2009;124(2):717-728.
comes. Acta Paediatr. 2017;106(9):1409-1437.
11. Johnson S, Hollis C, Kochhar P, Hennessy E, Wolke D, Marlow N.
Psychiatric disorders in extremely preterm children: longitudinal
finding at age 11 years in the EPICure study. J Am Acad Child Adolesc
Psychiatry. 2010;49(5):453-63.e1. How to cite this article: Pétursdóttir D, Holmström G, Larsson
12. Moster D, Lie RT, Markestad T. Long-term medical and social conse- E, Böhm B. Visual-motor functions are affected in young adults
quences of preterm birth. N Engl J Med. 2008;359(3):262-273.
who were born premature and screened for retinopathy of
13. Darlow BA, Elder MJ, Kimber B, Martin J, Horwood LJ. Vision in for-
mer very low birthweight young adults with and without retinopa- prematurity. Acta Paediatr. 2021;110:127–133. https://doi.
thy of prematurity compared with term born controls: the NZ 1986 org/10.1111/apa.15378
VLBW follow-up study. Br J Ophthalmol. 2018;102(8):1041-1046.
14. Holmstrom G, el Azazi M, Jacobson L, Lennerstrand G. A pop-
ulation based, prospective study of the development of ROP in
 | |
Received: 26 January 2020    Revised: 28 April 2020    Accepted: 29 April 2020

DOI: 10.1111/apa.15338

REGULAR ARTICLE

Preeclampsia was a risk factor for pulmonary interstitial

emphysema in preterm infants born ≤32 weeks of gestational
age

Judith Behnke1,2  | Anita Windhorst3 | Frank Oehmke4 | Lars D. Berthold5 |

Klaus-Peter Zimmer1 | Markus Waitz1 | Harald Ehrhardt1,2

1
Department of General Pediatrics &
Neonatology, Justus Liebig University and Abstract
Universities of Giessen and Marburg Lung Aim: This study determined the prenatal and postnatal risk factors for pulmonary
Center, Giessen, Germany
2 interstitial emphysema (PIE) in preterm infants born at up to 32 weeks of gestational
German Center for Lung Research, Giessen,
Germany age (GA) and their contribution to severe complications.
Methods: We studied 179 preterm infants, who had undergone chest X-rays during
3
Department of Medical Statistics, Justus
Liebig University of Giessen, Giessen,
Germany
the first five days of life at Justus Liebig University Giessen, Germany, between 2016
4
Department of Gynecology and Obstetrics, and 2017. Of these, 33 were retrospectively classified as PIE and 146 as non-PIE. The
Justus Liebig University of Giessen, PIE cases were also matched with 33 non-PIE cases by GA and gender. Risk factors
Germany
5 were identified by univariate analyses and multivariable logistic regression.
Department of Pediatric Radiology,
Institute for Diagnostic and Interventional Results: Previously known risk factors for pulmonary interstitial emphysema were
Radiology, Justus Liebig University of
confirmed, including GA and birthweight and the associations with adverse out-
Giessen, Giessen, Germany
comes like intraventricular haemorrhage and mortality. We identified preeclampsia
Correspondence
and haemolysis, elevated liver enzymes and low platelet count (HELLP) syndrome
Harald Ehrhardt, Department of General
Pediatrics and Neonatology, Justus Liebig as additional risk factors for PIE (P = .027), and lung impairment was associated with
University, Feulgenstrasse 12, Giessen
respiratory distress syndrome (P = .001), higher maximum inspired oxygen (P = .014)
35392, Germany.
Email: Harald.Ehrhardt@paediat.med.uni- and needing surfactant (P = .006).
giessen.de
Conclusion: Preeclampsia and HELLP syndrome were identified as possible addi-
tional risk factors for PIE in preterm infants. These conditions should be included in
future studies, to identify preterm infants at risk of PIE straight after birth.

KEYWORDS

air leak, pulmonary interstitial emphysema, prematurity, preeclampsia, mechanical ventilation

Abbreviations: BPD, bronchopulmonary dysplasia; FiO2, fractional inspired oxygen concentration; GA, gestational age; HELLP, haemolysis, elevated liver enzymes and low platelet
count; IUGR, intrauterine growth restriction; PIE, pulmonary interstitial emphysema; RDS, respiratory distress syndrome; VEGF, vascular endothelial growth factor.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in
any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2020 The Authors. Acta Paediatrica published by John Wiley & Sons Ltd on behalf of Foundation Acta Paediatrica

|
134     
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:134–140.
BEHNKE et al.
1 |  I NTRO D U C TI O N
Pulmonary interstitial emphysema (PIE) is an air leak syndrome, and
it is one of the most severe respiratory complications to affect pre-
term infants. It is characterised by the dissection of the interstitial or
perivascular lung tissue, which results in air being trapped along the
bronchovascular system. This leads to overdistension, decreased com-
1-3
pliance and restricted gas exchange. The radiological characteristics
of PIE are either cyst-like or linear radiolucencies radiating from the
pulmonary hilum to the surface of the lung, and these may be aggra-
vated by the development of pneumomediastinum, pneumothorax or
4,5
pneumopericardium. Chest radiography is the diagnostic gold stand-
ard for PIE, and lung ultrasound is a useful and non-invasive follow-up
tool.6 PIE typically affects extremely low birthweight infants who are
mechanically ventilated with respiratory distress syndrome (RDS), but
it has also been reported in non-invasive ventilated or non-ventilated
infants.2,7 It can occur unilaterally or bilaterally and is classified as local
or diffuse and acute or persistent PIE.
The incidence of PIE ranges from 5% to 40%, in extremely low
birthweight infants, despite advances in neonatal medicine. These
include the established use of antenatal corticosteroids, avoiding me-
chanical ventilation, the introduction of new, and more gentle, modes
of ventilation8 and the application of postnatal surfactant application.
PIE has been associated with higher mortality and high rates of severe
morbidities that are common in premature infants, such as intraventric-
ular haemorrhage and bronchopulmonary dysplasia (BPD).5,9,10 Other
suggested risk factors for PIE include the prenatal use of magnesium
sulphate,11 higher ventilation parameters, increased tidal volumes and
higher fractions of inspired oxygen (FiO2).12 However, one study re-
ported that infants who developed pulmonary interstitial emphysema
during the first days after birth had lower Apgar scores in the delivery
room and needed increased ventilator settings and FiO2, increased
rates of intubation and more frequent surfactant. These findings sug-
3
gest a pulmonary predisposition for PIE as early as birth. This assump-
tion has been further strengthened by reports of PIE occurring during
non-invasive ventilatory support.2
The aim of this study was to determine any associations between a
number of other prenatal and postnatal risk factors and the occurrence
of PIE. These included any diabetes during pregnancy, preeclampsia
and haemolysis, elevated liver enzymes and low platelet count (HELLP)
syndrome, intrauterine growth restriction (IUGR), the use of antenatal
corticosteroids, male gender and multiple birth. All of these conditions
are known to increase the risk of preterm infants developing broncho-
pulmonary dysplasia (BPD) and long-term pulmonary sequelae.13,14
2 |  M ATE R I A L A N D M E TH O DS
2.1 | Study design
This was a retrospective case-control study that focused on 226 pre-
term infants who were born at up to 32  weeks of gestational age
(GA) and were discharged from the neonatology unit of the Justus
|
      135
Key Notes
• This study focused on 179 preterm infants, who had un-
dergone chest X-rays during the first five days of life,
to determine the prenatal and postnatal risk factors for
pulmonary interstitial emphysema (PIE).
• We identified preeclampsia and haemolysis, elevated
liver enzymes and low platelet count syndrome as ad-
ditional risk factors in infants with PIE.
• In addition, lung impairment was associated with res-
piratory distress syndrome, higher maximum inspired
oxygen and needing surfactant.
Liebig University of Giessen, Germany, between 2016 and 2017. All
received chest X-ray examinations during the first 5 days of life. The
exclusion criteria included no chest X-ray being available and prena-
tal interventions like foetal paracentesis. We also excluded infants
with lower urinary tract obstructions, spina bifida, congenital heart
defects, apart from patent ductus arteriosus, atrial septum defects,
hypoplasia or sequestration of the lung, syndromal disease or oe-
sophageal atresia.
After the exclusion criteria were applied, we had 179 infants left
in the study (Figure 1). All the cases were retrospectively classified
by a neonatologist and a paediatric radiologist using an independent
double-blinded study design, and any differences of opinion were
jointly resolved. This classification showed that 33 of the cases had
PIE and 146 did not. The PIE cases were classified as unilateral or
bilateral and subdivided in central, global or lobar manifestations
(Figure 2). The first X-ray available was used to grade RDS.15 To com-
pensate for important confounders of pulmonary RDS severity, each
of the 33 PIE cases was matched with a non-PIE case by the clos-
est GA and gender (Figure  1, Appendix S1). A number of perinatal
variables were retrieved from the medical notes. These included the
infant's clinical characteristics, their birthweight, GA and gender. We
also collected data on any IUGR, premature rupture of the mem-
branes, histologically proven chorioamnionitis, preeclampsia/HELLP
syndrome,16,17 gestational diabetes, type I diabetes, antenatal cor-
ticosteroids, HELLP, maternal age, multiple pregnancy and Apgar
scores at one and five minutes. The factors that were considered
with regard to postnatal respiratory support included the following:
severe RDS (grades III-IV), invasive mechanical ventilation, applied
as assist controlled or synchronised intermittent positive pressure
ventilation, maximum ventilator settings during the first 24 hours of
life. We also included less invasive surfactant administration, if the
preterm infant was on non-invasive ventilatory support at the time
of surfactant application; otherwise, it was given via the endotra-
cheal tube. Postnatal corticosteroids were separated into early for
the first seven days of life and late if it was beyond that point. The
selected outcome parameters were intraventricular haemorrhage
(grade III-IV), BPD,18 necrotising enterocolitis, focal intestinal perfo-
ration, retinopathy of prematurity and mortality.
 |
136       BEHNKE et al.
Study population ≤32 GA
n = 226
Excluded Statistical evaluation
n = 47 n = 179
PIE Total non-PIE
n = 33 n = 146
Matched non-PIE
n = 33
F I G U R E 1   Study design. Flowchart of the study population and
analysed subgroups
2.2 | Statistical analysis
The statistical analyses were performed using R, version 3.5.2 (R
Foundation for Statistical Computing, Vienna, Austria). The demo-
graphics in Table 1 are shown as medians and interquartile ranges.
The absolute and relative frequencies of the parameters are given
for counted data. Comparisons were carried out using the using
Wilcoxon rank-sum test for unpaired, unmatched, metric data and
the Wilcoxon signed-rank test for matched metric data. Fisher's
exact test was used for unmatched categorical data, and conditional
logistic regression was used for matched data. To control for pos-
sible confounders, multivariable conditional logistic regression was
performed with PIE as the dependent variable and preeclampsia or
HELLP, birthweight, multiples, antenatal corticosteroids and severe
RDS as the independent variable (Table  2). Statistical significance
was defined as a P value of <.05. Appendix S1 provides a detailed
description of the complete statistical analysis, including the confi-
dence intervals and the area under the curve.
3 |   R E S U LT S
There were 226 preterm infants born at up to 32  weeks of gesta-
tion who were available for the analysis. Of these, 179 infants met
A B
the inclusion criteria and were included in the statistical evalua-
tion (Figure  1). The subjects’ demographic data are summarised in
Table 1, and the total data set can be found in Appendix S1. The inci-
dence of PIE in the study cohort was 18.4%, and the infants that we
included had a median GA of 28.7 (range 26.1-30.7) and birthweight
of 0.99kg (range 0.76-1.44) with a balanced gender distribution.
The detailed radiological classification showed that there
were 63.6% unilateral cases and 36.4% cases with bilateral PIE.
Focusing on the parameters of respiratory support within the first
24 hours of life enabled us to consider the early effects on postna-
tal manifestation. Just under a third (32.9%) of the preterm infants
underwent invasive mechanical ventilation, with a median posi-
tive end-expiratory pressure of 6.0 cmH2O and a peak inspiratory
pressure of 16.5 cmH2O. The majority (63.7%) received surfactant,
while 82.6% of the mothers were treated with antenatal cortico-
steroids, which is standard treatment for this high-risk population.
When the PIE and non-PIE cases were compared, this showed that
the established risk factors for PIE, including GA and birthweight,
were associated with the known outcomes of death and intraven-
tricular haemorrhage. In addition, severe BPD and retinopathy of
prematurity continued to present as complications following PIE
(Table 2). Important parameters for calculating risks for outcomes
after preterm birth, like antenatal corticosteroids, maternal age,
any diabetes during pregnancy or multiple births, did not continue
to pose risk factors for PIE. However, there was a trend towards
a higher incidence of preeclampsia and HELLP in the PIE group.
Therefore, we decided to carry out further evaluations of pre-
eclampsia and HELLP as risk factor for PIE by using a matched-
pairs analysis. Because the incidence of PIE correlates with RDS
severity, we excluded the main confounding factors of GA and
gender. The matched-pairs analysis showed that preeclampsia and
HELLP syndrome were the only statistically significant maternal
risk factor (P = .0499) in the conditional logistic regression model.
In contrast, other important contributors to pulmonary distress,
like chorioamnionitis and maternal diabetes, did not have any im-
pact on the incidence of PIE in our study. A conditional logistic
regression model that included the predictors of birthweight, mul-
tiple births, severe RDS and antenatal corticosteroids confirmed
that preeclampsia and HELLP syndrome were independent predic-
tors of PIE (P = .027) (Appendix S1). When it came to the severity
of pulmonary sequelae, maximum postnatal fraction of inspired
F I G U R E 2   Example of chest X-ray
with PIE. Anterior-posterior chest X-ray
with diagnosed PIE (A, unilateral left; B,
bilateral)
BEHNKE et al. |
      137

TA B L E 1   Demographics of 179 infants with and without PIE 4 | D I S CU S S I O N

included in the study

Data shown as median Pulmonary interstitial emphysema is a potentially life-threatening

Variables (interquartile range) or n (%) condition following premature birth. Understanding the aetiopatho-
Birth weight (kg) 0.99 (0.76-1.44) genesis of this pulmonary disease is crucial if we are to develop better

Gestational age (weeks) 28.71 (26.14-30.71)

Maternal age
Male
Multiple birth
Intrauterine growth restriction
Preeclampsia or HELLP
Diabetes (gestational/type I)
Antenatal corticosteroids
Premature rupture of the
membranes
Chorioamnionitis
Invasive mechanical ventilation
Surfactant
Postnatal corticosteroids (early)
Postnatal corticosteroids (late)
Pulmonary interstitial emphysema
treatment strategies for PIE cases that have not yet fully established
themselves. Prenatal steroids and postnatal surfactant replacement
31.00 (28.00-35.00)
have been reported to independently reduce mortality, the sever-
89 (49.72)
ity of RDS and air leaks in preterm infants.19 A high percentage of
75 (41.90)
our PIE cases were treated with antenatal corticosteroids and sur-
32 (17.88)
factant administration and therefore did not impact on the incidence
27 (15.08)
of PIE. Moreover, the timing of surfactant administration played a
15 (8.38) fundamental role. In our study cohort, surfactant application was
147 (82.58) standardised at FiO2  >  30%-40%. It is unclear whether mechanical
59 (33.15) ventilation or oxygen supply is the main drivers in the aetiology of PIE
or whether early PIE is the reason for invasive mechanical ventilation,
30 (17.86) FiO2 requirements and the need for surfactant administration. Only
59 (32.96) half of the PIE cases in our study underwent invasive mechanical ven-
114 (63.69) tilation during the first 24 hours of life, and the match-pairs analysis
8 (4.47) showed that ventilator settings had no significant impact on the in-
15 (8.38) cidence of PIE. Moreover, our cohort contained preterm infants with
33 (18.44) PIE who had not received invasive ventilation until they developed

Bronchopulmonary dysplasia 62 (34.64)

(36 wk of postmenstrual age)
Intraventricular haemorrhage
(III°-IV°)
Necrotising enterocolitis
Focal intestinal perforation
PIE and this finding has previously been described in low birthweight
twins.2 Conversely, anatomical and functional airway characteristics
15 (8.38) in the early course of severe RDS are different and may predispose
infants to developing PIE.20 Our data highlight the strong association
2 (1.12) between the development of PIE and RDS and between PIE and the
8 (4.47) well-known risk factors linked to immaturity, namely GA and birth-

Retinopathy of prematurity 49 (27.84)
Death 12 (6.70)
Abbreviation: HELLP, haemolysis, elevated liver enzymes and low
platelet count.
oxygen (FiO2) was higher in the PIE group (P  =  .014), according
to the Wilcoxon signed-rank test. Conditional logistic regres-
sion also showed that this was the case for the number of severe
RDS cases (P = .007) and the need for surfactant (P = .0281). The
matched-pairs analysis in the conditional logistic regression model
confirmed that the higher incidence of intraventricular haemor-
rhage was a relevant complication following PIE (P  =  .0499). As
birthweight plays a substantial role in calculating outcome param-
eters in preterm infants, we also analysed the total cohort in a
stepwise univariate multivariable regression model, starting with
birthweight and then adding GA, gender, multiple births, severe
RDS and antenatal corticosteroids. The calculated cut-off for the
strongest effect for preeclampsia and HELLP syndrome was a
birthweight of 790g based on these findings: birthweight ≤ 790g
(P = .008) versus birthweight ≥ 790g P = .054 (Appendix S1). These
data suggest that immature preterm infants had an increased pre-
disposition for PIE if their mothers had maternal preeclampsia or
HELLP syndrome.
weight.3,5 Our results also underline that a postnatal predisposition
to severe pulmonary disease severity was reflected by the need for
higher supplemental oxygen and more frequent application of sur-
factant. Previous studies had already reported that using higher oxy-
gen levels to resuscitate extremely premature infants had increased
oxidative stress and the incidence of BPD.21 Another study suggested
that higher oxygen needs during resuscitation were an independent
risk factor for PIE.3 Furthermore, PIE development has been related
to impaired oxygenation and ventilation efficiency and the oxygena-
tion index during the development of PIE has been shown to be a
predictor for death or BPD.22 Our matched-pairs analysis took all the
known major contributing factors to PIE into consideration, such as
birthweight, GA and gender. The matching procedure focused on the
main confounders of gestational age and gender, because prematu-
rity is the main risk factor for developing air leaks, especially for PIE, 23
and male gender has been associated with consistently worse pulmo-
nary outcomes.24 Once we excluded those confounding factors, the
new, and only relevant, maternal risk factors for PIE were revealed as
preeclampsia and HELLP syndrome. It had already been reported that
BPD was increased in infants exposed to preeclampsia.25 Placental
structural abnormalities related to IUGR and preeclampsia have also
been strongly associated with foetal growth restriction, and stud-
ies have linked antenatal stress with poor respiratory outcomes.26
Marked elevations of the soluble, endogenous vascular endothelial
138       | BEHNKE et al.

TA B L E 2   Clinical characteristics, postnatal ventilatory support and outcome parameters of preterm infants with and without PIE and the
matched cases subgroup drawn from the non-PIE sample

Total non-PIEa , Matched non-PIE P value matched-pairs

  PIEa , n = 33 n = 146 P value subgroupa , n = 33 comparison

Clinical characteristics
Birth weight (kg) 0.70 (0.58-1.16) 1.10 (0.83-1.46) .0002 0.77 (0.62-1.07) .0612
Gestational age (weeks) 27.29 (24.00-28.71) 29.00 (26.75-30.86) .0008 27.00 (25.00-28.71) .5994
Male 16 (48.48) 73 (50.00) 1.0000 16 (48.48) not applicable
Apgar score (1 min) 7.00 (5.00-8.00) 8.00 (6.00-8.00) .1026 7.00 (5.00-8.00) .5542
Apgar score (5 min) 9.00 (7.00-9.00) 9.00 (8.00-9.00) .4741 9.00 (8.00-9.00) .8531
Maternal age 31.00 (27.00-35.00) 31.00 (28.25-35.00) .8477 34.00 (28.00-35.00) .4579
Multiple birth 15 (45.45) 60 (41.10) .6981 13 (39.39) .4843
Premature rupture of the 12 (36.36) 47 (32.41) .6852 13 (39.39) .7633
membranes
Chorioamnionitis 8 (25.81) 22 (16.06) .2029 8 (25.81) 1.0000
Preeclampsia or HELLP 8 (24.24) 19 (13.01) .1119 1 (3.03) .0499
Diabetes (gestational/type I) 3 (9.09) 12 (8.22) 1.0000 2 (6.06) .5714
Antenatal corticosteroids 27 (81.82) 120 (82.76) 1.0000 29 (87.88) .4843
Intrauterine growth restriction 8 (24.24) 24 (16.44) .3166 7 (21.21) .7394
Postnatal respiratory support
Severe respiratory distress 17 (51.52) 26 (17.81) .0002 4 (12.12) .0074
syndrome (III°-IV°)
Invasive mechanical ventilation 18 (54.55) 41 (28.08) .0069 11 (33.33) .0674
Days of invasive ventilation 2 (0.00-6.00) 0 (0.00-1.00) <.0001 0.00 (0.00-8.00) .9899
Positive end-expiratory pressure 6.00 (6.00-6.00) 6.00 (6.00-7.00) .2834 6.50 (6.00-7.00) .1056
(cmH2O)
Peak inspiratory pressure (cmH2O) 19.00 (17.00-23.25) 16.00 (14.00-20.00) .0304 16.00 (14.00-21.00) .3963
Mean airway pressure (cmH2O) 9.00 (9.00-10.00) 9.00 (8.00-9.88) .3384 9.00 (9.00-10.00) .2785
Fraction of inspired oxygen 50.00 (40.00-81.00) 39.00 (28.25-50.00) <.0001 40.00 (25.00-55.00) .0143
Surfactant 32 (96.97) 82 (56.16) <.0001 23 (69.70) .0281
Early postnatal corticosteroids 5 (15.15) 3 (2.05) .0060 1 (3.03) .1418
Late postnatal corticosteroids 6 (18.18.) 9 (16.16) .0360 6 (18.18) 1
Outcome parameters
Intraventricular haemorrhage 8 (24.24) 7 (4.79) .0015 1 (3.03) .0499
Bronchopulmonary dysplasia 16 (48.48) 46 (31.51) .0713 21 (63.64) .1182
(36 wk postmenstrual age)
Severe bronchopulmonary 12 (36.36) 18 (12.33) .0031 10 (30.30) .5299
dysplasia
Necrotising enterocolitis 0 (0.00) 2 (1.37) 1.0000 1 (3.03) .9984
Focal intestinal perforation 3 (9.09) 5 (3.42) .1655 1 (3.03) .9985
Retinopathy of prematurity 15 (48.39) 34 (23.45) .0077 19 (57.58) .3719
Death 6 (18.18) 6 (4.11) .0103 1 (3.03) .9985

Note: Statistical significance was defined as P < .05.

P values for unmatched data were calculated using Fisher's exact test for count data and the Wilcoxon rank-sum test for metric data.
P values for matched data were calculated using conditional logistic regression for count data and the Wilcoxon signed-rank test with continuity
correction for metric data.
Abbreviation: HELLP, haemolysis, elevated liver enzymes and low platelet count.
a
Data shown as medians (interquartile ranges) or numbers (percentages).

growth factor (VEGF) receptor-1 inhibitor, have been reported in also been strongly associated with BPD in tracheal fluid samples from
the blood and amniotic fluid samples of mothers with preeclampsia. preterm infants with a low gestational age.27-30 These findings under-
Decreased VEGF and increased soluble VEGF receptor-1 levels have line the connection between alveolar and vascular lung development.
BEHNKE et al.
They also show that imbalanced circulating proangiogenic and antian-
giogenic factors could impair vasculogenesis in foetal lungs, leading
to early disruption in lung maturation. Hypothetically, the preeclamp-
tic intrauterine environment causes a primary insult in the develop-
ing foetal lung, predisposing it to further postnatal damage, including
PIE. Infants born following preeclampsia are more susceptible to fur-
ther lung impairment and need intensive treatment for respiratory
stability. Preeclampsia has been recognised as a risk factor for RDS
and BPD.27 The association between preeclampsia and PIE has never
been assessed, but is in line with the outcomes reported by those
studies.
Although the sample size of our study did not allow us to assess
the associations between preeclampsia and BPD, it was sufficient
to show an association between preeclampsia or HELLP and PIE.
Our analyses clearly indicate the need to include preeclampsia and
HELLP syndrome in future calculations, in order to identify high-
risk infants directly after birth. A limitation of our study was that
there are no uniform criteria for the radiological diagnosis of PIE,
and some cases represent a kind of a grey zone of cystic or linear ra-
diolucencies in the interstitium radiating from the hilum. Moreover,
there could be more unknown factors that have not been taken into
consideration. The incidence of PIE could also be influenced by other
important, unmeasured prenatal and postnatal risk factors related to
the mother or the preterm infant. The possible lack of generalisabil-
ity of the present study findings to other populations or other study
settings should be taken into account when interpreting the findings.
Due to the small study cohort, our data only suggest that there was
a trend towards an association between preeclampsia or HELLP and
PIE. It is important not to overestimate the possible importance of
the findings, as further studies are needed to confirm and enhance
the risk factors for PIE. The strengths of this study were the short
study period of two years, which ensured uniform prenatal and post-
natal treatment strategies and the radiological assessment by two
independent, blinded experts. Future large-scaled studies on risk
factors for PIE should include preeclampsia and HELLP to confirm
their impact on its incidence.
5 |  CO N C LU S I O N
Preeclampsia and HELLP syndrome were identified as possible ad-
ditional risk factor for PIE and should be taken into account in future
risk calculations. Infants who are delivered early due to preeclamp-
sia and HELLP syndrome need to receive improved treatment ap-
proaches that are dedicated to preventing PIE.
C O N FL I C T O F I N T E R E S T
The authors have no conflicts of interest to declare.
ORCID
Judith Behnke  https://orcid.org/0000-0003-4307-8759
|
      139
REFERENCES
1. Greenough A, Dixon AK, Roberton NR. Pulmonary interstitial em-
physema. Arch Dis Child. 1984;59(11):1046-1051.
2. Gronbach J, Ehrhardt H, Zimmer K-P, Waitz M. Early pulmonary in-
terstitial emphysema in preterm neonates-respiratory management
and case report in nonventilated very low birth weight twins. AJP
Rep. 2018;8(2):e99-e105.
3. Nuñez-Ramiro A, Aguar M, Cernada M, Parra-Llorca A, Vento M.
Oxygen needs during resuscitation and surfactant to achieve sta-
bilisation were independent risks factors for pulmonary interstitial
emphysema in preterm infants. Acta Paediatr. 2018;107(1):28-32.
4. Boothroyd AE, Barson AJ. Pulmonary interstitial emphysema
— a radiological and pathological correlation. Pediatr Radiol.
1988;18(3):194-199.
5. Hummler HD, Parys E, Mayer B, Essers J, Fuchs H, Schmid M. Risk
indicators for air leaks in preterm infants exposed to restrictive use
of endotracheal intubation. Neonatology. 2015;108(1):1-7.
6. Balcells C, Del Río R, Riaza L, Rebollo M, Rodriguez-Fanjul J,
Camprubí M. Lung ultrasound: a useful tool for the follow-up of neo-
natal localized interstitial emphysema. J Pediatr. 2015;166(6):1543.
7. Berk DR, Varich LJ. Localized persistent pulmonary interstitial em-
physema in a preterm infant in the absence of mechanical ventila-
tion. Pediatr Radiol. 2005;35(12):1243-1245.
8. Behnke J, Lemyre B, Czernik C, Zimmer K-P, Ehrhardt H, Waitz
M. Non-Invasive Ventilation In Neonatology. Dtsch Arztebl Int.
2019;116(11):177-183.
9. Yu VY, Wong PY, Bajuk B, Szymonowicz W. Pulmonary interstitial
emphysema in infants less than 1000 g at birth. Aust Paediatr J.
1986;22(3):189-192.
10. Morley CJ, Davis PG, Doyle LW, Brion LP, Hascoet J-M, Carlin JB.
Nasal CPAP or intubation at birth for very preterm infants. N Engl J
Med. 2008;358(7):700-708.
11. Verma RP, Chandra S, Niwas R, Komaroff E. Risk factors and clinical
outcomes of pulmonary interstitial emphysema in extremely low
birth weight infants. J Perinatol. 2006;26(3):197-200.
12. Lei X, Stangl O, Bösche C, Stuchlik K, Czorba R, Wieg C. Positional
treatment without mechanical ventilation in a very preterm infant
with unilateral pulmonary interstitial emphysema: case report and
review of the literature. BMC Pediatr. 2019;19(1):267.
13. Bancalari E, Jain D. Bronchopulmonary dysplasia: 50 years after the
original description. Neonatology. 2019;115(4):384-391.
14. Gronbach J, Shahzad T, Radajewski S, et al. The potentials and ca-
veats of mesenchymal stromal cell-based therapies in the preterm
infant. Stem Cells Int. 2018;2018:9652897.
15. Agrons GA, Courtney SE, Stocker JT, Markowitz RI. From the ar-
chives of the AFIP: Lung disease in premature neonates: radiolog-
ic-pathologic correlation. Radiographics. 2005;25(4):1047-1073.
16. Haram K, Svendsen E, Abildgaard U. The HELLP syndrome: clini-
cal issues and management. A Review. BMC Pregnancy Childbirth.
2009;9:8.
17. Steegers EAP, von Dadelszen P, Duvekot JJ, Pijnenborg R. Pre-
eclampsia. Lancet. 2010;376(9741):631-644.
18. Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit
Care Med. 2001;163(7):1723-1729.
19. Roberts D, Brown J, Medley N, Dalziel SR. Antenatal corticoste-
roids for accelerating fetal lung maturation for women at risk of
preterm birth. Cochrane Database Syst Rev. 2017;3:CD004454.
20. Ackerman NB, Coalson JJ, Kuehl TJ, et al. Pulmonary interstitial em-
physema in the premature baboon with hyaline membrane disease.
Crit Care Med. 1984;12(6):512-516.
21. Vento M, Moro M, Escrig R, et al. Preterm resuscitation with low
oxygen causes less oxidative stress, inflammation, and chronic lung
disease. Pediatrics. 2009;124(3):e439-e449.
 |
140       BEHNKE et al.

22. Williams E, Dassios T, Clarke P, Chowdhury O, Greenough A. 3 0. Tsao P-N, Wei S-C, Su Y-N, Chou H-C, Chen C-Y, Hsieh W-S.
Predictors of outcome of prematurely born infants with pulmonary Excess soluble fms-like tyrosine kinase 1 and low platelet counts
interstitial emphysema. Acta Paediatr. 2019;108(1):106-111. in premature neonates of preeclamptic mothers. Pediatrics.
23. Jeng M-J, Lee Y-S, Tsao P-C, Soong W-J. Neonatal air leak syndrome 2005;116(2):468-472.
and the role of high-frequency ventilation in its prevention. J Chin
Med Assoc. 2012;75(11):551-559.
24. O'Driscoll DN, McGovern M, Greene CM, Molloy EJ. Gender disparities S U P P O R T I N G I N FO R M AT I O N
in preterm neonatal outcomes. Acta Paediatr. 2018;107(9):1494-1499.
Additional supporting information may be found online in the
25. Hansen AR, Barnés CM, Folkman J, McElrath TF. Maternal pre-
eclampsia predicts the development of bronchopulmonary dyspla- Supporting Information section.
sia. J Pediatr. 2010;156(4):532-536.
26. Taglauer E, Abman SH, Keller RL. Recent advances in antenatal fac-
tors predisposing to bronchopulmonary dysplasia. Semin Perinatol. How to cite this article: Behnke J, Windhorst A, Oehmke F, et al.
2018;42(7):413-424.
Preeclampsia was a risk factor for pulmonary interstitial
27. Torchin H, Ancel P-Y, Goffinet F, et al. Placental complications and
bronchopulmonary dysplasia: EPIPAGE-2 cohort study. Pediatrics. emphysema in preterm infants born ≤32 weeks of gestational
2016;137(3):e20152163. age. Acta Paediatr. 2021;110:134–140. https://doi.org/10.1111/
28. Wallace B, Peisl A, Seedorf G, et al. Anti-sFlt-1 therapy preserves lung apa.15338
alveolar and vascular growth in antenatal models of bronchopulmo-
nary dysplasia. Am J Respir Crit Care Med. 2018;197(6):776-787.
29. Oak P, Pritzke T, Thiel I, et al. Attenuated PDGF signaling drives
alveolar and microvascular defects in neonatal chronic lung disease.
EMBO Mol Med. 2017;9(11):1504-1520.
 | |
Received: 13 March 2020    Revised: 9 May 2020    Accepted: 4 June 2020

DOI: 10.1111/apa.15408

REGULAR ARTICLE

Cholestasis after very preterm birth was associated with

adverse neonatal outcomes but no significant long-term liver
disease: A population-based study

Jonas Teng1,2  | Kajsa Bohlin1,4  | Antal Nemeth1 | Björn Fischler1,3

1
Division of Pediatrics, Department
of Clinical Science, Intervention and
Technology (CLINTEC), Karolinska Institutet,
Stockholm, Sweden
2 preterm infants.
Department of Pediatrics, Södertälje
Hospital, Södertälje, Sweden
3
Department of Pediatrics, Karolinska
University Hospital, Stockholm, Sweden
4
Department of Neonatology, Karolinska
University Hospital, Stockholm, Sweden
Correspondence
Jonas Teng, BUMM, Södertälje Sjukhus AB,
152 86 Södertälje, Sweden.
Email: jonas.teng@ki.se
Funding information
Funds were received from the Swedish
Order of Freemasons and Samaritan
Foundation for the study.
Abstract
Aim: To describe outcome linked to neonatal cholestasis in a defined cohort of very
Methods: Population-based retrospective case-control study of preterm infants,
gestational age <30 weeks, surviving for 28 days, in Stockholm County. Cholestasis
was defined as conjugated bilirubin ≥30 μmol/L exceeding 20% of total level at least
twice and graded as high if exceeding 100 μmol/L. Cholestatic cases were matched
on gestational week with two non-cholestatic controls.
Results: The incidence rate of cholestasis was 37/250 (14.8%), with increasing rates
in lower gestational weeks. Perinatal factors associated with cholestasis were pre-ec-
lampsia and being born small for gestational age. Cholestatic infants had three times
more bronchopulmonary dysplasia and eight times more retinopathy of prematurity.
The mortality was 13.5% in cholestatic infants versus 2.7% in controls (P = .040). All
deceased cholestatic infants had high-grade cholestasis. No surviving infants devel-
oped chronic liver disease by 10 years of age.
Conclusion: Cholestasis was common in very preterm infants and linked to disease
severity and adverse outcome. Cholestasis may be an independent risk factor for
bronchopulmonary dysplasia and retinopathy of prematurity and more severe chol-
estasis associated with increased mortality. Cholestasis was not associated with
chronic liver disease later in childhood.

KEYWORDS

bronchopulmonary dysplasia, chronic liver disease, neonatal intensive care, neonatal

mortality, retinopathy of prematurity

In preterm born and critically ill newborn infants, the incidence

1 |  I NTRO D U C TI O N of cholestasis is high and the cause is multifactorial. 2,3 The incidence
varies depending on selection criteria for the studied cohort, and
Cholestasis is defined as reduced bile flow resulting in retention there are few population-based studies. Risk factors reported are in-
within the liver of biliary substances, normally excreted into bile for testinal malformations, abdominal surgery, necrotising enterocolitis,
elimination into the intestinal lumen and is estimated to occur in 1 sepsis, low gestational age and birth weight. 2-6 Being small for ges-
1
out of 2500 term born infants. tational age has been identified as a risk factor in some studies, but
Abbreviations: BPD, bronchopulmonary dysplasia; CB, conjugated bilirubin; CI, confidence interval; IQR, interquartile range; OR, odds ratio; ROP, retinopathy of prematurity; TB, total
bilirubin.

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

Acta Paediatrica. 2021;110:141–148.  wileyonlinelibrary.com/journal/apa     141|

 |
142       TENG et al.
not in others.4,7-10 The development of cholestasis is closely linked to
prolonged use of parenteral nutrition, but despite much scientific in-
terest, the exact role of different fat sources used in parenteral lipid
emulsions regarding the development and outcome of cholestasis is
unclear.11
The aim of this study was to describe the perinatal factors linked
to cholestasis in very preterm infants on a population-based level
and to further evaluate the short- and long-term effects of cholesta-
sis in the neonatal period and beyond, including the development of
chronic liver disease later in childhood.
2 |  M E TH O DS
2.1 | Study design
We performed a population-based retrospective case-control
study of all preterm infants born and treated in Stockholm County
during a 2-year period. Infants born from March 2006 to February
2008 with gestational age less than 30 full weeks (≤29  +  6) and
surviving the first 28 days of life were included. The infants and
their mothers were identified through the Swedish neonatal qual-
ity register. Cholestatic infants were detected through medical
chart review. For every cholestatic case, two non-cholestatic con-
trol subjects born in the same gestational week were randomised
from the population.
Data from the perinatal period were primarily acquired from
medical charts review. Small for gestational age was defined
as birth weight less than 2 standard deviations from expected.
Bronchopulmonary dysplasia (BPD) was defined as need for oxygen
at 36  weeks' gestational age and/or continuous positive pressure
airway support. Death was defined as death before discharge from
hospital after neonatal or prolonged hospital care. Definitions and
data sources for the remaining variables are described in the online
supplementary table (Table S1).
Selected data from this cohort have previously been reported on
in a study comparing the use of two different types of fat emulsions
12
for total parenteral nutrition.
Ethical approval was received from the Stockholm regional ethi-
cal review board (ref 2008/1558-31/3).
2.2 | Definitions of cholestasis
Cholestasis was defined as conjugated serum bilirubin (CB)
≥30 μmol/L (≈1.8 mg/dL) and exceeding 20% of the total serum bili-
rubin (TB) on at least two occasions, a definition described in our
previously published work.12
The onset of cholestasis was defined as the age when the bili-
rubin levels fulfilled the cholestasis criteria for the first time. The
grade of cholestasis was arbitrarily dichotomised as low, if peak
conjugated bilirubin <100 μmol/L, and as high, if peak conjugated
bilirubin was ≥100 μmol/L. The duration of cholestasis was defined
Key Notes
• Cholestasis in very preterm infants is a common
clinical problem, but the long-term effects are poorly
known.
• In this population-based case-control study of very pre-
term infants, cholestasis developed in 14.8% of infants
and was associated with higher rates of retinopathy of
prematurity, bronchopulmonary dysplasia and death,
but not associated with chronic liver disease at 10 years
of age.
• Further prospective studies are needed to confirm these
results.
as the time elapsed from onset of cholestasis, as defined above, to
when conjugated bilirubin no longer exceeded 30 μmol/L.
2.3 | Development of chronic liver disease
The development of chronic liver disease of clinical significance was
defined as the infant being enrolled at any time, after discharge from
neonatal care, as a patient at any of the tertiary centres for paediatric
hepatology in Karolinska University Hospital in Stockholm or Queen
Silvia Children's Hospital in Gothenburg until the age of 10 years. These
centres are the only tertiary paediatric units for hepatology in Sweden.
All children with more severe liver disease, including progressive
chronic liver disease with or without decompensation and acute liver
failure, will be referred there from all paediatric units in the country.
2.4 | Statistical methods
The data analysis was performed using SPSS version 24.0 for Mac
(IBM). For the univariate analyses, Fisher's exact test was used for
proportions. For continuous variables, Mann-Whitney U test was
used, and medians were used to present the central tendency and
interquartile range (IQR) to report distribution. Mantel-Haenszel
test of trend was used for comparing cumulative incidence rates of
cholestasis in the different weeks of gestational age. For the mul-
tivariate analyses, binary logistic regression was used, and results
presented as crude and adjusted odds ratios (OR) and 95% confi-
dence intervals (CI). Statistical significance level was set at 5%.
3 | R E S U LT S
3.1 | Population characteristics
During the study period, 296 infants with a gestational age of less
than 30 weeks were born. A total of 46 infants were excluded, 22
TENG et al.
because of transfer to other regions and 24 because of short life
span, leaving a study population of 250 infants. Among the 46 ex-
cluded infants, five had already developed cholestasis. Out of these,
two were cholestatic before death and three before transfers to
other regions.
Cholestasis criteria were fulfilled in 37/250 (14.8%) of the in-
fants in the study population. There was a significant trend of higher
incidence rate of cholestasis with lower gestational age (P = .034
Mantel-Haenszel test of trend), as shown in Figure 1.
3.2 | Cholestatic versus non-cholestatic infants
When comparing the cholestatic infants with the non-cholestatic
controls in the univariate analysis, several differences were noted, as
detailed in Table 1 and below. The median maternal age was 32 years
in both groups. The frequency of caesarean delivery was 65% among
cholestatic infants and 62% among controls. The median Apgar score
at 5 minutes was 8 in both groups. The occurrence of singletons, pro-
longed rupture of membranes more than 24 hours, chorioamnionitis,
use of antenatal steroids and need for surfactant treatment did not
differ between groups (data not shown). The rate of pre-eclampsia
in the mothers of the non-cholestatic infants was 8/74 (11%). In the
mothers of the cholestatic infants, 12/37 (32%) had pre-eclampsia,
which was significantly higher than in the controls (P = .008).
With regard to neonatal characteristics, the cholestatic infants
were small for gestational age in 38% of cases, as compared to 16%
among the controls (P = .02). This was also reflected in a significant
difference in birth weight (Table  1). The occurrence of congenital
heart anomalies was 9/37 (24%) among the cholestatic infants and
2/74 (3%) among the controls (P = .001). This difference did not seem
to be due to different investigation rates. The proportion of infants
examined at least once with echocardiography was 34/37 (92%)
in cholestatic infants (89% in the low-grade and 95% high-grade
80
70
Cholestatic
60
50
40
30
20
F I G U R E 1   Proportion of cholestatic
cases per gestational week. Significant 10
trend of higher proportion of cholestasis 6 7
in lower gestational weeks (P = .034*, 0 1
Mantel-Haenszel test of trend) 23
|
      143
group) and 62/74 (84%) in controls (P = .38). The nine infants in the
cholestasis group with cardiac anomalies constituted five cases of
ventricular septal defects, one case of atrioventricular septal defect,
two cases of atrial septal defects and one case of combined aortic
stenosis and aortic insufficiency. Among controls, there were one
case of valvular pulmonary stenosis and one case of right ventric-
ular heart failure due to twin-twin transfusion syndrome. The rate
of patent ductus arteriosus, in need of treatment with ibuprofen or
surgically, did not differ significantly between groups.
The cholestatic infants needed more active neonatal treatment,
such as more mechanical ventilation, parenteral nutrition and eryth-
rocyte transfusions (Table  1). The use of surfactant treatment did
not differ between groups. The cholestatic infants regained their
birth weight faster than the control infants, but they had a lower
weight at 36 weeks of corrected age (Table 1).
The cholestatic infants had more neonatal morbidities than the
controls, including more necrotising enterocolitis and need for lap-
arotomy (Table  1). This was also reflected in the fact that choles-
tatic infants needed bowel resection surgery and stoma significantly
more often (data not shown).
The cholestatic infants also had a higher rate of culture veri-
fied sepsis (Table 1). This was still the case when clinically probable,
but blood culture negative infections were added (data not shown).
Also, multiple episodes of sepsis were more common among
cholestatic infants. The late-onset sepsis episodes, debuting at age
72 hours or later, were responsible for the difference in sepsis in-
cidence between the groups (Table 1). Early-onset sepsis occurred
in two infants in each group, which did not differ significantly. The
median age at onset of sepsis did not differ between groups; it was
10.5  days for those who had at least one culture verified sepsis
episode in each group. Of the cholestatic infants, 7/37 (19%) were
treated for culture verified fungal sepsis versus 5/74 (7%) of the
non-cholestatic (P = .10). One cholestatic and one non-cholestatic
infant were positive for cytomegalovirus in urine screening tests.
Cholestatic cases per gestational week
Non-cholestatic
64
41
43
30
24
5
8 9
3 5 4
24 25 26 27 28 29
144      | TENG et al.

TA B L E 1   Descriptive statistics and

Cholestasis Non-cholestasis
univariate analysis of cholestatic and non-
(n = 37) (n = 74) P
cholestatic very preterm infant controls,
Baseline characteristics matched for gestational age
Gestational age, week + days, median 27 + 4 27 + 2 (25 + 5-28 + 6) .91
(IQR)-matching variable (25 + 1-28 + 5)
Birth weight, kg, median (IQR) 0.810 0.983 (0.767-1.226) .040*
(0.672-1.085)
Small for gestational age 14/37 (37.8%) 12/74 (16.2%) .02*
Male gender 17/37 (45.9%) 52/74 (70.3%) .02*
Neonatal course
Mechanical ventilation, days, median 10 (3.5-20) 2 (0-11) .003*
(IQR)
Erythrocyte transfusions, median 15 (9-19.5) 6 (3-12) <.001*
(IQR)
Parenteral nutrition, days, median 32 (23-47) 11 (6-19) <.001*
(IQR)
No enteral feeding, days, median (IQR) 10 (3-19) 1 (0-3) <.001*
Age regaining birth weight, days, 8 (7-12)a  13 (9.5-15)b  <.001*
median (IQR)
Weight at 36 weeks gestational age, 1.978 2.200 (1.969-2.425)c  .003*
kg, median (IQR) (1.620-2.220)a 
Infections
Sepsis, ≥1 episode 24/37 (64.9%) 28/74 (37.8%) .009*
Sepsis, ≥2 episodes 11/37 (29.7%) 4/74 (5.4%) .001*
Late-onset sepsis (>72 h) 23/37 (62.2%) 26/74 (35.1%) .009*
Neonatal morbidities and outcomes
Respiratory distress syndrome 31/37 (83.8%) 54/74 (73.0%) .24
Necrotising enterocolitis 16/37 (43.2%) 4/74 (5.4%) <.001*
Laparotomy 11/37 (29.7%) 1/74 (1.4%) <.001*
ROP grade 3 or higher 13/36 (36.1%) 10/74 (13.5%) .011*
BPD 20/37 (54.1%) 22/74 (29.7%) .022*
Death 5/37 (13.5%) 2/74 (2.7%) .040*

Note: Data presented as absolute number/total number (per cent) if a proportion, median (IQR) for
continuous variables.
Abbreviations: BPD, bronchopulmonary dysplasia; IQR, Interquartile range; ROP, Retinopathy of
prematurity.
a
n = 34;
b
n = 73;
c
n = 70;
*Statistical significance.

Neonatal cholestasis was associated with poorer outcome.
Cholestatic infants had more retinopathy of prematurity (ROP) and BPD
than controls (Table 1). To further control for disease severity and other
risk factors for poor neonatal outcome, the results were tested in a mul-
tivariate model using logistic regression. The crude OR for ROP grade 3
or more, for cholestatic infants, was 3.6 (95% CI 1.4-9.4). When adjusted
for gestational age, gender and mechanical ventilation >2  weeks, the
adjusted OR was 8.4 (95% CI 2.1-33.4). For BPD, the crude OR was 2.8
(95% CI 1.2-6.3), and the adjusted OR was 3.2 (95% CI 1.2-8.6).
The rate of intraventricular haemorrhage was the same in both
groups (data not shown). Death before discharge was more common
in cholestatic infants, where 5/37 (13.5%) died, as compared with
2/74 (2.7%) in controls (P = .04). The median age at death was 65 days
(range 28-108) among the cholestatic infants. The two deceased
non-cholestatic infants died after 35 and 461 days, respectively. The
latter had severe BPD resulting in respiratory insufficiency, requiring
continuous positive airway pressure support and extra oxygen, and
was never able to leave the hospital.
TENG et al.
3.3 | Course and severity of cholestasis
Details regarding the course of cholestasis for all cholestatic infants
are shown in Table 2.
The two groups, defined as having high-grade (peak conjugated
bilirubin ≥100  μmol/L) and low-grade cholestasis (<100  μmol/L),
were compared. The cholestasis was graded as high in 19 and low in
18 infants using this cut-off level. Notably, the groups did not differ
in many aspects.
However, the mortality was higher in the high-grade group, since
5/19 (26%) of the high-grade cholestatic infants died versus none in
the low-grade group (P = .046). All of the deceased infants had ongo-
ing and considerable cholestasis at the time of death.
The baseline characteristics and the neonatal treatments showed
for all cholestatic infants in Table 1 showed no significant differences
between the subgroups (data not shown). Parenteral nutrition, be-
fore cholestasis and in total, did not differ between the groups (data
not shown). The time of total fasting tended to differ between the
groups. The median days of fasting were 12 days among the high-
grade cholestatic infants and 4 days among the low-grade (P = .051).
In the high-grade group, necrotising enterocolitis tended to be
more common, 11/19 developed necrotising enterocolitis versus
5/18 in the low-grade group (P  =  .10). The other morbidities pre-
sented in Table  1 were similar between the subgroups (data not
shown).
TA B L E 2   Characteristics of cholestasis in cholestatic infants
(n = 37)
Cholestasis onset, age, days, median (IQR) 24 (17-39)
Peak TB, μmol/L, median (IQR) 148 (103-245)
Peak CB, μmol/L, median (IQR) 104 (61-169)
Age at peak CB, days, median (IQR) 42 (31-65)
Cholestasis duration (CB ≥ 30 μmol/L), days, 57 (42-83)a  trols had moved abroad.
median (IQR)
Ursodeoxycholic acid treatment 25/37 (67.6%)
Parenteral nutrition before cholestasis, days, 22 (16-28)
median (IQR)
Total fasting before cholestasis, days, median 7 (2-12)
(IQR)
Necrotising enterocolitis before cholestasis onset 13/37 (35.1%)
Necrotising enterocolitis after cholestasis onset 3/37 (8.1%)
Infections before cholestasis onset
Sepsis, culture verified or clinical 32/37 (86.5%)
Sepsis, culture verified, any species 21/37 (56.8%)
Sepsis, coagulase negative staphylococci 17/37 (45.9%)
Fungal infection, culture verified 4/37 (10.8%)
Note: Data presented as absolute number/total number (per cent) if a
proportion, median (IQR) for continuous variables.
Abbreviations: CB, conjugated bilirubin; IQR, interquartile range; TB,
total bilirubin.
a
n = 30.
|
      145
The rates of infection were also similarly distributed. Neither the
total burden of infections nor their relationship in time with onset
of cholestasis seemed associated with the severity of cholestasis.
This result was not affected by adding clinically suspected sepsis ep-
isodes with negative blood cultures to the analyses (data not shown).
The infants with conjugated bilirubin peaking above 100 μmol/L
had a longer duration of cholestasis. Median duration of cholestasis
in this group was 87  days (IQR 45-112) as compared to 45.5  days
(IQR 18-62) among low-grade infants (P = .005). Data from five in-
fants in the former group were missing due to death before resolu-
tion of cholestasis and two in the latter group due to lack of blood
tests regarding bilirubin after discharge from the hospital. The rate
infants were treated with ursodeoxycholic acid were 79% and 56%
in the high- and low-grade groups, respectively (P = .17).
The rate of ROP and BPD did not differ significantly between
the groups.
3.4 | Long-term hepatic outcome following neonatal
cholestasis
Among those 101 out of 111 children surviving to discharge from
the NICU and remaining in Sweden, no child had developed clini-
cally significant chronic liver disease at the age of 10 years, neither
in the cholestasis group nor in the control group (P = 1.00). Two of
the cholestatic cases, male monozygotic twins, had developed small
gallstones during the neonatal period. They were asymptomatic and
followed by a general paediatrician up to 10  years of age and be-
yond. Repeated ultrasonography investigations showed gallstones
with unchanged small size, but no other hepatobiliary abnormalities.
The liver function tests were repeatedly normal.
Five children from each group were missing from this analysis.
During neonatal care, five out of 37 of the cholestatic infants and
two out of 74 non-cholestatic controls were deceased. Three con-
4 | D I S CU S S I O N
In this population-based cohort of very preterm infants, cholestasis
was common, with an overall incidence of 14.8%, being the high-
est in lower gestational ages. We found higher rates of inflamma-
tory complications such as BPD and ROP rates in cholestatic infants
than in non-cholestatic controls matched for gestational age. High-
grade cholestasis with maximum conjugated bilirubin exceeding
100 μmol/L was associated with a higher mortality than cholestasis
of lower grade. None of the surviving infants with neonatal choles-
tasis had developed chronic liver disease at 10 years of age.
Pre-eclampsia during pregnancy and low birth weight for ges-
tational age emerged as risk factors for cholestasis, based on the
differences between the cholestatic infants and non-cholestatic
controls. An association between cholestasis and growth restric-
tion has been described earlier,4,9,10 and small for gestational age is
 |
146       TENG et al.
associated with pre-eclampsia.13,14 Pre-eclampsia is an inflamma-
tory condition, and both pre-eclampsia and being born small for
gestational age have been associated with an increase in markers
for systemic inflammation at two to 4 weeks of age in very preterm
infants.15,16 Pre-eclampsia or being small for gestational age is also
17-19
associated with more ROP and BPD in very preterm infants.
20,21
Inflammation contributes to the development of BPD, and
there is a correlation between BPD and ROP. 22 Hence, inflamma-
tion in these susceptible organs, supposedly promoted by a gen-
erally proinflammatory environment, seems to play a role in the
multifactorial pathogenesis. Interestingly, inflammation may also
play a role in the development of cholestasis in these infants. For
example, DeMauro et al23 described that elevated c-reactive pro-
tein and interleukin-1β at 2 weeks of age in very low birth weight in-
fants was predictive of cholestasis. Alison et al10 showed that liver
stiffness was higher in infants with intra-uterine growth restriction
than in preterm infants with birth weight appropriate for gesta-
tional age, and even higher in growth-restricted infants who de-
veloped cholestasis. We therefore hypothesise that the increased
risk of cholestasis in infants with maternal pre-eclampsia, growth
restriction and being born small for gestational age in our cohort
might be linked to a proinflammatory state in these infants. This
warrants further studies as conclusions on any causational associa-
tion must be made with great caution because of the retrospective,
observational design of this study, and thereby the risk of unknown
bias or confounders.
The occurrence of adverse neonatal outcomes such as ROP, BPD
and death was significantly more frequent among cholestatic very
preterm infants than their non-cholestatic controls. We evaluated
the role of cholestasis in a multivariate model for BPD and ROP,
where the association remained after adjusting for gender, gesta-
tional age and duration of mechanical ventilation. It could be sug-
gested that cholestasis is linked to the development of ROP and BPD
in very preterm infants, possibly mediated by a common pathway of
increased inflammatory state. 21-24
The cholestatic infants were generally sicker, had a more compli-
cated neonatal period with prolonged need for parenteral nutrition,
and more episodes of sepsis than the non-cholestatic controls, in line
with reports by others.3,5 The sepsis episodes were predominantly
late-onset sepsis, possibly related to prolonged need for parenteral
nutrition and central line catheters, which is known to be associated
25
with increased risk for infection. Infections favour inflammation,
and sepsis has been shown to be associated with liver dysfunction in
different groups of neonates.5,26-28 Recent findings in a small group
of preterm infants on parenteral nutrition showed that the intesti-
nal microbiome was different even before cholestasis and shifted
towards gram-negative overweight in cholestatic infants compared
to the non-cholestatic infants receiving similar amount of parenteral
nutrition. 29 The possible role of the intestinal mibrobiome in associa-
tion to inflammation and cholestasis in this and other patient groups
warrants further investigation.
Mild cardiac anomalies were more common in the cholestatic
infants compared to controls. The finding was not explained by
selection bias, as the rate of infants in both groups examined by
echocardiography was high. Whether the development of cholesta-
sis in any way could be related to circulatory effects on the liver
remains to be established. However, direct hyperbilirubinaemia was
found in 17% of term and preterm neonates with congenital heart
disease by Fujishiro et al,30 which is an incidence rate much higher
than expected.
Somewhat surprisingly, the cholestatic infants regained their
birth weight faster than the control infants, but they had a signifi-
cantly lower weight at 36  weeks of corrected age. Infants in the
cholestatic group were generally sicker than controls, and this ini-
tially larger weight gain could tentatively be explained by more in-
tensive fluid therapy and parenteral nutrition.
We noted that all deaths in the cholestasis group occurred in
infants with peak conjugated bilirubin >100  μmol/L and that the
cholestasis in all these cases was ongoing without signs of resolution
at the time of death. All deceased infants died in episodes of multi-
ple organ failure, precipitated by sepsis or necrotising enterocolitis.
The fatal cases were too few to be analysed in a multivariate model.
However, in the univariate comparison, we can conclude that the
neonatal course, morbidities and complications were very similar
in the low-grade and high-grade cholestasis groups. However, our
results did suggest that higher grade of cholestasis may contribute
to higher mortality, a finding which needs to be verified in further
studies.
One important finding was that none of the very preterm
cholestatic infants, who survived the neonatal period, had de-
veloped clinically significant chronic liver disease at 10  years of
age. This type of very long-term follow-up has to our knowledge
not been reported earlier. It should be noted that this outcome
measurement did not include any test for liver function, only the
fact that the children had not been referred to the tertiary referral
centres of paediatric hepatology in Sweden. However, the devel-
opment of significant liver disease most likely would have been
discovered by this age and if so the tertiary centres of paediatric
hepatology would have been involved. Of the children still living in
Sweden at the age of 10 years, none had died between discharges
from the neonatal care until the age of 10 years. Only three infants
in the control group had moved abroad, but all surviving children
who were cholestatic as newborns still lived in Sweden at the age
of 10  years. We conclude that the rate of more severe chronic
liver disease, that manifests itself as cholestasis in very preterm
infants, is low. The sample size of 32 surviving preterm infants
may not be big enough to totally exclude that neonatal cholestasis
is associated with later liver disease, but it is large enough to sug-
gest that there is not a strong association. Prospective research
including liver function tests at follow-up would be needed to an-
swer this question more definitely. Even so, our finding supports
the current perception that cholestasis in preterm infants is a pri-
marily reversible condition in an otherwise healthy but immature
liver that resolves if the underlying conditions causing the need
for prolonged parenteral nutrition, and other risk factors, can be
managed.
TENG et al.
The strengths of this study are the population-based approach,
a relatively large cohort of very preterm infants, and the long-term
perspective. The limitations include the retrospective design, mak-
ing conclusions on causality difficult. Another limitation is that exact
doses of for example ursodeoxycholic acid and components of paren-
teral nutrition were not available. Also, since blood tests, diagnostic
procedures and most variables were based on clinical decisions from
the attending neonatologist rather than part of a prospective study
protocol, there was risk for misclassifications. However, this likely
has led to underestimation rather than overestimation of cholestasis
incidence.
5 | CO N C LU S I O N
We conclude that neonatal cholestasis is common in very preterm
infants and it is closely linked to disease severity and complications
in the neonatal period. Our results suggest that cholestasis may be
an independent risk factor for inflammatory complications such as
BPD and ROP in these patients and this needs be further addressed
in prospective studies. More severe cholestasis may be associated
with increased mortality, whereas cholestatic very preterm infants
who survive the neonatal period do not seem to be at great risk to
develop severe chronic liver disease later in childhood.
C O N FL I C T S O F I N T E R E S T
The authors have no conflicts of interest to declare.
ORCID
Jonas Teng  https://orcid.org/0000-0003-4351-9665
Kajsa Bohlin  https://orcid.org/0000-0003-3368-4149
Björn Fischler  https://orcid.org/0000-0002-8838-324X
REFERENCES
1. Fawaz R, Baumann U, Ekong U, et al. Guideline for the evaluation
of cholestatic jaundice in infants: joint recommendations of the
north american society for pediatric gastroenterology, hepatol-
ogy, and nutrition and the European society for pediatric gastro-
enterology, hepatology, and nutrition. J Pediatr Gastroenterol Nutr.
2017;64(1):154-168.
2. Koseesirikul P, Chotinaruemol S, Ukarapol N. Incidence and risk
factors of parenteral nutrition-associated liver disease in newborn
infants. Pediatr Int. 2012;54(3):434-436.
3. Lee HH, Jung JM, Nam SH, Lim G, Chung ML. Risk factor analysis of
parenteral nutrition-associated cholestasis in extremely low birth
weight infants. Acta Paediatr. 2016;105(7):e313-e319.
4. Lee SM, Namgung R, Park MS, et al. Parenteral nutrition associated
cholestasis is earlier, more prolonged and severe in small for gesta-
tional age compared with appropriate for gestational age very low
birth weight infants. Yonsei Med J. 2013;54(4):839-844.
5. Tufano M, Nicastro E, Giliberti P, Vegnente A, Raimondi F, Iorio R.
Cholestasis in neonatal intensive care unit: incidence, aetiology and
management. Acta Paediatr. 2009;98(11):1756-1761.
6. Christensen RD, Henry E, Wiedmeier SE, Burnett J, Lambert DK.
Identifying patients, on the first day of life, at high-risk of devel-
oping parenteral nutrition-associated liver disease. J Perinatol.
2007;27(5):284-290.
|
      147
7. Costa S, Maggio L, Sindico P, Cota F, De Carolis MP, Romagnoli
C. Preterm small for gestational age infants are not at higher
risk for parenteral nutrition-associated cholestasis. J Pediatr.
2010;156(4):575-579.
8. Robinson DT, Ehrenkranz RA. Parenteral nutrition-associ-
ated cholestasis in small for gestational age infants. J Pediatr.
2008;152(1):59-62.
9. Baserga MC, Sola A. Intrauterine growth restriction impacts tol-
erance to total parenteral nutrition in extremely low birth weight
infants. J Perinatol. 2004;24(8):476-481.
10. Alison M, Biran V, Tanase A, et al. Quantitative shear-wave elastog-
raphy of the liver in preterm neonates with intra-uterine growth
restriction. PLoS One. 2015;10(11):e0143220.
11. Kapoor V, Malviya MN, Soll R. Lipid emulsions for parenterally
fed term and late preterm infants. Cochrane Database Syst Rev.
2019;6:CD013171.
12. Teng J, Arnell H, Bohlin K, Nemeth A, Fischler B. Impact of paren-
teral fat composition on cholestasis in preterm infants: a population
based study. J Pediatr Gastroenterol Nutr. 2015;60(6):702-707.
13. Habli M, Levine RJ, Qian C, Sibai B. Neonatal outcomes in preg-
nancies with preeclampsia or gestational hypertension and in nor-
motensive pregnancies that delivered at 35, 36, or 37 weeks of
gestation. Am J Obstet Gynecol. 2007;197:406.e1-406.e7.
14. Wikstrom AK, Stephansson O, Cnattingius S. Previous preeclamp-
sia and risks of adverse outcomes in subsequent nonpreeclamptic
pregnancies. Am J Obstet Gynecol. 2011;204(2):148.e1-148.e6.
15. McElrath TF, Allred EN, Van Marter L, Fichorova RN, Leviton
A, Investigators ES. Perinatal systemic inflammatory re-
sponses of growth-restricted preterm newborns. Acta Paediatr.
2013;102(10):e439-e442.
16. Leviton A, Allred EN, Fichorova RN, et al. Antecedents of inflam-
mation biomarkers in preterm newborns on days 21 and 28. Acta
Paediatr. 2016;105(3):274-280.
17. Jensen EA, Foglia EE, Dysart KC, et al. Adverse effects of small for
gestational age differ by gestational week among very preterm in-
fants. Arch Dis Child Fetal Neonatal Ed. 2019;104(2):F192-F198.
18. Bose C, Van Marter LJ, Laughon M, et al. Fetal growth restriction
and chronic lung disease among infants born before the 28th week
of gestation. Pediatrics. 2009;124(3):e450-e458.
19. Hansen AR, Barnes CM, Folkman J, McElrath TF. Maternal pre-
eclampsia predicts the development of bronchopulmonary dyspla-
sia. J Pediatr. 2010;156(4):532-536.
20. Lal CV, Ambalavanan N. Biomarkers, early diagnosis, and clin-
ical predictors of bronchopulmonary dysplasia. Clin Perinatol.
2015;42(4):739-754.
21. Ambalavanan N, Carlo WA, D'Angio CT, et al. Cytokines associated
with bronchopulmonary dysplasia or death in extremely low birth
weight infants. Pediatrics. 2009;123(4):1132-1141.
22. Podraza W, Michalczuk B, Jezierska K, et al. Correlation of retinop-
athy of prematurity with bronchopulmonary dysplasia. Open Med
(Wars). 2018;13:67-73.
23. DeMauro SB, Kilpatrick LE, Gerdes JS, Abbasi S. Early inflammatory
markers for prediction of cholestasis in very-low-birth-weight in-
fants. Neonatology. 2012;102(3):229-234.
24. Pawlik D, Lauterbach R, Walczak M, Hurkala J, Sherman MP. Fish-
oil fat emulsion supplementation reduces the risk of retinopathy
in very low birth weight infants: a prospective, randomized study.
JPEN J Parenter Enteral Nutr. 2013;38(6):711-716.
25. Shah J, Jefferies AL, Yoon EW, Lee SK, Shah PS, Canadian NN.
Risk factors and outcomes of late-onset bacterial sepsis in
preterm neonates born at < 32 weeks' gestation. Am J Perinatol.
2015;32(7):675-682.
26. Yan W, Hong L, Wang Y, et al. Retrospective dual-center study of
parenteral nutrition-associated cholestasis in premature neonates:
15 years' experience. Nutr Clin Pract. 2017;32(3):407-413.
 |
148       TENG et al.

27. Wolf A, Pohlandt F. Bacterial infection: the main cause of acute S U P P O R T I N G I N FO R M AT I O N

cholestasis in newborn infants receiving short-term parenteral nu-
Additional supporting information may be found online in the
trition. J Pediatr Gastroenterol Nutr. 1989;8(3):297-303.
28. Khalil S, Shah D, Faridi MM, Kumar A, Mishra K. Prevalence and
Supporting Information section.
outcome of hepatobiliary dysfunction in neonatal septicaemia. J
Pediatr Gastroenterol Nutr. 2012;54(2):218-222.
29. Mokha JS, Davidovics ZH, Maas K, Caimano MJ, Matson A. Fecal How to cite this article: Teng J, Bohlin K, Nemeth A,
microbiomes in premature infants with and without parenteral Fischler B. Cholestasis after very preterm birth was associated
nutrition-associated cholestasis. J Pediatr Gastroenterol Nutr.
with adverse neonatal outcomes but no significant long-term
2019;69(2):224-230.
3 0. Fujishiro J, Sugiyama M, Ishimaru T, et al. Direct hyperbiliru- liver disease: A population-based study. Acta Paediatr.
binemia in infants with congenital heart disease. Pediatr Int. 2021;110:141–148. https://doi.org/10.1111/apa.15408
2018;60(2):179-182.
 | |
Received: 12 December 2019    Revised: 12 May 2020    Accepted: 18 May 2020

DOI: 10.1111/apa.15372

REGULAR ARTICLE

Controlled case series demonstrates how parents can be

trained to treat paediatric feeding disorders at home

Tessa Taylor1,2  | Neville Blampied2  | Nikolas Roglić1

1
Paediatric Feeding International, Sydney,
NSW, Australia
2
Department of Psychology, and the
School of Health Sciences, University of
Canterbury/Te Whare Wānanga o Waitaha,
Christchurch, New Zealand
Correspondence
Tessa Taylor, Department of Psychology,
Speech & Hearing, and the School of Health
Sciences, University of Canterbury/Te
Whare Wānanga o Waitaha, Private Bag
4800, Christchurch 8140, New Zealand.
Email: DrTaylor@PaediatricFeedingIntl.com
Abstract
Aim: Paediatric feeding disorders are normally managed by specialist clinics. We ex-
amined whether treatment gains were maintained when trained parents continued
the programme at home and during meals out.
Methods: This controlled consecutive case series recruited 26 children (22 boys) with
avoidant/restrictive food intake disorder, from a private paediatric feeding disorders
practice in New South Wales, Australia. Their mean age was six (2-13) years. All had
severe feeding problems and mealtime skill deficits, and most had autism and de-
velopmental delays or intellectual disabilities. The children received intensive, indi-
vidualised, behaviour-analytic treatment for 11 (6-21.5) days, and the parents were
trained to continue it at home. The primary treatment outcomes included the range
and amount of food eaten and mealtime behaviour.
Results: The children met all of the therapeutic goals agreed at the treatment out-
set. They ate a mean of 92 different foods and improved how they ate, drank and
behaved during mealtimes. The mean differences before and after treatment were
clinically and statistically significant, and the gains were maintained during follow-up
at a mean of 2.3 years. Parental satisfaction and treatment acceptability were high.
Conclusion: Specially trained parents successfully continued paediatric eating disor-
der treatment at home and maintained treatment gains.

KEYWORDS
avoidant/restrictive food intake disorder, home treatment, inappropriate mealtime behaviour
paediatric feeding disorders, parental training

1 |  I NTRO D U C TI O N
Chronic and severe paediatric feeding disorders can include being
dependent on a gastrostomy tube or liquids, refusing food and only
eating certain items. Feeding problems occur in wide-ranging popu-
lations, including children with typical development, medical com-
plications, such as food allergies and reflux,1 and developmental
Abbreviations: ARFID, avoidant/restrictive food intake disorder; ES, effect size; CI,
confidence interval.
disabilities. 2,3 Many meet the criteria for avoidant/restrictive food
intake disorder (ARFID), which includes failure to thrive, nutritional
deficiencies, dependence on artificial feeding methods and, or, di-
etary supplements and marked impairment in psychosocial function-
ing.4 When this occurs during critical neuro-developmental periods,
numerous adverse effects on behaviour, cognition, emotion and
health may occur and persist into adulthood.3-7
Over the past 40 years, treatment for paediatric feeding disor-
ders has been based on behaviour analysis principles and research

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

Acta Paediatrica. 2021;110:149–157.  |

wileyonlinelibrary.com/journal/apa     149
 |
150       TAYLOR et al.

has delivered a range of empirically supported, effective treatments.

However, these have only been generally available from specialised
multidisciplinary, hospital-based research clinics served by staff with
high levels of expertise in the treatment methods.4,8-11 Most children
do not have access to this treatment.7,12 In general, conventional
treatment methods do not have a high-quality evidence base world-
wide.4,12-17 Thus, lack of awareness of, and access to, effective treat-
ment may lead to continuing suffering and distress for children and
their families. Repeated treatment failure can also waste healthcare
resources and time.7 Translational research 18 is urgently needed to
demonstrate the feasibility of adapting specialist, in-hospital treat-
12,19,20
ment so that it can be delivered at home and in the community.
We also need research to evaluate treatment outcomes and show
that gains can be maintained over time and in different settings.
Our primary aim was to demonstrate what could be achieved
when effective behaviour-analytic treatment was translated from
specialist hospital settings and adapted so that specially trained par-
ents could use it at home and during meals out. It presents the first
analysis of the overall outcomes achieved with a consecutive, con-
trolled case series that received in-home, behaviour-analytic feeding
treatment without any additional food-related therapy.
2 |  M E TH O DS
The intensive, individualised treatments used in this translational
effectiveness trial study were adapted from Taylor et al19 The
study took place from May 2016 to March 2019 at a private clini-
cal practice specialising in treating paediatric feeding disorders in
New South Wales, Australia. It was performed in accordance with
the Declaration of Helsinki and the relevant ethical standards for
psychologists and behavioural analysts. The parent provided writ-
ten, informed consent for the therapy and the subsequent use of the
de-identified data.
Key Notes
• This study examined whether treatment gains were
maintained when specially trained parents treated pae-
diatric feeding disorders at home.
• We recruited 26 children aged 2-13 years with avoidant/
restrictive food intake disorder, and most had autism
and developmental delays or intellectual disabilities.
• Some 2 years after the programme started, the chil-
dren were eating a mean of 92 different foods and
had improved how they ate, drank and behaved during
mealtimes.
experienced unsuccessful treatment for a mean of 3.5 ± 3.0 years, and
the maximum was 11 years. Multiple treatment experiences were com-
mon, with a mean of 5 ± 4 treatments and a maximum of 12. Most chil-
dren were receiving ongoing services and monitoring, but these were
paused during the intensive behavioural treatment programme and the
previous feeding targets did not continue after the programme. The
therapists and parents agreed on a mean of 11 treatment goals (range
7-17), which were used to provide individualised treatment. The ses-
sions were held in settings with the necessary seating, equipment and
supplies. The primary setting was at home, and the other settings were
cafes, restaurants, day care centres and schools.
2.2 | Measurement, reliability, integrity and
acceptability
A trained observer computer-coded the children and therapists'
behaviours using specialised real-time data collection software
(BDataPro, Kennedy Krieger Institute; 21). The frequency keys
recorded the frequency of responses, including acceptance,

2.1 | Patients TA B L E 1   Characteristics of the 26 children in the study cohort

n (%)
We reviewed 29 consecutive cases and reported the outcomes re-
ported for 26 children who were diagnosed with ARFID by a registered Male 22 (85%)
clinical psychologist and doctoral-level certified behavioural analyst Autism, developmental delay 20 (77%)
with specialist training in paediatric feeding disorders. Three cases Growth impairments 14 (54%)
were excluded because they were receiving other therapies. Each Constipation 11 (42%)
child's physician approved their participation and provided medical in- Nutritional deficiencies 8 (31%)
formation, such as prior screening, growth and medical history, food Formula or baby bottle dependence 8 (31%)
allergies and any multidisciplinary specialist and testing referrals. The
Genetic/chromosomal abnormalities 5 (19%)
detailed patient characteristics can be found in Table 1, and the meth-
History of nasogastric tube 3 (12%)
odology is summarised here and described in more detail in Appendix
Eczema 3 (12%)
S1. The average age of the 26 children was six (range 2-13) years: 15
Gastrostomy tube 2 (8%)
were Caucasian Australian and the others were of Asian, Arabic and
Parenteral nutrition 1 (4%)
European ethnicities and nationalities. Most of the families had high so-
cio-economic status. At their initial assessment, the children displayed a Coeliac disease 1 (4%)

range of severe feeding problems and mealtime skill deficits.7 They had Note: Other factors not detailed here were prematurity and reflux.
TAYLOR et al. |
      151

swallowing, inappropriate mealtime behaviour and expulsion. The
duration keys measured delays in seconds before acceptance and
swallowing occurred and the duration of negative vocalisation.
Frequency of swallowing was divided by the number of therapist-
programmed bite presentations during the session to provide an
overall measure of the percentage of food consumed. The last
three data points in the initial phases, namely baseline one and
treatment one, were averaged, as were the repeated baseline two
and treatment two phases. The plates were photographed to re-
cord the variety of the food eaten, and the food was measured
before and after eating to determine the volume consumed. All
19
parents were provided with a questionnaire at the end of the
programme, and the responses to the five-point Likert scale items
were averaged to assess the 23 programme satisfaction items and
the 16 items on social acceptability of the treatment. All the ses-
sions were video-recorded to assess the interobserver agreement
and reliability, which was 96%. 21 The mean responses per minute
for lapses in procedural integrity for each minute were 0.1 lapses
for the therapist and 0.2 for the parents, and the mean interob-
server agreement was 99.5%.
2.3 | Experimental design and data analysis
The study included all cases that met the inclusion criteria, regard-
22
less of outcome, to reduce selection and publication bias. This
was consistent with the consecutive, controlled case series experi-
mental design. All the 26 cases also demonstrated experimental
control, using within-subjects, single-case experimental designs,
where each case served as its own control. 23 The most frequently
used design, for more than 80% of the children, was a reversal, or
withdrawal, design. This involved taking the treatment out, namely
baseline two, and putting it back in when needed, namely treat-
ment two. This was because replication showed that the treatment
was needed and was the reason for the improvement in eating.
The outcomes were examined using a combination of graphic tech-
niques24 and effect size estimation, using Cohen's dav, where the
standardiser was the average of the standard deviations before
and after treatment. 25 The graphs were predominantly modified
Brinley plots (Figure 1;26). These are scatterplots that display indi-
vidual changes over time, with each data point representing coor-
dinates of an individual's score at time one on the x-axis and time
two on the y-axis.
2.4 | Procedure
The mean waiting time for treatment was 3.5 (range 1-10) months.
Once it had started, the children received intensive continuous treat-
ment, which was provided for about 6 days per week for 2-4 weeks
during which they received treatment for an average of 11 (6-21.5)
days. The first author was the principal therapist and conducted the
sessions for approximately 7-8 hours per day with a trained assistant.
The involvement of the therapist reduced as the parents developed
the skills they needed and the responsibility for the intervention was
gradually transferred to them.

Modified Brinley plot interpretaon

Reducon = improvement Increase = improvement

Deterioriation Improvement
Final Treatment

ge
ch f
o
an
ne
Li
no

Improvement
Deterioriation

Baseline Baseline

F I G U R E 1   Illustrates the key features of modified Brinley plots and the interpretation of the graph space when a reduction in score
indicated clinical improvement (left panel) and when an increase indicated improvement (right panel). The + shows the coordinates of the
baseline and final treatment means for the lowest data points. Data points that fall near, or on, the 45° diagonal line of no change represent
individuals who showed little or no change between time one and time two. The magnitude of individual change is displayed by points lying
at greater distances above and below the line. Depending on the direction of the clinical and desired change for the measure, any change can
be classified as an improvement or deterioration. To aid interpretation, 26 the group mean coordinate is shown by a plus sign and the Cohen's
dav effect size 25 is displayed along with the 95% confidence Interval (95% CI), calculated using ESCII software developed by Cumming, La
Trobe University, Melbourne, Australia 28
152       | TAYLOR et al.

Percentage consumpon from baseline to final treatment

Baseline 1 vs Treatment 1 Baseline 1 vs Baseline 2 Baseline 2 vs Treatment 2

100

Treatment 2
Perc ent of bites c ons um ed

80 Treatment 1

Baseline 2 Baseline 2
60

40

20
Baseline 1

0
0 3 6 9 12 15 18 21 24 27 0 3 6 9 12 15 18 21 24 27 0 3 6 9 12 15 18 21 24 27
Parcipants (in order of case number)

F I G U R E 2   Horizontal dot plots show the percentage of bites consumed by each child in initial baseline one compared to initial treatment
one (left panel), replication of baseline two compared to initial baseline one (middle panel) and replication of final treatment two compared
to replicated baseline two (right panel). Pairs of data points are shown for each child in consistent case-number order from child one on the
extreme left of the x-axis to child 26 on the extreme right of each panel. Pairs of percentages (one above the other) are displayed for each
case sequentially for case one to case 26 in the same order over the three figure panels

2.5 | Assessment and treatment
The review of the intake assessment included the physician's report
and reports from other disciplines and service providers, meal obser-
vations, intake questionnaires, parental interviews, three-day diet
records, functional assessment, case-conceptualisation and therapy
goal specification. We directly assessed each child's food and reward
preferences.19,27 Foods from each food group were targeted and the
spoon size and food texture matched individual needs. Treatments
were individualised and data-driven, with components selected from
8-10,12
well-established, empirically supported treatments.
involved changing the mealtime environment, such as providing
incentives for swallowing and praising appropriate mealtime be-
haviour. Treatment also involved setting achievable mealtime re-
quirements, such as continuing to present food and liquids despite
rejection, replacing food or liquid that was spat out and providing
prompting instructions and guidance. We also taught the skills
needed to achieve goals, such as chewing, using utensils, self-feed-
ing and drinking, drinking with an open cup, biting from whole food
portions rather than pre-cut bite sizes and taking medicine. As the
treatment progressed, intensive parental behavioural skills training
was provided and we varied the cutlery and crockery provided, how
the children were placed at family meals and the setting where they
ate19 to ensure generalisation of therapy gains.
2.6 | Maintenance and follow-up
To help the children maintain their treatment gains, the parents were
given specific maintenance instructions. For two weeks after the
end of the programme, they were asked to offer their child the foods
on the discharge food list, record mealtime data, take photographs
of the plates and video the meals. The data were returned to the
first author for review and follow-up. The parents were contacted
2-3 years after treatment, at a mean of 2.3 years, and asked to com-
plete a five-point Likert satisfaction survey, which ranged from one
for worse than pre-treatment to five for resolved.
3 | R E S U LT S
Treatment
Figure 2 presents the primary outcome data, which was consump-
tion, and demonstrates the experimental control as horizontal dot
plots. Food consumption was very low at the initial baseline meas-
ure, baseline one, with most children not consuming any of the food
presented. At the end of the first treatment phase, treatment one,
children were completing 100% of the programmed bites consist-
ently, within a mean of 46 (2-134) minutes of treatment starting.
When treatment was withdrawn as a replication procedure, at base-
line two, consumption decreased for all children. When treatment
resumed in the second treatment phase, namely treatment two, con-
sumption increased again for all children, with no overlap of points
with the baseline, thus replicating the treatment effect. At the end
of the programme, the children were consuming an average of 158
(52-292) grams of solids per meal.
The number of foods the children ate increased from an av-
erage of six, mainly starch, to a mean of 92 from all food groups,
which equated to a very large Cohen dav effect size of about four
(Figure 3). One child could only eat 30 different foods due to aller-
gies and was awaiting tests because of vomiting (Figure 3). Examples
TAYLOR et al. |
      153

of the meals consumed are shown in Figure 4. Figure 5 shows that
for all children, inappropriate mealtime behaviour, expulsion and
negative vocalisation at the end of treatment were low or reduced.
The overall effects sizes for all three measures were large (Cohen's
dav −1.4 to −2.35). Before treatment, there were long delays to tak-
ing and swallowing bites which decreased after treatment (Figure 6).
Individual differences in the nature and texture of the food and how
skilled the children were at chewing and swallowing accounted for
the considerably lower Cohen's dav effect size for swallowing, of
about −1, than for accepting food when presented (−2.6).
The Cohen's dav effect sizes were large to very large (range
1.4-4) across the different measures and in no instance did the 95%
confidence interval (95% CI) indicate that the effect size was close
to zero, as the associated mean change was always statistically sig-
nificant (P < .05). When we measured the treatment effect, based on
the average percentage reduction from baseline to the end of the
treatment, it was 98% for non-consumption. Delays in acceptance
fell by 97% and 75% for delays to an empty mouth, but the figure
was 89% when we excluded the five children who had reached these
goal levels at baseline. Other reductions were 99% for food refusal,
97% for food expulsion and 97% for negative vocalisation. All the
children met 100% of the treatment goals, including learning to take
medication, chewing, feeding themselves with utensils, drinking
from a cup and sitting for meals.
At discharge, 20 parents (77%) completed the questionnaire.
They reported high satisfaction (mean 4.9, range 4.6-5, maximum
5.0) and social acceptability of the treatment (mean 4.9, range
4.6-5, maximum 5.0). During the immediate follow-up period of
two weeks, 17 parents reported on a mean of 25 (4-135) meals.
They reported high and stable consumption of the food presented
(mean 99%, range 90%-100%). Anecdotally, families reported im-
provements in other areas of life and well-being, for example in
height, weight and body mass index, bloodwork for nutritional
deficiencies, hair quality and growth, toothbrushing, sleeping
and toileting. They also reported improvements in their child's
behaviour, such as being able to change between activities and
follow instructions. Improvements in learning were also reported,
such as progress in therapy and school. Other benefits were

Variety of foods consumed from before treatment to final treatment

160
150 Number of Foods Consumed
Fi nal Treatment - Total Number of Foods

140
130
120
110
100
90
80
70
60
50
40
dav = 4.2 [3.1, 5.7]
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

Baseline - Total Number of Foods

45
Protein Starch Veges Fruit Combo
40
Final Treatment - Number consumed

35

30

25

F I G U R E 3   Modified Brinley plots 20

showing total number of foods consumed

15
before and at the end of treatment (top
figure) and truncated modified Brinley 10
plots showing the range of different
food groups consumed after treatment 5

compared to before treatment (bottom

0
figure). The diagonal line in every plot 0123456789 0123456789
0123456789 0123456789 0123456789
indicates the line of no change from
before treatment to the end of treatment Baseline - Number consumed
 |
154       TAYLOR et al.

Sample pictures of plates children consumed

F I G U R E 4   Sample pictures of meals the children consumed

related to energy, sports and social participation, travel, immunity,
skin colour and decreased dependence on items like supplements,
formulas and laxatives. One example of growth improvement was
an 11-year-old child who was overweight and had stunted growth
with a blood analysis that revealed 18 significant deficiencies and
abnormalities. After he was discharged, his parents reported that
he had rapidly grown 25 cm in height and this was confirmed by
his therapist.
At follow-up, 21 parents reported that their child's feeding prob-
lem was much better than before treatment and five said it had com-
pletely resolved. The mean Likert score was 4.12 (2-5). No parents
reported the problem was worse. All but one of the parents reported
that their children were eating from all food groups. The exception
was a child who had similar before and after results as moving home
meant their parents did not maintain the protocol.
4 | D I S CU S S I O N
The data analysis used in this study showed both individual and over-
all group patterns of change following therapy. The consistently large
size of change seen in the figures was confirmed by the standardised
mean difference effect size and 95% CI estimation, with Cohen's dav
values that were large to very large and statistically significant. This
TAYLOR et al. |
      155

Inappropriate mealme behaviour, expulsion, F I G U R E 5   Modified Brinley plots comparing before treatment
and negave vocalisaons from before treatment with the end of treatment. These show changes in the frequency
to final treatment of inappropriate mealtime behaviour (top panel), the rate at which
food was expelled by each child (middle panel) and the percentage
of sessions with negative vocalisations (bottom panel). The diagonal
Inappropriate Mealtime
35 line in every plot indicates the line of no change from before
Final Treatment - Refusals/minute

Behaviour
treatment to the end of treatment
(IMB)
30

dav = –2.1 highly positive outcome of the treatment was firstly the result of
25
matching the treatment to each child and, second, continuing with
[–2.85, –1.275]
20 treatment until all target goals were met. Confidence in conclusions
was enhanced because the interobserver agreement and procedural
15
integrity measurements showed that the therapy data were accu-

10 rately recorded and that the therapists accurately followed the pro-
tocol, as the data quality and treatment fidelity were high.
5 It was also a strength of this study that the parents were trained
to a high standard so that they could continue to use the meal pro-
0
tocols independently, rather than needing long-term professional
0 5 10 15 20 25 30 35 services. Treatment was tailored by directly working with each
Baseline - Refusals/minute child rather than adopting a generic approach. Parents were then
thoroughly trained on precisely established procedures that were
11 Rate of Food Expels specifically designed for their child and had already been shown
10
Final Treatment - Expels/minute

to work. The treatment was safe, ethical and posed no risks to the
9 dav = –1.44 [–2.1, –.79] child's health, as none of the therapy involved depriving children of
8 preferred foods, drinks or formula. It also encouraged children to
7 eat various nutritious foods from all food groups. Furthermore, the
6 children were taught the mealtime skills they needed for age-appro-
5 priate independence and so that they would accept different food
4 textures.

3 A limitation of this study was the lack of observational and sys-

2 tematic, standardised data during the follow-up period and the lack
of intake data while children were on the waiting list for treatment.
1
Another limitation was that their body mass index could not be cal-
0
culated during these two periods. However, the children increased
0 1 2 3 4 5 6 7 8 9 10 11 their consumption quickly, on the first day, so there were no con-
Baseline - Expels/minute cerns about any further weight loss during treatment, except for one
child with malabsorption and diarrhoea due to a genetic condition
100
Negative Vocalisations and vomiting. Future studies should include multi-method, multi-
90 trait measures that are comparable from the time of intake through
Final Treatment - % of Sessions

80 dav = –2.35 [–3.1, –1.5] to follow-up periods. More research is also needed on younger age

70 groups and early intervention and prevention methods. The current

study was carried out with mostly high socio-economic status fami-
60
lies, due to lack of government funding in this treatment, so it is not
50 clear whether the training and follow-up results would be the same
40 for low status families.

30 It should be noted that most of the children in our study were

not dependent on tube feeding. The severity of typical presenta-
20
tions would probably be higher in hospital settings than in our pri-
10 vate practice, because such clinics may have more patients with
0 long-term gastrostomy tubes and higher levels of more severe med-
0 10 20 30 40 50 60 70 80 90 100 ical complications and developmental issues. Future studies might
Baseline - % of Sessions replicate this treatment model with children who depend on tube
feeding, as they may need longer treatment, especially to teach
them to chew and drink from a cup and to accept different food
 |
156       TAYLOR et al.
Time taken (seconds) to accept and swallow bites comparing before treatment to final treatment
Latency to Acceptance
Final Treatment
180
Latency (sec)
120 dav = –2.59 [–3.5, –1.7]
60
0
0 60 120 180 240 300 360 420 480 540 600 0
Baseline Latency (sec)
F I G U R E 6   Truncated modified Brinley plots showing latency, in seconds, to food acceptance into the mouth upon presentation and
latency to showing a clean mouth, namely swallowing, on inspection following acceptance before treatment and at the end of treatment.
The diagonal line represents the line of no change
textures following tube weaning. Also, certain children may require
hospitalisation and multidisciplinary teams if they have swallowing
problems, severe food allergies, breathing issues, seizures and blood
sugar issues, such as glycogen storage disease. Treatment intensity
and duration needs to be individually calibrated to the severity of
cases and decisions about treatment should be data-driven and
based on progress towards goals. If not, there is a risk of iatrogenic
worsening of the feeding disorder.4,11,19 It is important to note that
the current study was not conducted in the same way as a conven-
tional outpatient service, as we provided continuous sessions for full
consecutive days. However, this intensive treatment produced rapid
improvements in feeding compared to typical, conventional outpa-
tient treatment. Finally, we note that proper expertise, notably years
of training at the hospital where this treatment was developed, is
needed to run this treatment programme.4,11,19
Another strength of the study was that quick improvements
were shown, even though many children had unsuccessful prior
treatment, often involving a range of disciplines and approaches.
In addition, no families withdrew because the parents thought the
feeding disorder had improved while they were on the waiting list.
On the contrary, parents typically reported that the feeding problem
had worsened. Feeding disorders become harder to treat as children
get older and treatment options become limited and more difficult.
That is why early, effective intervention is very important, both for
the well-being of the child and their family and to avoid wasting lim-
ited health system resources.7
5 |  CO N C LU S I O N
This study clearly demonstrated that an intensive, empirically valid
treatment for paediatric feeding disorders could be provided by
properly trained clinicians and continued by trained parents at home
or during meals out. The success of our programme is highly encour-
aging, as it suggests that these empirically valid behaviour-analytic
treatments can be used outside the few specialist settings where
they were developed. We demonstrated that specially trained par-
ents were able to continue the programme and continue to achieve
treatment outcomes several years after they took over responsibil-
ity for the programme. This could improve access to treatment and
Latency to Swallow (Mouth Clean)
dav = –1.3 [–1.95, –.65]
60 120 180 240 300 360 420 480 540 600
help many more families coping with severe and chronic paediatric
feeding disorders.
C O N FL I C T O F I N T E R E S T
The authors have no conflicts of interest to declare.
ORCID
Tessa Taylor  https://orcid.org/0000-0001-8643-3955
Neville Blampied  https://orcid.org/0000-0002-0158-4904
REFERENCES
1. Yeung KA, Taylor T, Scheimann A, et al. The prevalence of food al-
lergies in children referred to a multidisciplinary feeding program.
Clin Pediatr. 2015;54(11):1081-1086.
2. Taylor T, Kozlowski AM, Girolami PA. Comparing behavioral treat-
ment of feeding difficulties and tube dependence in children with
cerebral palsy and autism spectrum disorder. NeuroRehabilitation.
2017;41(2):395-402.
3. Kozlowski AM, Taylor T, González ML, Girolami PA. Feeding dis-
orders. In: Matson JL, Matson ML, eds. Comorbid Conditions in
Individuals with Intellectual Disabilities. Cham, Switzerland: Springer;
2015.
4. González ML, Mulderink TD, Girolami PA. Avoidant restrictive food
intake disorder. In: Maragakis A, O'Donohue WT, eds. Principle-
Based Stepped Care and Brief Psychotherapy for Integrated Care
Settings. Cham, Switzerland: Springer; 2018.
5. Zucker N, Copeland W, Franz L, et al. Psychological and psycho-
social impairment in preschoolers with selective eating. Pediatrics.
2015;136(3):e582-e590.
6. Schwarzenberg SJ, Georgieff MK. Advocacy for improving nutrition
in the first 1000 days to support childhood development and adult
health. Pediatrics. 2018;141(2):e20173716.
7. Taylor T, Taylor S. Let's not wait and see: the substantial risks of
paediatric feeding problems. Manuscript submitted for publication;
2020.
8. Kerwin MLE. Empirically supported treatments in pediatric psychol-
ogy: severe feeding problems. J Pediatr Psychol. 1999;24(3):193-214.
9. Volkert VM, Piazza CC. Pediatric feeding disorders. In: Sturmey
P, Hersen M, eds. Handbook of Evidence-Based Practice in Clinical
Psychology. Hoboken, NJ: John Wiley & Sons; 2012.
10. Sharp WG, Jaquess DL, Morton JF, Herzinger CV. Pediatric feed-
ing disorders: a quantitative synthesis of treatment outcomes. Clin
Child Fam Psychol Rev. 2010;13(4):348-365.
11. Piazza CC, Milnes SM, Shalev RA. A behavior-analytic approach to
the assessment and treatment of pediatric feeding disorders. In:
Roane HS, Ringdahl JE, Falcomata TS, eds. Clinical and Organizational
TAYLOR et al.
Applications of Applied Behavior Analysis. Cambridge, MA: Academic
Press; 2015.
12. Sharp WG, Volkert VM, Scahill L, McCracken CE, McElhanon B. A
systematic review and meta-analysis of intensive multidisciplinary
intervention for pediatric feeding disorders: how standard is the
standard of care? J Pediatr. 2017;181:116-124.
13. Peterson KM, Piazza CC, Volkert VM. A comparison of a modified
sequential oral sensory approach to an applied behavior-analytic
approach in the treatment of food selectivity in children with au-
tism spectrum disorder. J Appl Behav Anal. 2016;49(3):485-511.
14. Addison LR, Piazza CC, Patel MR, et al. A comparison of sensory in-
tegrative and behavioral therapies as treatment for pediatric feed-
ing disorders. J Appl Behav Anal. 2012;45(3):455-471.
15. Lukens CT, Silverman AH. Systematic review of psychological
interventions for pediatric feeding problems. J Pediatr Psychol.
2014;39(8):903-917.
16. Taylor SA, Virues-Ortega J, Anderson R. Transitioning children from
tube to oral feeding: a systematic review of current treatment ap-
proaches. Speech Language Hearing. 2019;22:1-14.
17. Gosa MM, Carden HT, Jacks CC, Threadgill AY, Sidlovsky TC.
Evidence to support treatment options for children with swallowing
and feeding disorders: a systematic review. J Pediatr Rehabil Med.
2017;10(2):107-136.
18. Lerman DC. From the laboratory to community application:
translational research in behavior analysis. J Appl Behav Anal.
2003;36(4):415-419.
19. Taylor T. Intensive paediatric feeding treatment and the use of a
side deposit for a clinical case within the home setting in Australia.
Eur J Behav Anal. 2018;19(2):208-227.
20. Taylor SA, Purdy SC, Jackson B, Phillips K, Virues-Ortega J.
Evaluation of a home-based behavioral treatment model for chil-
dren with tube dependency. J Pediatr Psychol. 2019;44(6):656-668.
21. Bullock CE, Fisher WW, Hagopian LP. Description and validation of
a computerized behavioral data program: “BDataPro”. Behav Anal.
2017;40(1):275-285.
|
      157
22. Hagopian LP. The consecutive controlled case series: design, da-
ta-analytics, and reporting methods supporting the study of gener-
ality. J Appl Behav Anal. 2020;53(2):596-619.
23. Cooper JO, Heron TE, Heward WL. Applied behavior analysis.
Hoboken, NJ: Pearson; 2020.
24. Wilkinson L. Graphic displays of data. In: Cooper H, ed. APA
Handbook of Research Methods in Psychology. Washington, DC:
American Logical Association; 2012.
25. Lakens D. Calculating and reporting effect sizes to facilitate cu-
mulative science: a practical primer for t-tests and ANOVAs. Front
Psychol. 2013;4:863.
26. Blampied NM. Analyzing therapeutic change using modified
Brinley plots: History, construction, and interpretation. Behav Ther.
2017;48(1):115-127.
27. Kunkel KR, Kozlowski AM, Taylor T, González ML. Validating a food
avoidance assessment for children with food selectivity. Behav Dev.
2018;23(2):89-105.
28. Cumming G. Understanding the new statistics: Effect sizes, confidence
intervals, and meta-analysis. New York, NY: Routledge; 2012.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
How to cite this article: Taylor T, Blampied N, Roglić N.
Controlled case series demonstrates how parents can be
trained to treat paediatric feeding disorders at home. Acta
Paediatr. 2021;110:149–157. https://doi.org/10.1111/
apa.15372
 | |
Received: 9 March 2020    Revised: 14 May 2020    Accepted: 20 May 2020

DOI: 10.1111/apa.15375

REGULAR ARTICLE

Influence of maternal region of birth on neonatal outcomes of

babies born small

Nicole E. Young1 | Miranda Davies-Tuck2 | Atul Malhotra2,3,4

1
Monash University, Melbourne, Vic.,
Australia
2
The Ritchie Centre, Hudson Institute of
Medical Research, Melbourne, Vic., Australia
3
Monash Newborn, Monash Children’s
Hospital, Melbourne, Vic., Australia
4
Department of Paediatrics, Monash
University, Melbourne, Vic., Australia
Correspondence
Atul Malhotra, Department of Paediatrics,
Monash University, 246 Clayton Road,
Clayton, Vic., Australia.
Email: atul.malhotra@monash.edu
Funding information
MDT received support from the National
Health and Medical Research Council of
Australia Fellowship program. AM received
funding from the Royal Australasian College
of Physicians. The funders had no role in
study design, data collection and analysis,
decision to publish or preparation of the
manuscript.
Abstract
Aim: To compare neonatal outcomes of small for gestational age (SGA) infants born
to South Asian (SA)-born women and Australian/New Zealand (ANZ)-born women.
Methods: Retrospective cohort study at a hospital network in Australia. Maternal
and neonatal data were collected for infants born SGA between 2013 and 2017 to
SA- or ANZ-born women. Rates of perinatal mortality and neonatal morbidities were
analysed between groups.
Results: A total of 1018 SA and 959 ANZ SGA infants were included. SA SGA babies
were older (median [IQR] 39 [38-40] weeks) and heavier (2590 [2310-2780] grams)
compared to ANZ SGA babies (38 [37-40] weeks and 2480 [2059-2740] grams;
P < .001 for both). After adjustment for differences in demographics, SA SGA babies
were 1.5 times more likely to develop hypothermia (CI: 1.16-1.88, P = .001), but 60%
less likely to be born with a major congenital malformation (CI: 0.24-0.67, P = .001)
and 36% less likely to need gavage feeding (CI: 0.43-0.93, P = .02) compared to ANZ
SGA babies.
Conclusion: Small for gestational age babies of SA-born women have different neo-
natal outcomes as compared to those born to ANZ-born women. Further research
into influence of maternal region of birth on placental function, organogenesis and
body composition of SGA babies is warranted.

KEYWORDS
foetal growth restriction, hypothermia, intrauterine growth restriction, small for gestational
age, South Asian

1 | I NTRO D U C TI O N
Small for gestational age (SGA) defines infants born with a birth-
1
weight below the 10th percentile for their gestational age. SGA
births can occur due to pathological processes such as intrauterine
or foetal growth restriction (IUGR/FGR), or infants can be consti-
tutionally small. 2 In both situations, SGA babies experience higher
rates of mortality and morbidity compared to appropriate for ges-
tational age babies.3 One group of women who are more likely to
give birth to SGA babies are those born in the South Asia region
(eg India, Sri Lanka, Bangladesh).4-9 This is important in many high-
income countries like Australia, where one in three births is to a
woman born overseas, and in 2017, 8.8% of births were to a South
Asian (SA)-born mother.10

Abbreviations: SGA, small for gestational age; SA, South Asian; ANZ, Australian/New Zealand; AGA, appropriate for gestational age; IUGR/FGR, intrauterine/foetal growth restriction;
ROB, region of birth; NICU, neonatal intensive care unit; SCN, special care nursery; NGT, nasogastric tube; HMD, hyaline membrane disease; CLD, chronic lung disease.

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

|
158     
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:158–165.
YOUNG et al.
South Asian-born women are 11 times more likely to give birth
to an SGA infant and it has been argued that this reflects a physio-
logically, rather than pathologically, small infant.11 Babies born to SA
women however are at higher risk of preterm delivery and stillbirth
at term.7,12,13 Interestingly, premature infants of SA region of birth
(ROB) seem to experience less severe neonatal outcomes and mor-
bidity when compared to other babies.14 It has been suggested that
this may relate a shortened gestational length and potentially earlier
foetal maturation in pregnancies of this population.5,9,13 Due to this,
they may present with different and more favourable outcomes in
the neonatal period; however, this is unknown.
The aim of this study was to compare perinatal and neonatal out-
comes of babies born SGA at all gestations, between women of SA
and Australian/New Zealand (ANZ) ROB. This would allow further
investigation into whether the mechanism and consequence of SGA
births in each population is of physiological or pathological signifi-
cance, for both preterm and term infants.
2 |  PATI E NT S A N D M E TH O DS
This was a retrospective cohort study performed at a large met-
ropolitan health service (Monash Health) in Melbourne, Australia.
Monash Health has three birthing and neonatal units in South East
Melbourne: Monash Medical Centre, Clayton; Dandenong Hospital;
and Casey Hospital. The three hospitals have a combined birth rate
of around 9000 babies per year. All birthing data for babies born
within the 5-year period of January 2013 to December 2017, under
the 10th centile of birthweight for gestational age, were extracted.
Birthweight centiles were assigned according to the Australian na-
tional birthweight charts.15
Data were collected from two electronic medical record data-
bases: the birthing outcomes system and scanned medical records.
The birthing outcomes system is the hospital's routine maternity da-
tabase, recording maternal and perinatal data on all births ≥20 weeks’
gestation. Data items extracted included self-reported maternal
country of birth, maternal age, pre-existing medical conditions, ob-
stetric complications (pregnancy-induced hypertension/pre-eclamp-
sia, gestational diabetes, cholestasis of pregnancy), smoking, alcohol
consumption and substance use during pregnancy. Birth character-
istics included baby birthweight, sex, gestation (sub-classified as
<32, 32-36, ≥37  weeks), birth type, onset of labour and mortality.
Mortality was further categorised into stillbirths, death after birth
with no resuscitation provided, and neonatal death after admission.
Scanned written medical records were used to study neonatal
outcomes, which included Apgar score at 5 minutes, resuscitation
(including need for intubation, continuous positive airway pressure,
intermittent positive pressure ventilation, oxygen, suction), and ad-
mission to neonatal intensive care unit (NICU) or special care nursery
(SCN). Respiratory morbidity was recorded as respiratory support for
greater than 4 hours, length of respiratory support required, intuba-
tion in NICU, hyaline membrane disease (HMD), transient tachypnoea
of newborn (respiratory support for >4 hours post-birth, with no
|
      159
Key notes
• Small for gestational age (SGA) infants have different
neonatal outcome profiles dependent on maternal re-
gion of birth.
• South Asian SGA infants are at higher risk of neonatal
hypothermia, while Australian/New Zealand SGA in-
fants are more likely to require nasogastric feeding or
have congenital abnormalities, and experience a longer
duration of hospital stay.
• These neonatal differences may be explained by physi-
ological and pathological aetiologies for low birthweight
in each population.
alternative diagnosis), meconium aspiration syndrome, chronic lung
disease (CLD) (oxygen or respiratory support required at 36 weeks
corrected gestation), discharge home on oxygen, caffeine, surfac-
tant and postnatal steroid therapy. Further data were collected on
hypothermia (temperature <36°C within first 24 hours of life), hy-
poglycaemia (blood glucose level reading of <2.6 mmol/L), hypo-
glycaemia requiring or while on intravenous therapy, the need for
naso/orogastric (NGT/gavage) feeding, hyperbilirubinaemia requiring
phototherapy, culture-positive sepsis, anaemia requiring treatment,
thrombocytopenia requiring platelet transfusion, necrotising entero-
colitis (demonstrated on abdominal X-ray and requiring medical or
surgical treatment), retinopathy of prematurity requiring (laser) treat-
ment, osteopaenia of prematurity requiring treatment, patent ductus
arteriosus requiring treatment (medical or surgical), brain injury, hy-
ponatraemia, hyperglycaemia, hypokalaemia, perinatal asphyxia, and
major congenital anomaly, malformation or syndrome. Length of stay
admitted to SCN/NICU and length of stay overall were also collected.
Mother–baby pairs were classified by maternal self-reported
country of birth. They were identified and grouped as ANZ (Australia
or New Zealand), or SA ROB (including Afghanistan, Bangladesh,
Bhutan, India, Iran, Maldives, Nepal, Pakistan and Sri Lanka), accord-
ing to the geographical subregions defined by the United Nations.16
All infants of other maternal ROBs were excluded.
Analyses were conducted using SPSS (IBM version 20;
IBM corp.). Continuous data were first assessed for normality.
Differences in maternal demographics, obstetric and maternal
characteristics, and birth characteristics were determined between
ANZ-born and SA-born women using t tests, Mann–Whitney U test
or chi-squared test as appropriate. Differences in neonatal out-
comes between SA and ANZ infants for the gestation groups <32,
32-36 and ≥37 weeks were also determined. The frequency distri-
bution for gestational age at birth for ANZ and SA infants was de-
termined. The association between maternal ROB, and outcomes
displaying differences between groups, was then assessed using
linear or logistic regression, adjusting for the potential confounders
that were shown to be significantly different between ANZ and SA-
born mother groups.
 |
160       YOUNG et al.

3 |   R E S U LT S ANZ-born women. The distribution of gestational age at birth by

Between 2013 and 2017, there were 43 234 births at Monash
Health, of which 2933 (6.8%) were classified as SGA. Of these, 959
(32.7%) were to ANZ-born women (ANZ babies), and 1018 (34.7%)
to SA-born women (SA babies). The differences in maternal and
obstetric characteristics are shown in Table 1. SA-born women
were slightly older than ANZ-born women (30.2 vs 29.5 years) and
less likely to report smoking (3.2% vs 39.5%), alcohol consumption
(0.7% vs 4.6%) and other substance use during pregnancy (0.2% vs
11.4%). SA-born women experienced more thyroid disease (14.6%
vs 3.3%) and gestational diabetes (18.2% vs 6.6%) compared to
maternal ROB is shown in Figure 1. Overall, SA babies were born
at a later gestation than ANZ babies (median IQR 39 [38-40] vs
38 [37-40] weeks) and had a higher birthweight (median IQR 2590
[2310-2780] vs 2480 [2059-2740] grams). Twenty-five percent of
ANZ babies were born preterm (<32 or 32-36 weeks), compared to
12.9% of SA babies. SA-born women were also significantly less
likely to experience stillbirth (2.6% vs 4.3%) but not overall neona-
tal death (0.5% vs 0.9%).
Table 2 presents the differences in neonatal outcomes of SGA
infants stratified by gestational age at birth. There were signifi-
cantly more preterm births in the ANZ SGA cohort than SA SGA

TA B L E 1   Maternal and birth characteristics of the population

Australian/ New Zealand South Asian P

n = 959 n = 1018 value

Maternal age (years) 29.5 (6) 30.2 (4.6) .006

Chronic medical condition
Pre-existing hypertension 16 (1.7%) 10 (1%) .18
Pre-existing diabetes 14 (1.5%) 13 (1.3%) .73
Thyroid disease 32 (3.3%) 149 (14.6%) <.001
Obstetric complications
Pregnancy-induced hypertension/ 88 (9.2%) 79 (7.8%) .26
pre-eclampsia
Gestational diabetes 63 (6.6%) 185 (18.2%) <.001
Cholestasis of pregnancy 4 (0.4%) 10 (1%) .13
Smoking during pregnancy 379 (39.5%) 33 (3.2%) <.001
Substance use during pregnancy 109 (11.4%) 2 (0.2%) <.001
Alcohol use during pregnancy 44 (4.6%) 7 (0.7%) <.001
Baby birthweight 2480 (2059-2740) 2590 (2310-2780) <.001
Gestation (Median) 38 (37-40) 39 (38-40) <.001
Preterm <32 wk 67 (7%) 27 (2.7%) <.001
Preterm 32-36 wk 172 (17.9%) 104 (10.2%)
Term ≥37 wk 720 (75%) 887 (87.1%)
Baby gender—female 736 (77%) 788 (77.4%) .73
Onset of labour
Spontaneous 354 (36.9%) 365 (35.9%) .052
Induced 389 (40.6%) 461 (45.3%)
No Labour 216 (22.5%) 192 (18.9%)
Birth type
Vaginal 604 (63%) 648 (64%) 0.85
Elective caesarean 115 (12%) 126 (12.4%)
Unplanned caesarean 240 (25%) 244 (24%)
Stillborn 41 (4.3%) 26 (2.6%) 0.04
Neonatal death overall 9 (0.94%) 5 (0.5%) 0.24
Death with no resuscitation offered 5 (0.5%) 2 (0.2%) -
Death after admission 4 (0.4%) 3 (0.3%) -

Note: Data expressed as number (%), mean (SD) or median (IQR).

YOUNG et al.
F I G U R E 1   Distribution of small for gestational age babies born
within each gestational window, separated according to Australian/
New Zealand and South Asian maternal region of birth
cohort. SA babies born <32 weeks required less ventilation sup-
port, and a lower rate of intubation (14% vs 69%) and surfactant
therapy (14% vs 62.1%). Among SGA babies born between 32
and 36 weeks, there were no significant differences between the
groups; however, length of stay admitted to SCN/NICU was in fact,
slightly longer for SA babies (15 [9-25] vs 14 days [8-20]) in this
group. Conversely, in the ≥37-week subgroup, length of stay admit-
ted to SCN/NICU was increased for ANZ compared to SA babies
(5 [4-8.5] vs 5 days [3-7]). ANZ babies in this subgroup also had
notably higher rates of major congenital anomaly (5% vs 2.3%) and
culture-positive sepsis (0.8% vs 0%). Neonatal hypothermia was
shown in 257 (29%) of SA babies but only 149 (21%) of ANZ babies
in the ≥37-week subgroup.
The association between maternal ROB and key neonatal out-
comes is presented in Table 3. Overall, SA SGA babies were sig-
nificantly less likely to experience mortality, require resuscitation
or admission to the SCN/NICU, respiratory support for >4 hours,
surfactant therapy or gavage feeding, experience HMD or CLD,
hyperbilirubinaemia, sepsis or anaemia, or have a major congenital
anomaly. SA babies also had a significantly shorter duration of stay,
but were more likely to experience hypothermia. After adjustment
for the possible confounding maternal and birth variables, SGA ba-
bies of SA-born women were 1.5 times more likely to develop hy-
pothermia (95% CI: 1.16-1.88; P = .001) and 2.5 times more likely
to die (95% CI: 1.12-5.7, P = .02). These infants were also 60% less
likely to be born with a major congenital anomaly (95% CI: 0.24-0.67,
P = .001) and 36% less likely to require NGT feeding (95% CI: 0.43-
0.93, P = .02). Overall length of stay was also 2.33 days shorter for
SA babies (95% CI: −4.1 to −0.57), P = .01).
4 |  D I S CU S S I O N
This is a large retrospective cohort study examining the neona-
tal outcomes of small babies, born to two diverse maternal popu-
lations in an Australian setting. To our knowledge, this is the first
study to conduct such a detailed analysis of a wide range of neonatal
|
      161
outcomes, for SGA babies with different maternal ROBs. Our most
significant findings were that SGA babies of SA-born women were
1.5 times more likely to experience neonatal hypothermia, had a 2.5
times greater risk of perinatal mortality, but were 60% less likely to
have a congenital anomaly, and 36% less likely to require gavage
feeding when compared to SGA babies of ANZ-born women. Apart
from this, no significant differences were demonstrated for the ma-
jority of perinatal and neonatal outcomes.
As previously reported, SA maternal ethnicity has a known asso-
ciation with SGA births,4-9 which was consistent with our study. The
characteristics of the SA and ANZ populations were significantly
different, with much higher rates of perinatal smoking and sub-
stance use reported by ANZ-born women, and higher rates of thy-
roid disorders for SA-born women. These findings are in line with
12,17,18
previous studies and are all factors known to be associated
with SGA and IUGR.19,20 The finding that SA infants were 60% less
likely to have a congenital abnormality, malformation or syndrome
may indeed be associated with the high rates of maternal smoking,
alcohol and substance use during pregnancies of ANZ-born women.
However, this difference remained statistically significant even
after adjusting for maternal factors, suggesting that maternal ROB
may also have an impact. After adjustment for potential confound-
ers, the rate of perinatal mortality in SA SGA babies was also higher
than ANZ SGA babies, consistent with recent studies.4,7,9,12
South Asian ethnicity has been linked to shortened gestational
length, accelerated foetal maturation and subsequent more favour-
able outcomes for neonates of this background. 21 One study noted
a median gestational length of 39 weeks for South Asians compared
to 40 weeks in Europeans and a shorter length of hospital stay for
Asian infants.9,13 These ethnic differences may be a result of earlier
placental maturation at a biological or genetic level.5 A recent study
from our group explored this theory, comparing outcomes of moder-
ate-late preterm infants of SA and ANZ-born women. Lower rates of
respiratory morbidity were noted in SA babies, suggesting earlier lung
maturation in utero.18 Our study did not discover statistically signif-
icant differences in neonatal respiratory outcomes between SA and
ANZ infants, which may indicate that the SGA-specific population has
a different neonatal profile or there was lack of power for the out-
comes. Interestingly, SA infants were 1.5 times less likely to require
NGT feeding, which could be associated with early maturation.
Despite the association between SA maternal ROB and both pre-
maturity and shortened gestational length, only 12.9% of SA SGA in-
fants were born preterm (<32 or 32-36 weeks), compared to 25% of
ANZ SGA infants in our cohort. This is illustrated in Figure 1, where
a larger proportion of SA SGA infants were born closer to term, com-
pared to ANZ SGA infants. A potential explanation for this may be
the different antenatal risk factor profiles of ANZ-born women re-
sulting in a higher number of preterm births. Conversely, SA-born
women may have more placental insufficiency near term or after
term age, leading to more numbers in that gestation subgroup. These
findings may indeed align with our hypothesis, and reflect a largely
physiologically or constitutionally small SA population, compared to
a pathologically small ANZ population.
TA B L E 2   Neonatal outcomes excluding stillbirths and neonatal death with no resuscitation offered
|

<32 wk 32-36 wk ≥37 wk

162      

ANZ SA ANZ SA ANZ SA

n = 29 n = 7 n = 166 n = 102 n = 718 n = 881

Resuscitation
Apgar at 5 min 8 (7-9) 9 (8-9) 9 (9-9) 9 (9-9) 9 (9-9) 9 (9-9)
Any resuscitation 29 (100%) 7 (100%) 94 (56.6%) 52 (51%) 194 (27%) 215 (24.5%)
Intubation 13 (44.8%) 1 (14.3%) 1 (0.6%) 0 (0%) 8 (1.1%) 9 (1%)
Continuous positive airway pressure 22 (76%) 7 (100%) 73 (44%) 35 (34%) 100 (14%) 102 (11.6%)
Oxygen 28 (97%) 7 (100%) 56 (34%) 28 (27.5%) 73 (10.2%) 72 (8.2%)
Intermittent positive pressure 26 (90%) 2 (29%) 55 (33.1%) 35 (34%) 82 (11.4%) 78 (8.9%)
ventilation
Suction 22 (76%) 3 (43%) 67 (40.4%) 38 (37.3%) 167 (23%) 188 (21.4)
Admitted to NICU/SCN 29 (100%) 7 (100%) 154 (93%) 93 (91.2%) 196 (27.3%) 233 (26.5%)
Respiratory outcomes
Length of respiratory support 61 (36-112) 17 (2-70) 0 (0-0) 0 (0-0) 0 (0-0) 0 (0-0)
Length of respiratory support if 61 (36-112) 17 (2-70) 0 (0-0) 0 (0-0) 0 (0-0) 0 (0-0)
admitted to NICU
Respiratory support >4 h 28 (97%) 6 (86%) 30 (18%) 17 (16.7%) 30 (4.2%) 24 (2.7%)
intubation 20 (69%) 1 (14%) 3 (1.8%) 2 (2%) 11 (1.5%) 10 (1.14%)
Hyaline membrane disease 28 (97%) 6 (86%) 7 (4.2%) 4 (3.9%) 0 (0%) 0 (0%)
Transient tachypnoea of newborn 0 (0%) 0 (0%) 20 (12%) 10 (9.8%) 12 (1.7%) 8 (0.9%)
Meconium aspiration 0 (0%) 0 (0%) 0 (0%) 0 (0%) 8 (1.1%) 6 (0.7%)
Chronic lung disease 15 (52%) 2 (29%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)
Discharged home on oxygen 8 (27.6%) 0 (0%) 0 (0%) 0 (0%) 3 (0.4%) 0 (0%)
Caffeine therapy 28 (97%) 6 (86%) 23 (14%) 7 (6.9%) 0 (0%) 0 (0%)
Surfactant therapy 18 (62.1%) 1 (14%) 1 (0.6%) 0 (0%) 1 (0.14%) 0 (0%)
Steroid therapy 3 (10.3%) 0 (0%) 0 (0%) 0 (0%) 2 (0.3%) 1 (0.1%)
Other neonatal outcomes
Hypothermia 18 (62%) 4 (57%) 88 (53%) 58 (57%) 149 (21%) 257 (29%)
Hypoglycaemia 21 (72.4%) 4 (57%) 102 (61.5%) 60 (59%) 216 (30%) 264 (30%)
Hypoglycaemia requiring IV therapy 20 (69%) 4 (57%) 69 (42%) 43 (42%) 38 (5.3%) 55 (6.2%)
NGT feeding 28 (96%) 7 (100%) 117 (70.5) 66 (64.7%) 58 (8%) 48 (5.5%)
Hyperbilirubinaemia 28 (96%) 7 (100%) 90 (54%) 47 (46%) 65 (9%) 78 (8.9%)
Sepsis (culture positive) 12 (41%) 2 (29%) 2 (1.2%) 1 (1%) 6 (0.8%) 0 (0%)
YOUNG et al.

(Continues)
 YOUNG et al.

TA B L E 2   (Continued)

<32 wk 32-36 wk ≥37 wk

ANZ SA ANZ SA ANZ SA

n = 29 n = 7 n = 166 n = 102 n = 718 n = 881

Anaemia requiring treatment 22 (76%) 4 (50%) 9 (5.5%) 3 (3%) 4 (0.6%) 2 (0.3%)

Thrombocytopaenia requiring 10 (34.5%) 2 (29%) 0 (0%) 0 (0%) 2 (0.3%) 4 (0.5%)
platelets
Necrotising enterocolitis 6 (21%) 0 (0%) 1 (0.6%) 0 (0%) 1 (0.1%) 0 (0%)
Retinopathy of prematurity 3 (10.3%) 0 (0%) 0 (0%) 0 (0)% 0 (0%) 0 (0%)
Osteopaenia of prematurity 14 (48%) 3 (43%) 3 (1.8%) 0 (0%) 0 (0%) 0 (0%)
Patent ductus arteriosus requiring 6 (21%) 0 (0%) 0 (0%) 0 (0%) 1 (0.1%) 1 (0.1%)
treatment
Brain injury 6 (20.7%) 3 (43%) 5 (3%) 2 (2%) 0 (0%) 2 (0.2%)
Hyponatraemia requiring treatment 17 (58.6%) 5 (71.4%) 4 (2.4%) 0 (0%) 3 (0.4%) 2 (0.2%)
Hyperglycaemia 6 (20.7%) 1 (14.3%) 1 (0.6%) 0 (0%) 1 (0.1%) 0 (0%)
Hypokalaemia 3 (10.3%) 1 (14.3%) 1 (0.6%) 0 (0%) 1 (0.1%) 3 (0.3%)
Perinatal asphyxia 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 (0.4%) 3 (0.3%)
Major congenital anomaly 6 (22%) 1 (14%) 10 (6%) 6 (6%) 39 (5%) 20 (2.3%)
Length of stay Admitted to NICU 72 (60-125) 48 (31-83) 14 (8-20) 15 (9-25) 5 (4-8.5) 5 (3-7)
Length of stay overall 72 (60-125) 48 (31-83) 12 (6-19) 14 (8-23) 3 (2-5) 3 (2-4)

Note: Data expressed as number (%) or median (IQR).

|
      163
 |
164       YOUNG et al.

TA B L E 3   Association between maternal region of birth and neonatal outcomes

Crude odds ratio/Regression coefficient P Adjusted odds ratio/Regression coefficient P

(95% CI) value (95% CI) value

Perinatal mortality 0.57 (0.36-0.90) .01 2.5 (1.12-5.7) .02

a
Overall length of stay   −4.1 (−5.7 to −2.4) <.001 −2.33 (−4.1 to −0.57) .01
Need for any resuscitation 0.72 (0.59-0.87) <.001 0.81 (0.63-1.02) .07
Admitted to SCN/NICU 0.71 (0.59-0.86) <.001 0.88 (0.97-1.01) .34
Hypothermia 1.2 (1.01-1.5) .03 1.48 (1.16-1.88) .001
Hypoglycaemia 0.84 (0.70-1.02) .08 0.96 (0.76-1.21) .72
Respiratory support >4 h 0.48 (0.33-0.69) <.001 0.69 (0.43-1.10) .12
Chronic lung disease 0.21 (0.07-0.64) .006 1.1 (0.19-6.24) .92
Home on oxygen 0.12 (0.03-0.53) .005 0.19 (0.04-1.02) .053
Hyaline membrane disease 0.26 (0.13-0.52) <.001 0.80 (0.22-2.91) .74
Surfactant therapy 0.09 (0.02-0.37) .001 0.26 (0.03-2.0) .20
NGT feeding 0.49 (0.39-0.63) <.001 0.64 (0.43-0.93) .02
Hyperbilirubinaemia 0.63 (0.49-0.80) <.001 0.97 (0.69-1.37) .87
Sepsis (culture positive) 0.20 (0.07-0.52) .001 0.51 (0.16-1.63) .26
Anaemia requiring treatment 0.27 (0.13-0.52) <.001 0.63 (0.25-1.63) .35
Major congenital anomaly 0.44 (0.27-0.69) .001 0.40 (0.24-0.67) .001

Note: Odds of each respective outcome for South Asian born women compared to Aus/NZ born women.
Adjusted for maternal age, gestation, birthweight, maternal thyroid disease, smoking, alcohol, substance and gestational diabetes.
a
Regression coefficient.

Our results showed that SA SGA babies were 1.5 times more
likely to develop neonatal hypothermia within the first 24 hours of
life. Low birthweight infants have a limited capacity for thermoregu-
lation during the initial neonatal period, where maintaining warmth is
22
critical to reduce morbidity and mortality. On exposure to the cold
extra-uterine environment, lipolysis of brown adipose tissue is acti-
vated to generate heat in the neonate. 22,23 We postulate that the in-
creased risk of hypothermia in SA babies may be explained by ethnic
differences in body composition and fat stores. Interestingly, multiple
studies in children and adolescents have shown ethnic variation in
body composition, 24 but that SA adolescents and adults have signifi-
cantly more body fat than white Europeans, despite having a lower
25,26
BMI. This highlights the importance of further research into vari-
ation in body composition specifically in the neonatal population.
There are a number of limitations to this study. Firstly, using
maternal self-reported country of birth as a proxy for ethnicity is
a limitation of perinatal data sets internationally, where ethnicity
is not explicitly recorded, and does not take into account pater-
nal factors. Furthermore, SA ethnicity itself is inherently diverse,
where women born in Afghanistan or Pakistan are biologically dif-
ferent to women born in Sri Lanka or India, for example. Future
studies, adequately powered to consider these differences will be
of interest. For SA-born women, it is also important to consider the
duration of residence in Australia, as migrant status itself is a risk
27
factor for adverse perinatal outcomes, which was not recorded
in our data. In addition to this, environmental factors including so-
cio-economic status and education level of families in both groups
could also influence foetal maturation and neonatal morbidity.
As our study included a large population, the contribution of so-
cio-economic status between groups is relatively controlled be-
tween the groups. Our study had low numbers in the very preterm
SGA group, particularly for SA-born women. This limits the abil-
ity to analyse and generalise the observed associations for this
subgroup. Finally, the hospital database uses national population
centile charts to classify and diagnose SGA. When using single
standard growth charts, SA ethnicity has higher rates of SGA, with
lower perinatal mortality risks, 28 indicating that many SA babies
are categorised as high risk when they are actually constitution-
ally small. This supports our hypothesis and explanation for more
favourable neonatal outcomes in the SA population. However, our
study also recognised that for most neonatal outcomes, SA and
ANZ babies behaved similar, and therefore SA SGA babies cannot
be presumed as constitutional or physiologically small. The as-
sumption of SGA being physiological due to maternal characteris-
tics (including ROB or ethnicity) could lead to the misidentification
of pathological, growth-restricted infants.
5 | CO N C LU S I O N S
Babies born SGA to SA-born women have a different neonatal out-
come profile as compared to babies born to ANZ-born women. ANZ
SGA infants experience more congenital abnormalities, gavage tube
requirements and a longer duration of hospital stay, whereas SA SGA
infants have a higher risk of neonatal hypothermia and perinatal
mortality. Apart from this, no significant differences in the rates of
YOUNG et al.
other neonatal outcomes were observed. These findings may be a
reflection of physiological and pathological causes of SGA, relating
to maternal ROB, and other maternal demographics. These discover-
ies also provide scope for further research into the influence of eth-
nicity on foetal and placental maturation, organogenesis and body
composition.
AC K N OW L E D G E M E N T S
The authors would like to thank Amanda Kendell and Michelle Knight
for help with data extraction from the hospital database.
C O N FL I C T O F I N T E R E S T
MDT has a secondment 1 day per week to Consultative Council on
Obstetric and Paediatric Mortality and Morbidity (CCOPMM). There
are no other conflicts to declare.
E T H I C A L A P P R OVA L
The study obtained ethics approval as a quality assurance project
from the Monash Health Human Research Ethics committee. The
collection of data was in accordance with ethical standards, and no
formal consent from participants was required.
ORCID
Atul Malhotra  https://orcid.org/0000-0001-9664-4182
T WITTER
Atul Malhotra  @Atul_Monash
REFERENCES
1. Battaglia FC, Lubchenco LO. A practical classification of newborn
infants by weight and gestational age. J Pediatr. 1967;71(2):159-163.
2. Ananth CV, Vintzileos AM. Distinguishing pathological from consti-
tutional small for gestational age births in population-based studies.
Early Human Dev. 2009;85(10):653-658.
3. Ewing AC, Ellington SR, Shapiro-Mendoza CK, Barfield WD,
Kourtis AP. Full-term small-for-gestational-age newborns in the
U.S.: characteristics, trends, and morbidity. Matern Child Health J.
2017;21(4):786-796.
4. Balchin I, Whittaker JC, Patel RR, Lamont RF, Steer PJ. Racial varia-
tion in the association between gestational age and perinatal mor-
tality: prospective study. BMJ. 2007;334(7598):833.
5. Ravelli AC, Tromp M, Eskes M, et al. Ethnic differences in still-
birth and early neonatal mortality in The Netherlands. J Epidemiol
Commun Health. 2011;65(8):696-701.
6. Ravelli Anita CJ, Schaaf Jelle M, Eskes M, Abu-Hanna A, de Miranda
E, Mol BWJ. Ethnic disparities in perinatal mortality at 40 and 41
weeks of gestation. J Perinatal Med. 2013;41:381.
7. Drysdale H, Ranasinha S, Kendall A, Knight M, Wallace EM.
Ethnicity and the risk of late-pregnancy stillbirth. Med J Australia.
2012;197(5):278-281.
8. Khalil A, Rezende J, Akolekar R, Syngelaki A, Nicolaides KH.
Maternal racial origin and adverse pregnancy outcome: a cohort
study. Ultrasound Obstet Gynecol. 2013;41(3):278-285.
9. Balchin I, Steer PJ. Race, prematurity and immaturity. Early Human
Dev. 2007;83(12):749-754.
10. Australian Institute of Health Welfare. Australia's Mothers and
Babies Data Visualisations. Canberra: AIHW; 2019.
11. Anderson NH, Sadler LC, Stewart AW, McCowan LM. Maternal and
pathological pregnancy characteristics in customised birthweight
|
      165
centiles and identification of at-risk small-for-gestational-age in-
fants: a retrospective cohort study. BJOG. 2012;119(7):848-856.
12. Davies-Tuck ML, Davey MA, Wallace EM. Maternal region of
birth and stillbirth in Victoria, Australia 2000–2011: a retro-
spective cohort study of Victorian perinatal data. PLoS One.
2017;12(6):e0178727.
13. Patel RR, Steer P, Doyle P, Little MP, Elliott P. Does gestation
vary by ethnic group? A London-based study of over 122,000
pregnancies with spontaneous onset of labour. Int J Epidemiol.
2004;33(1):107-113.
14. Schaaf JM, Mol BW, Abu-Hanna A, Ravelli AC. Ethnic disparities in
the risk of adverse neonatal outcome after spontaneous preterm
birth. Acta Obstet Gynecolo Scand. 2012;91(12):1402-1408.
15. Dobbins TA, Sullivan EA, Roberts CL, Simpson JM. Australian na-
tional birthweight percentiles by sex and gestational age, 1998–
2007. Med J Aust. 2012;197(5):291-294.
16. United Nations Statistics Division. Standard country or area codes
for statistical use (M49); 2019.
17. Nambiar V, Jagtap VS, Sarathi V, et al. Prevalence and impact of
thyroid disorders on maternal outcome in Asian-Indian pregnant
women. J Thyroid Res. 2011;2011:429097.
18. Cox AG, Narula S, Malhotra A, Fernando S, Wallace E, Davies-Tuck
M. The influence of maternal ethnicity on neonatal respiratory out-
come. Arch Dis Child Fetal Neonatal Ed. 2020;105:50-55.
19. Sahu MT, Das V, Mittal S, Agarwal A, Sahu M. Overt and sub-
clinical thyroid dysfunction among Indian pregnant women and
its effect on maternal and fetal outcome. Arch Gynecol Obstet.
2009;281(2):215.
20. Ko T-J, Tsai L-Y, Chu L-C, et al. Parental smoking during pregnancy
and its association with low birth weight, small for gestational age,
and preterm birth offspring: a birth cohort study. Pediatr Neonatol.
2014;55(1):20-27.
21. Ruan S, Abdel-Latif ME, Bajuk B, Lui K, Oei JL. The associations
between ethnicity and outcomes of infants in neonatal intensive
care units. Arch Dis Childhood Fetal Neonat Ed. 2012;97(2):F133
-F138.
22. Lunze K, Hamer DH. Thermal protection of the newborn in re-
source-limited environments. J Perinatol. 2012;32:317.
23. Dawkins MJR, Scopes JW. Non-shivering thermogenesis and
brown adipose tissue in the human new-born infant. Nature.
1965;206(4980):201-202.
24. Haroun D, Taylor SJC, Viner RM, et al. Validation of bioelectrical
impedance analysis in adolescents across different ethnic groups.
Obesity. 2010;18(6):1252-1259.
25. Ehtisham S, Crabtree N, Clark P, Shaw N, Barrett T. Ethnic differ-
ences in insulin resistance and body composition in United Kingdom
adolescents. J Clin Endocrinol Metab. 2005;90(7):3963-3969.
26. Nightingale CM, Rudnicka AR, Owen CG, Cook DG, Whincup PH.
Patterns of body size and adiposity among UK children of South
Asian, black African-Caribbean and white European origin: Child
Heart And health Study in England (CHASE Study). Int J Epidemiol.
2010;40(1):33-44.
27. Cacciani L, Asole S, Polo A, et al. Perinatal outcomes among immi-
grant mothers over two periods in a region of central Italy. BMC
Public Health. 2011;11(1):294.
28. Kierans WJ, Joseph K, Luo Z-C, Platt R, Wilkins R, Kramer MS.
Does one size fit all? The case for ethnic-specific standards of fetal
growth. BMC Pregnancy and Childbirth. 2008;8(1):1.
How to cite this article: Young NE, Davies-Tuck M, Malhotra
A. Influence of maternal region of birth on neonatal
outcomes of babies born small. Acta Paediatr. 2021;110:158–
165. https://doi.org/10.1111/apa.15375
DOI: 10.1111/apa.15427
BRIEF REPORT
Bronchopulmonary dysplasia outcome estimator in current
neonatal practice
Predicting an infant’s risk of bronchopulmonary dysplasia (BPD) is
challenging, particularly during the first weeks after birth. However,
strategies to prevent BPD may have most benefit if delivered during
this period. A tool assessing BPD risk would prove valuable to clini-
cians and researchers—aiding clinical decisions and developing eligi-
bility criteria to evaluate preventative BPD therapies.
The NICHD ‘BPD Outcome Estimator’ (BPD estimator) uses clin-
ical and demographic data to estimate, on postnatal days 1, 3, 7, 14,
21 and 28, an infant’s risk of mild, moderate and severe BPD, and the
competing outcome of death before 36 weeks’ postmenstrual age
(PMA).1 NICHD 2001 consensus definition2 and physiological chal-
3
lenge defined BPD. The validity of the BPD estimator outside the
United States has not been evaluated.
This retrospective cohort study assessed the BPD estima-
tor’s accuracy in predicting risk of severe BPD (sBPD) or death in
infants admitted to two perinatal centres in Victoria, Australia.
Infants admitted to Royal Women’s Hospital or Monash Children’s
Hospital born ≥23 weeks’ and ≤28 weeks’ gestational age (GA), be-
tween October 2016 and September 2017, birthweight >500 g and
<1251 g were included.
Infant characteristics (GA, birthweight, sex, race, postnatal day,
respiratory support mode and fraction of inspired oxygen (FiO2))
were entered into the BPD estimator to generate risk estimates for
death and sBPD on postnatal days 1, 3, 7, 14, 21 and 28. The bi-
nary outcome of sBPD or death before 36 weeks’ PMA, chosen for
its importance in both clinical practice and research, was generated
based on infants’ outcomes. Receiver operator characteristics (ROC)
curves evaluated the BPD estimator’s accuracy at distinguishing in-
fants with this outcome from those without. To compare estimated
risk with observed incidence of sBPD or death, infants were strat-
ified into estimated risk groups; <10%, 10%-19%, 20%-29%, 30%-
39%, 40%-49% and ≥50%.
On the day of interest, FiO2 was defined as the mean FiO2
of four time points and respiratory support as the ‘highest’ level
received on that day. Nasal high flow therapy (nHF) and non-in-
vasive positive pressure ventilation were classified as continuous
positive airway pressure. NICHD BPD definitions2 were applied
at 36 weeks’ PMA. Infants receiving nHF at 36 weeks’ PMA were
classified as having sBPD only if they required an FiO2 >0.21. If
A subset of this data was presented in poster format at the 2019 Annual Congress of the
Perinatal Society of Australia and New Zealand.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
|
166     
wileyonlinelibrary.com/journal/apa
infants' ethnicity was unknown or not available on the BPD esti-
mator (options ‘White’, ‘Black’ or ‘Hispanic’), they were assigned
‘White’.
During the study period, 215 eligible infants were admitted.
Twenty-six infants with birthweight <501 g or >1250 g and two in-
fants with unknown BPD outcomes were excluded; 187 infants were
included. Their mean (SD) GA was 26.6 (1.5) weeks’ and mean (SD)
birthweight 872  g (178). Fifty-two per cent of infants were male,
91% exposed to antenatal corticosteroids, 73% received surfactant,
and 31% received postnatal corticosteroids for BPD treatment.
The incidence of death before 36 weeks’ PMA or sBPD was
48.1%. Eighteen infants (9.6%) died before 36 weeks’ PMA, and 72
infants (38.5%) were diagnosed with sBPD.
In this population, the BPD estimator accurately distinguishes
infants who die before 36 weeks’ PMA or have sBPD from those
who do not, as evidenced by AUC of 0.81-0.84 at all postnatal time
points (Figure S1). The accuracy reported here is comparable to that
reported in the population used to model the estimator.1
However, predicted risk underestimates the observed incidence
of sBPD or death. The observed incidence of sBPD or death was
higher than the estimated risk in all risk strata except estimated risk
group <10% (Table 1). This underestimation may reflect practice
changes and population differences.
The BPD estimator, modelled on infants born between 2000 and
2004, uses mode of respiratory support to predict outcome. The use
of non-invasive respiratory support has evolved. Modes available in
the estimator no longer reflect practice and their influence on pre-
dicting outcome may have changed. Variation in the use of invasive
support modes, such as high frequency oscillatory ventilation, may
also erroneously influence risk estimates.
The use of postnatal corticosteroids has changed. Use declined
in very low birthweight infants from 20% in 1997-2000 to 8% in
2004.4 The BPD estimator was modelled on infants born during this
period. Thirty-one per cent of infants in our cohort received postna-
tal corticosteroids. The BPD estimator does not use postnatal corti-
costeroid exposure to estimate risk, though differences in use may
modify BPD evolution.
Multiple factors may contribute to the high incidence of sBPD or
death (48.1%) observed in our population. Intensive care is provided
to infants born at 23  weeks’ GA in centres included in this study.
Of the 10 infants born at 23 weeks’ GA included in this cohort, five
Acta Paediatrica. 2021;110:166–167.
BRIEF REPORT |
      167

TA B L E 1   Observed incidence of severe BPD (sBPD) or death before 36 wks’ postmenstrual age (PMA) stratified by estimated risk group.
Data displayed as the percentage of infants within the risk group who had severe BPD or died before 36 wks’ PMA

NICHD estimated risk of sBPD or death 36 before weeks’ PMA

<10% 10%-<20% 20%-<30% 30%-<40% 40%-<50% ≥50%

Day 1 7% (3/45) 45% (22/49) 54% (21/39) 80% (16/20) 69% (9/13) 93% (14/15)
Day 3 5% (2/43) 32% (14/44) 69% (25/36) 64% (16/25) 71% (12/17) 100% (14/14)
Day 7 12% (5/41) 27% (14/51) 55% (17/31) 78% (18/23) 73% (11/15) 100% (13/13)
Day 14 12% (3/25) 26% (19/72) 44% (12/27) 80% (16/20) 79% (11/14) 100% (16/16)
Day 21 7% (2/30) 32% (24/76) 62% (8/13) 74% (17/23) 87% (20/23) 100% (2/2)
Day 28 5% (1/21) 27% (21/77) 63% (17/27) 73% (16/22) 77% (10/13) 100% (6/6)

died prior to 36 weeks’ PMA and the five survivors had sBPD. The
lack of a physiological test of BPD severity and nHF use at 36 weeks’
PMA may have increased the diagnosis of sBPD. Sixty-two infants
received nHF at 36 weeks’ PMA; how these infants’ BPD is best cat-
egorised is currently unknown. Finally, clinical practices in our nurs-
eries may increase BPD rates.
The BPD estimator has limited ethnicity options. Those available,
White, Black or Hispanic, do not reflect our community with a sub-
stantial Asian population.
In the original population, equal numbers of infants died before
36 weeks’ PMA or had sBPD (13% for each outcome). In our popu-
lation, sBPD incidence was four-fold higher than that of death. This
differential may reflect recent increases in preterm infant survival
rates and an associated BPD increase.5
Infants ≤28  weeks’ GA were included in our cohort, whereas
the estimator was based on a population of infants <30 weeks’ GA.
Whilst limiting comparison at population level, this difference should
not influence individuals’ estimated risk.
Despite limitations, the BPD estimator distinguishes infants with
sBPD/death from those without in our population. However, the
BPD estimator needs recalibrating for local populations.
C O N FL I C T O F I N T E R E S T
The authors have no conflicts of interest to declare.
F U N D I N G I N F O R M AT I O N
National Health and Medical Research Council, Australia (Practitioner
Fellowship No. 1157782 to PGD).
Elizabeth K. Baker1,2,3
1,2
Peter G. Davis
1
Newborn Research, The Royal Women’s Hospital, Melbourne,
Vic., Australia
2
Department of Obstetrics and Gynaecology, The University of
Melbourne, Melbourne, Vic., Australia
3
Monash Newborn, Monash Children’s Hospital, Melbourne,
Vic., Australia
Correspondence
Elizabeth K. Baker, Newborn Research, The Royal Women’s
Hospital, 20 Flemington Road, Parkville, Vic. 3052, Australia.
Email: elizabeth.baker2@thewomens.org.au
ORCID
Elizabeth K. Baker  https://orcid.org/0000-0001-8523-951X
REFERENCES
1. Laughon MM, Langer JC, Bose CL, et al. Prediction of bronchopulmo-
nary dysplasia by postnatal age in extremely premature infants. Am J
Respir Crit Care Med. 2011;183(12):1715-1722.
2. Jobe A, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit
Care Med. 2001;163:1723-1729.
3. Walsh MC, Wilson-Costello D, Zadell A, Newman N, Fanaroff A.
Safety, reliability, and validity of a physiologic definition of broncho-
pulmonary dysplasia. J Perinatol. 2003;23(6):451-456.
4. Fanaroff AA, Stoll BJ, Wright LL, et al. Trends in neonatal morbidity
and mortality for very low birthweight infants. Am J Obstet Gynecol.
2007;196(2):147.e1-147.e8.
5. Stoll BJ, Hansen NI, Bell EF, et al. Trends in care practices, morbid-
ity, and mortality of extremely preterm neonates, 1993–2012. JAMA.
2015;314(10):1039-1051.
S U P P O R T I N G I N FO R M AT I O N
Additional supporting information may be found online in the
Supporting Information section.
 | |
Received: 13 April 2020    Revised: 20 July 2020    Accepted: 28 September 2020

DOI: 10.1111/apa.15604

BRIEF REPORT

Post-haemorrhagic hydrocephalus management: Delayed

neonatal transport negatively affects outcome

Alessandro Parodi1  | Ilaria Giordano1 | Laura De Angelis1  | Mariya Malova1 |

Maria Grazia Calevo2 | Deborah Preiti1 | Marcello Ravegnani3 | Armando Cama3 |
Carlo Bellini1  | Luca A. Ramenghi1,4

1
Neonatal Intensive Care Unit, IRCCS
Istituto Giannina Gaslini, Genoa, Italy
2
Epidemiology and Biostatistics Unit, IRCCS
Istituto Giannina Gaslini, Genoa, Italy
3
Paediatric Neurosurgery Unit, IRCCS
Istituto Giannina Gaslini, Genoa, Italy
4
Department of Neurosciences,
Rehabilitation, Ophthalmology, Genetics,
Maternal and Child Health (DINOGMI),
University of Genoa, Genoa, Italy
Correspondence
Alessandro Parodi, Neonatal Intensive Care
Unit, IRCCS Istituto Giannina Gaslini, Via
Gaslini 5, 16148 Genoa, Italy.
Email: alessandroparodimail@gmail.com
Abstract
Background: Post-haemorrhagic ventricular dilatation (PHVD) still represents an
important cause of brain injury in premature infants. Intervention for PHVD is rec-
ommended once Ventricular Index (VI) crosses the 97th percentile + 4 mm line ac-
cording to Levene.
Objectives: We aimed to compare PHVD severity, timing of intervention, and out-
come between outborn infants transferred to a level IV NICU in order to be treated
for PHVD and a control population of inborn infants.
Methods: Preterm infants with PHVD requiring treatment were divided into: out-
born infants (transferred to our NICU in order to be treated for PHVD) and inborn
infants (PHVD diagnosed at our NICU). Age at intervention, difference between
VI and the 97th percentile according to postmenstrual age (VI-p97), permanent
shunt rate, and developmental delay rate were compared between the two groups.
Neurodevelopmental outcome was assessed using the Vineland Adaptive Behavior
Scales II (VABS-II), a parental questionnaire investigating four domains of adaptive
behaviour and overall adaptive functioning. Developmental delay was defined as a
score <70 (−2 SD or less).
Results: Twelve outborn and 15 inborn infants were included. Age at intervention
(31.6 vs 17.4 days) and VI-p97 (left 10.0 vs 5.1 mm, right 7.7 vs 5.1 mm) were signifi-
cantly higher among outborn infants. A permanent shunt was inserted in 66.7% of
outborn and in 40.0% of inborn infants (p = 0.18). After excluding subjects with pa-
renchymal lesions, a significantly higher rate of developmental delay was observed at
5 years in outborn patients compared to inborn patients (66.7% of outborn vs 18.2%
of inborn patients with VABS-II composite score <70, p = 0.04).
Conclusions: Outborn infants reached a significantly more severe ventricular dila-
tation than inborn infants, largely exceeding the recommended cutoff for inter-
vention. Our follow-up data showed a trend towards a higher rate of permanent
shunt and developmental delay in outborn than in inborn patients. Infants requiring

Abbreviations: PHVD, post-haemorrhagic ventricular dilatation; VI, ventricular index; p97, 97th percentile; NICU, neonatal intensive care unit; VABS-II, Vineland Adaptive Behavior
Scales II.

©2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

|
168     
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:168–170.
PARODI et al.
treatment should be timely transported to centres with adequate expertise in PHVD
management.
KEYWORDS
preterm, posthemorrhagic hydrocephalus, neonatal transport
1 |  I NTRO D U C TI O N
Post-haemorrhagic ventricular dilatation (PHVD) remains a major
cause of brain injury in preterm infants. Cerebrospinal fluid
drainage is often needed to decrease dilatation and prevent fur-
ther damage.1 Like most European neonatal intensive care units
(NICUs), we treat infants once the ventricular index (VI) has
crossed the 97th percentile (p97) and 4mm line, 2,3 as recently
recommended.1,4
We had noticed that neonates transferred to our level-four NICU
had extremely severe PHVD frequently surpassing the treatment
threshold. Our aim was to assess the degree of ventricular dilatation,
the timing of surgery and neurodevelopmental outcome in inborn
and outborn patients. The hypothesis was that infants were being
referred to our centre later than recommended.
The clinical database at the Gaslini Institute of Genoa, Italy, was
retrospectively searched for all preterm infants treated for postna-
tal-onset PHVD from January 2012 to December 2018. Our NICU
has a longstanding practice to treat PHVD once the VI has crossed
3
the p97 + 4 mm line according to Levene by external ventricular
drain (EVD) placement. This temporary drain progressively reduces
ventricular dilatation by continuously draining 10–15 ml/kg of cere-
brospinal fluid daily, according to ventricular size changes and aim-
ing for VI < p97 over the next two weeks. Once the cerebrospinal
fluid protein concentration is below 1 g/L, EVD is removed and ven-
tricular size is closely monitored to identify patients with recurrent
progressive dilatation who require a permanent ventriculoperitoneal
shunt.
Study patients were divided into outborn infants transferred to
the unit for PHVD treatment and inborn infants admitted before
PHVD onset. Outborn infants who had already received PHVD
treatment were excluded.
The clinical data collected included the admission age of outborn
patients, the age when EVD placement was decided and provided
and any need for a ventriculoperitoneal shunt. The ultrasound data
included the prospectively measured VI in each lateral ventricle on
the day that we decided to treat inborn and outborn infants. VI was
retrospectively plotted on Levene's graph to calculate the difference
between the VI and the p97 (VI-p97), according to postmenstrual
3
age.
Neurodevelopmental outcome was assessed using the Vineland
Adaptive Behavior Scales II (VABS-II), a parental questionnaire in-
vestigating four domains of adaptive behaviour (communication,
daily living skills, socialisation and motor skills) and overall adaptive
functioning. Developmental delay was defined as a score <70 (−2
SD or less).
|
      169
Demographic, clinical and ultrasound data were compared be-
tween the two groups. Measurements of non-dilated lateral ventri-
cles were excluded if there was unilateral hydrocephalus.
Parameters were compared using the Mann-Whitney test for
continuous variables and the chi-square or Fisher's exact test for
categorical variables. The statistical analysis was performed using
SPSS for Windows, version 20 (IMB Corp).
Of the 27 preterm infants treated for postnatal-onset PHVD, 12
were outborn and 15 were inborn. Decisions about EVD treatment
were made the same day that outborn infants were admitted. They
were significantly older than inborn infants at the time of the deci-
sion and actual surgery (Table 1).
We included 47/54 lateral ventricle measurements for 21 out-
born and 26 inborn patients: six measurements were missing due
to porencephaly involving the frontal horn, following periventricular
haemorrhagic infarction, and we excluded a non-dilated lateral ven-
tricle in one patient with unilateral PHVD. The VI-p97 values were
significantly higher in the outborn group (Table 1).
A ventriculoperitoneal shunt was inserted in 66.7% of outborn
and 40.0% of inborn patients (P = .18). A higher rate of developmen-
tal delay in all domains was observed in outborn patients compared
to inborn patients, although statistically not significant (Table 1). A
similar trend was observed after excluding subjects with parenchy-
mal lesions (3 patients with infarction in each group and 1 inborn
patient with cystic periventricular leukomalacia), reaching statisti-
cal significance for overall adaptive functioning (66.7% of outborn
vs 18.2% of inborn patients with VABS-II composite score  <  70,
P = .04).
Infants transferred from other centres with progressive PHVD
were treated later and reached more severe ventricular dilatation
than inborn infants. Their ventricular measurements were sig-
nificantly higher before surgery and most widely exceeded the
p97  +  4  mm cut-off commonly used in Europe. 2 Interestingly, our
follow-up data showed a trend towards a higher rate of permanent
shunt and developmental delay in outborn than in inborn patients,
although the sample was too small to reach statistical significance
for most outcomes. These results agree with data showing impaired
neurodevelopmental outcomes for EVD treatment after 25 days of
life compared to earlier treatment.5 Furthermore, the p97 + 4 mm
treatment threshold is supported by evidence suggesting that early
PHVD management (ie based on ventricular measurements) results
in improved neurodevelopmental outcome and in reduced shunt rate
when compared to late treatment (ie based on clinical signs of in-
creased intracranial pressure).1,4
Delays in transferring infants to our centre may have been due to
their poor clinical conditions related to extreme prematurity and local
 |
170       PARODI et al.

TA B L E 1   Comparison of demographic
Inborn Outborn
characteristics, clinical characteristics,
Means and standard deviations (±) unless patients patients
neurodevelopmental outcome and
otherwise stated (n = 15) (n = 12) P value
ventricular measurements between
Birth weight (grams) 1157 ± 402 1359 ± 575 0.37 inborn and outborn groups (EVD: external
Gestational age (weeks) 28.4 ± 2.1 28.2 ± 3.4 0.87 ventricular drain; VABS-II: Vineland
Adaptive Behavior Scales II; VI-p97:
Male sex, n (%) 8 (53.3) 8 (66.7) 0.70
difference between the Ventricular Index
Admissions age (days) - 30.3 ± 20.5 - and the 97th percentile according to
Age at decision about EVD surgery (days) 16.1 ± 5.4 30.3 ± 20.5 0.04 postmenstrual age)
Age at EVD surgery (days) 17.4 ± 5.4 31.6 ± 20.2 0.03
Days between decision for surgery and 1.3 ± 0.8 1.3 ± 0.9 1
EVD surgery
Postmenstrual age at EVD surgery (weeks) 30.9 ± 2.0 32.8 ± 3.7 0.07
Age at VABS-II administration (years) 5.9 ± 1.7 5.2 ± 1.8 0.30
Communication score < 70, n (%) 5 (33.3) 7 (58.3) 0.18
Daily Living Skills score < 70, n (%) 5 (33.3) 6 (50.0) 0.31
Socialisation score < 70, n (%) 3 (20.0) 4 (33.3) 0.36
Motor Skills score < 70, n (%) 9 (60.0) 9 (75.0) 0.34
Composite (overall) score < 70, n (%) 6 (40.0) 9 (75.0) 0.08
Inborn lateral Outborn lateral p value
ventricles ventricles
(N = 26) (N = 21)
Right VI-p97 (mm) 5.1 ± 0.9 7.7 ± 1.5 <0.0001
Left VI-p97 (mm) 5.1 ± 1.1 10.0 ± 3.6 <0.0001

deficiencies in the regionalisation of perinatal care. Transporting an REFERENCES

unstable preterm infant is challenging, but delayed transport means
delayed treatment.
Our findings suggest that patients with progressive PHVD
should be transferred to specialist care by skilled neonatal emer-
gency transport services when ventricular dilatation is approaching
the intervention threshold: it might be reasonable to plan transport
once VI has reached the p97, following a progressive increase over
the last week. The risk of subsequent neurological disabilities caused
by delayed PHVD treatment and the risk to move unstable preterm
infants should be carefully weighed.
C O N FL I C T O F I N T E R E S T
The authors have no conflicts of interest to declare.
ORCID
Alessandro Parodi  https://orcid.org/0000-0002-3463-3004
Laura De Angelis  https://orcid.org/0000-0002-0201-257X
Carlo Bellini  https://orcid.org/0000-0002-4199-7412
1. Leijser LM, Miller SP, van Wezel-Meijler G, et al. Posthemorrhagic
ventricular dilatation in preterm infants: when best to intervene?
Neurology. 2018;90(8):e698-e706.
2. Brouwer AJ, Brouwer MJ, Groenendaal F, et al. European perspective
on the diagnosis and treatment of posthaemorrhagic ventricular dila-
tation. Arch Dis Child Fetal Neonatal Ed. 2012;97:F50-55.
3. Levene MI. Measurement of the growth of the lateral ventricles in preterm
infants with real-time ultrasound. Arch Dis Child. 1981;56:900-904.
4. De Vries LS, Groenendaal F, Liem KD, et al. Treatment thresholds for
intervention in posthaemorrhagic ventricular dilation: a randomised
controlled trial. Arch Dis Child Fetal Neonatal Ed. 2019;104(1):F70-F75.
5. Bassan H, Eshel R, Golan I, et al. Timing of external ventricular drain-
age and neurodevelopmental outcome in preterm infants with pos-
themorrhagic hydrocephalus. Eur J Paediatr Neurol. 2012;16:662-670.
How to cite this article: Parodi A, Giordano I, De Angelis L, et
al. Post-haemorrhagic hydrocephalus management: Delayed
neonatal transport negatively affects outcome. Acta Paediatr.
2021;110:168–170. https://doi.org/10.1111/apa.15604
DOI: 10.1111/apa.15517
BRIEF REPORT
Breastfeeding practices and weight gain predicted head
circumference in young Amazonian children
The human brain experiences intensive connectivity and size trans-
formations after birth.1 Head circumference (HC) shows important
increments in size, specifically in the first year; then, this growth rate
decelerates over time. 2 Identifying the different elements required
for proper physical growth in children during critical periods is rel-
evant for primary health care.
This study investigated the HC predictors in the first year using
data from the Maternal and Child Health and Nutrition in Acre
(MINA-Brazil) study carried out in Cruzeiro do Sul, Acre State,
Western Brazilian Amazon. We enrolled 1246 mother-infant pairs at
3
birth, as previously described, and 774 (52% female) took part in the
1-year follow-up study from August 2016 to July 2017, when they
were 10-15 months of age. The baseline survey at birth provided
data on the mother's and infant's health conditions. The follow-up
visit detailed the child's health profile and anthropometric status.
Structured questionnaires, administered by trained researchers,
collected data on the participants’ sociodemographic information,
the child's health conditions and morbidity and feeding practices in
early childhood. Perinatal and birth data were collected from hospi-
tal records. The participants, and those lost to follow-up at 1 year,
had similar profiles with regard to maternal education, primiparous
mothers, delivery methods, low birthweight and the children's sex.3
Body weight change in the first year of life was estimated using
a conditional weight (variable that expressed the weight change, but
was not correlated with the birthweight values).4 Normality assump-
tions for the HC standard deviation scores (z-scores) were confirmed
using the Shapiro-Wilk test. To report the distribution of HC z-score
values, according to the independent variables, we used the Student
t test for dichotomous variables. The regression coefficients (β) and
their 95% confidence intervals (95% CIs) were calculated by multiple
linear regression, to assess the independent associations with HC
z-score values at the follow-up visit. All reported P values are two-
tailed. The statistical analyses were performed using Stata version
16.0 (StataCorp).
At the 1-year follow-up visit, we observed lower HC z-score
values among children from socially vulnerable mothers. Maternal
age ranged from 14 to 43 years: 22% were under 20, and 11%
A full list of members of the MINA-Brazil Study Group is presented in the
Acknowledgements.
© 2020 Foundation Acta Paediatrica. Published by John Wiley & Sons Ltd
Acta Paediatrica. 2021;110:171–173. 
were over 35. We found that 9% of the infants had experienced
dengue, malaria or pneumonia at least once in their first year, and
these were significantly associated with lower HC z-score (mean
−0.18  ± 0.91, P  < 0.01). Overall, 2% of the infants were stunted
(height for age ≤ 2 z-score), 8% were overweight (weight for age ≥ 2
z-score) and 1% had an HC measure of ≤2 z-score (data not shown).
After multiple adjustments, there were number of variables that
were positive predictors of the HC z-score in the first year of life, in
addition to the child's age, primiparity, type of delivery and birth-
weight. These were as follows: higher maternal education, mother
living with a partner, exclusive breastfeeding for at least 90 days
of life and conditional weight. The final model had a HC z-score
variation value of 30% (Table 1).
We believe this was the first birth cohort study to determine
whether conditional weight was a predictor of HC and to address
social and nutritional factors. Our results also showed that exclusive
breastfeeding in the first 90 days of life was an independent pre-
dictor of HC in the first year. Although HC may be a significant way
of estimating brain volume, other studies using different methods
may reinforce the need for deeper investigation of the role of diet
on brain development. An American cross-sectional study examined
magnetic resonance imaging scans and the dietary feeding patterns
of 133 infants aged 10 months to 4 years. The results showed that
longer breastfeeding duration significantly improved brain develop-
ment and that receiving breast milk exclusively for at least the first
90 days had a positive impact on white matter microstructure.5
The study limitations included the high dropout rate and the
possible unreliability of self-reported data. In addition, the in-
formation on other possible predictors of HC was not extensive,
limiting the generalisability of our findings, due to potential selec-
tion bias. Another limitation was using just one timepoint for HC,
which precluded estimating long-term predictors and trajectories.
However, the anthropometric data were collected by trained re-
searchers, which improved reliability, and this was the first popula-
tion-based cohort study in the Western Brazilian Amazon to assess
the predictors of HC during a critical period of the children's lives.
Our findings highlight the importance of comprehensively un-
derstanding the role of nutritional status and breastfeeding prac-
tices on HC in the first year of life and reinforce the need to support
effective strategies for childcare.
|
wileyonlinelibrary.com/journal/apa     171
172      | BRIEF REPORT

TA B L E 1   Predictors of head circumference z-scores (zHC) in the first year of life (n = 774)

Z-scores
Maternal and child characteristics n (%) meana  SD β b  95% CI P value
Mother's education (n = 759)
≤9 y schooling 226 (29.8) −0.07 0.06 1
>9 y schooling 533 (70.2) 0.14 0.04 0.10 0.02-0.18 0.02
Living with partner (n = 762)
No 162 (21.3) −0.06 0.97 1
Yes 600 (78.7) 0.12 0.91 0.14 0.00-0.28 0.05
Primiparity (n = 762)
No 441 (58.9) 0.19 0.91 1
Yes 321 (41.1) 0.16 0.94 0.09 0.03-0.21 0.14
Type of delivery (n = 774)
Vaginal 401 (51.8) −0.03 0.96 1
Caesarean 373 (48.2) 0.20 0.87 0.04 −0.02-0.10 0.18
Birthweight (n = 773)
Low (<2500 g) 49 (6.3) −0.45 0.98 1
Adequate (≥2500 g) 724 (93.7) 0.11 0.91 0.24 0.18-0.30 <0.01
Exclusive breastfeeding in the first 90 d (n = 773)
<90 d 480 (62.1) 0.04 0.04 1
≥90 d 293 (37.9) 0.13 0.06 0.11 0.00-0.23 0.05
Conditional weight change - 0.28 1.20 0.41 0.35-0.46 <0.01
(continuous z-score) (n = 773)c 

Note: Totals differ due to missing values.

a
Student t test, P < 0.05.
b
Adjusted for child age (mo); R 2 = 0.299. Factors associated with zHC were selected following a hierarchical conceptual model with these levels of
determination: (a) distal (maternal education and mother living with partner), (b) intermediate (primiparity, delivery method and birthweight) and (c)
proximal (breastfeeding practices and conditional weight change in first year). At each level of determination, covariates were retained in the multiple
linear regression model if they were associated with the HC z-score in the crude analysis at P < 0.10 or if their inclusion changed the adjusted R 2 by at
least 10%.
c
Body weight change in first year was estimated as the standardised residuals from linear regression of weight at the 1-y follow-up visit on
birthweight.

AC K N OW L E D G E M E N T S São Paulo, São Paulo, Brazil

2
We are grateful to the families, health professionals and MINA- Hospital Regional do Juruá, Cruzeiro do Sul, Brazil
Brazil Study Group members: Alicia Matijasevich Manitto, Bárbara
Hatzlhoffer Lourenço, Maíra Barreto Malta, Jenny Abanto, Suely Correspondence
Godoy Agostinho Gimeno, Marly Augusto Cardoso, Ana Alice Marly A. Cardoso, Department of Nutrition, School of Public
Damasceno, Bruno Pereira da Silva, Rodrigo Medeiros de Souza, Health, University of São Paulo, Avenida Doutor Arnaldo
Simone Ladeia-Andrade and Marcia Caldas de Castro. 715, São Paulo, SP, 01246-904, Brazil.
Email: marlyac@usp.br
C O N FL I C T S O F I N T E R E S T
None. ORCID
Isabel Giacomini  https://orcid.org/0000-0003-3460-7196
Isabel Giacomini1 Lalucha Mazzucchetti  https://orcid.
1
Lalucha Mazzucchetti org/0000-0002-9649-5727
Thainara A. B. Lima2 Maíra B. Malta  https://orcid.org/0000-0003-4993-1589
Maíra B. Malta1 Bárbara H. Lourenço  https://orcid.org/0000-0002-2006-674X
Bárbara H. Lourenço1 Marly A. Cardoso  https://orcid.org/0000-0003-0973-3908
Marly A. Cardoso1
the MINA-Brazil Study Group REFERENCES
1. van Dyck LI, Morrow EM. Genetic control of postnatal human brain
1
Department of Nutrition, School of Public Health, University of growth. Curr Opin Neurol. 2017;30:114-124.
BRIEF REPORT
2. Sacco R, Gabriele S, Persico AM. Head circumference and brain size
in autism spectrum disorder: a systematic review and meta-analysis.
Psychiatry Res. 2015;234:239-251.
3. Cardoso MA, Matijasevich A, Malta MB, et al. Cohort profile: the
Maternal and Child Health and Nutrition in Acre, Brazil, birth cohort
study (MINA-Brazil). BMJ Open. 2020;10:e034513.
4. Adair LS, Fall CH, Osmond C, et al. Associations of linear growth and
relative weight gain during early life with adult health and human
|
      173
capital in countries of low and middle income: findings from five birth
cohort studies. Lancet. 2013;382:525-534.
5. Deoni SCL, Dean DC, Piryatinsky I, et al. Breastfeeding and early
white matter development: a cross-sectional study. NeuroImage.
2013;82:77-86.
DOI: 10.1111/apa.15518

BRIEF REPORT

Managing mother’s own milk for very preterm infants in

neonatal units in 11 European countries

Feeding with human milk has been recognised as an essential com-
ponent of newborn care and is especially important for infants born
very preterm (VPT, below 32 weeks of gestation) who face higher
risks of adverse outcomes. WHO recommends that infants who can-
not be fed mother's own milk (MOM) should receive donor human
milk. Direct feeding at the breast takes time to establish after VPT
birth and procedures are required for expressing, collecting, stor-
ing and administering breast milk that respect microbiological safety
rules and ensure nutritional and immunological quality. However, as
recommendations are scarce, these procedures appear to be depen-
dent on organisational structures and policies of the units.1-3
Breast milk is one of the most common modes of postnatal
human cytomegalovirus (HCMV) transmission in preterm infants
with rates for fresh or frozen MOM ranging from 8% to 37% and
about 10% probability of infant infection. 2 The debate on the risk of
neonatal morbidity and long-term neurodevelopmental impairment
in VPT infants with HCMV infection is ongoing,4 raising additional
challenges for establishing best practices for managing MOM.1,2 In
2012, the American Academy of Paediatrics stated that the benefits
of fresh MOM from HCMV-seropositive mothers outweighed the
risk of HCMV infection. In 2018, the working group of the French
Neonatal Society on fresh human milk use in preterm infants ad-
vised pasteurisation of MOM from HCMV-seropositive mothers
2
for infants born <28 weeks or <1000 grams. In addition to HCMV
transmission, also bacterial contamination of expressed milk is sus-
pected to lead to significant problems in VPT infants. 2
Published reports have revealed differences in MOM manage-
1,3
ment among neonatal intensive care units (NICUs) and countries.
We aimed to compare practices for handling MOM for VPT infants
using standardised questions in European NICUs.
Data were collected as part of the area-based EPICE (Effective
Perinatal Intensive Care in Europe) study set-up to investigate the
use of evidence-based practices for the care of VPT infants in 19
regions in 11 countries.5
In 2012, structured questionnaires were sent to the head of all
NICUs with at least 10 VPT admissions during the study period. The
unit questionnaire recorded information on the units’ characteristics
and policies, including procedures for managing MOM. Of the 135
eligible NICUs, 134 (99.3%) replied.
Descriptive statistics are presented as numbers and valid
proportions.
Ethical approval was obtained in each region from appropriate
ethics committees, as required by national legislation, and for the
European database by the French data protection committee.
As shown in Table 1, 72% of the 134 neonatal units were level III
and 91% had a written protocol (developed at unit or regional level)
for breastfeeding/human milk use; when MOM was not available,
34% used human banked milk for all VPT infants (range: 0%-100% in
Sweden) and 13% for some groups only (range: 0%-50%).
Management of MOM varied widely across the countries.
Overall, 56% of units reported using fresh MOM (not frozen and
unpasteurised milk) without restrictions regarding gestational age,
birthweight or risk of HCMV transmission (range: 0% in Germany to
100% in Netherlands and Denmark). Different practices were also
observed between units in the same country, particularly in France
and Italy. In most units (70%), MOM was not pasteurised (range:
23%-100%); and in 29 (22%) units, all VPT infants received pasteur-
ised MOM (range: 0%-73% in France).
HCMV serology on all mothers who express their milk (or
HCMV polymerase chain reaction of the milk) was required in 29%
of units (range: 0% in 6 countries to 93% in Germany). Wide differ-
ences were found between NICUs in France and Italy. Among these
units (n = 36), 47% used human bank/MOM pasteurised, 26% fro-
zen-thawed MOM, 3% untreated fresh MOM and 24% formula in the
case of HCMV-seropositive mothers.
Systematic bacteriological analyses of the fresh or fro-
zen-thawed MOM were not performed in 76% of units (range: 29%-
100%), performed in <10% for the first milk feeding only, and every
week in 7%.
These large variations in practice between countries could reflect
differences in local regulations or guidelines, as well as lack of strong
recommendations at international and national level. Heterogeneity

EPICE (Effective Perinatal Intensive Care in Europe) Research Group: BELGIUM: Flanders (E Martens, G Martens, P Van Reempts); DENMARK: Eastern Region (K Boerch, A Hasselager,
LD Huusom, O Pryds, T Weber); ESTONIA (L Toome, H Varendi); FRANCE: Burgundy, Ile-de France and Northern Region (PY Ancel, B Blondel, A Burguet, PH Jarreau, P Truffert);
GERMANY: Hesse (RF Maier, B Misselwitz, S Schmidt), Saarland (L Gortner); ITALY: Emilia Romagna (D Baronciani, G Gargano), Lazio (R Agostino, I Croci, F Franco), Marche (V Carnielli),
M Cuttini, D DiLallo; NETHERLANDS: Eastern & Central (C Koopman-Esseboom, A van Heijst, J Nijman); POLAND: Wielkopolska (J Gadzinowski, J Mazela); PORTUGAL: Lisbon and
Tagus Valley (LM Graça, MC Machado), Northern region (Carina Rodrigues, T Rodrigues), H Barros; SWEDEN: Stockholm (AK Bonamy, M Norman, E Wilson); UK: East Midlands and
Yorkshire and Humber (E Boyle, ES Draper, BN Manktelow), Northern Region (AC Fenton, DWA Milligan); INSERM, Paris (J Zeitlin, M Bonet, A Piedvache).

© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd

Acta Paediatrica. 2021;110:123–126.  wileyonlinelibrary.com/journal/apa     123|

 |
124       BRIEF REPORT

TA B L E 1   Management of mother's own milk for very preterm infants in the neonatal unit, by countryb

Belgium Denmark Estonia France Germany

(n = 9) (n = 8) (n = 4) (n = 23) (n = 14)

n (%) n (%) n (%) n (%) n (%)

Level of care
Level III 8 (88.9) 2 (25.0) 3 (75.0) 22 (95.6) 13 (92.9)
Written protocol for breastfeeding and human milk use
Yes, regional or network protocols 0 (0.0) 0 (0.0) 3 (75.0) 3 (14.3) 0 (0.0)
Yes, unit protocols 6 (66.7) 8 (100.0) 1 (25.0) 15 (71.4) 12 (85.7)
No 3 (33.3) 0 (0.0) 0 (0.0) 3 (14.3) 2 (14.3)
Use of human bank milk/donor's milk to feed VPT infants whose mothers do not express their milk
Yes, for all infants less than 32 wk GA 0 (0.0) 7 (87.5) 1 (25.0) 19 (82.6) 0 (0.0)
Yes, for some infants 0 (0.0) 1 (12.5) 0 (0.0) 2 (8.7) 2 (14.3)
No 9 (100.0) 0 (0.0) 3 (75.0) 2 (8.7) 12 (85.7)
a
Infants whose mothers express breast milk receive fresh milk  
Yes, with no restrictions 4 (50.0) 8 (100.0) 3 (75.0) 6 (28.6) 0 (0.0)
Yes, but with restrictions 2 (25.0) 0 (0.0) 1 (25.0) 11 (52.4) 14 (100.0)
No, all milk is frozen or pasteurised 2 (25.0) 0 (0.0) 0 (0.0) 4 (19.0) 0 (0.0)
Infants whose mothers express breast milk receive frozen-thawed milk
Yes, with no restrictions 6 (66.7) 8 (100.0) 1 (25.0) 3 (13.0) 2 (14.3)
Yes, but with restrictions 2 (22.2) 0 (0.0) 1 (25.0) 1 (4.4) 11 (78.6)
No 1 (11.1) 0 (0.0) 2 (50.0) 19 (82.6) 1 (7.1)
Infants whose mothers express breast milk receive own mother's pasteurised milk
Yes, all babies <32 wk GA 2 (22.2) 1 (12.5) 1 (25.0) 16 (72.7) 1 (7.1)
Yes, only some babies <32 wk GA 0 (0.0) 0 (0.0) 0 (0.0) 1 (4.6) 3 (21.4)
No, own mother's milk is not 7 (77.8) 7 (87.5) 3 (75.0) 5 (22.7) 10 (71.4)
pasteurised
Human cytomegalovirus (HCMV) serology on all mothers of VPT infants (or HCMV PCR of the milk) is required, if they express their milk
Yes, according to GA or weight 1 (11.1) 0 (0.0) 0 (0.0) 5 (31.2) 13 (92.9)
Other criteria 0 (0.0) 0 (0.0) 0 (0.0) 3 (18.8) 0 (0.0)
No 8 (88.9) 7 (100.0) 4 (100.0) 8 (50.0) 1 (7.1)
Type of milk given to VPT infants born to HCMV-seropositive mothers, if HCMV serology performed
n = 1 n = 0 n = 0 n = 8 n = 13
Human bank or MOM pasteurised 1 (100.0) 7 (100.0) 3 (23.1)
Frozen-thawed breast milk 0 (0.0) 0 (0.0) 4 (30.8)
Untreated fresh breast milk 0 (0.0) 0 (0.0) 0 (0.0)
Only formula 0 (0.0) 0 (0.0) 6 (46.1)
Systematic bacteriological analysis of the mother's fresh or frozen-thawed milk is performed
Yes, for the first milk feeding 0 (0.0) 0 (0.0) 0 (0.0) 6 (28.6) 4 (28.6)
Yes, every week 0 (0.0) 0 (0.0) 0 (0.0) 9 (42.9) 0 (0.0)
Another frequency 1 (11.1) 1 (12.5) 2 (50.0) 0 (0.0) 1 (7.1)
No 8 (88.9) 7 (87.5) 2 (50.0) 6 (28.6) 9 (64.3)

Note: Differences in the total number of units for each item are due to missing data.
Abbreviations: GA, gestational age; HCMV, human cytomegalovirus; MOM, mother's own milk; PCR, polymerase chain reaction; UK,
United Kingdom; VPT, very preterm.
a
Fresh breast milk corresponds to not frozen and unpasteurised milk, including refrigerated milk.
b
Regions included in the EPICE study by country: Flanders in Belgium; the Eastern Region of Denmark; Estonia (entire country); Burgundy,
Ile-de France and the Northern regions in France; Hesse and Saarland in Germany; Emilia-Romagna, Lazio and Marche regions in Italy; the
Central and Eastern regions of The Netherlands; Wielkopolska in Poland; the Lisbon and Northern regions of Portugal; and the East Midlands,
Northern and Yorkshire and Humber regions in the UK; and the Stockholm region in Sweden.
BRIEF REPORT |
      125

Italy Netherlands Poland Portugal Sweden UK Total

(n = 22) (n = 2) (n = 4) (n = 17) (n = 4) (n = 27) (n = 134)

n (%) n (%) n (%) n (%) n (%) n (%) n (%)

16 (72.7) 2 (100.0) 1 (25.0) 15 (88.2) 2 (50.0) 12 (44.4) 96 (71.6)

3 (13.6) 0 (0.0) 3 (75.0) 5 (29.4) 1 (25.0) 1 (3.7) 19 (14.4)

18 (81.8) 2 (100.0) 1 (25.0) 10 (58.8) 3 (75.0) 25 (92.6) 101 (76.5)
1 (4.6) 0 (0.0) 0 (0.0) 2 (11.8) 0 (0.0) 1 (3.7) 12 (9.1)

9 (40.9) 0 (0.0) 0 (0.0) 1 (5.9) 4 (100.0) 4 (14.8) 45 (33.6)

2 (9.1) 1 (50.0) 0 (0.0) 1 (5.9) 0 (0.0) 8 (29.6) 17 (12.7)
11 (50.0) 1 (50.0) 4 (100.0) 15 (88.2) 0 (0.0) 15 (55.6) 72 (53.7)

11 (50.0) 2 (100.0) 2 (66.7) 11 (64.7) 2 (50.0) 24 (88.9) 73 (56.1)

7 (31.8) 0 (0.0) 1 (33.3) 5 (29.4) 2 (50.0) 3 (11.1) 46 (35.4)
4 (18.2) 0 (0.0) 0 (0.0) 1 (5.9) 0 (0.0) 0 (0.0) 11 (8.5)

16 (72.7) 2 (100.0) 1 (25.0) 9 (52.9) 3 (75.0) 24 (88.9) 75 (56.0)

3 (13.6) 0 (0.0) 0 (0.0) 1 (5.9) 1 (25.0) 3 (11.1) 23 (17.2)
3 (13.6) 0 (0.0) 3 (75.0) 7 (41.2) 0 (0.0) 0 (0.0) 36 (26.9)

6 (27.3) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (7.4) 29 (21.8)

6 (27.3) 0 (0.0) 1 (25.0) 0 (0.0) 0 (0.0) 0 (0.0) 11 (8.3)
10 (45.4) 2 (100.0) 3 (75.0) 17 (100.0) 4 (100.0) 25 (92.6) 93 (69.9)

6 (27.3) 0 (0.0) 0 (0.0) 2 (11.8) 0 (0.0) 0 (0.0) 27 (21.4)

5 (22.7) 0 (0.0) 0 (0.0) 1 (5.9) 0 (0.0) 0 (0.0) 9 (7.1)
11 (50.0) 2 (100.0) 4 (100.0) 14 (82.4) 4 (100.0) 27 (100.0) 90 (71.4)

n = 11 n = 0 n = 0 n = 3 n = 0 n = 0 n = 36

5 (45.4) 0 (0.0) 16 (47.1)
4 (36.4) 1 (50.0) 9 (26.5)
1 (9.1) 0 (0.0) 1 (2.9)
1 (9.1) 1 (50.0) 8 (23.5)

3 (13.6) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 13 (9.8)

0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 9 (6.8)
3 (13.6) 0 (0.0) 0 (0.0) 1 (5.9) 0 (0.0) 1 (3.7) 10 (7.6)
16 (72.7) 2 (100.0) 4 (100.0) 16 (94.1) 4 (100.0) 26 (96.3) 100 (75.8)
 |
126       BRIEF REPORT
was also observed between units within the same country, revealing
that different options can operate locally, such as milk bank availabil-
ity which may differ between regions within countries. Study limita-
tions include data collected in 2012, which may underestimate current
human milk bank availability, and a smaller number of units in par-
ticipating regions in some countries; however, no more recent com-
parable information on practices is available from so many European
countries.
While it is recognised that MOM should be encouraged as the
primary feeding method for VPT infants, this variation indicates
substantial differences in attitudes about what constitutes best
practices for the management of MOM among European neonatol-
ogists. To guide practice, further studies are needed to reinforce the
evidence base on ensuing outcomes of HCMV infection as well as
practices for managing these risks while ensuring that VPT infants
can benefit from MOM.
AC K N OW L E D G E M E N T
The authors are grateful to the staff from the participating neonatal
units.
C O N FL I C T O F I N T E R E S T
None to declare.
F U N D I N G I N FO R M AT I O N
The research received funding from the European Union's Seventh
Framework Programme (grant 259882) and from the Foundation
for Science and Technology - FCT (Portuguese Ministry of Science,
Technology and Higher Education), the Operational Programmes
Competitiveness and Internationalization (COMPETE 2020) and
Human Capital (POCH), Portugal 2020, under the EPIUnit (POCI-01-
0145-FEDER-006862; UID/DTP/04750/2019) and the PhD grant
SFRH/BD/111794/2015 (CR).
Carina Rodrigues1
Jennifer Zeitlin2
Emilija Wilson3
Liis Toome 4,5
Marina Cuttini6
Rolf F. Maier7
Véronique Pierrat 2
Henrique Barros1,8
the EPICE Research Group
1
EPIUnit–Instituto de Saúde Pública, Universidade do Porto,
Porto, Portugal
2
Obstetrical, Perinatal and Pediatric Epidemiology Research
Team (EPOPé), CRESS, INSERM, INRA, Université de Paris, Paris,
France
3
Department of Women’s and Children’s Health, Karolinska
Institutet, Stockholm, Sweden
4
Department of Neonatal and Infant Medicine, Tallinn Children's
Hospital, Tallinn, Estonia
5
Department of Pediatrics, University of Tartu, Tartu, Estonia
6
Clinical Care and Management Innovation Research Area,
Bambino Gesù Children’s Hospital, IRCCS, Rome, Italy
7
Children´s Hospital, University Hospital, Philipps University,
Marburg, Germany
8
Departamento de Ciências da Saúde Pública e Forenses e
Educação Médica, Faculdade de Medicina, Universidade do
Porto, Porto, Portugal
Correspondence
Carina Rodrigues, Instituto de Saúde Pública da Universidade
do Porto, Rua das Taipas, 135, 4050-600 Porto, Portugal.
Email: carina.rodrigues@ispup.up.pt
ORCID
Carina Rodrigues  https://orcid.org/0000-0003-1720-2808
Jennifer Zeitlin  https://orcid.org/0000-0002-9568-2969
Marina Cuttini  https://orcid.org/0000-0002-3284-6874
Henrique Barros  https://orcid.org/0000-0003-4699-6571
REFERENCES
1. Klotz D, Jansen S, Gebauer C, Fuchs H. Handling of breast milk by
neonatal units: large differences in current practices and beliefs.
Front Pediatr. 2018;6:235.
2. Picaud JC, Buffin R, Gremmo-Feger G, Rigo J, Putet G, Casper C.
Review concludes that specific recommendations are needed to har-
monise the provision of fresh mother's milk to their preterm infants.
Acta Paediatr. 2018;107(7):1145-1155.
3. Bonet M, Forcella E, Blondel B, et al. Approaches to supporting lacta-
tion and breastfeeding for very preterm infants in the NICU: a qualita-
tive study in three European regions. BMJ Open. 2015;5(6):e006973.
4. Weimer KED, Kelly MS, Permar SR, Clark RH, Greenberg RG.
Association of adverse hearing, growth, and discharge age outcomes
with postnatal cytomegalovirus infection in infants with very low
birth weight. JAMA Pediatr. 2020;174(2):133-140.
5. Zeitlin J, Maier RF, Cuttini M, et al. Cohort profile: Effective Perinatal
Intensive Care in Europe (EPICE) very preterm birth cohort. Int J
Epidemiol. 2020;49(2):372-386.