Professional Documents
Culture Documents
ACTA PÆDIATRICA
N U R T U R I N G T H E C H I L D
www.actapaediatrica.com
EDITORIAL OFFICE
Acta Pædiatrica
Tomtebodavägen 9
171 65 Solna, Sweden
T +46 73 2024447
E mail@actapaediatrica.se
C ON TE N TS
HIGHLIGHTS IN THIS ISSUE
EDITORIALS
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
PAEDIATRIC ESSAY
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
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
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
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
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
DOI: 10.1111/apa.15463
REGULAR ARTICLE
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
©2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
|
68
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:68–71.
CAVALLIN et al. |
69
2 | M E TH O DS
Key notes
2.1 | Study design
• In low-resource settings, oxygen administration at birth
is exclusively driven by clinical observation, but it is un-
This study includes the secondary analysis of infant colour and oxygen
known whether assessing infant colour may determine
saturation data from a main study (ClinicalTrials.gov NCT03854435)
when to give supplemental oxygen.
on heart rate assessment in a low-resource setting. This analysis was
• There was an association between oxygen saturation
included in the study protocol and was approved by the institutional
and infant colour after birth, but a considerable pro-
review board of the St. Luke Catholic Hospital. All parents gave writ-
portion of infants were difficult to classify as pink or
ten informed consent to participate in the study.
cyanotic.
• In low-resource settings, the availability of a pulse oxi-
meter may improve clinical evaluation and guide oxy-
2.2 | Settings
genation administration.
The study was conducted at the St. Luke Catholic Hospital in Wolisso
(Ethiopia), which is a tertiary hospital with around 3600 deliveries per year.
2.7 | Data collection
2.3 | Patients
An external observer, who was not involved in the care of the new-
Patients enrolled in the main study were inborn infants at risk of as- born, was responsible for data collection. Maternal and infant char-
phyxia (ie suspected foetal distress assessed by intermittent foetal acteristics included maternal age, attendance to antenatal clinic,
auscultation, shoulder dystocia, suspected placenta abruption, cord pro- antenatal steroids, HIV, intrauterine growth restriction (IUGR),
lapse, foetal malpresentation, vacuum- or forceps-assisted vaginal deliv- meconium-stained amniotic fluid, foetal bradycardia, mode of deliv-
ery and meconium-stained amniotic fluid). Exclusion criteria were major ery, gestational age, sex, birthweight and Apgar score at 1 and 5 min-
congenital malformations or parental refusal to participate to the study. utes. Infant colour and oxygen saturation were recorded at 60, 90,
120 and 300 seconds after birth.
2.4 | Procedure
2.8 | Statistical analysis
During the main study, an external observer, who was not involved in
the care of the newborn, positioned a pulse-oximeter sensor (LNCS® This is a secondary analysis of data from the main study; thus, no formal
SENSORS, Masimo Radical, Masimo) on the right hand of the baby. The sample size calculation was performed. Continuous variables were ex-
maximum sensitivity setting was chosen. Time of birth was the ‘time 0’, pressed as median and interquartile range (IQR), while categorical data
and oxygen saturation was evaluated at 60, 90, 120 and 300 seconds by variables were expressed as frequency and percentage. Oxygen satura-
the external observer with a stopwatch. The midwife involved in the care tion during the first 5 minutes of life was evaluated with a mixed re-
of the baby was asked to evaluate infant colour at the same time points. gression model including infant colour (cyanotic, unclear, pink), time and
Infant colour was evaluated by looking at perioral area including lips and interaction colour*time. All tests were 2-sided, and a P-value <.05 was
tongue, and was classified as pink, cyanotic or unclear. Colour was defined considered statistically significant. Statistical analysis was performed
as unclear when the midwife defined the baby as neither pink nor cyanotic. using R 3.5 (R Foundation for Statistical Computing, Vienna, Austria).6
2.5 | Blinding 3 | R E S U LT S
The midwife assessing infant colour was masked to the pulse oxime- The analysis included 60 neonates who were enrolled in the main
try data. The statistician who analysed the data was masked to the study from February to April 2019. Participant characteristics are
classification of infant colour. 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%).
2.6 | Outcome measure 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
The outcome measure was the association between infant colour 90 seconds, 35/60 (59%) at 120 seconds and 22/60 (37%) at 300 sec-
and oxygen saturation (%) at 60, 90, 120 and 300 seconds after birth. 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 The differences of oxygen saturation among infant colours (cyanotic,
and 18/60 (30%) at 300 seconds after birth. Infants who looked pink unclear, pink) persisted over time (interaction colour*time P = .28).
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 4 | D I S CU S S I O N
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%) Our findings showed an association between oxygen saturation and
infants at 90-120 seconds. Median oxygen saturation in infants who infant colour after birth, with higher saturation levels in pink infants
looked cyanotic was 60% (IQR 45-70) at 60 seconds, 64% (IQR 52-69) compared to cyanotic infants. A considerable proportion of infants
at 90 seconds, 63% (IQR 56-68) at 120 seconds and 66% (IQR 62-74) at were difficult to classify as pink or cyanotic.
300 seconds. Median oxygen saturation in infants with unclear colour To our knowledge, this is the first study investigating the rela-
was 57% (IQR 56-60) at 60 seconds, 78% (IQR 71-81) at 90 seconds, tionship between clinical assessment of infant colour and oxygen-
81% (IQR 79-88) at 120 seconds and 80% (IQR 76-84) at 300 seconds. ation in a low-resource setting. According to guidelines on neonatal
In infants who looked pink, oxygen saturation was 87% (only one in- resuscitation, assessment of oxygenation should be performed by
fant) at 60 seconds, and median of oxygen saturation was 90% (IQR pulse oximetry in high-resource settings,1,2 while clinical evaluation
83-91) at 90 seconds, 86% (IQR 81-94) at 120 seconds and 93% (IQR remains the only option in low-resource settings due to guidelines
90-96) at 300 seconds. Overall, oxygen saturation increased over time recommendations and equipment constraints.7 Two previous studies
(P < .0001) and was different according to infant colour (P < .0001). 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
F I G U R E 1 Clinical assessment of infant colour at 60, 90, 120 high-resource settings, where oxygen supplementation is based on
and 300 s after birth pulse oximetry. In our study, higher saturation levels were found in
pink infants compared to cyanotic infants, thus suggesting the clin-
TA B L E 1 Participant characteristics
ical importance of assessing infant colour in low-resource settings.
N 60 However, there was some variability in oxygen saturation levels
Maternal age, ya 28 (24-30) within colour groups, in agreement with data from a high-resource
At least one antenatal visit 55 (92) setting.3 While previous studies classified infant colour as pink or
Antenatal steroids 12 (20) 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. |
71
midwives and to all the local CUAMM staff of the St. Luke Catholic
Hospital in Wolisso (Ethiopia) for their support during this study.
How to cite this article: Cavallin F, Cori MS, Negash S, et al.
Limited agreement between clinical assessment of infant
C O N FL I C T O F I N T E R E S T
colour at birth and oxygen saturation in a hospital in Ethiopia.
All the authors have no conflicts of interest to disclose.
Acta Paediatr. 2021;110:68–71. https://doi.org/10.1111/
apa.15463
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
| |
Received: 30 January 2020 Revised: 5 April 2020 Accepted: 6 April 2020
DOI: 10.1111/apa.15303
REGULAR ARTICLE
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
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. 73|
interventions, and international guidelines suggest the electro-
cardiogram (ECG) may be used in the delivery room to monitor
Key notes
it.4 However, electrode application can be difficult, ECG monitors
• Wireless, continuous heart rate (HR) monitoring of
are not always available and electrical cardiac activity may not
newborns using a cap mounted device (fhPPG) could be
always equate to effective cardiac output. 5,6 Pulse oximetry (PO)
beneficial during delayed cord clamping, especially in
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
• Using electrocardiogram HR as a gold standard, a novel
the delay of a few minutes to obtain a reliable HR after birth due
fhPPG performed similarly to pulse oximetry in the neo-
to poor peripheral perfusion and motion. 8 This could explain why
natal intensive care and immediately after caesarean
PO underestimates HR when compared to ECG during the first
section birth.
minutes of life.9,10
• The fhPPG is an accurate, wireless alternative to current
Delayed cord clamping (DCC) for term infants is widely practised
methods of HR monitoring in newborn infants.
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 2 | M ATE R I A L S A N D M E TH O DS
the attending team to be confident of the baby's condition including
an acceptable HR and avoidance of hypothermia. A non-intrusive, 2.1 | Study population
wireless, easy to apply HR monitoring system, which avoids the risk
of hypothermia, could be advantageous in this setting. This prospective observational study was conducted at the
We have previously described a prototype wired, fore- Nottingham University Hospitals NHS Trust, UK following ethi-
head-mounted sensor studied in NICU patients that utilises reflec- cal approval (NHS Health Research Authority Yorkshire & The
tance-mode green light photoplethysmography (PPG) to measure Humber—Sheffield Research Ethics Committee 15/YH/0522).
HR.13 This has the advantages that this wavelength of light (525 nm) is Informed parental consent was obtained prior to infants being en-
optimised for reflection-mode detection of blood flow and captures rolled in the study, and evaluations were performed in the NICU
this at the forehead where brain perfusion shares arterial pathways and on babies born by elective caesarean section (ECS). For the
via the carotid artery and so is less susceptible to poor peripheral NICU phase, infants were included if they were >22 weeks gesta-
perfusion.14,15 This original device was held in place on the forehead tion, required HR monitoring and were deemed clinically stable
by a spun bond laminate headband and successfully demonstrated by the attending team. Inclusion criteria for the ECS phase were
that HR measurement by this method was feasible but had a Bland- infants ≥37 weeks gestation with no obvious requirement for re-
Altman limit of agreement (LOA) of up to ± 12bpm. In addition, it was suscitation, based on antenatal history, and delivered by planned
not compatible with current neonatal care due to its inability to ac- caesarean section.
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 2.2 | Study design
bedside stabilisation of preterm infants during DCC.
This device has now been substantially modified (Video S1, There were two phases to this study, the NICU and ECS. On the
Figure S1 and Figure S2) in terms of its practicality, and several major NICU, an appropriately sized fhPPG cap was fitted to the infant's
modifications have taken place including the following: head and connected wirelessly to a bespoke synchronised data-
logging system via the inbuilt wireless BT module. As part of
(i) A T-shape cap compatible with respiratory equipment fittings. their routine care, infants had 3 neonatal ECG monitoring elec-
(ii) A miniaturised and ruggedised sensor (10 mm diameter). trodes (PD50-F4C, SKINTACT, Leonhard Lang GmbH) attached
(iii) Bluetooth (BT) wireless connection removing the need for trail- and connected to a CARESCAPE Monitor B450 (General Electric
ing wires. Healthcare) with Masimo SET® for SpO2 (Masimo). A transmission
mode PO sensor (LNCS Neo, Masimo) was attached to the wrist
The primary aim of this study was to assess, via a clinical trial, the of the right hand and connected to the same monitor. An in-house
accuracy and reliability of this new cap mounted newborn HR device designed system using LabVIEW 2014 (National Instruments) si-
the SurePulse VS (fhPPG) (SurePulse Medical Ltd). The study was multaneously collected synchronised data from all devices and
designed to satisfy the regulatory requirements of the CE Medical displayed waveforms from each device in real time on a laptop
Devices Directive 93/42 and its 2007 amendment for measuring HR (Lenovo Y50 20378, Lenovo Group Ltd). In keeping with normal
compared to PO, with ECG as a gold standard. A secondary aim was practice, if any device presented a poor signal output the elec-
to investigate acquisition time. 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.
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.
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
150 150
PO HR (bpm)
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
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. |
77
7. Mann C, Ward C, Grubb M, et al. Marked variation in newborn 19. Japanese Industrial Standards JT. 分べん(娩)監視装置 Fetal moni-
resuscitation practice: a national survey in the UK. Resuscitation. tors. Tokyo: Japanese Standards Association; 2005 (revised 2018).
2012;83(5):607-611. Japanese Industrial Standards T1303.
8. Kamlin CO, Dawson JA, O'Donnell CP, et al. Accuracy of pulse oxim- 20. Bustos R, Bejar R, Arroyave H, et al. Heart rate in fetuses and neo-
etry measurement of heart rate of newborn infants in the delivery nates in normal conditions and with mild depression. J Perinat Med.
room. J Pediatr. 2008;152(6):756-760. 1975;3(3):172-179.
9. van Vonderen JJ, Hooper SB, Kroese JK, et al. Pulse oximetry mea- 21. Cohen WR, Ommani S, Hassan S, et al. Accuracy and reliability of
sures a lower heart rate at birth compared with electrocardiogra- fetal heart rate monitoring using maternal abdominal surface elec-
phy. J Pediatr. 2015;166(1):49-53. trodes. Acta Obstet Gynecol Scand. 2012;91(11):1306-1313.
10. Murphy MC, De Angelis L, McCarthy LK, O'Donnell CPF. 22. Bland JM, Altman DG. Agreement between methods of measure-
Randomised study comparing heart rate measurement in newly ment with multiple observations per individual. J Biopharm Stat.
born infants using a monitor incorporating electrocardiogram and 2007;17(4):571-582.
pulse oximeter versus pulse oximeter alone. Arch Dis Child Fetal 23. Altman DG, Bland JM. Measurement in medicine: the analysis of
Neonatal Ed. 2019;104:F547-F550. method comparison studies. Statistician. 1983;307-317.
11. Tarnow-Mordi W, Morris J, Kirby A, et al. Delayed versus im- 24. Singh JK, Kamlin CO, Morley CJ, O'Donnell CP, Donath SM, Davis
mediate cord clamping in preterm infants. N Engl J Med. PG. Accuracy of pulse oximetry in assessing heart rate of in-
2017;377(25):2445-2455. fants in the neonatal intensive care unit. J Paediatr Child Health.
12. Manley BJ, Owen LS, Hooper SB, et al. Towards evi- 2008;44(5):273-275.
dence-based resuscitation of the newborn infant. Lancet. 25. Pickup L, Lang A, Shipley L, et al. Development of a clinical interface
2017;389(10079):1639-1648. for a novel newborn resuscitation device: human factors approach
13. Grubb MR, Carpenter J, Crowe JA, et al. Forehead reflectance pho- to understanding cognitive user requirements. JMIR Hum Factors.
toplethysmography to monitor heart rate: preliminary results from 2019;6(2):e12055.
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. S U P P O R T I N G I N FO R M AT I O N
2007;16(2):146-152. Additional supporting information may be found online in the
15. Berkenbosch JW, Tobias JD. Comparison of a new forehead reflec- Supporting Information section.
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
How to cite this article: Henry C, Shipley L, Ward C, et al.
delay in obtaining pulse oximetry during neonatal resuscitation.
J Pediatr. 2005;147(5):698-699. Accurate neonatal heart rate monitoring using a new
17. O'Donnell CP, Kamlin CO, Davis PG, Morley CJ. Obtaining pulse wireless, cap mounted device. Acta Paediatr. 2021;110:72–78.
oximetry data in neonates: a randomised crossover study of https://doi.org/10.1111/apa.15303
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.
| |
Received: 13 November 2019 Revised: 7 April 2020 Accepted: 16 April 2020
DOI: 10.1111/apa.15321
REGULAR ARTICLE
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
1 | I NTRO D U C TI O N albicans and Candida parapsilosis are the most common fungal spe-
cies in this population. 2 The incidence of NIC has decreased since
The incidence of proven neonatal invasive candidiasis (NIC) has the 2000s with the control of identified risk factors, improved infec-
ranged from 2.6% to 13.2% in children with birth weight of <1500 g tion control and the use of fluconazole prophylaxis.1-5 However, NIC
and from 6.6% to 26.0% with birth weight of <1000 g.1,2 Candida 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,
Abbreviations: BDG, (1 → 3)-β-D-glucan; CI, confidence interval; IQR, interquartile with a risk of false-negative findings, the diagnosis is difficult. That's
range; NIC, neonatal invasive candidiasis; ROC, receiver operating characteristic.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
the results of fungal and bacterial cultures were collected. The time with continuity correction, if needed. We graphically represented
interval between obtaining the BDG assay, blood culture and anti- both cut-off values in terms of pre-test and post-test probabilities
fungal therapy was assessed. 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).
2.5 | Statistical analysis
Positive Negative
Newborn infant
included (n = 55) Proven (n = 2) Probable (n = 5) Possible (n = 54)
negative NIC, respectively. We did not observe any NIC in the ex- to the recommended cut-off of 80 pg/mL were 85.7% (95% CI 42.1%-
cluded infants (no BDG available). Treatments used and some NIC 99.6%) and 51.9% (37.8%-65.7%), respectively. The positive and
risk factors are described in Table S2. The most frequent first-line negative likelihood ratios were 1.78 (1.18-2.68) and 0.28 (0.04-1.72),
empirical antifungal treatment used for NIC was fluconazole. The respectively.
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 3.3 | ROC analysis
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 The area under the ROC curve for the BDG assay for detecting NIC
surgery (transposition of the great arteries). Blood and urine cultures was 0.88 (0.75-1.00; Figure S1). The optimal cut-off was 174 pg/mL,
were positive for C parapsilosis, and the BDG value was 7840 pg/mL. for sensitivity 85.7% (42.1%-99.6%) and specificity 77.8% (64.4%-
Of the five episodes classified as probable NIC, no bacteria were 88.0%). The positive and negative likelihood ratios were 3.86 (2.15-
found in blood culture, no alternative diagnosis was suspected and 6.91) and 0.18 (0.03-1.13), respectively. The sensitivity was similar
confirmed, and C albicans was the only species identified from urine, (85.7%) but the specificity was significantly higher with the opti-
stool or trachea culture. Among the 54 NIC-negative episodes, 37 mal cut-off of 174 pg/mL than the recommended cut-off of 80 pg/
(69%) had an alternative diagnosis: 19 (35%) had a positive blood cul- mL (77.8% vs 51.9%; P < .01). The pre-test probability of NIC was
ture for bacteria, six (11%) were associated with necrotising entero- seven (12%) in this population of newborn infants at risk of NIC.
colitis, two (4%) were associated with pulmonary haemorrhage, two With the optimal cut-off of 80 pg/mL, the negative and positive like-
(4%) were associated with virus and eight (15%) had other explana- lihood ratios yielded post-test probabilities of 18.8% (13.3%-25.9%)
tions for clinical deterioration. 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%
3.2 | Diagnostic accuracy of serum BDG (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 4 | D I S CU S S I O N
blood culture sample collection. The median BDG value was sig-
nificantly higher in NIC-positive episodes (proven and probable NIC In a high-risk population of newborn infants under fluconazole
combined: 605 pg/mL, IQR 338-4448; proven NIC: 16 420 pg/mL, prophylaxis, we founded that the BDG assay showed good sensitiv-
IQR 12 130-20 710; and probable NIC: 500 pg/mL, IQR 176-605) ity for the diagnosis of NIC with the recommended cut-off of 80 pg/
than in NIC-negative episodes (75 pg/mL, IQR 31-165) (P < .01). The mL (85.7%) but poor specificity (51.9%). Even with a good area under
sensitivity and specificity of serum BDG for detecting NIC according the ROC curve value (0.88), the specificity was improved with the
CLIQUENNOIS et al. |
83
cut-off of 174 pg/mL but remained poor (77.8%) in terms of our ob- hospital of Dijon, France, for his biological expertise. We thank
jective to rule out the diagnosis of NIC in order to stop empirical Laura Smales for correcting the English language several times in the
antifungal treatments. manuscript.
In comparison with previous findings, the median values of BDG
increased with the clinical probability of NIC, such as 16 420 pg/mL C O N FL I C T O F I N T E R E S T
with proven NIC, 500 pg/mL with probable NIC and 75 pg/mL with The authors declare no conflicts of interest.
possible NIC, which agrees with previous studies.10-13,16 A system-
atic review of newborn infants reported a sensitivity from 75% to ORCID
100% and a specificity from 47% to 100% using the same BDG assay Pauline Cliquennois https://orcid.org/0000-0001-5525-9931
16
(Fungitell) with the recommended cut-off of 80 pg/mL. Moreover, Pauline Scherdel https://orcid.org/0000-0003-4517-9430
some studies suggested that the cut-off value used in adults was too Rose-Anne Lavergne https://orcid.org/0000-0001-9339-1562
low for children 21,22 because many factors in neonates could increase Cyril Flamant https://orcid.org/0000-0002-2665-8244
the BDG values. Indeed, the different cut-off values suggested in neo- Florent Morio https://orcid.org/0000-0002-5859-6105
11,12,23
nate infants in previous studies ranged from 106 to 305 pg/mL. Jeremie F. Cohen https://orcid.org/0000-0003-3572-8985
Otherwise, the diagnostic performance may vary depending on the Elise Launay https://orcid.org/0000-0002-4081-4653
BDG assay kit used such as Fungitell, Dynamiker Fungus or GKT-5M Christele Gras Le Guen https://orcid.org/0000-0001-9060-2344
and the studied population.13,24 Indeed, some studied populations in
previous publications did not reflect neither current clinical practices REFERENCES
nor the characteristics of newborn infants at high risk of NIC admitted 1. Aliaga S, Clark RH, Laughon M, et al. Changes in the inci-
to neonatal intensive care units. dence of candidiasis in neonatal intensive care units. Pediatrics.
2014;133:236-242.
Despite some strengths, especially the choice of a targeted
2. Benjamin DK, Stoll BJ, Gantz MG, et al. Neonatal candidia-
population with clinically suspected NIC reflecting the current sis: epidemiology, risk factors, and clinical judgment. Pediatrics.
clinical practices in neonatal intensive care units, our study has 2010;126:e865-e873.
limitations. First, it was performed in a single centre. Second, sev- 3. Fisher BT, Ross RK, Localio AR, Prasad PA, Zaoutis TE. Decreasing
rates of invasive candidiasis in pediatric hospitals across the United
eral eligible newborn infants with clinically suspected NIC during
States. Clin Infect Dis. 2014;58:74-77.
the study period could not be included because the BDG assay 4. Cleminson J, Austin N, McGuire W. Prophylactic systemic anti-
was not performed. Finally, the scarcity of proven and probable fungal agents to prevent mortality and morbidity in very low birth
episodes of NIC in this population limits the statistical power and weight infants. Cochrane Database Syst Rev. 2015;(10):CD003478.
5. Kaufman DA, Morris A, Gurka MJ, Kapik B, Hetherington S.
the external validity of the study. The proposed optimal cut-off
Fluconazole prophylaxis in preterm infants: a multicenter case-con-
of 174 pg/mL should be validated in other populations and set- trolled analysis of efficacy and safety. Early Hum Dev. 2014;90(Suppl
tings before any recommendation can be made. We did not include 1):S87-S90.
enough children to analyse environmental factors associated with 6. Stoll BJ, Hansen NI, Adams-Chapman I, et al. Neurodevelopmental
false-positive BDG results as observed in previous studies.7,8,25 and growth impairment among extremely low-birth-weight infants
with neonatal infection. JAMA. 2004;292:2357-2365.
Such factors included the use of cotton gauze during surgery,
7. Cuenca-Estrella M, Verweij PE, Arendrup MC, et al. ESCMID* guide-
blood product transfusions, broad-spectrum antibiotics or antifun- line for the diagnosis and management of Candida diseases 2012: di-
8,17,18
gal therapy and systemic bacterial infections. For example, agnostic procedures. Clin Microbiol Infect. 2012;18(Suppl 7):9-18.
Goudjil et al reported that blood product transfusion were associ- 8. Pappas PG, Kauffman CA, Andes DR, et al. Clinical practice guideline
for the management of candidiasis: 2016 update by the Infectious
ated with increased BDG value in neonates.17
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-
5 | CO N C LU S I O N 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-
The BDG assay seems potentially useful for optimising antifungal view and meta-analysis. Clin Microbiol Infect. 2020;26(3):291-298.
treatments by decreasing the treatment duration after ruling out 11. Cornu M, Goudjil S, Kongolo G, et al. Evaluation of the (1,3)-be-
the diagnosis of NIC in case of negative results. The recommended ta-D-glucan assay for the diagnosis of neonatal invasive yeast in-
fections. Med Mycol. 2018;56:78-87.
cut-off of 80 pg/mL for the BDG assay is probably too low for a
12. Goudjil S, Kongolo G, Dusol L, et al. (1–3)-beta-D-glucan levels in
population of newborn infants at high risk of NIC. Higher cut-off candidiasis infections in the critically ill neonate. J Matern Fetal
values may be more appropriate to help the clinician in evaluating Neonatal Med. 2013;26:44-48.
the risk of NIC and deciding to stop empirical antifungal treatments. 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.
AC K N OW L E D G E M E N T S 14. Mackay CA, Ballot DE, Perovic O. Serum 1,3-beta-D-Glucan assay
We thank all the personal of neonatal intensive care unit who con- in the diagnosis of invasive fungal disease in neonates. Pediatr Rep.
tributed to the study. We also thank Dr Frédéric Dalle, university 2011;3:e14.
|
84 CLIQUENNOIS et al.
15. Montagna MT, Coretti C, Lovero G, et al. Diagnostic performance children: preliminary data for diagnostic use of the beta-glucan
of 1→3-beta-d-glucan in neonatal and pediatric patients with can- assay in a pediatric setting. Clin Vaccine Immunol. 2007;14:924-925.
didemia. Int J Mol Sci. 2011;12:5871-5877. 23. Calitri C, Caviglia I, Cangemi G, et al. Performance of 1,3-be-
16. Ramos JT, Villar S, Bouza E, et al. Performance of a quantitative ta-D-glucan for diagnosing invasive fungal diseases in children.
PCR-based assay and beta-d-glucan detection for diagnosis of inva- Mycoses. 2017;60:789-795.
sive candidiasis in very-low-birth-weight preterm neonatal patients 24. Shabaan AE, Elbaz LM, El-Emshaty WM, Shouman B. Role of
(CANDINEO Study). J Clin Microbiol. 2017;55:2752-2764. serum (1,3)-beta-d-glucan assay in early diagnosis of invasive fun-
17. Goudjil S, Chazal C, Moreau F, Leke A, Kongolo G, Chouaki T. Blood gal infections in a neonatal intensive care unit. J Pediatr (Rio J).
product transfusions are associated with an increase in serum 2018;94:559-565.
(1–3)-beta-d-glucan in infants during the initial hospitalization in 25. Koo S, Bryar JM, Page JH, Baden LR, Marty FM. Diagnostic per-
neonatal intensive care unit (NICU). J Matern Fetal Neonatal Med. formance of the (1→3)-beta-D-glucan assay for invasive fungal dis-
2017;30:933-937. ease. Clin Infect Dis. 2009;49:1650-1659.
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. S U P P O R T I N G I N FO R M AT I O N
19. Cohen JF, Korevaar DA, Altman DG, et al. STARD 2015 guidelines Additional supporting information may be found online in the
for reporting diagnostic accuracy studies: explanation and elabora- Supporting Information section.
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
How to cite this article: Cliquennois P, Scherdel P, Lavergne
Group and the National Institute of Allergy and Infectious Diseases R-A, et al. Serum (1 → 3)-β-D-glucan could be useful to rule out
Mycoses Study Group (EORTC/MSG) Consensus Group. Clin Infect invasive candidiasis in neonates with an adapted cut-off. Acta
Dis. 2008;46:1813-1821. Paediatr. 2021;110:79–84. https://doi.org/10.1111/apa.15321
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
| |
Received: 23 December 2019 Revised: 22 April 2020 Accepted: 24 April 2020
DOI: 10.1111/apa.15331
REGULAR ARTICLE
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
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
1 | I NTRO D U C TI O N
Keynotes
Neonatal hypoxic-ischemic encephalopathy (HIE) in term infants
• Cooling is standard treatment for neonatal hypoxic-
constitutes a serious public health problem which often has life-
ischemic encephalopathy; however, there are limited
long neurological consequences. An estimated 3-5 out of every
resources for cooling equipment in low-middle income
1000 live term births are affected,1 a quarter of which have severe
countries.
symptoms; 10%-30% of affected children do not survive; 25-30%
• We show that phase-changing material (PCM) can in-
of HIE survivors will have long-term neurodevelopmental dis-
duce hypothermia effectively and maintain temperature
abilities that include cerebral palsy, seizure disorder and mental
at target range during 80% of the treatment period.
retardation. 2 The incidence may be 10-fold higher in low-income
• Future studies are needed to determine the feasibility of
countries. 3
PCM during transportation.
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 using a PCM-based cooling device was feasible and safe when
neonatal encephalopathy and that it is safe in an intensive care set- practiced in level 3 NICUs. A PCM-based device was comparable
ting. A prerequisite is that the treatment is started within 6 hours to standard servo-controlled equipment in maintaining the target
after birth and, ideally, as soon as possible.9,10 temperature.16,17
There is still limited experience with hypothermia treatment Based on results from previous studies, we proposed PCMs as a
in low- and middle-income countries, and a great heterogeneity in cooling technique in order to achieve hypothermia without the need
study protocols limits the evaluation.11 Current methods of cooling for electricity or water, and without risk of overcooling in order to
are expensive, need continuous power supply and a range of expen- treat more patients who otherwise would be at risk for HIE damages
sive consumables, and are very difficult to use during transporta- of the brain, and lower the amount of injuries due to asphyxia or
tion,12 a major issue because it is hard to reach a neonatal unit with HIE. The aim of this study was to determine the feasibility of whole
adequate resources for cooling treatment in time. Phase-changing body cooling to 33.5-34.5°C for 72 hours by use of a PCM mattress
materials (PCMs) appear to be ideal for such purposes. PCMs do in a low-resource setting, such as at Vietnam National Children's
not require electricity; they are biologically safe for humans, are Hospital (VNCH).
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 2 | M E TH O DS
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 2.1 | Study participants
is capable of storing or releasing large amounts of energy per unit
weight (or volume) while maintaining a given temperature.14,15 Like This non-randomised intervention study took place in Hanoi be-
water, a PCM has three phases: solid, liquid and steam, although only tween September 2014 and September 2016. Asphyxiated term
the solid-liquid change is used. A PCM substance with a set ‘melting newborn infants from 8 provincial hospitals with an average 2-hour
point’ of, for example 33.5°C will be solid at room temperatures of transport distance from Vietnam National Children's Hospital
~25°C. (Hanoi Obstetrics and Gynecology Hospital, National Hospital
When the human body (37°C) is in contact with a PCM, the PCM of Obstetrics and Gynecology, Ha Dong General Hospital, Thai
will absorb body heat, and the temperature of the PCM itself will Binh Provincial Hospital of Obstetrics and Pediatrics, Hai Duong
rise. When the PCM reaches its melting point, a great deal of addi- Provincial Hospital of Pediatrics, Bac Giang Provincial Hospital
tional energy is absorbed from the body in contact with the PCM, of Obstetrics and Pediatrics, Vinh Phuc Provincial Hospital of
which remains at its specific set temperature until the phase change Obstetrics and Pediatrics, Hung Yen Provincial Hospital) who ful-
eventually occurs. It is this natural property of PCMs that effectively filled criteria for hypothermia treatment after birth asphyxia and
cools contacting bodies for extended periods of time. A PCM works referred to VNCH were included in the study, as there was no in-
as a heat sink and thus stabilises the temperature of whatever it is born delivery at VNCH.
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. 2.2 | Study procedures
Results from the multicentre randomised controlled trial in India
concluded that therapeutic hypothermia of neonates with HIE The study process is summarised in Figure 1.18,19
TRAN et al. |
87
2.3 | Patient recruitment and consent a patient was enrolled, the complete obstetric history was obtained,
and the degree of HIE was determined using Sarnat's clinical stag-
2.3.1 | Referring hospitals ing system (Appendix 1). Basic patient data were recorded, includ-
ing respiratory symptoms, oxygen dependence, oxygen saturation,
Doctors at the obstetric hospitals identified patients potentially capillary refill time, convulsions (clinical and/or seizures on aEEG),
20
eligible for cooling using eligibility criteria, as in the TOBY study vitality (scored as spontaneous activity, alertness, reaction to stimu-
(Table 1): upon being granted verbal consent (yes/no) in which the lation and muscular tonus), need for immediate shock therapy and
parents were explained about the newborn's condition and thera- time from birth until hypothermia treatment was started, seizures
peutic hypothermia treatment, and once contact was made with the before hypothermia treatment and highest registered rectal tem-
on-call doctor at VNCH NICU for a transfer. During transportation, perature before treatment.
passive cooling was used (removing clothes, no heater). 21 Basic data The infant was put on the PCM mattress (Medical Cooling
regarding the mother (pregnancy complications, mother having any Sweden AB; by TST AB) and nursed in a cot naked or lightly covered
sign of infection, type of delivery, time of ruptures of membranes, with a cotton sheet. The PCM mattress that was used contained 8
amniotic fluid check and contact information) and the child (place of PCM packs with a specific melting point of 32°C, built into two lay-
birth, gestational age, weight, resuscitation Y/N, Apgar score, first ers of 4 × 2 PCM sheets (illustrated in Figure 2). The mattress was
gasp in min, whether the infant was heated) was collected. 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-
2.3.2 | At VNCH tum and was connected to a multiparameter monitor to monitor core
temperature. The target rectal temperature was 33.5°C-34.5°C. The
On arrival at VNCH, all patients were evaluated again for eligibility temperature was continuously monitored and recorded every hour
by the NICU doctor. Written consent form was acquired from any for 72 hours. If the temperature decreased below 33.5°C, a folded
one of the parents who was transferred with the infant; in all cases, cloth sheet was inserted between the infant and the PCM mattress,
it was the father because the mother was still at the hospital. Once 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
TA B L E 1 Criteria for cooling 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
Infants ≥ 36 wk gestational age and ≤6 h after birth, and at least one
A criterion and one B criterion radiant warmer. The radiant warmer was used in manual mode with
an output of 10% to begin, and was then adjusted in increments or
Criteria A (perinatal factors)
decrements of 5%, depending on the infant's temperature.
Apgar score ≤ 5 at 10 min
After 72 hours of cooling, the infant was rewarmed by removing
Continued need for resuscitation, 10 min
the mattress and then naturally rewarming in room temperature to
pH < 7.0 and/or base deficit >16 mmol/L (Blood gas within 60 min)
a normal temperature by no more than 0.5°C/h. The radiant warmer
Criteria B (encephalopathy)
was added for rewarming if the targeted temperature of 36°C was
Altered consciousness not reached.
Abnormal tone (hypertonia or hypotonia)
Abnormal primitive reflexes
Exclusion 2.3.3 | Treatment and monitoring
>6 h after birth at time of referral/evaluation; coagulopathy with
active bleeding; prenatally diagnosed syndromes, malformations or During the hypothermia treatment, standard medical care and treat-
metabolic disorders not compatible with survival.
ment of the NICU were given to all patients. During cooling, infants
|
88 TRAN et al.
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)
were sedated with an intravenous infusion of low-dose morphine Demographic and clinical variables were compared using unpaired
10 μg/kg/hour as soon as cooling started. Phenobarbital was admin- comparisons obtained from chi-square for categorical variables and
istered if clinical and/or an aEEG seizure were detected (loading dose t tests for continuous variables.
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 2.6 | Ethics
(TPN) 40 mL/kg/day and nil by mouth for 72 hours. Adjustments for
TPN on successive days were made individually based on patients’ This study followed its procedures in accordance with the ethical
kidney functions, and serum sodium and fluid balances. Kidney and standards of the responsible committee on human experimenta-
liver function, serum sodium, potassium, calcium and magnesium tion and with the Helsinki Declaration of 1975, as revised in 1983,
were tested daily during treatment. The severity of the infant's en- and was approved by the Ethical Review Board of National Hospital
cephalopathy was assessed and scored daily during hospitalisation of Pediatrics Research Institute for Child Health (RICH) (NHP
and at discharge using encephalopathy scores. - RICH- 13-002).
Verbal consent was elicited at the local hospital by the referring
doctor regarding the transfer for possible cooling therapy. Official
2.4 | Outcomes written consent was obtained from the parents when they came to
the neonatal ward of VNCH by the admitting doctor.
2.4.1 | Primary outcome
TA B L E 2 Maternal baseline characteristics of enrolled infants was 5 ± 1.23 hours. The temperature profiles of all 52 infants are
(n = 52) shown in Figures 3 and 4.
Baseline characteristic No % Among the 30 infants that presented with seizures on admis-
sion, seizure was controlled with phenobarbital in 25 cases and 5
Age – y 27 ± 6
remained with seizures until rewarming phase. Cooling had to be dis-
Intrapartum complications
continued for 4 (7.7%) infants because parents decided to withdraw
Foetal heart rate deceleration 10 19
treatment due to poor neurological prognosis based on cerebral
Cord prolapse 2 4
function monitoring (aEEG) and seizure control. Forty-eight infants
Shoulder dystocia 2 4
who completed cooling therapy received brain MRI at 7 days of age.
Emergency ± cesarean delivery 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.
TA B L E 3 Baseline characteristics of enrolled infants (n = 52)
The median hospitalisation time was 12 days (IQR 8-18).
Baseline characteristic No %
IQR 8-18 inborn or transported. The study also pointed out about that PCM
a
cooling was effective only when the ambient temperature was
Even though in the protocol Apgar score and blood gas at 60 min were
<28°C; in this study, we always set the room temperature inside the
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 nursery at 26-27°C and continuously measured and regulated during
infant fulfilled cooling criteria. Blood gas was only available for 32/52 treatment. In the current study, hypothermia treatment was initiated
infants, in 3 out of 6 hospitals possibility for analysing blood gas not in all infants within 6 hours, although it would have been even better
available.
if we could have started within 3 hours, due to reasons mentioned
by Thoresen et al.10
24 (80%) were treated with an anticonvulsant agent, either pheno- Though not a servo-controlled system, it was possible to in-
barbital or midazolam, or both, at referring hospitals. duce hypothermia to reach the target temperature using the
A total of 3744 temperature readings were recorded for 52 PCM mattress within 2.5 (IQR 2-3) hours. Compared to Thayyil
patients during 72 hours of therapeutic hypothermia treatment. et al trial,17 the time taken to reach the target temperature in
Median temperature at initiation of cooling was 35.3 (IQR 34.5- our study was longer, but their temperature at initiation was also
35.9)°C, which was also the temperature on arrival at VNCH as the lower than ours, and the PCM used in their study had lower melt-
patients were immediately placed on PCM. Median time for reaching ing points compared to ours (29°C vs 32°C),22 which could have
the target temperature of 33.5°C was 2.5 (IQR 2-3) hours. caused the differences in effectivity. Both studies required more
The mean ± SD temperature during the cooling phase was time to reach the target temperature compared to the trial using
33.95 ± 0.2°C. Throughout the cooling phase, the target tempera- servo-controlled equipment, as in that conducted by the NICHD
ture range (33.5-34.5°C) was maintained 80% of the time. The oc- (within 90 minutes). 5 On the other hand, we feel that our way of
currence of temperatures reading below 33.5°C, above 34.5° and cooling offers more stability and less of fluctuation than some
below 32°C was 10.21%, 9.23% and 0.58%, respectively. No reading servo-controlled systems. This is because our PCM mattress is in
above 35.5° was recorded within 72 hours for all infants. very close contact with the skin of the patient and PCMs do not
As for the rewarming phase, the rate of rewarming was have a swing in phase as a servo-controlled system does, where
0.5 ± 0.14°C/hour, and the average time for infants to reach 36.5°C the temperature is fed back to the processor of the servo-control
|
90 TRAN et al.
system, which adjusts the temperature of the cooling system to target temperature. Interventions mentioned in the study proce-
reach the intended temperature of the patient, who will then try dures such as inserting a folded cloth sheet between the infant and
to compensate while getting a lower temperature then that which the PCM mattress, covering the infant with another sheet or switch-
is fed back to the servo-control system. Normally, the adjustment ing on the radiant warmer, were required in all cases. In the study by
at the servo-control system is performed every minute.12 Thomas et al, 22 different types of PCM blocks with different melt-
In terms of the ability to maintain the target temperature of ing points (29°C and 21°C) were used for interventions similar to
33.5°C-34.5°C, our results show that the PCM mattress provided those of our study, but with PCM blocks added or removed to adjust
stable body temperatures during cooling and that infants had tem- temperatures. Many other ‘low-tech’ cooling methods, such as ice or
peratures below 32°C only 1% of the time. However, since the PCM frozen gel packs, are also labour intensive, and may result in marked
mattress is a low-technology cooling device, it requires trained temperature fluctuations and shivering, with a potential loss of neu-
nurses and pre-specified interventions in order to maintain the roprotective efficacy. 23–25
TRAN et al. |
91
The rewarm of cooling infants was initiated in a slow and con- it is valuable if used correctly with the right timing. 29,30 However, it
trolled manner (0.5°C/h) using PCMs. The advantage of PCMs over should be used during the whole treatment to be comparable with
cool gel packs is that there is no need for frequent changes, there other studies.
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- 5 | CO N C LU S I O N
ture might be possible using a blanket with PCMs integrated into the
fabric; some non-medical fabrics are now on the market for other Use of a PCM mattress is an effective and low-cost method of cool-
temperature spans; however, their use in a study such as this would ing infants in low- to middle-income countries, and probably during
26
require a medically approved version with our target temperature transportation anywhere. However, as this is not a servo-controlled
Working in a low- to middle-income setting in a country on a system, careful monitoring and good nursing care are needed, es-
fast track towards modern health care in the most advanced hospi- pecially during the induction and rewarming phases. However, this
tals is challenging. Vietnam is one of the fastest growing economies, caveat applies to conventional servo-controlled systems as well.
and technology and knowledge are rapidly changing. The staff size Further studies are needed to look at the influence of the environ-
at hospitals is always limited, and the staff members need to learn mental temperature on the performance of PCM mattresses and to
how to operate new equipment and to increase general medical determine their feasibility during transportation. Larger studies in
knowledge at the same time as they are busy handling the ward. other clinical settings are also needed to detect more unusual com-
Also, what was needed and recorded in medical journals at the be- plications and handling problems with the method. Further, the op-
ginning of the study is not the same as in the later phases. This is timal usage of PCM mattresses (if used for other brief terms, such
even more noticeable in referral hospitals. The theoretic knowledge as for short-distance transportation, which is a potential follow-up
at the referral hospitals and the receiving hospital, and the amount study to this one) and the optimal melting point for newborn infants
of patient history obtained at admission, improved during the study. need to be addressed in depth.
Examples of other challenges are that the ward where the study
was performed had a severe shortage of doctors and nurses who AC K N OW L E D G E M E N T S
were caring for 150-200 neonatal patients in a small and crowded This work was undertaken at VNCH, which received a proportion
space designed for 80 patients. The ward also had various types of funding from the Swedish Research Council (VR), the Swedish
of medical equipment from different donors that were often being Foundation for International Cooperation in Research and Higher
used without appropriate training, which created handling prob- Education, STINT (through the TRAC Collaboration Sweden-
lems, a problem frequently observed in NICUs around the globe. Vietnam). The authors would like to thank Medical Cooling Sweden
The situation improved when a part of one room was reserved for for supplying the PCM mattress used in this study and all the clini-
hypothermia that could be isolated from air conditioning units and cal staff, from the director of the hospital to the nurses and clean-
fans. The standard practice of care in Vietnam is that severely ill ing staff of the Neonatal and Neurology departments at VNCH, for
neonates are isolated from their parents due to practical reasons, all their hard work. We would also like to acknowledge the support
mainly limited space and crowded wards. The goal is to have care- of Karolinska Institutet, the Research Institute for Child Health and
takers more involved and closer to their newborns, as many studies Vietnam National Children's Hospital for giving us time to do im-
have shown that the first connection between mother and child is portant research that might save lives or help newborns with HIE to
of critical importance 27,28 enjoy a better long-term outcome.
The study has some limitations. The nursing interventions during
which it was discovered that the temperature readings were out of C O N FL I C T O F I N T E R E S T
target range were not objectively measured. Our study was not de- None of the authors have any conflict of interest to present.
signed to incorporate variations in endogenous thermogenesis or
environmental temperatures. A different ambient temperature, es- ORCID
pecially in cases of diurnal temperature swing, may also affect the Hang T. T. Tran https://orcid.org/0000-0003-0254-2535
efficacy, stability and life expectancy of PCMs. These factors have Tobias Alfvén https://orcid.org/0000-0002-2328-3512
to be carefully considered when interpreting the findings. Further, Linus Olson https://orcid.org/0000-0003-0046-6348
the effective cooling duration of the PCM mattress and its optimal
melting point for newborn infants needs to be addressed more thor- REFERENCES
oughly; shielding the patients with a transparent cover that keeps 1. Kurinczuk JJ, White-Koning M, Badawi N. Epidemiology of neona-
the infants on the mattress and also reflects outside temperature tal encephalopathy and hypoxic-ischaemic encephalopathy. Early
Hum Dev. 2010;86(6):329-338.
swings better will be done during follow-up studies.
2. Vannucci RC, Perlman JM. Interventions for perinatal hypoxic-isch-
We could not use the eEEG/CFM as an inclusion criteria, since it emic encephalopathy. Pediatrics. 1997;100(6):1004-1014.
was not available at the local hospitals and could not be used during 3. Lawn JE, Cousens S, Zupan J. 4 million neonatal deaths: when?
transportation. The evidence of eEEG/CFM from other studies shows Where? Why? Lancet. 2005;365(9462):891-900.
|
92 TRAN et al.
4. Gluckman PD, Wyatt JS, Azzopardi D, et al. Selective head cooling 20. Azzopardi D, Brocklehurst P, Edwards D, et al. The TOBY study.
with mild systemic hypothermia after neonatal encephalopathy: Whole body hypothermia for the treatment of perinatal asphyx-
multicentre randomised trial. Lancet. 2005;365(9460):663-670. ial encephalopathy: a randomised controlled trial. BMC Pediatr.
5. Shankaran S, Laptook AR, Ehrenkranz RA, et al. Whole-body hypo- 2008;8:17.
thermia for neonates with hypoxic-ischemic encephalopathy. N Engl 21. Hallberg B, Olson L, Bartocci M, Edqvist I, Blennow M. Passive in-
J Med. 2005;353(15):1574-1584. duction of hypothermia during transport of asphyxiated infants: a
6. Jacobs SE, Berg M, Hunt R, Tarnow-Mordi WO, Inder TE, Davis risk of excessive cooling. Acta Paediatr. 2009;98:942-946.
PG. Cooling for newborns with hypoxic ischaemic encephalopathy. 22. Thomas N, Abiramalatha T, Bhat V, et al. Phase changing ma-
Cochrane Database Syst Rev. 2013;(1):Cd003311. terial for therapeutic hypothermia in neonates with hypoxic
7. Schulzke SM, Rao S, Patole SK. A systematic review of cooling for ischemic encephalopathy - a multi-centric study. Indian Pediatr.
neuroprotection in neonates with hypoxic ischemic encephalopa- 2018;55(3):201-205.
thy - are we there yet? BMC Pediatr. 2007;7:30. 23. Bharadwaj SK, Bhat BV. Therapeutic hypothermia using gel packs
8. Shah PS, Ohlsson A, Perlman M. Hypothermia to treat neonatal for term neonates with hypoxic ischaemic encephalopathy in re-
hypoxic ischemic encephalopathy: systematic review. Arch Pediatr source-limited settings: a randomized controlled trial. J Trop Pediatr.
Adolesc Med. 2007;161(10):951-958. 2012;58(5):382-388.
9. Thoresen M. Cooling the asphyxiated brain - ready for clinical tri- 24. Horn AR, Woods DL, Thompson C, Eis I, Kroon M. Selective ce-
als? Eur J Pediatr. 1999;158(Suppl 1):S5-S8. rebral hypothermia for post-hypoxic neuroprotection in neonates
10. Thoresen M, Tooley J, Liu X, et al. Time is brain: starting therapeu- using a solid ice cap. S Afr Med J. 2006;96(9 Pt 2):976-981.
tic hypothermia within three hours after birth improves motor out- 25. Shabeer MP, Abiramalatha T, Smith A, et al. Comparison of two low-
come in asphyxiated newborns. Neonatology. 2013;104(3):228-233. cost methods of cooling neonates with hypoxic ischemic encepha-
11. Montaldo P, Pauliah SS, Lally PJ, Olson L, Thayyil S. Cooling in a lopathy. J Trop Pediatr. 2017;63(3):174-181.
low-resource environment: lost in translation. Semin Fetal Neonatal 26. Shanghai Cloud Chemical Co. L. [cited 2019. PCM fabric]. https://
Med. 2015;20(2):72-79. cloudchem.en.made-in-china.com/produc t/NjzxvnEhEKk Z/China
12. Robertson NJ, Kendall GS, Thayyil S. Techniques for therapeutic -Ice-Cold-Instan t-Coolin g-Towel- Runnin g-Sports-Phase- Chang
hypothermia during transport and in hospital for perinatal asphyxial e-Material.html. Accessed May 13, 2018.
encephalopathy. Semin Fetal Neonatal Med. 2010;15(5):276-286. 27. Angelhoff C, Blomqvist YT, Sahlén Helmer C, et al. Effect of
13. Prashantha YN, Suman Rao PN, Nesargi S, Chandrakala BS, Balla skin-to-skin contact on parents' sleep quality, mood, parent-in-
KC, Shashidhar A. Therapeutic hypothermia for moderate and se- fant interaction and cortisol concentrations in neonatal care
vere hypoxic ischaemic encephalopathy in newborns using low-cost units: study protocol of a randomised controlled trial. BMJ Open.
devices – ice packs and phase changing material. Paediatr Int Child 2018;8(7):e021606.
Health. 2019;39(4):234-239. 28. Csaszar-Nagy N, Bokkon I. Mother-newborn separation at birth
14. Iwata S, Iwata O, Olson L, et al. Therapeutic hypothermia can be in hospitals: a possible risk for neurodevelopmental disorders?
induced and maintained using either commercial water bottles or a Neurosci Biobehav Rev. 2018;84:337-351.
"phase changing material" mattress in a newborn piglet model. Arch 29. Bruns N, Blumenthal S, Meyer I, Klose-Verschuur S, Felderhoff-
Dis Child. 2009;94(5):387-391. Muser U, Muller H. Application of an amplitude-integrated
15. Yang G, Yim YJ, Lee JW, Heo YJ, Park SJ. Carbon-filled organic EEG monitor (Cerebral function monitor) to neonates. J Vis Exp.
phase-change materials for thermal energy storage: a review. 2017;127.
Molecules. 2019;24(11):E2055. 3 0. de Vries LS, Hellstrom-Westas L. Role of cerebral function monitor-
16. Higgins RD, Raju T, Edwards AD, et al. Hypothermia and other ing in the newborn. Arch Dis Child Fetal Neonatal Ed. 2005;90(3):F2
treatment options for neonatal encephalopathy: an executive sum- 01-F207.
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-
How to cite this article: Tran HTT, Le HTT, Tran HTP, et al.
natal encephalopathy using phase changing material. Arch Dis Child.
2013;98(3):F280-F281.
Hypothermic treatment for neonatal asphyxia in low-resource
18. Age & Stages questionnaires. https://agesandstages.com. Accessed settings using phase-changing material—An easy to use and
October 9, 2018. low-cost method. Acta Paediatr. 2021;110:85–93. https://doi.
19. Haataja L, Mercuri E, Regev R, et al. Optimality score for the neuro- org/10.1111/apa.15331
logic examination of the infant at 12 and 18 months of age. J Pediatr.
1999;135(2 Pt 1):153-161.
TRAN et al. |
93
APPENDIX 1
DOI: 10.1111/apa.15339
REGULAR ARTICLE
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
1 | I NTRO D U C TI O N between the relevant parties so that they can exchange information,
ask questions and share views.
Communication between doctors and their patients and families The environment of the neonatal intensive care (NICU) presents
1
is one of the most important elements of medical care for three healthcare staff and parents with particular challenges with regard
key reasons. Firstly, it enables shared decision-making.1,2 Secondly, to communication and information. It is important that parents are
it creates good inter-personal relationships, which is important be- kept informed while their baby is in the NICU, because they are con-
cause the perceived quality of these relationships can be regarded fronted with feelings of shock and stress and need to cope with fear
as a prerequisite for medical care. Thirdly, it promotes interaction and to deal with anticipation and uncertainty.3-5 The parents’ emo-
tions have consequences with regard to their ability to concentrate,
Abbreviations: NICU, neonatal intensive care unit.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
|
94
wileyonlinelibrary.com/journal/apa Acta Paediatrica. 2021;110:94–100.
GAELLE et al. |
95
2.1 | Design 3 | R E S U LT S
This was a qualitative exploratory study, conducted with doctors A total of five doctors (three women) with an average age of 43 years
practising in one French NICU. Each physician used a digital dicta- old agreed to participate: three had more than 10 years of experience
phone to record their usual discussions with parents of newborn in the NICU, and the mean NICU experience of the five doctors was
infants hospitalised in the unit over a period of four months. The 14 years. A total of 40 recordings were made between October 1,
reasons for the discussions were left to the discretion of the doc- 2015, and January 31, 2016. There was an average of 1.6 ± 1.0 re-
tors, who did not use a standardised discussion grid. The aim was to cordings for each of the 26 children, according to the duration of their
capture the doctors’ natural communication. 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
2.2 | Data analysis 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 recorded discussions were fully transcribed verbatim and an- the characteristics of interviews in Table 2.
onymised. Several items were collected: the place of the discussion, The content analysis (Table 3) shows that five main themes
the parents’ socio-professional category, which parents partici- emerged from the analysis. These were as follows: building under-
pated in the discussion, the gestational age of the newborn infant, standing, how the communication was constructed, the role of the
the infant's clinical status at the time of any discussion and their doctor, and of the parents, in the overall care of the newborn infant
clinical status at the end of their NICU stay. Each discussion was 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 6 (23.1)
did you understand about the situation?’
The doctors use the information given upstream and sum-
Gestational age, mean (SD) (weeks) 31 (± 4.6)
marise the essentials. ‘What we said with your husband was the
Male gender, n (%) 13 (50)
context of the birth and the fact that she was resuscitated at birth
Singletons, n (%) 15 (57.6)
with the need to give her a cardiac massage, which is very bad
Maternal age, mean (SD) (years) 30.5 (± 5.2)
actually for babies as small. I think that's the information you had
Parents in low socio-professional category, n (%) 17/22 (77.2)
from Dr…’.
Length of stay (days), median (IQR) 13 (8-23) Throughout the interview, the doctor pauses for a good under-
Clinical status at the end of the stay, n (%) standing of the information given. It will then be able to validate or
Alive 20 (76.9) correct the parent's understanding and continue to deliver the infor-
a
Died 6 (23.1) mation (feedback). ‘Exactly, and now we've seen a brain bleed’.
Number of meeting per child, mean (SD) 1.6 (± 1.02)
Duration of meeting (minimum), median, (IQR) 11 (7.5-18)
a 3.2 | Construction of communication
In five cases, death occurred after treatment was limited or withdrawn;
SD, standard deviation; IQR, interquartile range.
This theme was of concern to doctors throughout the discussions.
Several sub-themes were found to be a necessary part of this element.
TA B L E 2 Characteristics of interviews
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
Place, n (%) doctor said: ‘You had questions already about what was explained
Stable 12 (30)
more. Okay.’
The first two themes were included in all of the discussions, the role 3.3 | Role of the doctor in the global care of the
of the doctor was discussed in three-quarters and the role of the newborn infant
parents appeared in about half of the discussions. The notion of pro-
gressivity of discourse figured in all the discussions. 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-
3.1 | Building understanding ing: ‘We are very careful about how she is, how comfortable she is,
that she does not feel pain, that she does not suffer.’
This theme was of concern to doctors throughout the discussions. The doctor role also appeared to place the medical knowl-
Several sub-themes were found to be a necessary part of this edge and experience of the child and his family in similar situa-
element. tions. ‘Because in these terms, we know, that's what I tell you,
GAELLE et al. |
97
TA B L E 3 Content analysis
Build understanding Upstream knowledge ‘Did she explain to you a little bit about what happened? What did you
understand about the situation?’
Summary of information already ‘What we said with your husband was the context of the birth and the
given 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…’
Make sure you understand You understand? ‘Well, I was clear?’
Validation/ correction of ‘Exactly. There is risk already because he is born at 25 weeks. Okay?’
understanding
Construction of Question search ‘You had questions already about what was explained to you?’
communication Clarity ‘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’.
Reflection time ‘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’
Maintaining the link/availability ‘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?’
Role of the doctor in the Do not harm ‘We are very careful about how she is, how comfortable she is, that she does
global care of the newborn not hurt, that she does not suffer. We are very careful’.
infant Medical knowledge/experience ‘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’ (…)’
Collegiality/equip ‘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’.
Compassion/reassurance/hope/ ‘It's hard too but it's good… we had you explained a little at first… but it's sure
empathy 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…’
Loyalty ‘What I said to you, the decision, you expressed your will, but it's a decision
we had… that will be medical.’
Limit of medicine/humility ‘I sincerely think that the situation… the situation escapes us completely. It
does not make sense to go on…’
Parents' place in the project Role/involvement of parents ‘For you it's really about being patient and wise and coming to see your child
of care of their child every day. It's important.’
Consent/notice wanted ‘What I told you, the decision, you made us by your will but… it was a decision
we had… medical’.
Progressivity of information (Figure 1)
on this kind of very serious haemorrhage, we know there's a bad we do, we will keep you informed about everything that's happening
prognosis’. to your baby …’.
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- 3.4 | Role of the parent in the global care of the
cult situations. One doctor said: ‘Our role is to think about what's newborn infant
best for the baby… With this kind of very serious haemorrhage,
we involve several doctors, because complicated decisions have to This theme was the least represented in the discussions and, when it
be made. These are decisions that can have serious consequences.’ was discussed, it seemed to focus on the role the parents played in
The doctors made it clear that they wanted to provide the most being there for their child and being patient. As one doctor explained
accurate and regular information for the children in their care. As ‘For you it's really about being patient and wise and coming to see
one doctor said: ‘Whatever happens, whatever you see on the exams your child every day. It's important.’
|
98 GAELLE et al.
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 family feel more in control, especially when it came to making deci-
parents needed to be there in just an advisory capacity, but the final sions. Delivering information is a key element of providing neonatal
decisions had to be made by the clinical staff on medical grounds. counselling. Indeed, providing medical information and follow-up
‘What I said to you, the decision, you expressed your will, but it's a allows parents to be involved in their child's decision-making and
decision we had… that will be medical.’ 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
3.5 | Progression of the information been reported that the perception of shared decision-making has
been associated with lower grief scores in the long term, compared
The progression of how the medical information was delivered to the to paternalistic decision-making. 20 However, the results of our study
parents, which was identified during the content analysis process, indicated that the doctors tended to place the greatest emphasis on
is demonstrated in Figure 1. This has led us to propose a number of the information that explained the current medical situation. They
approaches to communication. Although these approaches do not focussed less on delivering this information by exploring the history
cover all eventualities, the basic communication principals are the of the child and parents or their belief systems or cultural or religious
same. references.
The role the doctor played in caring for the child, which was
the third theme of our analysis, was based on non-maleficence,
4 | D I S CU S S I O N 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-
To our knowledge, this was one of the first studies to examine the scribed by Hippocrate, which translates as first do no harm. But,
discussion between doctors to parents in the natural setting of a unlike adult patients, newborn infants cannot express their pref-
NICU. The analysis allowed us to highlight existing communication erences. Its legal representatives are its parents or guardians who
practices or strategies by exploring what information was provided have parental authority. Thus, the ways that doctors communicate
and how it was delivered. with parents and what they tell them are key to establishing a part-
Our analysis showed that how the discussions were constructed nership of care for the child. The language that doctors use when
by the doctors and understood by the parents was very important as discussing medical conditions differs from that of parent and some
it shaped how any information was provided. The goals were to build doctors protect their emotions by using very technical, rational-
trusting relationships, kindness and empathy, and help to make the ising language. Actively listening to parents is essential. Parents
GAELLE et al. |
99
may be more receptive to a doctor's explanations if they feel that discussions and information and their autonomy will be weakened,
their emotions are understood and their patenting role is rec- or even destroyed, if doctors do not make allowances for their
ognised. 23 Trust is built by paying attention to others. It is not just changing views. The way that they react is partly shaped by what
about the team's medical skills, but about their intentions towards medical professionals say to them on an ongoing basis.
the child and its parents and their ability to predict the child's fu- The progression of how information was provided was a key el-
ture and support the child. It would have been very interesting to ement of our analysis. In some ways, it was a bit like a movie script,
explore the parents' views on how this relationship of trust was which tells a story, but also contains technical and organisational de-
built. These situations have a strong emotional impact on medical tails. This progression did not seem to be a matter of chance and the
staff. In addition to providing medical care, they have to take into doctors in our study used it as a real communication strategy. It was
account the values and emotions of the parents when they make used as a way to fight against the uncertainty inherent in the situa-
decisions. There is a close relationship between emotions and val- tion by implementing approaches such as paraclinical examinations.
ues, because the emotions that are felt in a situation reveal the It was also a way of creating a bond with the family and organising
value of a person's ethical principles. 24 Compassion or empathy future meetings in the hope that the clinical picture would become
reveals the principles of doing good to others, doing no harm and clearer with fewer possible scenarios. Lastly, delivering information
autonomy, together with fear and respect. In difficult situations, at various stages seemed to be used as a tool for mitigating informa-
several emotions can be felt at the same time, such as compassion tion, especially when it was considered difficult. The phenomenon
and fear. These can lead to conflicting values between doing good of psychologisation, which corresponds to attributing psychologi-
to others and doing no harm and the best interests of the child cal states to others, can provide the basis of this communicational
often move between these two principles. Possible divergent de- strategy. Progressivity, by essence, implies that the final information
cisions can lead to argumentative discussions, and doctors need cannot be delivered straight away, because it could have a negative
to listen to the views of others. That is why doctors are advised to emotional or psychological effect on the recipient. The information
involve other colleagues when difficult decisions need to be made, needs to be watered down or diluted over the course of several dis-
especially when cases raise ethical issues or they need to make cussions to make it more bearable.
decisions about limiting or withdrawing active treatment. This study had some limitations. The constraints of the medical
Our fourth theme, on the role parents played in the care of activity in the NICU meant that the five doctors who took part over
their child, was less developed in medical discussions, contrary the four-month period spoke to the parents of the same 26 children
to current policies. Until recently, doctors saw their primary role more than once during the 40 discussions. In addition, the reason
as providing parents with detailed, objective information about for the discussions was at the discretion of the individual doctors
treatment choices and likely outcomes. We now have a newer and we cannot neglect the fact that they may have modified how
approach, which recommends that doctors help parents discover the information was delivered because they were recording the con-
their own values and ethical commitments as they face an unan- versations. Finally, the parent's responses were a vital part of the
ticipated situation and a series of life-altering decisions. 25 Parents discussions, but we were unable to use these data.
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- 5 | CO N C LU S I O N
tense emotions following the shock of a birth that did not happen
as expected. In addition, they are suddenly immersed in their new Our study focused on the information that doctors provided parents
role as parents, while most of the time the life of their child de- during their child's NICU stay and also explored how it was provided.
pends on complex interventions and technologies,. 6,26-28 Although This revealed five key themes that can help to inform future commu-
shared decision-making is the ethically preferred approach, as op- nication strategies. These were as follows: building understanding,
posed to pure parental autonomy or pure medical paternalism in how the communication was constructed, the role of the doctor, and
29
the grey zone of ethical ambiguity, most neonatologists do not of the parents, in the overall care of the newborn infant and how
share decision-making well. 30 the information given to the parents developed over time. The latter
The shared decision-making model may seem ideal, but rigid theme, and the model of shared decision-making, is of particular in-
schemes are not adapted to all situations, even when it comes to terest for future communication strategies. The findings also under-
20
reaching this type of decision together. Therefore, it is important lined the importance of making healthcare professionals feel more
that doctors give parents the opportunity to express themselves at comfortable in these sometimes difficult situations. A new study is
different times and after reflection. The parents also need to decide currently underway, concerning the medical information delivered in
what degree of involvement they want. It is important to recognise critical care situations.
that parents are in unfamiliar territory when their baby is in the NICU
and their thoughts and feelings may change as current and future C O N FL I C T S O F I N T E R E S T
situations unfold. Their positions can also be shaped by ongoing 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 16. Koh TH, Butow PN, Coory M, et al. Provision of taped conversa-
tions with neonatologists to mothers of babies in intensive care:
Both physicians have given their consent to participate in this study.
randomised controlled trial. BMJ. 2007;334:28.
An information leaflet was given to the parents of the newborn 17. Flick U. Thematic coding and content analysis. In: Metzler K, ed.
infants by the doctors. They were they could withdraw from the An Introduction to Qualitative Research, 5th edn. London: Sage
research at any time and assured that the identities would remain Publications; 2014:420–438.
18. Arockiasamy V, Holsti L, Albersheim S. Fathers’ experiences in
confidential. Ethical approval for this research was provided by the
the neonatal intensive care unit: a search for control. Pediatrics.
university ethics committee of Aix-Marseille University, France 2008;121:215-222.
(number 2015-06-03-001). 19. Macdonell K, Omrin D, Pytlik K, Pezzullo S, Bracht M, Diambomba
Y. An effective communication initiative: using parents’ experiences
ORCID to improve the delivery of difficult news in the NICU. Journal of
Neonatal Nursing. 2015;21:142-149.
Gaelle Sorin https://orcid.org/0000-0002-5283-9030
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
REFERENCES response. Arch Dis Child Fetal Neonatal Ed. 2013;98:26-31.
1. Ong LM, De Haes JC, Hoos AM, Lammes FB. Doctor-patient commu- 21. Dehan M, Gold F, Grassin M, et al. Ethical dilemmas in the peri-
nication: a review of the literature. Soc Sci Med. 1995;40:903-918. natal period: guidelines for end-of-life decisions. Arch Pediatr.
2. Beisecker AE, Beisecker TD. Patient information-seeking behaviors 2001;8:407-419.
when communicating with doctors. Med Care. 1990;28:19-28. 22. Moriette G, Rameix S, Azria E, et al. Very premature births:
3. Brazy JE, Anderson BM, Becker PT, Becker M. How parents of Dilemmas and management. Second part: Ethical aspects and rec-
premature infants gather information and obtain support. Neonat ommendations. Arch Pediatr. 2010;17:527-533.
Netw. 2001;20:41-48. 23. Boss RD, Hutton N, Sulpar LJ, West AM, Donohue PK. Values par-
4. Charchuk M, Simpson C. Hope, disclosure, and control in the neo- ents apply to decision-making regarding delivery room resuscita-
natal intensive care unit. Health Commun. 2005;17:191-203. tion for high-risk newborns. Pediatrics. 2008;122:583-589.
5. Ward K. Perceived needs of parents of critically ill infants in a neo- 24. Beauchamp TL, Childress JF. Principles of Biomedical Ethics, 6th edn.
natal intensive care unit. Pediatr Nurs. 2001;27:281-286. New York, NY: Oxford University Press; 2009.
6. Hall SL, Cross J, Selix NW, et al. Recommendations for enhancing 25. Gaucher N, Nadeau S, Barbier A, Janvier A, Payot A. Personalized
psychosocial support of NICU parents through staff education and antenatal consultations for preterm labor: responding to mothers’
support. J Perinatol. 2015;35:529-536. expectations. J Pediatr. 2016;178:130-134.
7. De Wit S, Donohue P, Shepard J, Boss RD. Mother–clinician dis- 26. Haines DKJ, Remedois L, Berney SC, Knott DC, Denehy L. The con-
cussions in the neonatal intensive care unit: agree to disagree? J ceptualisation of health-related quality of life in decision-making
Perinatol. 2013;33:278-281. by intensive care physicians: a qualitative inquiry. Aust Crit Care.
8. Bass LS. What do parents need when their infant is a patient in the 2017;30:152-159.
NICU? Neonatal Netw. 1991;10:25-33. 27. Ainbinder JG, Blanchard LW, Singer GH, et al. A qualitative study
9. Brinchman BS, Forde R, Nortvedt P. Parents' experiences with life- of Parent to Parent support for parents of children with special
and-death decisions concerning their premature infants. Tidsskr Nor needs. Consortium to evaluate Parent to Parent. J Pediatr Psychol.
Laegeforen. 2002;122:2098-2101. 1998;23:99-109.
10. Wocial LD. Life support decisions involving imperiled infants. J 28. Bourque CJ, Dahan S, Mantha G, Robson K, Reichherzer M, Janvier
Perinat Neonatal Nurs. 2000;14:73-86. A. Improving neonatal care with the help of veteran resource par-
11. Alderson P, Hawthorne J, Killen M. Parents’ experiences of sharing ents: an overview of current practices. Semin Fetal Neonatal Med.
neonatal information and decisions: consent, cost and risk. Soc Sci 2018;23:44-51.
Med. 2006;62:1319-1329. 29. Rosenbaum L. The paternalism preference choosing unshared deci-
12. Vermeulen E. Dealing with doubt: making decisions in a neonatal sion making. N Engl J Med. 2015;373:589-592.
ward in The Netherlands. Soc Sci Med. 2004;59:2071-2085. 3 0. Lantos J. Ethical problems in decision making in the neonatal ICU.
13. De Rouck S, Leys M. Information needs of parents of children ad- Engl J Med. 2018;379:1851-1860.
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. How to cite this article: Sorin G, Dany L, Vialet R, et al. How
Perceptions of lethal fetal abnormality among perinatal profes- doctors communicated with parents in a neonatal intensive
sionals and the challenges of neonatal palliative care. J Palliat Med.
care: Communication and ethical issues. Acta Paediatr.
2014;17:924-930.
15. Orzalesi M, Aite L. Communication with parents in neonatal inten- 2021;110:94–100. https://doi.org/10.1111/apa.15339
sive care. J Matern Fetal Neonatal Med. 2011;24:135-137.
| |
Received: 11 December 2019 Revised: 30 April 2020 Accepted: 5 May 2020
DOI: 10.1111/apa.15345
REGULAR ARTICLE
1
NICU, Fondazione IRCCS Policlinico San
Matteo, Pavia, Italy Abstract
2
Child Neurology and Psychiatry Unit, IRCCS Aim: Although parenting is key to promoting healthy development of at-risk preterm
Mondino Foundation, Pavia, Italy
infants, parents have often restricted access to neonatal intensive care units (NICUs).
3
0-3 Center for the at-Risk Infant, Scientific
Institute IRCCS E. Medea, Lecco, Italy
This study aimed to assess the effect of an early parenting intervention on the psy-
4
Department of Brain and Behavioral chomotor outcome in preterm children at 24 months of corrected age.
Sciences, University of Pavia, Pavia, Italy Methods: Forty-two preterm children and their parents were consecutively recruited
Correspondence at a level III NICU in Northern Italy and randomly allocated to early intervention
Livio Provenzi, Child Neurology and (two educational peer-group sessions and four individual infant observation ses-
Psychiatry Unit, IRCCS Mondino
Foundation, via Mondino 2, Pavia 27100, sions) or care as usual (no educational or infant observation sessions). During NICU
Italy. stay, parents provided information on daily holding and skin-to-skin. Psychomotor
Email: livio.provenzi@mondino.it
development was measured at 24 months of corrected age using the Griffith Mental
Funding information Development Scales.
This study was supported by funds from the
Italian Ministry of Health, Ricerca Corrente Results: There were no significant differences in socio-demographic and clinical vari-
2016, 2017, 2018, 2019, 2020, and Cinque ables between early intervention (n = 21; 13 females) and care as usual (n = 21; 12
per Mille, 2017.
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
1 | I NTRO D U C TI O N
Key notes
Preterm birth is one of the major challenges for the healthcare sys-
• Preterm-born children are at high risk for suboptimal
tems worldwide, representing the leading cause of long-lasting mor-
psychomotor development in the first years of life.
bidities in infants and children.1 Preterm infants are especially at risk
• An early parenting intervention was successful in pro-
for detrimental developmental consequences, mainly due to their
moting better psychomotor outcomes in preterm chil-
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
• Parenting interventions are needed to be implemented
intensive care unit (NICU) often results in early separation from the
precociously in order to improve the psychomotor out-
parents, thus making it difficult to establish immediate bonding and
comes of preterm children.
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- control group exposed to usual NICU care without any dedicated
velopmental outcomes.4 parenting support, the children of parents who received the early
The NICU environment can be especially stressful for the par- intervention would show better psychomotor development.
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 2 | M E TH O DS
early separation from the infant and the need of specialised intensive
medical and nursing actions may lead the mothers and the fathers to 2.1 | Participants
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 Between January 2015 and December 2016, 42 preterm children
approaches are promoted to increase the physical proximity and from the NICU of the Fondazione IRCCS Policlinico San Matteo,
emotional closeness between parents and infants, and to facilitate Pavia, Italy, and their parents were enrolled in this study. All preterm
10-12
attachment processes that are crucial for healthy development. infants with gestational age ≤32 weeks or birthweight ≤1500 grams
Specifically, the parental caregiving ability to read infants’ signals, were eligible for the study. Subjects were not considered eligible to
respond contingently and provide support to emotional regulation is the study if at least one parent was unable to converse in Italian,
early interactive behaviours that may be challenged in the NICU.13-15 had psychiatric morbidities or reported alcohol/drug abuse, was less
Moreover, it should be highlighted that the needs of both mothers than 18 years old, and if the infants had any genetic syndrome. A
and fathers have be addressed in NICU to support their engagement complete flowchart of enrolment is reported in Figure 1. Twenty-
in the early caregiving and to promote an adaptive post-discharge one children and their parents were allocated to the early interven-
transition to home.16-18 tion arm, whereas twenty-one were allocated to the care as usual
As such, early interventions that focus on providing information arm. In order to minimise the risk of biases related to the interactions
and psychoeducational guidelines for parents should be prioritised among parents belonging to different groups, allocation occurred
in the NICU environment.19 Mother-focused parenting programmes in two blocks: care as usual allocation occurred between January
have been found to be successful in improving the cognitive and 2015 and May 2016, whereas early intervention allocation occurred
behavioural outcomes of preterm-born children up to five years between July and December 2016, with a wash-out period of one
of age. 20 When maternal involvement in caregiving is greatly sup- month during June 2016. Enrolment occurred as soon as the infant
ported, this can lead to increased scores in the psychomotor profile was not considered at risk for survival. The study was approved
of preterm children at 18 months of age. 21 A meta-analytic study by the Ethical Committee of the Fondazione IRCCS Policlinico San
suggests that early parenting interventions in the NICU are suc- Matteo and the IRCCS Mondino Foundation in Pavia, Italy. All the
cessful in promoting the psychomotor developmental outcomes of parents signed an informed consent.
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- 2.2 | Study design and procedures
sults are promising, previous research in this field has mainly focused
on parenting interventions delivered to mothers rather than to both Figure 2 reports an overview of the study design. The parents of chil-
the parents of preterm infants. 23,24 dren allocated to the early intervention group participated in group
In the present study, we assessed the effects of an early and brief and individual sessions (Table 1). A detailed description of the interven-
parenting intervention that includes both psychoeducational support tion is reported in Table S2, according to the Template for Intervention
on the psychomotor developmental outcomes of preterm children at Description and Replication (TIDieR) checklist. Two psychoeducational
24 months of corrected age. We hypothesised that, compared to a 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 communicative signals, signs of distress and regulation processes. The
provided information on typical characteristics of preterm infants, spe- joint observation sessions were informed by the methodological ap-
cial needs in caregiving and the procedures used by the NICU staff. proach of Dr Boukydis, namely the collaborative consultation with
The group sessions were interspersed with up to four individual ses- Parents and Infants in the Perinatal Period,25 which integrates princi-
sions that were led by a trained neurodevelopmental therapist. These ples from evidence-based infant research and well-validated neurobe-
sessions lasted approximately 45 minutes and consisted of a joint havioural assessments. These principles included both guidelines for
observation of the infant by focusing on behaviours, temperament, the joint observation of the infant and communicative skills to support
|
104 PISONI et al.
parenting. On the one hand, the therapists were trained to focus on two groups. Mothers and fathers completed the diary separately.
maternal regulatory behaviours, infant communicative signals, mutual The parental closeness diary allows the collection of quantitative
dyadic processes, sensitive caregiving and intrusive behaviours. On data on the presence of parents, time spent holding the infant and
the other hand, specific relational skills included asking examples from time spent in skin-to-skin contact. This instrument has been used in
the parents, elicit their own view of the infant and themselves in the previous research with NICU parents. 26 The holding time score was
interaction, active listening by the therapist, and stimulating parents’ computed as the percentage of the total time spent by parents hold-
curiosity about infants’ signals and behaviour. The parents allocated to ing the infant with respect to the time they were present in the unit.
the care as usual group received no structured parenting support, and The skin-to-skin time score was computed as the percentage of the
there were no differences in how the NICU staff cared for the infants. total time spent by parents in skin-to-skin contact with respect to
Any parents of the care as usual group who reported high levels of psy- the time they were present in the unit. The absolute values (in min-
chological distress could be referral for a psychotherapist consultation. utes) for presence, holding and skin-to-skin are reported in Table S1.
chi-square test (dummy variables). General linear models were may speculate that promoting parental knowledge through a psy-
used to assess the presence of significant group differences in choeducational training aimed at improving caregiving skills may
the total and domain DQ scores of the GMDS. Statistical analyses benefit the quality of early parent-infant relationship so that the
were carried out using SPSS Statistics for Windows, version 25 infant receives adequate stimulations and contingent responses.
(SPSS Inc.) setting P < .05. 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
3 | R E S U LT S performance domains further highlights the need to develop more
integrated programmes capable of acting on multiple levels of the
Descriptive statistics for the socio-demographic and clinical charac- developmental trajectories of the preterm child. Consistently, future
teristics as well as for holding and skin-to-skin percentage scores are research should explore the role of parental engagement in physical
reported in Table 2. There were no significant differences between therapies for preterm infants and children.
the two groups for any of these variables. Apgar score at the first Notably, this intervention contributes to the existing literature
(W = 381.00, P > .05) and fifth minute (W = 432.50, P > .05) did not on the importance of early family-centred interventions in the
significantly differ between early intervention. NICU.12,29,31 The intervention reported here moves from theoreti-
A significant difference emerged for the total GMDS score cal and methodological principles of the collaborative consultation
(t(40) = 2.40, P = .02) (Figure 3). Children from parents in the early approach and adopted both psychoeducational group sessions and
intervention arm had higher global DQ compared to those in the tailored joint infant observation sessions with parents. The main goal
care as usual arm. Moreover, a significant multivariate effect was de- of the intervention was to increase parental knowledge of factors
tected (F (5,36) = 4.49, P = .003, η2p = 0.38). Significant univariate dif- related to infant's development and to improve both mothers and
ferences emerged for two domains: personal-social (F (1,40) = 14.93, fathers’ skills in reading the baby's signals, responding contingently
2
P < .001, η p = 0.27) and hearing-speech (F (1,40) = 5.36, P = .026, and supporting emotional regulation. As previously suggested,33,33
2
η p = 0.12). For both the domains, children of parents in the early investing resources in improving parental involvement in NICU is a
intervention group had higher DQ scores compared to those in the critical protective factor for the development of preterm infants and
care as usual group. These effects persisted even when the analyses children. Although these findings promisingly support the relevant
were restricted to children without cerebral palsy. 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-
4 | D I S CU S S I O N tive preterm samples in future research.
This study had limitations. First, the sample size was relatively
In the present study, we report on the effects of an early NICU small. As such, we could not include more complex models in our
psychoeducational parenting intervention including both moth- statistical plan to control for potential confounders such as life
ers and fathers on the psychomotor outcome of preterm infants at events, parental educational style and socio-economic conditions.
24 months of CA. Children of parents who took part in the early Nonetheless, the two groups were not significantly different in
intervention programme had greater global DQ and higher scores terms of socio-demographic and clinical characterisation, which
in the personal-social and hearing-speech domains of the GMDS indirectly indicates the robustness of the study and the reliability
assessment. of findings. Second, parenting a preterm infant is often associated
Previous research has showed that early interventions that ac- with increased levels of stress, depression and anxiety.30,34,35 We
tively involve the parents in informative and/or psychoeducational could not exclude that affective problems may have had a role in
programmes have the greatest long-term efficacy in reducing det- moderating the effects of the intervention on children outcomes.
rimental psychomotor and cognitive outcomes. 27 Moreover, the For example, parents with significant depressive and/or anxious
effect of maternal caregiving support may be especially observed symptomatology may be less able to engage in psychoeducational
between 24 and 36 months, with the greatest benefits reported for programmes focused on parent-child interaction. Third, the selected
socio-emotional and behavioural domains. 22 Our findings further sample included low-risk families, as single parents and those with
extend previous evidence and suggest that engaging both parents psychiatric morbidities were not eligible for the study. As such, our
in an early and brief parenting intervention by providing informa- findings only have partial generalisability to the broader population
tive support and psychoeducational guidance in understanding the of families of preterm infants hospitalised in the NICU. Finally, the
infant's communicative and behavioural signals may be significantly domain specificity of these findings indirectly suggests that a po-
beneficial for the psychomotor development of preterm children. tential process underlining the effectiveness of the intervention may
The finding that parental support may be more effective in im- be ascribed to the improved daily quality of parent-child interaction
proving social-related domains, rather than motor or physical com- or at least to sensitive and protective parenting behaviour. 22 As
28
petences, resonates with previous reports in older children. This such, we suggest that future studies assessing the psychomotor out-
specific effect may suggest pathways of neuroprotection. First, one 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
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. |
107
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
url?eid=2-s2.0-84888148811&partne rID=40&md5=893a07b4ba
video-recording of parent-child dyadic and triadic interactions to d2d617c8ac831e5dd62cd1
2. Lester BM, Miller RJ, Hawes K, et al. Infant neurobehavioral devel-
better highlight potential relational mechanisms of action.
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-
5 | CO N C LU S I O N 282. Available from http://www.scopus.com/inward/record.
url?eid=2-s2.0-0036581688& partne rID=40&md5=5d70e8 ec34
daa6f5552e4a94aa6a11cd
We found that a brief and early psychoeducational parenting inter- 4. Montirosso R, Provenzi L. Implications of epigenetics and stress
vention during the NICU stay was significantly associated with im- regulation on research and developmental care of preterm infants.
proved psychomotor outcome in children at 24 months of corrected JOGNN - J Obstet Gynecol Neonatal Nurs. 2015;44:174-182.
age. Social-related domains were significantly improved, suggesting 5. Muller-Nix C, Forcada-Guex M, Pierrehumbert B, Jaunin L, Borghini
A, Ansermet F. Prematurity, maternal stress and mother–child
that the potential pathways of protective action may be related to
interactions. Early Hum Dev. 2004;79:145-158. Available from
improvements in the quality of parent-child relationships in the first ht tps://linkin ghub.elsevi er.com/retrie ve/pii/S 03783 7820
months of life. As parenting is advocated to be the major preventive 4000775
factor for the child health and well-being,36 these findings suggest 6. Pisoni C, Spairani S, Fauci F, Ariaudo G, Tzialla C, Tinelli C, et al.
Effect of maternal psychopathology on neurodevelopmental out-
that investing in early parenting engagement and caregiving skills
come and quality of the dyadic relationship in preterm infants: an
may be crucial to further promote better developmental outcomes explorative study. J Matern Neonatal Med. 2020;33(1):103-112.
in preterm children. 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://
C O N FL I C T S O F I N T E R E S T
www.scopus.com/inward /record.url?eid=2-s2.0-848557 3344
The authors have no conflicts of interest to declare. 2&partnerID=MN8TOARS
8. Aftyka A, Rozalska I, Rybojad B, Samardakiewicz M. Polish version
ORCID of the Parental Stressor Scale: neonatal intensive care unit. Ann
Agric Environ Med. 2019;26:67-72. Available from http://www.ncbi.
Camilla Pisoni https://orcid.org/0000-0001-9292-9682
nlm.nih.gov/pubmed/30922032
Livio Provenzi https://orcid.org/0000-0001-7424-8744 9. Franck LS, Cox S, Allen A, Winter I. Measuring neonatal intensive
Rosario Montirosso https://orcid.org/0000-0002-2175-3082 care unit-related parental stress. J Adv Nurs. 2005;49:608-615.
Renato Borgatti https://orcid.org/0000-0001-8165-4994 10. Hoffenkamp HN, Tooten A, Hall RAS, et al. The impact of prema-
ture childbirth on parental bonding. Evol Psychol. 2012;10:542-561.
Simona Orcesi https://orcid.org/0000-0001-8947-9342
11. Provenzi L, Fumagalli M, Bernasconi F, et al. Very preterm and full-
term infants’ response to socio-emotional stress: the role of post-
REFERENCES natal maternal bonding. Infancy. 2016;1-18. Available from http://
1. Blencowe H, Cousens S, Chou D, et al. Too Soon: The global epide- doi.wiley.com/10.1111/infa.12175
miology of 15 million preterm births. Reprod Health. 2013;10(Suppl 12. Welch MG, Hofer MA, Brunelli SA, et al. Family nurture interven-
1). Available from http://www.scopus.com/inward/record. 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.
pubmedcentr al.nih.gov/articlerender.fcgi?artid=3394087&tool=p- 27. Spittle A, Orton J, Anderson P, Boyd R, Doyle LW. Early develop-
mcentrez&render type=abstract mental intervention programmes post-hospital discharge to pre-
13. Fumagalli M, Provenzi L, De Carli P, et al. From early stress to vent motor and cognitive impairments in preterm infants. Spittle
12-month development in very preterm infants: preliminary find- A, ed. Cochrane database Syst Rev. 2012;12:CD005495 http://doi.
ings on epigenetic mechanisms and brain growth. PLoS ONE. wiley.com/10.1002/14651858.CD005495.pub3
2018;13:e0190602. Available from http://www.ncbi.nlm.nih.gov/ 28. Johnson S, Ring W, Anderson P, Marlow N. Randomised trial of pa-
pubmed/29304146 rental support for families with very preterm children: outcome at 5
14. Feldman R. Infant-mother and infant-father synchrony: the co- years. Arch Dis Child. 2005;90:909-915.
regulation of positive arousal. Infant Ment Health J. 2003;24:1- 29. Als H, McAnulty GB. The Newborn Individualized Developmental
23. Available from http://www.scopus.com/inward/record. Care and Assessment Program (NIDCAP) with Kangaroo Mother
url?eid=2-s2.0-354308 3158& partne rID=40&md5=55b1ad 3a63 Care (KMC): comprehensive care for preterm infants. Curr Womens
086f071486188b9fc312e5 Health Rev. 2011;7:288-301. http://www.ncbi.nlm.nih.gov/pubme
15. Tronick E, Beeghly M. Infants' meaning-making and the develop- d/25473384
ment of mental health problems. Am Psychol. 2011;66:107-119. 3 0. Ahlqvist-Björkroth S, Axelin A, Korja R, Lehtonen L. An educational
Available from http://psycnet.apa.org/journals/amp/66/2/107/ intervention for NICU staff decreased maternal postpartum de-
16. Provenzi L, Santoro E. The lived experience of fathers of preterm pression. Pediatr Res. 2019;85(7):982-986.
infants in the neonatal intensive care unit: a systematic review 31. Provenzi L, Barello S. Behavioral Epigenetics of Family-Centered
of qualitative studies. J Clin Nurs. 2015;24(13-14):1784-1794. Care in the Neonatal Intensive Care Unit. JAMA Pediatr.
Available from http://doi.wiley.com/10.1111/jocn.12828 2015;169(7):697-698. https://jamanetwork.com/journals/jamap
17. Guillaume S, Michelin N, Amrani E, et al. Parents' expectations of ediatrics/article-abstract/2337233
staff in the early bonding process with their premature babies in the 32. Montirosso R, Casini E, Del Prete A, et al. Neonatal developmental
intensive care setting: a qualitative multicenter study with 60 par- care in infant pain management and internalizing behaviours at 18
ents. BMC Pediatr. 2013;13:18. Available from http://www.biome months in prematurely born children. Eur J Pain (United Kingdom).
dcentral.com/1471-2431/13/18 2016;20:1010-1021.
18. Obeidat HM, Bond EA, Callister LC. The parental experience of 33. Montirosso R, Del Prete A, Bellu R, et al. Level of NICU quality
having an infant in the newborn intensive care unit. J Perinat Educ. of developmental care and neurobehavioral performance in very
2009;18:23-29. preterm infants. Pediatrics. 2012;129:e1129-e1137 http://www.
19. Hall SL, Cross J, Selix NW, et al. Recommendations for enhanc- pubmedcentr al.nih.gov/articlerender.fcgi?artid= 4074610&tool=p-
ing psychosocial support of NICU parents through staff educa- mcentrez&render type=abstract
tion and support. J Perinatol. 2015;35(Suppl 1):S29-S36. Available 3 4. Shields-Poë D, Pinelli J. Variables associated with parental stress
from http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid in neonatal intensive care units. Neonatal Netw. 1997;16:29-
=4660046&tool=pmcentrez&render type=abstract 37. Available from http://www.scopus.com/inward/record.
20. Nordhov SM, Rønning JA, Ulvund SE, Dahl LB, Kaaresen PI. Early url?eid=2-s2.0-0031062943& partne rID=40&md5=6be9349e67
intervention improves behavioral outcomes for preterm infants: 1dde375f2d854cdb6c95aa
randomized controlled trial. Pediatrics. 2012;129(1):e9-e16. 35. Pisoni C, Spairani S, Manzoni F, Ariaudo G, Naboni C, Moncecchi
21. Lester BM, Hawes K, Abar B, et al. Single-family room care and neu- M, et al. Depressive symptoms and maternal psychological distress
robehavioral and medical outcomes in preterm infants. Pediatrics. during early infancy: A pilot study in preterm as compared with
2014;134:754-760. Available from http://pediatrics.aappublica term mother–infant dyads. J Affect Disord. 2019;257:470-476.
tions.org/cgi/doi/10.1542/peds.2013-4252 36. Perrin EC, Leslie LK, Boat T. Parenting as Primary Prevention. JAMA
22. Vanderveen J a, Bassler D, Robertson C m t, Kirpalani H. Early Pediatr. 2016;170:637. Available from http://www.ncbi.nlm.nih.
interventions involving parents to improve neurodevelopmen- gov/pubmed/27182902
tal outcomes of premature infants: a meta-analysis. J Perinatol. 37. ProvenziLivio, Barello Serena. Behavioral Epigenetics of Family-
2009;29(5):343-351. Available from 10.1038/jp.2008.229%5Cnht Centered Care in the Neonatal Intensive Care Unit. JAMA Pediatrics.
tp://www.nature.com/doifinder/10.1038/jp.2008.229 2015;169(7):697–698. https://jamanetwork.com/journals/jamap
23. Fisher D, Khashu M, Adama EA, et al. Fathers in neonatal units: ediatrics/article-abstract/2337233
improving infant health by supporting the baby-father bond and
mother-father coparenting. J Neonatal Nurs. 2018;24:306-312.
Available from https://linkinghub.elsevier.com/retrieve/pii/S1355 S U P P O R T I N G I N FO R M AT I O N
184118300930
Additional supporting information may be found online in the
24. Feeley N, Waitzer E, Sherrard K, Boisvert L, Zelkowitz P. Fathers'
perceptions of the barriers and facilitators to their involvement Supporting Information section.
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 How to cite this article: Pisoni C, Provenzi L, Moncecchi M,
the perinatal period. Infant Ment Health J. 2015;36:240-241. et al. Early parenting intervention promotes 24-month
26. Raiskila S, Axelin A, Toome L, et al. Parents' presence and parent–
psychomotor development in preterm children. Acta Paediatr.
infant closeness in 11 neonatal intensive care units in six European
countries vary between and within the countries. Acta Paediatr Int J 2021;110:101–108. https://doi.org/10.1111/apa.15345
Paediatr. 2017;106:878-888.
| |
Received: 5 December 2019 Revised: 30 April 2020 Accepted: 12 May 2020
DOI: 10.1111/apa.15354
REGULAR ARTICLE
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.
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
1 | I NTRO D U C TI O N
Key Notes
Very premature birth is associated with a high risk of serious com-
• Early diagnosis of enteropathies in very preterm neo-
plications, including gastrointestinal diseases, that is enteropathies,
nates is challenging, leading to unnecessary enteral
which the most severe form is necrotising enterocolitis (NEC), one
feeding discontinuation and prolonged courses of
of the leading cause of morbidity and mortality in neonatal intensive
antibiotics.
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
in children and adults, showed highly variable levels
enteropathy (Bell's stages Ia and Ib, suspected NEC) can be as high
in very preterm neonates, limiting its use to predict
as 44%.3 The severe form, that is definite NEC (Bell's stages IIb to
enteropathy.
IIIb), has an incidence depending on the country, from 2% to about
• By contrast with necrotising enterocolitis, the most se-
10% in extremely preterm neonates, associated with a mortality up
vere form of enteropathies, gut microbiota did not show
to 50%.1,4
a specific profile linked to milder forms.
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 2 | PATI E NT S A N D M E TH O DS
uncertainty leads to frequent unnecessary and extended periods of
cessation of enteral feeding that could impair the intestinal matu- 2.1 | Patients
ration, and prolonged courses of parenteral nutrition and antibiot-
ics. New tools are thus required to facilitate an early and precise This prospective multicentre study was conducted in three French
diagnosis and to optimise the management of preterm neonates with Neonatal Intensive Care Units (NICUs). Eligible preterm neonates
intestinal symptoms. had a gestational age (GA) of 33 weeks or less, and did not present
Biomarkers are molecular indicators of disease process, diag- with any relevant malformation. They were enrolled at birth, and var-
nosis, and/or prognosis. There is currently a paucity of reliable and ious perinatal parameters were recorded: sex, GA, mode of delivery,
5
robust biomarkers for neonatal enteropathies, including NEC. Apgar score, intra-uterine growth retardation, acute foetal distress,
Noninvasive biomarkers have obvious advantages in preterm in- respiratory distress, antenatal and per partum antibiotic therapy, and
fants. Among the noninvasive faecal biomarkers, the most studied is suspicion of maternofoetal infection. Neonates underwent a weekly
faecal calprotectin (FC), a protein mainly expressed in the cytoplasm clinical follow-up until hospital discharge, including weight, height,
of neutrophils and highly resistant to proteolysis, which is now rou- head circumference, abdominal circumference, collateral venous ab-
tinely measured to evaluate mucosal inflammation in patients with dominal circulation and mode of feeding. Digestive tolerance was
inflammatory bowel diseases.6-8 Physiological values of neonatal FC evaluated through stool frequency and characteristics according to
are higher than adult ones with a large inter-individual variability, ‘Amsterdam’ stool form scale,12 stool bleeding, bloating, type and
reflecting the permeability of neonatal digestive tract, the immune volume of feeding, and residual gastric volume. Enteropathy was
stimulation induced by the microbiota settlement and the first con- suspected when occurred an intestinal distress requiring the inter-
tacts with feeding allergens.9 Several studies have shown increased ruption of enteral feeding for more than 48 hours, the commonly
FC levels in infants with enterocolitis. However, there is still no con- used treatment for a suspected enterocolitis (Bell's stage ≥ Ia). When
sensus about cut-off levels in the neonatal population and regarding neonates experienced such events, a daily monitoring was set-up in-
its use in current clinical practice.10 cluding all relevant previously described parameters, and biological
In a retrospective study, we previously proposed a FC thresh- parameters (C-reactive protein (CRP), procalcitonin). Medical files
old at 350 µg/g faeces (sensitivity 0.65, specificity 0.82) using en- were further checked by both a gastropaediatrician and a public
zyme-linked immunosorbent assay (ELISA) for early diagnosis of health specialist to ensure that the event had a digestive origin and
neonatal enteropathy requiring strict enteral feeding interruption in was not secondary to another infectious disease. Cases, that is defi-
preterm neonates.11 The current prospective cohort study was de- nite enteropathies, were defined by the concurrent presence of an
signed to establish a prognosis and diagnosis score of enteropathy abdominal distension (either clinically diagnosed by visual observa-
based on combined criteria including neonatal characteristics and tion and/or measurement of the abdominal circumference, or radio-
FC concentration. Furthermore, the study aimed at establishing the logically diagnosed, or both) and either rectal bleeding or elevated
agreement between the ELISA method considered as the reference gastric retention (the threshold was higher than one third of the vol-
assay performed in laboratories and a point of care testing (POCT), ume of milk ingested per 24 hours) or both. All cases were classified
that is an immunochromatographic (IC) assay, more convenient for according to the Bell's scale. 2
rapid diagnosis. In addition, faecal microbiota composition during Stools were collected at enrolment, then weekly until discharge,
the digestive event was analysed. for FC assays and gut microbiota analysis, and immediately stored at
CAMPEOTTO et al. |
111
−20°C and −80°C, respectively. During the event, complementary was 9% in neonates with FC < 350 µg/g. For a RR equal to 4 with
stool samples were collected daily from the first digestive symp- alpha risk 5% and power 90%, it was needed to include forty-seven
toms. The expected quantity of stools was over 2g to make it possi- infants per class group (FC ≥ 350 µg/g, FC < 350 µg/g). Ninety-four
ble high-quality grade assays. infants were required, and the sample size was extended to 125 in-
This study was approved by an Ethic Committee (CPP Ile-de- fants overall to account for other possible risk factors.
France I, number 2012-Janv.-12813) and registered on https://clini All statistical analyses were undertaken using R 2.3.3 software
caltrials.gov/ (NCT02010268). Parental written consent was signed (https://cran.rproject.org/). Statistical tests were two-sided with
prior to any study procedure. 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
2.2 | Calprotectin assays variables, and frequencies (%) for qualitative variables. Univariate anal-
yses were carried out to determine the association between enterop-
Stool samples were assayed for FC by Calprest ELISA method athy and FC levels using ELISA, clinical parameters and the occurrence
(Eurospital, Italy) and Calfast IC assay (Eurospital, Italy)—as the of an enteropathy. Since a prognosis score was seeking, this analysis
POCT—according to manufacture recommendations. Assays were was performed on neonates with at least one sample available before
performed at the end of the sampling period and simultaneously day 7 of life (D7) and before the event when it occurred before D7. The
for both methods to avoid any bias in the measurements. Hence, agreement between faecal calprotectin concentrations determined by
samples were stored at −20°C before the assay no more than ELISA and IC assay was evaluated by intra-class correlation coefficient
12 months, which is known to prevent from significant alteration (ICC) and Bland and Altman plot.19 Gut microbiota was analysed using
13
of FC levels. Wilcoxon test and permutational multivariate analysis of variance
(PERMANOVA).
2.4 | Statistical analysis
3.2 | Enteropathies
Cases were defined as infants with intestinal distress leading to
interruption of enteral feeding for more than 48 hours, associated During their study follow-up (median duration 41 days [1-83]), 21
with an abdominal distension and either rectal bleeding or elevated neonates (17%) experienced at least one episode of enteropathy
gastric retention or both. The sample size was calculated to show (one event for 19 neonates and two events for 2 neonates. The
that a FC threshold (ELISA) higher than 350 µg/g faeces is a risk fac- first enteropathy occurred at a median age of 15 days [1-38], and
tor of feeding interruption in preterm neonates as previously dem- the second at 21 and 71 days, respectively. The enteropathy was
onstrated.11,18 We hypothesised that the prevalence of enteropathy 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).
Overall, 681 samples were assayed for FC levels. One sample was Prognostic value of FC-ELISA was evaluated on the 84 neonates for
excluded for an aberrant value (5250 µg/g faeces), and 634 samples whom we collected at least one sample before D7 and before the
(93%) were studied by both ELISA and IC assay. The median FC-ELISA event (when it occurred before D7). Among these neonates, 14 (17%)
was 88 µg/g faeces (range [8-798], inter-quartile range [44.3-140]). presented with an enteropathy. Univariate analysis did not show any
FC-ELISA ≥ 350 µg/g faeces were found only in 12 patients (10%) significant association between the onset of enteropathy, FC-ELISA
from the cohort. Among them, 7 had no enteropathies during hospi- at inclusion and/or before the event, and the clinical parameters re-
tal stay, 4 had a high FC-ELISA during or after the event, and 1 before corded prior inclusion, precluding the establishment of a prognostic
the event. Thus, the threshold for enteropathy of 350 µg/g faeces score (supplementary file, Table S1, which demonstrates the ab-
was not confirmed in the current cohort. 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- enteropathy including perinatal parameters and FC level. Univariate
rence of NEC cases in our cohort. 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.
3.5 | Gut microbiota characteristics Literature reported wide variations in FC levels in premature in-
fants.10 In our cohort, FC levels were of the same order than those
To identify microbiota signatures of enteropathy, microbiota com- described recently by MacQueen et al20 and Ho et al21 but lower
position was analysed at various phylogenetic levels. Two main than those of Van Zoonen et al22 or Nakayuenyongsuk et al. 23 These
profiles with either Proteobacteria or Firmicutes predominance discrepancies could be related to either neonates’ characteristics as
were observed in both enteropathy and control groups (Figure 2A). a recent study indicated a relationship between FC and corrected
Comparison of species diversity and relative bacterial abundance GA, 24 or to the use of different methods of assay, which are known
profiles between infants with or without enteropathy revealed no to give numerical values which differ substantially. 25 This out-
specific alterations associated with the onset of enteropathy, except come points out that results between cohorts are not interchange-
a trend towards a higher diversity in infants without enteropathy able. Moreover, it should necessitate the use of adapted clinical
(Figure 2B,C). No correlations with perinatal factors were identified. 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
4 | D I S CU S S I O N We failed to determine a prognosis score of enteropathy. Of
note, the prevalence of enteropathies in our cohort (23 enteropa-
This multicentre prospective study on 121 very preterm neonates thies in 21 patients, including only 5 with Bell's stage II) was lower
aimed at defining a combined prognostic and diagnostic score of 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 Recent studies also failed to establish such relationship. Van Zoonen
to improvements in the care of very preterm neonates that may et al did not find any differences between control infants and the ten
have led to a decrease in the risk of enteropathy and NEC, with ones who further developed NEC (Bell’ stages 2 or 3). 22 Interestingly,
an incidence of NEC in France of 3.2% in very preterm infants in MacQueen et al20 found more frequently lower calprotectin levels in
26
a recent nationwide cohort. Moreover, only 14 cases out of the preterm infants with bloody stools (Bell's stage Ib), but who did not
23 with enteropathy could be included in the prognosis score cal- progress to NEC, suggesting that calprotectin level may be a marker
culation. The hypothesised threshold of 350 µg/g faeces to char- of the severity of the enteropathy. However, this relationship was
acterise enteropathy requiring feeding interruption was reached in imperfect, with neonates with elevated FC levels and without NEC,
very few neonates with and without enteropathy. Several cut-off and neonates with levels within the reference range and with NEC.
values have been proposed, especially for prediction of severe form The usefulness of point-of-care test to determine FC in preterm
of NEC.10 However, these cut-off levels remain within the range of infants is obvious, providing results rapidly available to the paedia-
FC levels found in infants without digestive symptom enteropathies. tricians. POCT assay was carried out centrally and compared with
CAMPEOTTO et al. |
115
FC-ELISA as the reference assay performed on the same samples. advice. The sponsor was Assistance Publique—Hôpitaux de Paris
We have previously shown a correlation between the two tech- (Clinical Research and Innovation Department).
niques in adult patients with IBD. 27 The expected agreement was
not observed in our very preterm neonate population, even if the C O N FL I C T O F I N T E R E S T
individual plots overlaying the results of both techniques presented None declared.
slightly similar evolution whatever the digestive status. A recent
study using a different POCT confirmed the high calprotectin lev- ORCID
els with major fluctuations, and nonspecific increase in some infants Marie-José Butel https://orcid.org/0000-0003-4249-2130
that developed or not NEC. 20
There are few information on microbiota in mild-to-moderate REFERENCES
enteropathies by contrast to NEC. We found no specific microbi- 1. Battersby C, Santhalingam T, Costeloe K, Modi N. Incidence
ota profiles by comparison with infants without digestive events, of neonatal necrotising enterocolitis in high-income coun-
tries: a systematic review. Arch Dis Child Fetal Neonatal Ed.
by contrast with that we26 and others previously reported for
2018;103:F182-F189.
NEC. 28 However, the high heterogeneity of microbiota of very 2. Kliegman RM, Walsh MC. Neonatal necrotizing enterocolitis:
preterm infants that we and others described, and the low num- pathogenesis, classification, and spectrum of illness. Curr Probl
ber of digestive events in this study may have limited the analysis Pediatr. 1987;17:219-288.
3. Campeotto F, Kalach N, Lapillonne A, Butel MJ, Dupont C, Kapel N.
power.
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.
4.1 | Strengths and limitations 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
This prospective study on preterm neonates born at less than stages and biomarkers for NEC. Semin Pediatr Surg. 2018;27:3-10.
33 weeks of GA gave detailed information on their characteristics 6. Fagerhol MK. Calprotectin, a faecal marker of organic gastrointesti-
that may help to define population of further studies. Neonates with nal abnormality. Lancet. 2000;356:1783-1784.
enteropathy leading to interruption of enteral feeding were fully de- 7. van Rheenen PF, Van de Vijver E, Fidler V. Faecal calprotectin for
screening of patients with suspected inflammatory bowel disease:
scribed. The strength of our study also lies in the repeated meas-
diagnostic meta-analysis. BMJ. 2010;341:c3369.
urements of the faecal calprotectin. However, in our cohort, the 8. Heida A, Van de Vijver E, van Ravenzwaaij D, et al. Predicting in-
occurrence of enteropathy was lower than expected and reached flammatory bowel disease in children with abdominal pain and diar-
only half of the one observed in our previous study used to sup- rhoea: calgranulin-C versus calprotectin stool tests. Arch Dis Child.
2018;103:565-571.
port the hypothesis for sample size calculation,9 thus precluding the
9. Kapel N, Campeotto F, Kalach N, Baldassare M, Butel MJ, Dupont
power of the study to define a prognosis score. 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
5 | CO N C LU D I N G R E M A R K S
of the literature. Inflamm Res. 2016;65:847-852.
11. Campeotto F, Baldassarre M, Butel MJ, et al. Fecal calprotectin:
This study did not achieve its goal to propose a prognostic score of en- cutoff values for identifying intestinal distress in preterm infants. J
teropathy mainly due to the large inter- and intra-individual variability Pediatr Gastroenterol Nutr. 2009;48:507-510.
12. Bekkali N, Hamers SL, Reitsma JB, Van Toledo L, Benninga
of faecal calprotectin levels that limit its use as a biomarker of enter-
MA. Infant stool form scale: development and results. J Pediatr.
opathy in very preterm infants. However, the interest of a prognostic 2009;154(521–6):e1.
score in this population at high risk of mild-to-severe enteropathy that 13. Ton H, Brandsnes DS, Holtlund J, et al. Improved assay for fecal
will help to take appropriate management of the disease as early as calprotectin. Clin Chim Acta. 2000;292:41-54.
14. Rousseau C, Levenez F, Fouqueray C, Dore J, Collignon A, Lepage
possible deserves to design a larger multicentric study which would
P Clostridium difficile colonization in early infancy is accompanied
take into account the difficulties faced in this study. by changes in intestinal microbiota composition. J Clin Microbiol.
2011;49:858-865.
AC K N OW L E D G E M E N T S 15. Escudie F, Auer L, Bernard M, et al. FROGS: Find, Rapidly,
The authors would like to thank the support of all the staff of the OTUs with Galaxy Solution. Bioinformatics (Oxford, England).
2018;34:1287-1294.
neonatal units participating in this project and the parents for their
16. Federhen S. The NCBI Taxonomy database. Nucleic Acids Res.
agreement in the study. They would also like to thank the staff 2012;40:D136-D143.
Clinical Research Unit of Necker-Enfants Malades for the monitor- 17. Cole JR, Wang Q, Fish JA, et al. Ribosomal Database Project: data
ing of the clinical trial, Robert Campos Oriola from AUXESIA for and tools for high throughput rRNA analysis. Nucleic Acids Res.
2014;42:D633-D642.
the manuscript preparation, Julie Salomon, gastropaediatrician at
18. Yang Q, Smith PB, Goldberg RN, Cotten CM. Dynamic change of
Necker-Enfants Malades hospital and Maria-Elisabetta Baldassarre, fecal calprotectin in very low birth weight infants during the first
neonatologist at Aldo Moro University, Bari, Italy, for their helpful month of life. Neonatology. 2008;94:267-271.
|
116 CAMPEOTTO et al.
19. Bland JM, Altman DG. Statistical methods for assessing agree- 26. Roze JC, Ancel PY, Lepage P, et al. Nutritional strategies and gut
ment between two methods of clinical measurement. Lancet. microbiota composition as risk factors for necrotizing enterocolitis
1986;1:307-310. in very-preterm infants. Am J Clin Nutr. 2017;106:821-830.
20. MacQueen BC, Christensen RD, Yost CC, et al. Reference intervals 27. Benahmed NA, Manene D, Barbot-Trystram L, Kapel N. Evaluation
for stool calprotectin in preterm neonates and their utility for the di- of Calfast(R) immunochromatographic quantitative assay for
agnosis of necrotizing enterocolitis. J Perinatol. 2018;38:1379-1385. the measurement of calprotectin in faeces. Clin Chem Lab Med.
21. Ho TTB, Groer MW, Kane B, et al. Enteric dysbiosis and fecal 2014;52:e143-e145.
calprotectin expression in premature infants. Pediatr Res. 28. Baldassarre ME, Di Mauro A, Capozza M, et al. Dysbiosis and pre-
2019;85:361-368. maturity: is there a role for probiotics? Nutrients. 2019;11(6):1273.
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. S U P P O R T I N G I N FO R M AT I O N
2019;54:455-459.
Additional supporting information may be found online in the
23. Nakayuenyongsuk W, Christofferson M, Stevenson DK, Sylvester
K, Lee HC, Park KT. Point-of-Care Fecal Calprotectin Monitoring Supporting Information section.
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- How to cite this article: Campeotto F, Elie C, Rousseau C,
tectin concentration in neonatal necrotizing enterocolitis. Korean J
et al. Faecal calprotectin and gut microbiota do not predict
Pediatr. 2014;57:351-356.
25. Labaere D, Smismans A, Van Olmen A, et al. Comparison of six
enteropathy in very preterm infants. Acta Paediatr.
different calprotectin assays for the assessment of inflammatory 2021;110:109–116. https://doi.org/10.1111/apa.15354
bowel disease. United European Gastroenterol J. 2014;2:30-37.
DOI: 10.1111/apa.15448
BRIEF REPORT
Education on neonatal resuscitation is most urgent in settings with The outcome measures were the proportion of infants who re-
poor access to intrapartum obstetric care, where immediate postna- ceived suctioning, the characteristics of suctioning and the propor-
tal mortality can be reduced by 30% with basic training in neonatal tion of infants who needed PPV.
1
resuscitation. Management of newborns at birth includes different Since this study presents a secondary analysis of data collected
interventions based on progressive steps (initial steps, ventilation, during a prospective, a convenience sample comprising all 150 vid-
chest compressions and medications). Initial steps include oropha- eos of the original study was evaluated. Continuous data were sum-
ryngeal suctioning, which is recommended under vision in infants marised using median and interquartile range (IQR), and categorical
who have obvious obstruction to spontaneous breathing or who re- data as number and percentage.
quire positive pressure ventilation (PPV). 2 The Manual of Neonatal Of the 150 video recordings, suctioning was performed in 117
Resuscitation describes when and how to perform oropharyngeal neonates (78%). There were 77 males (66%) and 40 females (34%).
suctioning, 2 but information on how the procedure is actually per- Median gestational age was 38 weeks (IQR 36-40), and median birth
formed in low-resource settings is lacking. weight was 2900 g (IQR 2350-3200). Median Apgar score at 5 min-
The aim of this study was to evaluate the occurrence and the utes was 6 (IQR 4-7). Suctioning resolved the airway obstruction in
characteristics of suctioning at birth in newly born infants in a 40 neonates (34%), while ventilation was needed after suctioning in
low-resource setting. 70 neonates (60%) who displayed persisting bradycardia and/or ap-
The study was performed at Beira Central Hospital (Beira, noea. In seven neonates (6%), suctioning was performed due to fail-
Mozambique) where about 4500 deliveries occur every year. Beira ure of ventilation for suspected airway block. Information on suction
Central Hospital is the referral hospital for a geographical area that is shown in Table 1. Suction started at median 70 seconds (IQR 40-
covers about 7 million people, with large referral services for mater- 120) and lasted median 100 seconds (IQR 50-140). A catheter was
nal and neonatal care in the province. used in 99 neonates (85%) and a bulb syringe in 18 (15%). The correct
This study presents a secondary analysis of data from a pro- sequence (suctioning the mouth before the nose) was performed in
spective study on education in neonatal resuscitation using video 106 neonates (91%), but suctioning the stomach was observed in 78
recording.3 The National Committee of Bioethics and the Minister of neonates (67%).
Health of the Republic of Mozambique approved the research pro- Overall, our findings showed that suctioning was performed in
tocol.3 Parental consent to record neonatal delivery room manage- three out of four neonates receiving resuscitation, with further PPV
ment and to use the data was obtained before every delivery. needed in the majority of them. The video review underlined late
Among 150 neonates from the original study, who required suc- initiation and long duration of the procedure, with preference for
tioning were included in this secondary analysis. Lack of parental deep suctioning and the use of a catheter. This study has the merit of
consent was the only exclusion criterion. describing magnitude and quality of suctioning in neonates receiving
Neonatal resuscitation was performed routinely under ra- resuscitation in a low-resource setting by using video recording, but
diant warmers in the delivery room or in the operating room. could not provide information on the impact of suctioning on clinical
Suctioning was performed with bulb syringe or suction catheter outcomes.
by the attending caregiver in infants who have obvious obstruc- While oro/nasopharyngeal suctioning is no longer recommended
tion to spontaneous breathing or who require PPV. 2,3 A camera as routine care immediately after birth, this procedure can be useful
installed above the radiant warmers video-recorded all interven- in clearing the airway or when PPV is required. 2 Our data suggested
3
tions as previously described. Two researchers (SA and SC) re- that, after receiving drying and stimulation, suctioning was suffi-
viewed and evaluated all 150 videos of neonatal resuscitation, cient to avoid further resuscitation procedures in one out of three
with a third researcher (DT) resolving any conflicts. Time of birth neonates. Although suctioning may activate vagal response induc-
was defined as the time the Apgar clock was started or birth was ing bradycardia and apnoea, we can speculate that deep suctioning
announced. acted as some form of stimulation in our neonates.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
TA B L E 1 Characteristics of suctioning F U N D I N G I N F O R M AT I O N
No. of neonates receiving suction 117 Associazione Pulcino, Italy, supported the study providing funds
Time of suctioning initiation, seconds a
70 (40-120) for video cameras. The funders had no role in the study design or
Number of suctions a
1 (1-2)
conduct, the management, analysis or interpretation of the data, the
preparation, review or approval of the report, or the decision to sub-
Duration of first suction, secondsa 65 (43-105)
mit the manuscript for publication.
Total duration of suction, secondsa 100 (50-140)
Bulb syringe 18 (15)
Francesco Cavallin1
Catheter 99 (85)
Shaimaa Abuelnoor Ahmed Abdelghany2
Suctioning the mouth before the nose 106 (91)
Serena Calgaro2,3
Suctioning the stomach 78 (67) Amir Hussein Abubacar Seni4
Note: Data expressed as No. (%) or amedian (IQR). Bonifacio Rodriguez Cebola4
Giovanni Putoto3
Our data disclosed low adherence to international guidelines on Daniele Trevisanuto2
neonatal resuscitation in terms of start and duration of the proce-
1
dure. When indicated, suctioning should be performed within the Independent Statistician, Solagna, Italy
first minute of life and should last few seconds. 2 However, at least 2
Department of Woman's and Child's Health, University of
half of our procedures began beyond the first minute of life and Padua, Padua, Italy
3
lasted almost two minutes. Previous research suggested that many Doctors with Africa CUAMM, Padua, Italy
4
non-breathing infants may be in primary apnoea with a heart rate Beira Central Hospital, Beira, Mozambique
and may initiate spontaneous breathing in response to the initial
steps of resuscitation (including drying, stimulation and suctioning) Correspondence
only when timely implemented.4 Delayed start and prolonged dura- Daniele Trevisanuto, Department of Woman's and Child's
tion of suctioning represent an additional hazard for neonates, be- Health, University of Padua, Via Giustiniani 3, 35128, Padua,
cause of the delay of PPV initiation. The correct order of suctioning Italy.
(mouth before the nose) was followed in the majority of cases, in Email: daniele.trevisanuto@unipd.it
order to reduce the risk of aspiration. 2 However, the high prevalence
of deep suctioning had the potential risk of opening the oesophagus ORCID
and leading to more air leaks into the stomach. Daniele Trevisanuto https://orcid.org/0000-0002-6462-0079
In conclusion, suctioning was largely performed in neonates re-
ceiving resuscitation, but adherence to official recommendations REFERENCES
was limited in terms of late initiation, prolonged duration and deep 1. Lawn JE, Blencowe H, Oza S, et al. Lancet Every Newborn Study
suctioning. Appropriate training is warranted to improve the quality Group. Progress, priorities, and potential beyond survival. Lancet.
2014;384:189-205.
and the timing of this procedure. On the other hand, further evi-
2. American Academy of Pediatrics and American Heart Association.
dence is required to shed lights on benefits and/or harms of suction- NRP Neonatal Resuscitation Textbook, 7th edn. Chicago, IL: American
ing in neonates needing resuscitation at birth.5 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
AC K N OW L E D G E M E N T S
low-resource setting. Neonatology. 2018;114:294-302.
We are very grateful to the midwifery and medical staff of the Beira 4. Ersdal HL, Mduma E, Svensen E, Perlman JM. Early initiation of
Central Hospital, Beira, Mozambique, for their participation and in- basic resuscitation interventions including face mask ventilation
valuable cooperation in this study. may reduce birth asphyxia related mortality in low-income coun-
tries: a prospective descriptive observational study. Resuscitation.
2012;83:869-873.
C O N FL I C T O F I N T E R E S T
5. Foster JP, Dawson JA, Davis PG, Dahlen HG. Routine oro/nasopha-
The authors declare that they have no conflict of interest. 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
Improved obstetric and neonatal care have reduced the preva- 4 hours after this injury could reduce neuropathology in those pups
lence of severe hypoxic-ischaemic-encephalopathy (HIE). However, surviving into juvenile age.
1-3/1000 newborns in the developed world1 suffer death or neuro- We report here that the injection of hypoxia-derived conditioned
developmental disability from HIE. The normal development of the media to healthy pups causes a modest loss of parvalbumin neurons
brain during gestation can also be altered by placental reprogram- in the thalamic reticular nucleus (TRN), the hippocampus and the
ming under oxidative stress. Under these conditions, the placenta cortex at P30 of survival (Figure 1). The glial response to neuronal
releases DNA-damaging molecules, bone morphogenic proteins, loss was assessed by GFAP immunofluorescence and showed a
2
microRNAs and glutamate. At present, one is unable to diagnose or marked increase in astrocytes in addition to an activated morphol-
treat these factors. ogy. The injury also affected dendrite lengths, a proxy measure for
We have previously applied Xenon, a rare noble gas used in an- degree of arborisation/connectivity, and this was consistent with
aesthesia, at 50% using a closed-circuit system with and without hy- previous in vitro findings.8 There were no changes in overall neu-
3,4
pothermia in the newborn piglet and rodent models of HIE. Unlike ronal counts, but dopaminergic neurons process lengths decreased
other inhalational anaesthetics, Xenon did not induce neuroapopto- in some areas. Importantly, Xenon treatment conferred some resis-
sis in the immature brain5 and improved cardiovascular control after tance to the increase in glial numbers (P < 0.05) in the cortex and
hypoxia-ischaemia (HI). A clinical feasibility and ongoing randomised hippocampus (Figure 1). Most strikingly, we observed Xenon treat-
50%
phase-two study are testing the effects of breathing Xenon in ment to be protective against the loss of parvalbumin cells in the
term infants undergoing therapeutic hypothermia (TH) against those TRN caused by the injury (P < 0.05). Interestingly, Xenon treatment
undergoing TH alone. Experimentally, inhaling Xenon50% improves did not protect dendritic arborisations/complexity but did greatly in-
motor function and cognition after long-term survival in rats post-in- crease the lengths of dopaminergic (tyrosine hydroxylase) processes
4
jury. In rat models of HI brain injury, Xenon is neuroprotective by and overall neuronal numbers post-injury.
upregulating neurotrophic factors and anti-apoptotic proteins,6 by These promising results albeit limited in scope support that Xenon
inducing hypoxia-inducible factor (HIF-1α) pathways allowing for treatment after mild injury due to maternal hypoxia does offer some
pre-conditioning,7 by suppressing the astroglial response to injury protection. Our findings are consistent with previously reported effects
and limiting glutamate release to counter excitotoxicity thereby im- of Xenon in toning down gliosis in neonatal rat cortex, hippocampus
proving neuronal survival.4 and thalamus in a classical HI rodent model.9,10 Most neurons including
Xenon's neuroprotective properties may be extended to treat parvalbumin neurons are sensitive to hypoxia and are lost in the 2%-8%
brain injury arising from placental reprogramming under oxida- condition in most areas. Xenon acts as an anti-apoptotic agent, and as
tive stress. To test this hypothesis, we used a rat injection model neurogenesis is still very active in the early developing postnatal rodent
whereby media obtained directly from human placentae under oxi- brain, Xenon may be promoting a compensatory neuroblast differen-
dative stress were injected into postnatal day 4(P4) rat brain (human tiation response in vulnerable areas explaining the higher number of
gestational age 29-31 weeks equivalent). In brief, media were col- densities we observed compared with 21%. This is speculative, and we
lected from human first trimester placentae cultured under 21% O2 do not know the underlying mechanisms. We also cannot speculate on
(CM + 21%) and 2%-8% O2 (CM + 2%-8%). An additional group was the behavioural significance of an overall increased number of neurons
injected with saline (Sal) and constituted the sham condition. The without further study. Xenon is thought to provide partial protection
pups were then allowed to survive into the juvenile age, brains were of dopaminergic cells by acting as a trophic factor in conditions of ex-
culled, and neuropathology was examined (see Supplementary files citotoxicity and by suppressing the astroglial response.11 This improves
for more information). cell survival but may also result in outgrowth around the area of dam-
We have tested (a) the effects of hypoxic injury modelled by the age. Fibre outgrowth is linked with altered connectivity in the brain and,
injection of hypoxia-derived conditioned media from the placenta therefore, may not necessarily beneficial. Behavioural work is needed
into P4 rat pup brains and (b) whether breathing 50% Xenon for to test how motor skills have been affected with/ without Xenon after
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
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
1
injury. Although we did not perform cognitive tests on the rodents, Translational Health Sciences, Bristol Medical School,
no significant neuropathology in the tissues was detected. Previously, University of Bristol, Bristol, UK
2
when Xenon was administered shortly after a carotid ligation + hypoxic UK Dementia Research Institute, Cardiff University, Cardiff, UK
3
insult on P7 rat pups (near-term equivalent), which survived into adult- Biological Sciences, Faculty of Environmental and Life Sciences,
hood, both neuropathology and behaviural testing were improved by University of Southampton, Southampton, UK
Xenon50%.4 In this 10-week survival to adulthood model, we defined 4
Nuffield Department of Clinical Neurosciences, University of
that the time-window for neuroprotection with Xenon was 5 hours, Oxford, Oxford, UK
5
which is the basis for our clinical trial of administering Xenon within Department of Obstetrics and Gynaecology, Southmead
5 hours after birth since delayed Xenon significantly improved outcome. Hospital, Bristol, UK
Another clinical trial delivered Xenon30% starting at 10 hours of age and 6
Department of Medical Education, King's College London,
short-term outcome (using Magnetic Resonance Spectroscopy) after London, UK
10 days did not show any difference between TH and TH + Xenon.12 7
Neonatal Neuroscience, Translational Health Sciences, Bristol
It is likely that any effect of Xenon on cognition needs long-term fol- Medical School, University of Bristol, Bristol, UK
8
low-up. We are currently undertaking full IQ testing at 3-5 years in the Institute of Basic Medical Sciences, Section for Physiology,
CoolXenon trial (ending 010321). University of Oslo, Oslo, Norway
Several clinical conditions present with hypoxic insults to the
placenta including pre-eclampsia, maternal gestational diabetes13 Correspondence
and stress.14 Placental reprogramming can alter fetal neurodevelop- Marianne Thoresen, Neonatal Neuroscience, Translational
ment due to maternal hypoxia and Xenon inhalation may yet offer a Health Sciences, Bristol Medical School, University of
promising therapeutic strategy. Bristol, Bristol, UK.
Email: marianne.thoresen@bristol.ac.uk
K E Y WO R D S
brain injury, cooling, hypoxic-ischaemic encephalopathy, Thomas Phillips and David A. Menassa equal contribution.
neurodevelopment, placental reprogramming, Xenon
ORCID
AC K N OW L E D G E M E N T S David A. Menassa https://orcid.org/0000-0002-5984-8407
We acknowledge the financial support of the JP Moulton founda- Marianne Thoresen https://orcid.org/0000-0002-9615-9109
tion, the Perivoli Trust and SPARKS-The Children's Medical Research
Charity. We also acknowledge Dr Charles Patrick Case for his input REFERENCES
in this study and Dr John Dingley for providing Xenon-delivery 1. Lai M-C, Yang S-N. Perinatal hypoxic-ischemic encephalopathy. J
expertise. 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-
C O N FL I C T O F I N T E R E S T ment. Sci Rep. 2017;7(1):9079.
The authors declare no conflict of interest. 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.
Thomas Phillips1,2
4. Thoresen M, Hobbs CE, Wood T, Chakkarapani E, Dingley J. Cooling
David A. Menassa1,3,4 combined with immediate or delayed xenon inhalation provides
Simon Grant5 equivalent long-term neuroprotection after neonatal hypoxia-isch-
Nicki Cohen6 emia. J Cereb Blood Flow Metab. 2009;29(4):707-714.
Marianne Thoresen7,8
|
122 PHILLIPS et al.
5. Sabir H, Walløe L, Dingley J, Smit E, Liu X, Thoresen M. Combined 11. Lavaur J, Le Nogue D, Lemaire M, et al. The noble gas xenon pro-
treatment of xenon and hypothermia in newborn rats–additive or vides protection and trophic stimulation to midbrain dopamine neu-
synergistic effect? PLoS One. 2014;9(10):e109845. rons. J Neurochem. 2017;142(1):14-28.
6. Fan X, Kavelaars A, Heijnen CJ, Groenendaal F, van Bel F. 12. Azzopardi D, Robertson NJ, Bainbridge A, et al. Moderate hypo-
Pharmacological neuroprotection after perinatal hypoxic-ischemic thermia within 6 h of birth plus inhaled xenon versus moderate
brain injury. Curr Neuropharmacol. 2010;8(4):324-334. hypothermia alone after birth asphyxia (TOBY-Xe): a proof-of-
7. Ma D, Lim T, Xu J, et al. Xenon preconditioning protects against concept, open-label, randomised controlled trial. Lancet Neurol.
renal ischemic-reperfusion injury via HIF-1alpha activation. J Am 2016;15(2):145-153.
Soc Nephrol. 2009;20(4):713-720. 13. Gauster M, Desoye G, Tötsch M, Hiden U. The placenta and gesta-
8. Curtis DJ, Sood A, Phillips TJ, et al. Secretions from placenta, after tional diabetes mellitus. Curr Diab Rep. 2012;12(1):16-23.
hypoxia/reoxygenation, can damage developing neurones of brain 14. Gheorghe CP, Goyal R, Mittal A, Longo LD. Gene expression in the
under experimental conditions. Exp Neurol. 2014;261:386-395. placenta: maternal stress and epigenetic responses. Int J Dev Biol.
9. Dingley J, Tooley J, Porter H, Thoresen M. Xenon provides short- 2010;54(2–3):507-523.
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 S U P P O R T I N G I N FO R M AT I O N
cognitive function, reduces neuronal loss and chronic neuroinflam- Additional supporting information may be found online in the
mation, and improves survival after traumatic brain injury in mice.
Supporting Information section.
Br J Anaesth. 2019;123(1):60-73.
| |
Received: 28 February 2020 Revised: 27 April 2020 Accepted: 25 May 2020
DOI: 10.1111/apa.15378
REGULAR ARTICLE
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
1 | I NTRO D U C TI O N births has remained relatively constant, while milder neurocogni-
tive deficits have become more prominent.1-3 A higher prevalence
Preterm birth can affect a child´s development in many ways. The of low intelligence quotient (IQ) among very low birth weight
immature central nervous system is vulnerable, not only to injury preterm individuals than in term controls has been reported.4
but also to disruptions of normal development. With improve- Some studies have shown that children born preterm have more
ments in neonatal intensive care and survival of more immature difficulties with nonverbal reasoning and visuospatial tasks than
infants, the rate of major disabilities associated with preterm 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
7 died
1 emigrated
4 excluded
1 died
6 emigrated
1 immigrated
1 protected identy
214 located
25-29 years
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 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,
Descriptive data were analysed using IBM SPSS Statistics version 25 binocular near VA and contrast sensitivity in the better eye. Better eye
(IBM Corp). Mann-Whitney and Kruskal-Wallis tests were used to com- was defined as the eye with the best VA. Thereafter, multivariable lin-
pare the results of the two study groups on the three-part Beery VMI ear regression analyses were performed to identify predictors for the
test. Post hoc analyses were carried out using Dunn´s test. Since there three parts of the Beery VMI. The multivariable analyses included ROP,
was a difference in proportions of females and males in the preterm yes or no, treated ROP, yes or no, neurology at 2.5 years of age, VA
and control groups, linear regression models adjusting for gender, also and contrast sensitivity, together with GA in weeks in one and BW in
were used to estimate the differences between the groups on Beery one, since GA and BW were highly correlated. The analyses were per-
VMI. Simple regression analyses were performed within the preterm formed using R version 3.6.1 (R Foundation for Statistical Computing,
group, and the variables were ROP, yes or no, in the most affected eye, 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
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- In the present study, an extensive ophthalmological examination
ported on the visual function13 and social outcome21 in a large cohort was performed,17 which made it possible to analyse the association
of preterm young adults from New Zealand, born before the introduc- between the results of the VMI test and visual function, assessed as
tion of treatment for ROP. Molloy et al reported on the visual function VA and contrast sensitivity. Best-corrected distance VA and near VA
and visual perception in a cohort from Australia of individuals aged 14 did not correlate with the results of the Beery VMI in the preterm
to 20 born extremely preterm22 and found significantly poorer perfor- group. One can hypothesise that the VA was generally sufficiently
mance in visual perceptual tasks in the preterm adolescents compared good to accomplish the Beery VMI, even if VA was significantly
with the controls. In addition, the authors showed that those with lower among those born preterm compared with the controls.17 This
previous severe ROP performed more poorly regarding visual-motor emphasises that VA is only one of the many factors involved in the
integration,8 which was in accordance with the present study. Molloy integration of visual perception and hand movements. Moreover, it
et al concluded that this association was largely mediated by other relates to the finding that the preterm individuals with neurological
neonatal risk factors. It may reflect the fact that children who develop complications in the present study had significantly lower scores on
treatment-requiring ROP are the most immature and vulnerable.23 the main test, the Beery VMI, a test that is not only visually demand-
In a cohort from Stockholm, Sweden, Lundequist et al found that ing, but also cognitively.19
extremely preterm (GA < 28 weeks) individuals had significantly Contrast sensitivity, which is an important part of visual func-
lower scores than controls on the Beery VMI, but the results of the tion, may be affected in persons with brain lesions. 24 In the current
very premature group (GA 28-31 weeks) did not differ significantly study, a correlation between contrast sensitivity and performance
from those of the controls.7 In the current study, we found differ- on the visual perception test was found in the preterm group, con-
ences in all parts of the VMI test compared with controls (Figure 2), firmed in the multivariable analysis. It cannot be concluded if this
but did not further divide the VLBW group into subgroups of imma- finding was due to minor changes in the brain.
turity. However, within the preterm group we found no correlations A lower proportion of preterm individuals than controls had or was
with BW or GA. 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.
AC K N OW L E D G E M E N T S
4.1 | Strengths and limitations We thank Eva Nuija, study nurse, for her skilful help with the study,
and Fabian Söderdahl MSc, who assisted with the statistical analysis.
The strength of this project is its long-term prospective ophthalmo-
logical follow-up from birth to young adult age of a population-based C O N FL I C T O F I N T E R E S T
group of preterm subjects, screened for ROP in the neonatal period. The authors have no conflicts of interest to declare.
The ophthalmological examination in this study was extensive and
performed on all participants by the same ophthalmologist and the ORCID
Beery VMI by the same psychologist. Dýrleif Pétursdóttir https://orcid.org/0000-0002-9757-1373
A limitation of the study was the low follow-up attendance rate, Gerd Holmström https://orcid.org/0000-0002-5600-7186
as can be the case with longitudinal studies. However, those who at- Eva Larsson https://orcid.org/0000-0001-9674-0094
tended at the 25-year follow-up did not differ regarding GA, BW, VA Birgitta Böhm https://orcid.org/0000-0002-9406-0930
PÉTURSDÓTTIR et al. |
133
DOI: 10.1111/apa.15338
REGULAR ARTICLE
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
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. |
135
1 | I NTRO D U C TI O N
Key Notes
Pulmonary interstitial emphysema (PIE) is an air leak syndrome, and
• This study focused on 179 preterm infants, who had un-
it is one of the most severe respiratory complications to affect pre-
dergone chest X-rays during the first five days of life,
term infants. It is characterised by the dissection of the interstitial or
to determine the prenatal and postnatal risk factors for
perivascular lung tissue, which results in air being trapped along the
pulmonary interstitial emphysema (PIE).
bronchovascular system. This leads to overdistension, decreased com-
1-3 • We identified preeclampsia and haemolysis, elevated
pliance and restricted gas exchange. The radiological characteristics
liver enzymes and low platelet count syndrome as ad-
of PIE are either cyst-like or linear radiolucencies radiating from the
ditional risk factors in infants with PIE.
pulmonary hilum to the surface of the lung, and these may be aggra-
• In addition, lung impairment was associated with res-
vated by the development of pneumomediastinum, pneumothorax or
4,5 piratory distress syndrome, higher maximum inspired
pneumopericardium. Chest radiography is the diagnostic gold stand-
oxygen and needing surfactant.
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 Liebig University of Giessen, Germany, between 2016 and 2017. All
infants.2,7 It can occur unilaterally or bilaterally and is classified as local received chest X-ray examinations during the first 5 days of life. The
or diffuse and acute or persistent PIE. exclusion criteria included no chest X-ray being available and prena-
The incidence of PIE ranges from 5% to 40%, in extremely low tal interventions like foetal paracentesis. We also excluded infants
birthweight infants, despite advances in neonatal medicine. These with lower urinary tract obstructions, spina bifida, congenital heart
include the established use of antenatal corticosteroids, avoiding me- defects, apart from patent ductus arteriosus, atrial septum defects,
chanical ventilation, the introduction of new, and more gentle, modes hypoplasia or sequestration of the lung, syndromal disease or oe-
of ventilation8 and the application of postnatal surfactant application. sophageal atresia.
PIE has been associated with higher mortality and high rates of severe After the exclusion criteria were applied, we had 179 infants left
morbidities that are common in premature infants, such as intraventric- in the study (Figure 1). All the cases were retrospectively classified
ular haemorrhage and bronchopulmonary dysplasia (BPD).5,9,10 Other by a neonatologist and a paediatric radiologist using an independent
suggested risk factors for PIE include the prenatal use of magnesium double-blinded study design, and any differences of opinion were
sulphate,11 higher ventilation parameters, increased tidal volumes and jointly resolved. This classification showed that 33 of the cases had
higher fractions of inspired oxygen (FiO2).12 However, one study re- PIE and 146 did not. The PIE cases were classified as unilateral or
ported that infants who developed pulmonary interstitial emphysema bilateral and subdivided in central, global or lobar manifestations
during the first days after birth had lower Apgar scores in the delivery (Figure 2). The first X-ray available was used to grade RDS.15 To com-
room and needed increased ventilator settings and FiO2, increased pensate for important confounders of pulmonary RDS severity, each
rates of intubation and more frequent surfactant. These findings sug- of the 33 PIE cases was matched with a non-PIE case by the clos-
3
gest a pulmonary predisposition for PIE as early as birth. This assump- est GA and gender (Figure 1, Appendix S1). A number of perinatal
tion has been further strengthened by reports of PIE occurring during variables were retrieved from the medical notes. These included the
non-invasive ventilatory support.2 infant's clinical characteristics, their birthweight, GA and gender. We
The aim of this study was to determine any associations between a also collected data on any IUGR, premature rupture of the mem-
number of other prenatal and postnatal risk factors and the occurrence branes, histologically proven chorioamnionitis, preeclampsia/HELLP
of PIE. These included any diabetes during pregnancy, preeclampsia syndrome,16,17 gestational diabetes, type I diabetes, antenatal cor-
and haemolysis, elevated liver enzymes and low platelet count (HELLP) ticosteroids, HELLP, maternal age, multiple pregnancy and Apgar
syndrome, intrauterine growth restriction (IUGR), the use of antenatal scores at one and five minutes. The factors that were considered
corticosteroids, male gender and multiple birth. All of these conditions with regard to postnatal respiratory support included the following:
are known to increase the risk of preterm infants developing broncho- severe RDS (grades III-IV), invasive mechanical ventilation, applied
pulmonary dysplasia (BPD) and long-term pulmonary sequelae.13,14 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
2 | M ATE R I A L A N D M E TH O DS preterm infant was on non-invasive ventilatory support at the time
of surfactant application; otherwise, it was given via the endotra-
2.1 | Study design cheal tube. Postnatal corticosteroids were separated into early for
the first seven days of life and late if it was beyond that point. The
This was a retrospective case-control study that focused on 226 pre- selected outcome parameters were intraventricular haemorrhage
term infants who were born at up to 32 weeks of gestational age (grade III-IV), BPD,18 necrotising enterocolitis, focal intestinal perfo-
(GA) and were discharged from the neonatology unit of the Justus ration, retinopathy of prematurity and mortality.
|
136 BEHNKE et al.
A B
Retinopathy of prematurity 49 (27.84) weight.3,5 Our results also underline that a postnatal predisposition
to severe pulmonary disease severity was reflected by the need for
Death 12 (6.70)
higher supplemental oxygen and more frequent application of sur-
Abbreviation: HELLP, haemolysis, elevated liver enzymes and low
factant. Previous studies had already reported that using higher oxy-
platelet count.
gen levels to resuscitate extremely premature infants had increased
oxidative stress and the incidence of BPD.21 Another study suggested
oxygen (FiO2) was higher in the PIE group (P = .014), according that higher oxygen needs during resuscitation were an independent
to the Wilcoxon signed-rank test. Conditional logistic regres- risk factor for PIE.3 Furthermore, PIE development has been related
sion also showed that this was the case for the number of severe to impaired oxygenation and ventilation efficiency and the oxygena-
RDS cases (P = .007) and the need for surfactant (P = .0281). The tion index during the development of PIE has been shown to be a
matched-pairs analysis in the conditional logistic regression model predictor for death or BPD.22 Our matched-pairs analysis took all the
confirmed that the higher incidence of intraventricular haemor- known major contributing factors to PIE into consideration, such as
rhage was a relevant complication following PIE (P = .0499). As birthweight, GA and gender. The matching procedure focused on the
birthweight plays a substantial role in calculating outcome param- main confounders of gestational age and gender, because prematu-
eters in preterm infants, we also analysed the total cohort in a rity is the main risk factor for developing air leaks, especially for PIE, 23
stepwise univariate multivariable regression model, starting with and male gender has been associated with consistently worse pulmo-
birthweight and then adding GA, gender, multiple births, severe nary outcomes.24 Once we excluded those confounding factors, the
RDS and antenatal corticosteroids. The calculated cut-off for the new, and only relevant, maternal risk factors for PIE were revealed as
strongest effect for preeclampsia and HELLP syndrome was a preeclampsia and HELLP syndrome. It had already been reported that
birthweight of 790g based on these findings: birthweight ≤ 790g BPD was increased in infants exposed to preeclampsia.25 Placental
(P = .008) versus birthweight ≥ 790g P = .054 (Appendix S1). These structural abnormalities related to IUGR and preeclampsia have also
data suggest that immature preterm infants had an increased pre- been strongly associated with foetal growth restriction, and stud-
disposition for PIE if their mothers had maternal preeclampsia or ies have linked antenatal stress with poor respiratory outcomes.26
HELLP syndrome. 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
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
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. |
139
They also show that imbalanced circulating proangiogenic and antian- REFERENCES
giogenic factors could impair vasculogenesis in foetal lungs, leading 1. Greenough A, Dixon AK, Roberton NR. Pulmonary interstitial em-
to early disruption in lung maturation. Hypothetically, the preeclamp- physema. Arch Dis Child. 1984;59(11):1046-1051.
2. Gronbach J, Ehrhardt H, Zimmer K-P, Waitz M. Early pulmonary in-
tic intrauterine environment causes a primary insult in the develop-
terstitial emphysema in preterm neonates-respiratory management
ing foetal lung, predisposing it to further postnatal damage, including and case report in nonventilated very low birth weight twins. AJP
PIE. Infants born following preeclampsia are more susceptible to fur- Rep. 2018;8(2):e99-e105.
ther lung impairment and need intensive treatment for respiratory 3. Nuñez-Ramiro A, Aguar M, Cernada M, Parra-Llorca A, Vento M.
Oxygen needs during resuscitation and surfactant to achieve sta-
stability. Preeclampsia has been recognised as a risk factor for RDS
bilisation were independent risks factors for pulmonary interstitial
and BPD.27 The association between preeclampsia and PIE has never emphysema in preterm infants. Acta Paediatr. 2018;107(1):28-32.
been assessed, but is in line with the outcomes reported by those 4. Boothroyd AE, Barson AJ. Pulmonary interstitial emphysema
studies. — a radiological and pathological correlation. Pediatr Radiol.
1988;18(3):194-199.
Although the sample size of our study did not allow us to assess
5. Hummler HD, Parys E, Mayer B, Essers J, Fuchs H, Schmid M. Risk
the associations between preeclampsia and BPD, it was sufficient indicators for air leaks in preterm infants exposed to restrictive use
to show an association between preeclampsia or HELLP and PIE. of endotracheal intubation. Neonatology. 2015;108(1):1-7.
Our analyses clearly indicate the need to include preeclampsia and 6. Balcells C, Del Río R, Riaza L, Rebollo M, Rodriguez-Fanjul J,
HELLP syndrome in future calculations, in order to identify high- Camprubí M. Lung ultrasound: a useful tool for the follow-up of neo-
natal localized interstitial emphysema. J Pediatr. 2015;166(6):1543.
risk infants directly after birth. A limitation of our study was that
7. Berk DR, Varich LJ. Localized persistent pulmonary interstitial em-
there are no uniform criteria for the radiological diagnosis of PIE, physema in a preterm infant in the absence of mechanical ventila-
and some cases represent a kind of a grey zone of cystic or linear ra- tion. Pediatr Radiol. 2005;35(12):1243-1245.
diolucencies in the interstitium radiating from the hilum. Moreover, 8. Behnke J, Lemyre B, Czernik C, Zimmer K-P, Ehrhardt H, Waitz
M. Non-Invasive Ventilation In Neonatology. Dtsch Arztebl Int.
there could be more unknown factors that have not been taken into
2019;116(11):177-183.
consideration. The incidence of PIE could also be influenced by other 9. Yu VY, Wong PY, Bajuk B, Szymonowicz W. Pulmonary interstitial
important, unmeasured prenatal and postnatal risk factors related to emphysema in infants less than 1000 g at birth. Aust Paediatr J.
the mother or the preterm infant. The possible lack of generalisabil- 1986;22(3):189-192.
10. Morley CJ, Davis PG, Doyle LW, Brion LP, Hascoet J-M, Carlin JB.
ity of the present study findings to other populations or other study
Nasal CPAP or intubation at birth for very preterm infants. N Engl J
settings should be taken into account when interpreting the findings. Med. 2008;358(7):700-708.
Due to the small study cohort, our data only suggest that there was 11. Verma RP, Chandra S, Niwas R, Komaroff E. Risk factors and clinical
a trend towards an association between preeclampsia or HELLP and outcomes of pulmonary interstitial emphysema in extremely low
birth weight infants. J Perinatol. 2006;26(3):197-200.
PIE. It is important not to overestimate the possible importance of
12. Lei X, Stangl O, Bösche C, Stuchlik K, Czorba R, Wieg C. Positional
the findings, as further studies are needed to confirm and enhance treatment without mechanical ventilation in a very preterm infant
the risk factors for PIE. The strengths of this study were the short with unilateral pulmonary interstitial emphysema: case report and
study period of two years, which ensured uniform prenatal and post- review of the literature. BMC Pediatr. 2019;19(1):267.
13. Bancalari E, Jain D. Bronchopulmonary dysplasia: 50 years after the
natal treatment strategies and the radiological assessment by two
original description. Neonatology. 2019;115(4):384-391.
independent, blinded experts. Future large-scaled studies on risk 14. Gronbach J, Shahzad T, Radajewski S, et al. The potentials and ca-
factors for PIE should include preeclampsia and HELLP to confirm veats of mesenchymal stromal cell-based therapies in the preterm
their impact on its incidence. 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.
5 | CO N C LU S I O N 16. Haram K, Svendsen E, Abildgaard U. The HELLP syndrome: clini-
cal issues and management. A Review. BMC Pregnancy Childbirth.
Preeclampsia and HELLP syndrome were identified as possible ad- 2009;9:8.
17. Steegers EAP, von Dadelszen P, Duvekot JJ, Pijnenborg R. Pre-
ditional risk factor for PIE and should be taken into account in future
eclampsia. Lancet. 2010;376(9741):631-644.
risk calculations. Infants who are delivered early due to preeclamp- 18. Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit
sia and HELLP syndrome need to receive improved treatment ap- Care Med. 2001;163(7):1723-1729.
proaches that are dedicated to preventing PIE. 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.
C O N FL I C T O F I N T E R E S T 20. Ackerman NB, Coalson JJ, Kuehl TJ, et al. Pulmonary interstitial em-
The authors have no conflicts of interest to declare. 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
ORCID
oxygen causes less oxidative stress, inflammation, and chronic lung
Judith Behnke https://orcid.org/0000-0003-4307-8759 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
1
Division of Pediatrics, Department
of Clinical Science, Intervention and Abstract
Technology (CLINTEC), Karolinska Institutet, Aim: To describe outcome linked to neonatal cholestasis in a defined cohort of very
Stockholm, Sweden
2 preterm infants.
Department of Pediatrics, Södertälje
Hospital, Södertälje, Sweden Methods: Population-based retrospective case-control study of preterm infants,
3
Department of Pediatrics, Karolinska gestational age <30 weeks, surviving for 28 days, in Stockholm County. Cholestasis
University Hospital, Stockholm, Sweden
4
was defined as conjugated bilirubin ≥30 μmol/L exceeding 20% of total level at least
Department of Neonatology, Karolinska
University Hospital, Stockholm, Sweden twice and graded as high if exceeding 100 μmol/L. Cholestatic cases were matched
on gestational week with two non-cholestatic controls.
Correspondence
Jonas Teng, BUMM, Södertälje Sjukhus AB, Results: The incidence rate of cholestasis was 37/250 (14.8%), with increasing rates
152 86 Södertälje, Sweden. in lower gestational weeks. Perinatal factors associated with cholestasis were pre-ec-
Email: jonas.teng@ki.se
lampsia and being born small for gestational age. Cholestatic infants had three times
Funding information more bronchopulmonary dysplasia and eight times more retinopathy of prematurity.
Funds were received from the Swedish
Order of Freemasons and Samaritan The mortality was 13.5% in cholestatic infants versus 2.7% in controls (P = .040). All
Foundation for the study. 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
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
2.1 | Study design
We performed a population-based retrospective case-control as the time elapsed from onset of cholestasis, as defined above, to
study of all preterm infants born and treated in Stockholm County when conjugated bilirubin no longer exceeded 30 μmol/L.
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 2.3 | Development of chronic liver disease
their mothers were identified through the Swedish neonatal qual-
ity register. Cholestatic infants were detected through medical The development of chronic liver disease of clinical significance was
chart review. For every cholestatic case, two non-cholestatic con- defined as the infant being enrolled at any time, after discharge from
trol subjects born in the same gestational week were randomised neonatal care, as a patient at any of the tertiary centres for paediatric
from the population. hepatology in Karolinska University Hospital in Stockholm or Queen
Data from the perinatal period were primarily acquired from Silvia Children's Hospital in Gothenburg until the age of 10 years. These
medical charts review. Small for gestational age was defined centres are the only tertiary paediatric units for hepatology in Sweden.
as birth weight less than 2 standard deviations from expected. All children with more severe liver disease, including progressive
Bronchopulmonary dysplasia (BPD) was defined as need for oxygen chronic liver disease with or without decompensation and acute liver
at 36 weeks' gestational age and/or continuous positive pressure failure, will be referred there from all paediatric units in the country.
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 2.4 | Statistical methods
supplementary table (Table S1).
Selected data from this cohort have previously been reported on The data analysis was performed using SPSS version 24.0 for Mac
in a study comparing the use of two different types of fat emulsions (IBM). For the univariate analyses, Fisher's exact test was used for
12
for total parenteral nutrition. proportions. For continuous variables, Mann-Whitney U test was
Ethical approval was received from the Stockholm regional ethi- used, and medians were used to present the central tendency and
cal review board (ref 2008/1558-31/3). 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-
2.2 | Definitions of cholestasis tivariate analyses, binary logistic regression was used, and results
presented as crude and adjusted odds ratios (OR) and 95% confi-
Cholestasis was defined as conjugated serum bilirubin (CB) dence intervals (CI). Statistical significance level was set at 5%.
≥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 3 | R E S U LT S
The onset of cholestasis was defined as the age when the bili-
rubin levels fulfilled the cholestasis criteria for the first time. The 3.1 | Population characteristics
grade of cholestasis was arbitrarily dichotomised as low, if peak
conjugated bilirubin <100 μmol/L, and as high, if peak conjugated During the study period, 296 infants with a gestational age of less
bilirubin was ≥100 μmol/L. The duration of cholestasis was defined than 30 weeks were born. A total of 46 infants were excluded, 22
TENG et al. |
143
because of transfer to other regions and 24 because of short life group) and 62/74 (84%) in controls (P = .38). The nine infants in the
span, leaving a study population of 250 infants. Among the 46 ex- cholestasis group with cardiac anomalies constituted five cases of
cluded infants, five had already developed cholestasis. Out of these, ventricular septal defects, one case of atrioventricular septal defect,
two were cholestatic before death and three before transfers to two cases of atrial septal defects and one case of combined aortic
other regions. stenosis and aortic insufficiency. Among controls, there were one
Cholestasis criteria were fulfilled in 37/250 (14.8%) of the in- case of valvular pulmonary stenosis and one case of right ventric-
fants in the study population. There was a significant trend of higher ular heart failure due to twin-twin transfusion syndrome. The rate
incidence rate of cholestasis with lower gestational age (P = .034 of patent ductus arteriosus, in need of treatment with ibuprofen or
Mantel-Haenszel test of trend), as shown in Figure 1. surgically, did not differ significantly between groups.
The cholestatic infants needed more active neonatal treatment,
such as more mechanical ventilation, parenteral nutrition and eryth-
3.2 | Cholestatic versus non-cholestatic infants rocyte transfusions (Table 1). The use of surfactant treatment did
not differ between groups. The cholestatic infants regained their
When comparing the cholestatic infants with the non-cholestatic birth weight faster than the control infants, but they had a lower
controls in the univariate analysis, several differences were noted, as weight at 36 weeks of corrected age (Table 1).
detailed in Table 1 and below. The median maternal age was 32 years The cholestatic infants had more neonatal morbidities than the
in both groups. The frequency of caesarean delivery was 65% among controls, including more necrotising enterocolitis and need for lap-
cholestatic infants and 62% among controls. The median Apgar score arotomy (Table 1). This was also reflected in the fact that choles-
at 5 minutes was 8 in both groups. The occurrence of singletons, pro- tatic infants needed bowel resection surgery and stoma significantly
longed rupture of membranes more than 24 hours, chorioamnionitis, more often (data not shown).
use of antenatal steroids and need for surfactant treatment did not The cholestatic infants also had a higher rate of culture veri-
differ between groups (data not shown). The rate of pre-eclampsia fied sepsis (Table 1). This was still the case when clinically probable,
in the mothers of the non-cholestatic infants was 8/74 (11%). In the but blood culture negative infections were added (data not shown).
mothers of the cholestatic infants, 12/37 (32%) had pre-eclampsia, Also, multiple episodes of sepsis were more common among
which was significantly higher than in the controls (P = .008). cholestatic infants. The late-onset sepsis episodes, debuting at age
With regard to neonatal characteristics, the cholestatic infants 72 hours or later, were responsible for the difference in sepsis in-
were small for gestational age in 38% of cases, as compared to 16% cidence between the groups (Table 1). Early-onset sepsis occurred
among the controls (P = .02). This was also reflected in a significant in two infants in each group, which did not differ significantly. The
difference in birth weight (Table 1). The occurrence of congenital median age at onset of sepsis did not differ between groups; it was
heart anomalies was 9/37 (24%) among the cholestatic infants and 10.5 days for those who had at least one culture verified sepsis
2/74 (3%) among the controls (P = .001). This difference did not seem episode in each group. Of the cholestatic infants, 7/37 (19%) were
to be due to different investigation rates. The proportion of infants treated for culture verified fungal sepsis versus 5/74 (7%) of the
examined at least once with echocardiography was 34/37 (92%) non-cholestatic (P = .10). One cholestatic and one non-cholestatic
in cholestatic infants (89% in the low-grade and 95% high-grade infant were positive for cytomegalovirus in urine screening tests.
70
Cholestatic Non-cholestatic
60
50
40 64
30
41
43
20 30
F I G U R E 1 Proportion of cholestatic 24
cases per gestational week. Significant 10 5
trend of higher proportion of cholestasis 6 7 8 9
3 5 4
in lower gestational weeks (P = .034*, 0 1
Mantel-Haenszel test of trend) 23 24 25 26 27 28 29
144 | TENG et al.
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. The rate of intraventricular haemorrhage was the same in both
Cholestatic infants had more retinopathy of prematurity (ROP) and BPD groups (data not shown). Death before discharge was more common
than controls (Table 1). To further control for disease severity and other in cholestatic infants, where 5/37 (13.5%) died, as compared with
risk factors for poor neonatal outcome, the results were tested in a mul- 2/74 (2.7%) in controls (P = .04). The median age at death was 65 days
tivariate model using logistic regression. The crude OR for ROP grade 3 (range 28-108) among the cholestatic infants. The two deceased
or more, for cholestatic infants, was 3.6 (95% CI 1.4-9.4). When adjusted non-cholestatic infants died after 35 and 461 days, respectively. The
for gestational age, gender and mechanical ventilation >2 weeks, the latter had severe BPD resulting in respiratory insufficiency, requiring
adjusted OR was 8.4 (95% CI 2.1-33.4). For BPD, the crude OR was 2.8 continuous positive airway pressure support and extra oxygen, and
(95% CI 1.2-6.3), and the adjusted OR was 3.2 (95% CI 1.2-8.6). was never able to leave the hospital.
TENG et al. |
145
3.3 | Course and severity of cholestasis The rates of infection were also similarly distributed. Neither the
total burden of infections nor their relationship in time with onset
Details regarding the course of cholestasis for all cholestatic infants of cholestasis seemed associated with the severity of cholestasis.
are shown in Table 2. This result was not affected by adding clinically suspected sepsis ep-
The two groups, defined as having high-grade (peak conjugated isodes with negative blood cultures to the analyses (data not shown).
bilirubin ≥100 μmol/L) and low-grade cholestasis (<100 μmol/L), The infants with conjugated bilirubin peaking above 100 μmol/L
were compared. The cholestasis was graded as high in 19 and low in had a longer duration of cholestasis. Median duration of cholestasis
18 infants using this cut-off level. Notably, the groups did not differ in this group was 87 days (IQR 45-112) as compared to 45.5 days
in many aspects. (IQR 18-62) among low-grade infants (P = .005). Data from five in-
However, the mortality was higher in the high-grade group, since fants in the former group were missing due to death before resolu-
5/19 (26%) of the high-grade cholestatic infants died versus none in tion of cholestasis and two in the latter group due to lack of blood
the low-grade group (P = .046). All of the deceased infants had ongo- tests regarding bilirubin after discharge from the hospital. The rate
ing and considerable cholestasis at the time of death. infants were treated with ursodeoxycholic acid were 79% and 56%
The baseline characteristics and the neonatal treatments showed in the high- and low-grade groups, respectively (P = .17).
for all cholestatic infants in Table 1 showed no significant differences The rate of ROP and BPD did not differ significantly between
between the subgroups (data not shown). Parenteral nutrition, be- the groups.
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- 3.4 | Long-term hepatic outcome following neonatal
grade cholestatic infants and 4 days among the low-grade (P = .051). cholestasis
In the high-grade group, necrotising enterocolitis tended to be
more common, 11/19 developed necrotising enterocolitis versus Among those 101 out of 111 children surviving to discharge from
5/18 in the low-grade group (P = .10). The other morbidities pre- the NICU and remaining in Sweden, no child had developed clini-
sented in Table 1 were similar between the subgroups (data not cally significant chronic liver disease at the age of 10 years, neither
shown). 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-
TA B L E 2 Characteristics of cholestasis in cholestatic infants
yond. Repeated ultrasonography investigations showed gallstones
(n = 37)
with unchanged small size, but no other hepatobiliary abnormalities.
Cholestasis onset, age, days, median (IQR) 24 (17-39)
The liver function tests were repeatedly normal.
Peak TB, μmol/L, median (IQR) 148 (103-245)
Five children from each group were missing from this analysis.
Peak CB, μmol/L, median (IQR) 104 (61-169) During neonatal care, five out of 37 of the cholestatic infants and
Age at peak CB, days, median (IQR) 42 (31-65) two out of 74 non-cholestatic controls were deceased. Three con-
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) 4 | D I S CU S S I O N
median (IQR)
Total fasting before cholestasis, days, median 7 (2-12) In this population-based cohort of very preterm infants, cholestasis
(IQR)
was common, with an overall incidence of 14.8%, being the high-
Necrotising enterocolitis before cholestasis onset 13/37 (35.1%)
est in lower gestational ages. We found higher rates of inflamma-
Necrotising enterocolitis after cholestasis onset 3/37 (8.1%) tory complications such as BPD and ROP rates in cholestatic infants
Infections before cholestasis onset than in non-cholestatic controls matched for gestational age. High-
Sepsis, culture verified or clinical 32/37 (86.5%) grade cholestasis with maximum conjugated bilirubin exceeding
Sepsis, culture verified, any species 21/37 (56.8%) 100 μmol/L was associated with a higher mortality than cholestasis
Sepsis, coagulase negative staphylococci 17/37 (45.9%) of lower grade. None of the surviving infants with neonatal choles-
Fungal infection, culture verified 4/37 (10.8%) tasis had developed chronic liver disease at 10 years of age.
Pre-eclampsia during pregnancy and low birth weight for ges-
Note: Data presented as absolute number/total number (per cent) if a
proportion, median (IQR) for continuous variables. tational age emerged as risk factors for cholestasis, based on the
Abbreviations: CB, conjugated bilirubin; IQR, interquartile range; TB, differences between the cholestatic infants and non-cholestatic
total bilirubin. controls. An association between cholestasis and growth restric-
a
n = 30. 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- selection bias, as the rate of infants in both groups examined by
tory condition, and both pre-eclampsia and being born small for echocardiography was high. Whether the development of cholesta-
gestational age have been associated with an increase in markers sis in any way could be related to circulatory effects on the liver
for systemic inflammation at two to 4 weeks of age in very preterm remains to be established. However, direct hyperbilirubinaemia was
infants.15,16 Pre-eclampsia or being small for gestational age is also found in 17% of term and preterm neonates with congenital heart
17-19
associated with more ROP and BPD in very preterm infants. disease by Fujishiro et al,30 which is an incidence rate much higher
20,21
Inflammation contributes to the development of BPD, and than expected.
there is a correlation between BPD and ROP. 22 Hence, inflamma- Somewhat surprisingly, the cholestatic infants regained their
tion in these susceptible organs, supposedly promoted by a gen- birth weight faster than the control infants, but they had a signifi-
erally proinflammatory environment, seems to play a role in the cantly lower weight at 36 weeks of corrected age. Infants in the
multifactorial pathogenesis. Interestingly, inflammation may also cholestatic group were generally sicker than controls, and this ini-
play a role in the development of cholestasis in these infants. For tially larger weight gain could tentatively be explained by more in-
example, DeMauro et al23 described that elevated c-reactive pro- tensive fluid therapy and parenteral nutrition.
tein and interleukin-1β at 2 weeks of age in very low birth weight in- We noted that all deaths in the cholestasis group occurred in
fants was predictive of cholestasis. Alison et al10 showed that liver infants with peak conjugated bilirubin >100 μmol/L and that the
stiffness was higher in infants with intra-uterine growth restriction cholestasis in all these cases was ongoing without signs of resolution
than in preterm infants with birth weight appropriate for gesta- at the time of death. All deceased infants died in episodes of multi-
tional age, and even higher in growth-restricted infants who de- ple organ failure, precipitated by sepsis or necrotising enterocolitis.
veloped cholestasis. We therefore hypothesise that the increased The fatal cases were too few to be analysed in a multivariate model.
risk of cholestasis in infants with maternal pre-eclampsia, growth However, in the univariate comparison, we can conclude that the
restriction and being born small for gestational age in our cohort neonatal course, morbidities and complications were very similar
might be linked to a proinflammatory state in these infants. This in the low-grade and high-grade cholestasis groups. However, our
warrants further studies as conclusions on any causational associa- results did suggest that higher grade of cholestasis may contribute
tion must be made with great caution because of the retrospective, to higher mortality, a finding which needs to be verified in further
observational design of this study, and thereby the risk of unknown studies.
bias or confounders. One important finding was that none of the very preterm
The occurrence of adverse neonatal outcomes such as ROP, BPD cholestatic infants, who survived the neonatal period, had de-
and death was significantly more frequent among cholestatic very veloped clinically significant chronic liver disease at 10 years of
preterm infants than their non-cholestatic controls. We evaluated age. This type of very long-term follow-up has to our knowledge
the role of cholestasis in a multivariate model for BPD and ROP, not been reported earlier. It should be noted that this outcome
where the association remained after adjusting for gender, gesta- measurement did not include any test for liver function, only the
tional age and duration of mechanical ventilation. It could be sug- fact that the children had not been referred to the tertiary referral
gested that cholestasis is linked to the development of ROP and BPD centres of paediatric hepatology in Sweden. However, the devel-
in very preterm infants, possibly mediated by a common pathway of opment of significant liver disease most likely would have been
increased inflammatory state. 21-24 discovered by this age and if so the tertiary centres of paediatric
The cholestatic infants were generally sicker, had a more compli- hepatology would have been involved. Of the children still living in
cated neonatal period with prolonged need for parenteral nutrition, Sweden at the age of 10 years, none had died between discharges
and more episodes of sepsis than the non-cholestatic controls, in line from the neonatal care until the age of 10 years. Only three infants
with reports by others.3,5 The sepsis episodes were predominantly in the control group had moved abroad, but all surviving children
late-onset sepsis, possibly related to prolonged need for parenteral who were cholestatic as newborns still lived in Sweden at the age
nutrition and central line catheters, which is known to be associated of 10 years. We conclude that the rate of more severe chronic
25
with increased risk for infection. Infections favour inflammation, liver disease, that manifests itself as cholestasis in very preterm
and sepsis has been shown to be associated with liver dysfunction in infants, is low. The sample size of 32 surviving preterm infants
different groups of neonates.5,26-28 Recent findings in a small group may not be big enough to totally exclude that neonatal cholestasis
of preterm infants on parenteral nutrition showed that the intesti- is associated with later liver disease, but it is large enough to sug-
nal microbiome was different even before cholestasis and shifted gest that there is not a strong association. Prospective research
towards gram-negative overweight in cholestatic infants compared including liver function tests at follow-up would be needed to an-
to the non-cholestatic infants receiving similar amount of parenteral swer this question more definitely. Even so, our finding supports
nutrition. 29 The possible role of the intestinal mibrobiome in associa- the current perception that cholestasis in preterm infants is a pri-
tion to inflammation and cholestasis in this and other patient groups marily reversible condition in an otherwise healthy but immature
warrants further investigation. liver that resolves if the underlying conditions causing the need
Mild cardiac anomalies were more common in the cholestatic for prolonged parenteral nutrition, and other risk factors, can be
infants compared to controls. The finding was not explained by managed.
TENG et al. |
147
The strengths of this study are the population-based approach, 7. Costa S, Maggio L, Sindico P, Cota F, De Carolis MP, Romagnoli
C. Preterm small for gestational age infants are not at higher
a relatively large cohort of very preterm infants, and the long-term
risk for parenteral nutrition-associated cholestasis. J Pediatr.
perspective. The limitations include the retrospective design, mak- 2010;156(4):575-579.
ing conclusions on causality difficult. Another limitation is that exact 8. Robinson DT, Ehrenkranz RA. Parenteral nutrition-associ-
doses of for example ursodeoxycholic acid and components of paren- ated cholestasis in small for gestational age infants. J Pediatr.
2008;152(1):59-62.
teral nutrition were not available. Also, since blood tests, diagnostic
9. Baserga MC, Sola A. Intrauterine growth restriction impacts tol-
procedures and most variables were based on clinical decisions from erance to total parenteral nutrition in extremely low birth weight
the attending neonatologist rather than part of a prospective study infants. J Perinatol. 2004;24(8):476-481.
protocol, there was risk for misclassifications. However, this likely 10. Alison M, Biran V, Tanase A, et al. Quantitative shear-wave elastog-
has led to underestimation rather than overestimation of cholestasis raphy of the liver in preterm neonates with intra-uterine growth
restriction. PLoS One. 2015;10(11):e0143220.
incidence.
11. Kapoor V, Malviya MN, Soll R. Lipid emulsions for parenterally
fed term and late preterm infants. Cochrane Database Syst Rev.
2019;6:CD013171.
5 | CO N C LU S I O N 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.
We conclude that neonatal cholestasis is common in very preterm 13. Habli M, Levine RJ, Qian C, Sibai B. Neonatal outcomes in preg-
infants and it is closely linked to disease severity and complications nancies with preeclampsia or gestational hypertension and in nor-
in the neonatal period. Our results suggest that cholestasis may be motensive pregnancies that delivered at 35, 36, or 37 weeks of
gestation. Am J Obstet Gynecol. 2007;197:406.e1-406.e7.
an independent risk factor for inflammatory complications such as
14. Wikstrom AK, Stephansson O, Cnattingius S. Previous preeclamp-
BPD and ROP in these patients and this needs be further addressed sia and risks of adverse outcomes in subsequent nonpreeclamptic
in prospective studies. More severe cholestasis may be associated pregnancies. Am J Obstet Gynecol. 2011;204(2):148.e1-148.e6.
with increased mortality, whereas cholestatic very preterm infants 15. McElrath TF, Allred EN, Van Marter L, Fichorova RN, Leviton
A, Investigators ES. Perinatal systemic inflammatory re-
who survive the neonatal period do not seem to be at great risk to
sponses of growth-restricted preterm newborns. Acta Paediatr.
develop severe chronic liver disease later in childhood. 2013;102(10):e439-e442.
16. Leviton A, Allred EN, Fichorova RN, et al. Antecedents of inflam-
C O N FL I C T S O F I N T E R E S T mation biomarkers in preterm newborns on days 21 and 28. Acta
Paediatr. 2016;105(3):274-280.
The authors have no conflicts of interest to declare.
17. Jensen EA, Foglia EE, Dysart KC, et al. Adverse effects of small for
gestational age differ by gestational week among very preterm in-
ORCID fants. Arch Dis Child Fetal Neonatal Ed. 2019;104(2):F192-F198.
Jonas Teng https://orcid.org/0000-0003-4351-9665 18. Bose C, Van Marter LJ, Laughon M, et al. Fetal growth restriction
Kajsa Bohlin https://orcid.org/0000-0003-3368-4149 and chronic lung disease among infants born before the 28th week
of gestation. Pediatrics. 2009;124(3):e450-e458.
Björn Fischler https://orcid.org/0000-0002-8838-324X
19. Hansen AR, Barnes CM, Folkman J, McElrath TF. Maternal pre-
eclampsia predicts the development of bronchopulmonary dyspla-
REFERENCES sia. J Pediatr. 2010;156(4):532-536.
1. Fawaz R, Baumann U, Ekong U, et al. Guideline for the evaluation 20. Lal CV, Ambalavanan N. Biomarkers, early diagnosis, and clin-
of cholestatic jaundice in infants: joint recommendations of the ical predictors of bronchopulmonary dysplasia. Clin Perinatol.
north american society for pediatric gastroenterology, hepatol- 2015;42(4):739-754.
ogy, and nutrition and the European society for pediatric gastro- 21. Ambalavanan N, Carlo WA, D'Angio CT, et al. Cytokines associated
enterology, hepatology, and nutrition. J Pediatr Gastroenterol Nutr. with bronchopulmonary dysplasia or death in extremely low birth
2017;64(1):154-168. weight infants. Pediatrics. 2009;123(4):1132-1141.
2. Koseesirikul P, Chotinaruemol S, Ukarapol N. Incidence and risk 22. Podraza W, Michalczuk B, Jezierska K, et al. Correlation of retinop-
factors of parenteral nutrition-associated liver disease in newborn athy of prematurity with bronchopulmonary dysplasia. Open Med
infants. Pediatr Int. 2012;54(3):434-436. (Wars). 2018;13:67-73.
3. Lee HH, Jung JM, Nam SH, Lim G, Chung ML. Risk factor analysis of 23. DeMauro SB, Kilpatrick LE, Gerdes JS, Abbasi S. Early inflammatory
parenteral nutrition-associated cholestasis in extremely low birth markers for prediction of cholestasis in very-low-birth-weight in-
weight infants. Acta Paediatr. 2016;105(7):e313-e319. fants. Neonatology. 2012;102(3):229-234.
4. Lee SM, Namgung R, Park MS, et al. Parenteral nutrition associated 24. Pawlik D, Lauterbach R, Walczak M, Hurkala J, Sherman MP. Fish-
cholestasis is earlier, more prolonged and severe in small for gesta- oil fat emulsion supplementation reduces the risk of retinopathy
tional age compared with appropriate for gestational age very low in very low birth weight infants: a prospective, randomized study.
birth weight infants. Yonsei Med J. 2013;54(4):839-844. JPEN J Parenter Enteral Nutr. 2013;38(6):711-716.
5. Tufano M, Nicastro E, Giliberti P, Vegnente A, Raimondi F, Iorio R. 25. Shah J, Jefferies AL, Yoon EW, Lee SK, Shah PS, Canadian NN.
Cholestasis in neonatal intensive care unit: incidence, aetiology and Risk factors and outcomes of late-onset bacterial sepsis in
management. Acta Paediatr. 2009;98(11):1756-1761. preterm neonates born at < 32 weeks' gestation. Am J Perinatol.
6. Christensen RD, Henry E, Wiedmeier SE, Burnett J, Lambert DK. 2015;32(7):675-682.
Identifying patients, on the first day of life, at high-risk of devel- 26. Yan W, Hong L, Wang Y, et al. Retrospective dual-center study of
oping parenteral nutrition-associated liver disease. J Perinatol. parenteral nutrition-associated cholestasis in premature neonates:
2007;27(5):284-290. 15 years' experience. Nutr Clin Pract. 2017;32(3):407-413.
|
148 TENG et al.
DOI: 10.1111/apa.15372
REGULAR ARTICLE
1
Paediatric Feeding International, Sydney,
NSW, Australia Abstract
2
Department of Psychology, and the Aim: Paediatric feeding disorders are normally managed by specialist clinics. We ex-
School of Health Sciences, University of
amined whether treatment gains were maintained when trained parents continued
Canterbury/Te Whare Wānanga o Waitaha,
Christchurch, New Zealand the programme at home and during meals out.
Methods: This controlled consecutive case series recruited 26 children (22 boys) with
Correspondence
Tessa Taylor, Department of Psychology, avoidant/restrictive food intake disorder, from a private paediatric feeding disorders
Speech & Hearing, and the School of Health
practice in New South Wales, Australia. Their mean age was six (2-13) years. All had
Sciences, University of Canterbury/Te
Whare Wānanga o Waitaha, Private Bag severe feeding problems and mealtime skill deficits, and most had autism and de-
4800, Christchurch 8140, New Zealand.
velopmental delays or intellectual disabilities. The children received intensive, indi-
Email: DrTaylor@PaediatricFeedingIntl.com
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 disabilities. 2,3 Many meet the criteria for avoidant/restrictive food
intake disorder (ARFID), which includes failure to thrive, nutritional
Chronic and severe paediatric feeding disorders can include being deficiencies, dependence on artificial feeding methods and, or, di-
dependent on a gastrostomy tube or liquids, refusing food and only etary supplements and marked impairment in psychosocial function-
eating certain items. Feeding problems occur in wide-ranging popu- ing.4 When this occurs during critical neuro-developmental periods,
lations, including children with typical development, medical com- numerous adverse effects on behaviour, cognition, emotion and
plications, such as food allergies and reflux,1 and developmental health may occur and persist into adulthood.3-7
Over the past 40 years, treatment for paediatric feeding disor-
Abbreviations: ARFID, avoidant/restrictive food intake disorder; ES, effect size; CI, ders has been based on behaviour analysis principles and research
confidence interval.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
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 control, using within-subjects, single-case experimental designs,
duration keys measured delays in seconds before acceptance and where each case served as its own control. 23 The most frequently
swallowing occurred and the duration of negative vocalisation. used design, for more than 80% of the children, was a reversal, or
Frequency of swallowing was divided by the number of therapist- withdrawal, design. This involved taking the treatment out, namely
programmed bite presentations during the session to provide an baseline two, and putting it back in when needed, namely treat-
overall measure of the percentage of food consumed. The last ment two. This was because replication showed that the treatment
three data points in the initial phases, namely baseline one and was needed and was the reason for the improvement in eating.
treatment one, were averaged, as were the repeated baseline two The outcomes were examined using a combination of graphic tech-
and treatment two phases. The plates were photographed to re- niques24 and effect size estimation, using Cohen's dav, where the
cord the variety of the food eaten, and the food was measured standardiser was the average of the standard deviations before
before and after eating to determine the volume consumed. All and after treatment. 25 The graphs were predominantly modified
19
parents were provided with a questionnaire at the end of the Brinley plots (Figure 1;26). These are scatterplots that display indi-
programme, and the responses to the five-point Likert scale items vidual changes over time, with each data point representing coor-
were averaged to assess the 23 programme satisfaction items and dinates of an individual's score at time one on the x-axis and time
the 16 items on social acceptability of the treatment. All the ses- two on the y-axis.
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 2.4 | Procedure
for the therapist and 0.2 for the parents, and the mean interob-
server agreement was 99.5%. 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
2.3 | Experimental design and data analysis during which they received treatment for an average of 11 (6-21.5)
days. The first author was the principal therapist and conducted the
The study included all cases that met the inclusion criteria, regard- sessions for approximately 7-8 hours per day with a trained assistant.
22
less of outcome, to reduce selection and publication bias. This The involvement of the therapist reduced as the parents developed
was consistent with the consecutive, controlled case series experi- the skills they needed and the responsibility for the intervention was
mental design. All the 26 cases also demonstrated experimental gradually transferred to them.
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.
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 end of the programme, they were asked to offer their child the foods
on the discharge food list, record mealtime data, take photographs
The review of the intake assessment included the physician's report of the plates and video the meals. The data were returned to the
and reports from other disciplines and service providers, meal obser- first author for review and follow-up. The parents were contacted
vations, intake questionnaires, parental interviews, three-day diet 2-3 years after treatment, at a mean of 2.3 years, and asked to com-
records, functional assessment, case-conceptualisation and therapy plete a five-point Likert satisfaction survey, which ranged from one
goal specification. We directly assessed each child's food and reward for worse than pre-treatment to five for resolved.
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 3 | R E S U LT S
8-10,12
well-established, empirically supported treatments. Treatment
involved changing the mealtime environment, such as providing Figure 2 presents the primary outcome data, which was consump-
incentives for swallowing and praising appropriate mealtime be- tion, and demonstrates the experimental control as horizontal dot
haviour. Treatment also involved setting achievable mealtime re- plots. Food consumption was very low at the initial baseline meas-
quirements, such as continuing to present food and liquids despite ure, baseline one, with most children not consuming any of the food
rejection, replacing food or liquid that was spat out and providing presented. At the end of the first treatment phase, treatment one,
prompting instructions and guidance. We also taught the skills children were completing 100% of the programmed bites consist-
needed to achieve goals, such as chewing, using utensils, self-feed- ently, within a mean of 46 (2-134) minutes of treatment starting.
ing and drinking, drinking with an open cup, biting from whole food When treatment was withdrawn as a replication procedure, at base-
portions rather than pre-cut bite sizes and taking medicine. As the line two, consumption decreased for all children. When treatment
treatment progressed, intensive parental behavioural skills training resumed in the second treatment phase, namely treatment two, con-
was provided and we varied the cutlery and crockery provided, how sumption increased again for all children, with no overlap of points
the children were placed at family meals and the setting where they with the baseline, thus replicating the treatment effect. At the end
ate19 to ensure generalisation of therapy gains. 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-
2.6 | Maintenance and follow-up 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
To help the children maintain their treatment gains, the parents were (Figure 3). One child could only eat 30 different foods due to aller-
given specific maintenance instructions. For two weeks after the 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 goal levels at baseline. Other reductions were 99% for food refusal,
for all children, inappropriate mealtime behaviour, expulsion and 97% for food expulsion and 97% for negative vocalisation. All the
negative vocalisation at the end of treatment were low or reduced. children met 100% of the treatment goals, including learning to take
The overall effects sizes for all three measures were large (Cohen's medication, chewing, feeding themselves with utensils, drinking
dav −1.4 to −2.35). Before treatment, there were long delays to tak- from a cup and sitting for meals.
ing and swallowing bites which decreased after treatment (Figure 6). At discharge, 20 parents (77%) completed the questionnaire.
Individual differences in the nature and texture of the food and how They reported high satisfaction (mean 4.9, range 4.6-5, maximum
skilled the children were at chewing and swallowing accounted for 5.0) and social acceptability of the treatment (mean 4.9, range
the considerably lower Cohen's dav effect size for swallowing, of 4.6-5, maximum 5.0). During the immediate follow-up period of
about −1, than for accepting food when presented (−2.6). two weeks, 17 parents reported on a mean of 25 (4-135) meals.
The Cohen's dav effect sizes were large to very large (range They reported high and stable consumption of the food presented
1.4-4) across the different measures and in no instance did the 95% (mean 99%, range 90%-100%). Anecdotally, families reported im-
confidence interval (95% CI) indicate that the effect size was close provements in other areas of life and well-being, for example in
to zero, as the associated mean change was always statistically sig- height, weight and body mass index, bloodwork for nutritional
nificant (P < .05). When we measured the treatment effect, based on deficiencies, hair quality and growth, toothbrushing, sleeping
the average percentage reduction from baseline to the end of the and toileting. They also reported improvements in their child's
treatment, it was 98% for non-consumption. Delays in acceptance behaviour, such as being able to change between activities and
fell by 97% and 75% for delays to an empty mouth, but the figure follow instructions. Improvements in learning were also reported,
was 89% when we excluded the five children who had reached these such as progress in therapy and school. Other benefits were
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
45
Protein Starch Veges Fruit Combo
40
Final Treatment - Number consumed
35
30
25
related to energy, sports and social participation, travel, immunity, that their children were eating from all food groups. The exception
skin colour and decreased dependence on items like supplements, was a child who had similar before and after results as moving home
formulas and laxatives. One example of growth improvement was meant their parents did not maintain the protocol.
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 4 | D I S CU S S I O N
he had rapidly grown 25 cm in height and this was confirmed by
his therapist. The data analysis used in this study showed both individual and over-
At follow-up, 21 parents reported that their child's feeding prob- all group patterns of change following therapy. The consistently large
lem was much better than before treatment and five said it had com- size of change seen in the figures was confirmed by the standardised
pletely resolved. The mean Likert score was 4.12 (2-5). No parents mean difference effect size and 95% CI estimation, with Cohen's dav
reported the problem was worse. All but one of the parents reported 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.
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
Time taken (seconds) to accept and swallow bites comparing before treatment to final treatment
Latency to Acceptance Latency to Swallow (Mouth Clean)
Final Treatment
180
Latency (sec)
120 dav = –2.59 [–3.5, –1.7] dav = –1.3 [–1.95, –.65]
60
0
0 60 120 180 240 300 360 420 480 540 600 0 60 120 180 240 300 360 420 480 540 600
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 help many more families coping with severe and chronic paediatric
hospitalisation and multidisciplinary teams if they have swallowing feeding disorders.
problems, severe food allergies, breathing issues, seizures and blood
sugar issues, such as glycogen storage disease. Treatment intensity C O N FL I C T O F I N T E R E S T
and duration needs to be individually calibrated to the severity of The authors have no conflicts of interest to declare.
cases and decisions about treatment should be data-driven and
based on progress towards goals. If not, there is a risk of iatrogenic ORCID
worsening of the feeding disorder.4,11,19 It is important to note that Tessa Taylor https://orcid.org/0000-0001-8643-3955
the current study was not conducted in the same way as a conven- Neville Blampied https://orcid.org/0000-0002-0158-4904
tional outpatient service, as we provided continuous sessions for full
consecutive days. However, this intensive treatment produced rapid REFERENCES
improvements in feeding compared to typical, conventional outpa- 1. Yeung KA, Taylor T, Scheimann A, et al. The prevalence of food al-
tient treatment. Finally, we note that proper expertise, notably years lergies in children referred to a multidisciplinary feeding program.
Clin Pediatr. 2015;54(11):1081-1086.
of training at the hospital where this treatment was developed, is
2. Taylor T, Kozlowski AM, Girolami PA. Comparing behavioral treat-
needed to run this treatment programme.4,11,19 ment of feeding difficulties and tube dependence in children with
Another strength of the study was that quick improvements cerebral palsy and autism spectrum disorder. NeuroRehabilitation.
were shown, even though many children had unsuccessful prior 2017;41(2):395-402.
3. Kozlowski AM, Taylor T, González ML, Girolami PA. Feeding dis-
treatment, often involving a range of disciplines and approaches.
orders. In: Matson JL, Matson ML, eds. Comorbid Conditions in
In addition, no families withdrew because the parents thought the Individuals with Intellectual Disabilities. Cham, Switzerland: Springer;
feeding disorder had improved while they were on the waiting list. 2015.
On the contrary, parents typically reported that the feeding problem 4. González ML, Mulderink TD, Girolami PA. Avoidant restrictive food
intake disorder. In: Maragakis A, O'Donohue WT, eds. Principle-
had worsened. Feeding disorders become harder to treat as children
Based Stepped Care and Brief Psychotherapy for Integrated Care
get older and treatment options become limited and more difficult. Settings. Cham, Switzerland: Springer; 2018.
That is why early, effective intervention is very important, both for 5. Zucker N, Copeland W, Franz L, et al. Psychological and psycho-
the well-being of the child and their family and to avoid wasting lim- social impairment in preschoolers with selective eating. Pediatrics.
ited health system resources.7 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.
5 | CO N C LU S I O N 7. Taylor T, Taylor S. Let's not wait and see: the substantial risks of
paediatric feeding problems. Manuscript submitted for publication;
2020.
This study clearly demonstrated that an intensive, empirically valid
8. Kerwin MLE. Empirically supported treatments in pediatric psychol-
treatment for paediatric feeding disorders could be provided by ogy: severe feeding problems. J Pediatr Psychol. 1999;24(3):193-214.
properly trained clinicians and continued by trained parents at home 9. Volkert VM, Piazza CC. Pediatric feeding disorders. In: Sturmey
or during meals out. The success of our programme is highly encour- P, Hersen M, eds. Handbook of Evidence-Based Practice in Clinical
Psychology. Hoboken, NJ: John Wiley & Sons; 2012.
aging, as it suggests that these empirically valid behaviour-analytic
10. Sharp WG, Jaquess DL, Morton JF, Herzinger CV. Pediatric feed-
treatments can be used outside the few specialist settings where ing disorders: a quantitative synthesis of treatment outcomes. Clin
they were developed. We demonstrated that specially trained par- Child Fam Psychol Rev. 2010;13(4):348-365.
ents were able to continue the programme and continue to achieve 11. Piazza CC, Milnes SM, Shalev RA. A behavior-analytic approach to
the assessment and treatment of pediatric feeding disorders. In:
treatment outcomes several years after they took over responsibil-
Roane HS, Ringdahl JE, Falcomata TS, eds. Clinical and Organizational
ity for the programme. This could improve access to treatment and
TAYLOR et al. |
157
Applications of Applied Behavior Analysis. Cambridge, MA: Academic 22. Hagopian LP. The consecutive controlled case series: design, da-
Press; 2015. ta-analytics, and reporting methods supporting the study of gener-
12. Sharp WG, Volkert VM, Scahill L, McCracken CE, McElhanon B. A ality. J Appl Behav Anal. 2020;53(2):596-619.
systematic review and meta-analysis of intensive multidisciplinary 23. Cooper JO, Heron TE, Heward WL. Applied behavior analysis.
intervention for pediatric feeding disorders: how standard is the Hoboken, NJ: Pearson; 2020.
standard of care? J Pediatr. 2017;181:116-124. 24. Wilkinson L. Graphic displays of data. In: Cooper H, ed. APA
13. Peterson KM, Piazza CC, Volkert VM. A comparison of a modified Handbook of Research Methods in Psychology. Washington, DC:
sequential oral sensory approach to an applied behavior-analytic American Logical Association; 2012.
approach in the treatment of food selectivity in children with au- 25. Lakens D. Calculating and reporting effect sizes to facilitate cu-
tism spectrum disorder. J Appl Behav Anal. 2016;49(3):485-511. mulative science: a practical primer for t-tests and ANOVAs. Front
14. Addison LR, Piazza CC, Patel MR, et al. A comparison of sensory in- Psychol. 2013;4:863.
tegrative and behavioral therapies as treatment for pediatric feed- 26. Blampied NM. Analyzing therapeutic change using modified
ing disorders. J Appl Behav Anal. 2012;45(3):455-471. Brinley plots: History, construction, and interpretation. Behav Ther.
15. Lukens CT, Silverman AH. Systematic review of psychological 2017;48(1):115-127.
interventions for pediatric feeding problems. J Pediatr Psychol. 27. Kunkel KR, Kozlowski AM, Taylor T, González ML. Validating a food
2014;39(8):903-917. avoidance assessment for children with food selectivity. Behav Dev.
16. Taylor SA, Virues-Ortega J, Anderson R. Transitioning children from 2018;23(2):89-105.
tube to oral feeding: a systematic review of current treatment ap- 28. Cumming G. Understanding the new statistics: Effect sizes, confidence
proaches. Speech Language Hearing. 2019;22:1-14. intervals, and meta-analysis. New York, NY: Routledge; 2012.
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. S U P P O R T I N G I N FO R M AT I O N
2017;10(2):107-136.
Additional supporting information may be found online in the
18. Lerman DC. From the laboratory to community application:
translational research in behavior analysis. J Appl Behav Anal. Supporting Information section.
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. How to cite this article: Taylor T, Blampied N, Roglić N.
Eur J Behav Anal. 2018;19(2):208-227. Controlled case series demonstrates how parents can be
20. Taylor SA, Purdy SC, Jackson B, Phillips K, Virues-Ortega J.
trained to treat paediatric feeding disorders at home. Acta
Evaluation of a home-based behavioral treatment model for chil-
dren with tube dependency. J Pediatr Psychol. 2019;44(6):656-668. Paediatr. 2021;110:149–157. https://doi.org/10.1111/
21. Bullock CE, Fisher WW, Hagopian LP. Description and validation of apa.15372
a computerized behavioral data program: “BDataPro”. Behav Anal.
2017;40(1):275-285.
| |
Received: 9 March 2020 Revised: 14 May 2020 Accepted: 20 May 2020
DOI: 10.1111/apa.15375
REGULAR ARTICLE
1
Monash University, Melbourne, Vic.,
Australia Abstract
2
The Ritchie Centre, Hudson Institute of Aim: To compare neonatal outcomes of small for gestational age (SGA) infants born
Medical Research, Melbourne, Vic., Australia
to South Asian (SA)-born women and Australian/New Zealand (ANZ)-born women.
3
Monash Newborn, Monash Children’s
Hospital, Melbourne, Vic., Australia
Methods: Retrospective cohort study at a hospital network in Australia. Maternal
4
Department of Paediatrics, Monash and neonatal data were collected for infants born SGA between 2013 and 2017 to
University, Melbourne, Vic., Australia SA- or ANZ-born women. Rates of perinatal mortality and neonatal morbidities were
Correspondence analysed between groups.
Atul Malhotra, Department of Paediatrics, Results: A total of 1018 SA and 959 ANZ SGA infants were included. SA SGA babies
Monash University, 246 Clayton Road,
Clayton, Vic., Australia. were older (median [IQR] 39 [38-40] weeks) and heavier (2590 [2310-2780] grams)
Email: atul.malhotra@monash.edu compared to ANZ SGA babies (38 [37-40] weeks and 2480 [2059-2740] grams;
Funding information P < .001 for both). After adjustment for differences in demographics, SA SGA babies
MDT received support from the National were 1.5 times more likely to develop hypothermia (CI: 1.16-1.88, P = .001), but 60%
Health and Medical Research Council of
Australia Fellowship program. AM received less likely to be born with a major congenital malformation (CI: 0.24-0.67, P = .001)
funding from the Royal Australasian College and 36% less likely to need gavage feeding (CI: 0.43-0.93, P = .02) compared to ANZ
of Physicians. The funders had no role in
study design, data collection and analysis, SGA babies.
decision to publish or preparation of the Conclusion: Small for gestational age babies of SA-born women have different neo-
manuscript.
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 rates of mortality and morbidity compared to appropriate for ges-
tational age babies.3 One group of women who are more likely to
Small for gestational age (SGA) defines infants born with a birth- give birth to SGA babies are those born in the South Asia region
1
weight below the 10th percentile for their gestational age. SGA (eg India, Sri Lanka, Bangladesh).4-9 This is important in many high-
births can occur due to pathological processes such as intrauterine income countries like Australia, where one in three births is to a
or foetal growth restriction (IUGR/FGR), or infants can be consti- woman born overseas, and in 2017, 8.8% of births were to a South
tutionally small. 2 In both situations, SGA babies experience higher 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. |
159
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-
F I G U R E 1 Distribution of small for gestational age babies born
stance use reported by ANZ-born women, and higher rates of thy-
within each gestational window, separated according to Australian/
roid disorders for SA-born women. These findings are in line with
New Zealand and South Asian maternal region of birth
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
cohort. SA babies born <32 weeks required less ventilation sup- likely to have a congenital abnormality, malformation or syndrome
port, and a lower rate of intubation (14% vs 69%) and surfactant may indeed be associated with the high rates of maternal smoking,
therapy (14% vs 62.1%). Among SGA babies born between 32 alcohol and substance use during pregnancies of ANZ-born women.
and 36 weeks, there were no significant differences between the However, this difference remained statistically significant even
groups; however, length of stay admitted to SCN/NICU was in fact, after adjusting for maternal factors, suggesting that maternal ROB
slightly longer for SA babies (15 [9-25] vs 14 days [8-20]) in this may also have an impact. After adjustment for potential confound-
group. Conversely, in the ≥37-week subgroup, length of stay admit- ers, the rate of perinatal mortality in SA SGA babies was also higher
ted to SCN/NICU was increased for ANZ compared to SA babies than ANZ SGA babies, consistent with recent studies.4,7,9,12
(5 [4-8.5] vs 5 days [3-7]). ANZ babies in this subgroup also had South Asian ethnicity has been linked to shortened gestational
notably higher rates of major congenital anomaly (5% vs 2.3%) and length, accelerated foetal maturation and subsequent more favour-
culture-positive sepsis (0.8% vs 0%). Neonatal hypothermia was able outcomes for neonates of this background. 21 One study noted
shown in 257 (29%) of SA babies but only 149 (21%) of ANZ babies a median gestational length of 39 weeks for South Asians compared
in the ≥37-week subgroup. to 40 weeks in Europeans and a shorter length of hospital stay for
The association between maternal ROB and key neonatal out- Asian infants.9,13 These ethnic differences may be a result of earlier
comes is presented in Table 3. Overall, SA SGA babies were sig- placental maturation at a biological or genetic level.5 A recent study
nificantly less likely to experience mortality, require resuscitation from our group explored this theory, comparing outcomes of moder-
or admission to the SCN/NICU, respiratory support for >4 hours, ate-late preterm infants of SA and ANZ-born women. Lower rates of
surfactant therapy or gavage feeding, experience HMD or CLD, respiratory morbidity were noted in SA babies, suggesting earlier lung
hyperbilirubinaemia, sepsis or anaemia, or have a major congenital maturation in utero.18 Our study did not discover statistically signif-
anomaly. SA babies also had a significantly shorter duration of stay, icant differences in neonatal respiratory outcomes between SA and
but were more likely to experience hypothermia. After adjustment ANZ infants, which may indicate that the SGA-specific population has
for the possible confounding maternal and birth variables, SGA ba- a different neonatal profile or there was lack of power for the out-
bies of SA-born women were 1.5 times more likely to develop hy- comes. Interestingly, SA infants were 1.5 times less likely to require
pothermia (95% CI: 1.16-1.88; P = .001) and 2.5 times more likely NGT feeding, which could be associated with early maturation.
to die (95% CI: 1.12-5.7, P = .02). These infants were also 60% less Despite the association between SA maternal ROB and both pre-
likely to be born with a major congenital anomaly (95% CI: 0.24-0.67, maturity and shortened gestational length, only 12.9% of SA SGA in-
P = .001) and 36% less likely to require NGT feeding (95% CI: 0.43- fants were born preterm (<32 or 32-36 weeks), compared to 25% of
0.93, P = .02). Overall length of stay was also 2.33 days shorter for ANZ SGA infants in our cohort. This is illustrated in Figure 1, where
SA babies (95% CI: −4.1 to −0.57), P = .01). 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-
4 | D I S CU S S I O N sulting in a higher number of preterm births. Conversely, SA-born
women may have more placental insufficiency near term or after
This is a large retrospective cohort study examining the neona- term age, leading to more numbers in that gestation subgroup. These
tal outcomes of small babies, born to two diverse maternal popu- findings may indeed align with our hypothesis, and reflect a largely
lations in an Australian setting. To our knowledge, this is the first physiologically or constitutionally small SA population, compared to
study to conduct such a detailed analysis of a wide range of neonatal a pathologically small ANZ population.
TA B L E 2 Neonatal outcomes excluding stillbirths and neonatal death with no resuscitation offered
|
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)
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 As our study included a large population, the contribution of so-
likely to develop neonatal hypothermia within the first 24 hours of cio-economic status between groups is relatively controlled be-
life. Low birthweight infants have a limited capacity for thermoregu- tween the groups. Our study had low numbers in the very preterm
lation during the initial neonatal period, where maintaining warmth is SGA group, particularly for SA-born women. This limits the abil-
22
critical to reduce morbidity and mortality. On exposure to the cold ity to analyse and generalise the observed associations for this
extra-uterine environment, lipolysis of brown adipose tissue is acti- subgroup. Finally, the hospital database uses national population
vated to generate heat in the neonate. 22,23 We postulate that the in- centile charts to classify and diagnose SGA. When using single
creased risk of hypothermia in SA babies may be explained by ethnic standard growth charts, SA ethnicity has higher rates of SGA, with
differences in body composition and fat stores. Interestingly, multiple lower perinatal mortality risks, 28 indicating that many SA babies
studies in children and adolescents have shown ethnic variation in are categorised as high risk when they are actually constitution-
body composition, 24 but that SA adolescents and adults have signifi- ally small. This supports our hypothesis and explanation for more
cantly more body fat than white Europeans, despite having a lower favourable neonatal outcomes in the SA population. However, our
25,26
BMI. This highlights the importance of further research into vari- study also recognised that for most neonatal outcomes, SA and
ation in body composition specifically in the neonatal population. ANZ babies behaved similar, and therefore SA SGA babies cannot
There are a number of limitations to this study. Firstly, using be presumed as constitutional or physiologically small. The as-
maternal self-reported country of birth as a proxy for ethnicity is sumption of SGA being physiological due to maternal characteris-
a limitation of perinatal data sets internationally, where ethnicity tics (including ROB or ethnicity) could lead to the misidentification
is not explicitly recorded, and does not take into account pater- of pathological, growth-restricted infants.
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 5 | CO N C LU S I O N S
studies, adequately powered to consider these differences will be
of interest. For SA-born women, it is also important to consider the Babies born SGA to SA-born women have a different neonatal out-
duration of residence in Australia, as migrant status itself is a risk come profile as compared to babies born to ANZ-born women. ANZ
27
factor for adverse perinatal outcomes, which was not recorded SGA infants experience more congenital abnormalities, gavage tube
in our data. In addition to this, environmental factors including so- requirements and a longer duration of hospital stay, whereas SA SGA
cio-economic status and education level of families in both groups infants have a higher risk of neonatal hypothermia and perinatal
could also influence foetal maturation and neonatal morbidity. mortality. Apart from this, no significant differences in the rates of
YOUNG et al. |
165
other neonatal outcomes were observed. These findings may be a centiles and identification of at-risk small-for-gestational-age in-
fants: a retrospective cohort study. BJOG. 2012;119(7):848-856.
reflection of physiological and pathological causes of SGA, relating
12. Davies-Tuck ML, Davey MA, Wallace EM. Maternal region of
to maternal ROB, and other maternal demographics. These discover- birth and stillbirth in Victoria, Australia 2000–2011: a retro-
ies also provide scope for further research into the influence of eth- spective cohort study of Victorian perinatal data. PLoS One.
nicity on foetal and placental maturation, organogenesis and body 2017;12(6):e0178727.
composition. 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.
AC K N OW L E D G E M E N T S 2004;33(1):107-113.
The authors would like to thank Amanda Kendell and Michelle Knight 14. Schaaf JM, Mol BW, Abu-Hanna A, Ravelli AC. Ethnic disparities in
for help with data extraction from the hospital database. 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-
C O N FL I C T O F I N T E R E S T tional birthweight percentiles by sex and gestational age, 1998–
MDT has a secondment 1 day per week to Consultative Council on 2007. Med J Aust. 2012;197(5):291-294.
Obstetric and Paediatric Mortality and Morbidity (CCOPMM). There 16. United Nations Statistics Division. Standard country or area codes
for statistical use (M49); 2019.
are no other conflicts to declare.
17. Nambiar V, Jagtap VS, Sarathi V, et al. Prevalence and impact of
thyroid disorders on maternal outcome in Asian-Indian pregnant
E T H I C A L A P P R OVA L women. J Thyroid Res. 2011;2011:429097.
The study obtained ethics approval as a quality assurance project 18. Cox AG, Narula S, Malhotra A, Fernando S, Wallace E, Davies-Tuck
M. The influence of maternal ethnicity on neonatal respiratory out-
from the Monash Health Human Research Ethics committee. The
come. Arch Dis Child Fetal Neonatal Ed. 2020;105:50-55.
collection of data was in accordance with ethical standards, and no 19. Sahu MT, Das V, Mittal S, Agarwal A, Sahu M. Overt and sub-
formal consent from participants was required. clinical thyroid dysfunction among Indian pregnant women and
its effect on maternal and fetal outcome. Arch Gynecol Obstet.
2009;281(2):215.
ORCID
20. Ko T-J, Tsai L-Y, Chu L-C, et al. Parental smoking during pregnancy
Atul Malhotra https://orcid.org/0000-0001-9664-4182 and its association with low birth weight, small for gestational age,
and preterm birth offspring: a birth cohort study. Pediatr Neonatol.
T WITTER 2014;55(1):20-27.
Atul Malhotra @Atul_Monash 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
REFERENCES -F138.
1. Battaglia FC, Lubchenco LO. A practical classification of newborn 22. Lunze K, Hamer DH. Thermal protection of the newborn in re-
infants by weight and gestational age. J Pediatr. 1967;71(2):159-163. source-limited environments. J Perinatol. 2012;32:317.
2. Ananth CV, Vintzileos AM. Distinguishing pathological from consti- 23. Dawkins MJR, Scopes JW. Non-shivering thermogenesis and
tutional small for gestational age births in population-based studies. brown adipose tissue in the human new-born infant. Nature.
Early Human Dev. 2009;85(10):653-658. 1965;206(4980):201-202.
3. Ewing AC, Ellington SR, Shapiro-Mendoza CK, Barfield WD, 24. Haroun D, Taylor SJC, Viner RM, et al. Validation of bioelectrical
Kourtis AP. Full-term small-for-gestational-age newborns in the impedance analysis in adolescents across different ethnic groups.
U.S.: characteristics, trends, and morbidity. Matern Child Health J. Obesity. 2010;18(6):1252-1259.
2017;21(4):786-796. 25. Ehtisham S, Crabtree N, Clark P, Shaw N, Barrett T. Ethnic differ-
4. Balchin I, Whittaker JC, Patel RR, Lamont RF, Steer PJ. Racial varia- ences in insulin resistance and body composition in United Kingdom
tion in the association between gestational age and perinatal mor- adolescents. J Clin Endocrinol Metab. 2005;90(7):3963-3969.
tality: prospective study. BMJ. 2007;334(7598):833. 26. Nightingale CM, Rudnicka AR, Owen CG, Cook DG, Whincup PH.
5. Ravelli AC, Tromp M, Eskes M, et al. Ethnic differences in still- Patterns of body size and adiposity among UK children of South
birth and early neonatal mortality in The Netherlands. J Epidemiol Asian, black African-Caribbean and white European origin: Child
Commun Health. 2011;65(8):696-701. Heart And health Study in England (CHASE Study). Int J Epidemiol.
6. Ravelli Anita CJ, Schaaf Jelle M, Eskes M, Abu-Hanna A, de Miranda 2010;40(1):33-44.
E, Mol BWJ. Ethnic disparities in perinatal mortality at 40 and 41 27. Cacciani L, Asole S, Polo A, et al. Perinatal outcomes among immi-
weeks of gestation. J Perinatal Med. 2013;41:381. grant mothers over two periods in a region of central Italy. BMC
7. Drysdale H, Ranasinha S, Kendall A, Knight M, Wallace EM. Public Health. 2011;11(1):294.
Ethnicity and the risk of late-pregnancy stillbirth. Med J Australia. 28. Kierans WJ, Joseph K, Luo Z-C, Platt R, Wilkins R, Kramer MS.
2012;197(5):278-281. Does one size fit all? The case for ethnic-specific standards of fetal
8. Khalil A, Rezende J, Akolekar R, Syngelaki A, Nicolaides KH. growth. BMC Pregnancy and Childbirth. 2008;8(1):1.
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 How to cite this article: Young NE, Davies-Tuck M, Malhotra
Dev. 2007;83(12):749-754.
A. Influence of maternal region of birth on neonatal
10. Australian Institute of Health Welfare. Australia's Mothers and
Babies Data Visualisations. Canberra: AIHW; 2019. outcomes of babies born small. Acta Paediatr. 2021;110:158–
11. Anderson NH, Sadler LC, Stewart AW, McCowan LM. Maternal and 165. https://doi.org/10.1111/apa.15375
pathological pregnancy characteristics in customised birthweight
DOI: 10.1111/apa.15427
BRIEF REPORT
Predicting an infant’s risk of bronchopulmonary dysplasia (BPD) is infants' ethnicity was unknown or not available on the BPD esti-
challenging, particularly during the first weeks after birth. However, mator (options ‘White’, ‘Black’ or ‘Hispanic’), they were assigned
strategies to prevent BPD may have most benefit if delivered during ‘White’.
this period. A tool assessing BPD risk would prove valuable to clini- During the study period, 215 eligible infants were admitted.
cians and researchers—aiding clinical decisions and developing eligi- Twenty-six infants with birthweight <501 g or >1250 g and two in-
bility criteria to evaluate preventative BPD therapies. fants with unknown BPD outcomes were excluded; 187 infants were
The NICHD ‘BPD Outcome Estimator’ (BPD estimator) uses clin- included. Their mean (SD) GA was 26.6 (1.5) weeks’ and mean (SD)
ical and demographic data to estimate, on postnatal days 1, 3, 7, 14, birthweight 872 g (178). Fifty-two per cent of infants were male,
21 and 28, an infant’s risk of mild, moderate and severe BPD, and the 91% exposed to antenatal corticosteroids, 73% received surfactant,
competing outcome of death before 36 weeks’ postmenstrual age and 31% received postnatal corticosteroids for BPD treatment.
(PMA).1 NICHD 2001 consensus definition2 and physiological chal- The incidence of death before 36 weeks’ PMA or sBPD was
3
lenge defined BPD. The validity of the BPD estimator outside the 48.1%. Eighteen infants (9.6%) died before 36 weeks’ PMA, and 72
United States has not been evaluated. infants (38.5%) were diagnosed with sBPD.
This retrospective cohort study assessed the BPD estima- In this population, the BPD estimator accurately distinguishes
tor’s accuracy in predicting risk of severe BPD (sBPD) or death in infants who die before 36 weeks’ PMA or have sBPD from those
infants admitted to two perinatal centres in Victoria, Australia. who do not, as evidenced by AUC of 0.81-0.84 at all postnatal time
Infants admitted to Royal Women’s Hospital or Monash Children’s points (Figure S1). The accuracy reported here is comparable to that
Hospital born ≥23 weeks’ and ≤28 weeks’ gestational age (GA), be- reported in the population used to model the estimator.1
tween October 2016 and September 2017, birthweight >500 g and However, predicted risk underestimates the observed incidence
<1251 g were included. of sBPD or death. The observed incidence of sBPD or death was
Infant characteristics (GA, birthweight, sex, race, postnatal day, higher than the estimated risk in all risk strata except estimated risk
respiratory support mode and fraction of inspired oxygen (FiO2)) group <10% (Table 1). This underestimation may reflect practice
were entered into the BPD estimator to generate risk estimates for changes and population differences.
death and sBPD on postnatal days 1, 3, 7, 14, 21 and 28. The bi- The BPD estimator, modelled on infants born between 2000 and
nary outcome of sBPD or death before 36 weeks’ PMA, chosen for 2004, uses mode of respiratory support to predict outcome. The use
its importance in both clinical practice and research, was generated of non-invasive respiratory support has evolved. Modes available in
based on infants’ outcomes. Receiver operator characteristics (ROC) the estimator no longer reflect practice and their influence on pre-
curves evaluated the BPD estimator’s accuracy at distinguishing in- dicting outcome may have changed. Variation in the use of invasive
fants with this outcome from those without. To compare estimated support modes, such as high frequency oscillatory ventilation, may
risk with observed incidence of sBPD or death, infants were strat- also erroneously influence risk estimates.
ified into estimated risk groups; <10%, 10%-19%, 20%-29%, 30%- The use of postnatal corticosteroids has changed. Use declined
39%, 40%-49% and ≥50%. in very low birthweight infants from 20% in 1997-2000 to 8% in
On the day of interest, FiO2 was defined as the mean FiO2 2004.4 The BPD estimator was modelled on infants born during this
of four time points and respiratory support as the ‘highest’ level period. Thirty-one per cent of infants in our cohort received postna-
received on that day. Nasal high flow therapy (nHF) and non-in- tal corticosteroids. The BPD estimator does not use postnatal corti-
vasive positive pressure ventilation were classified as continuous costeroid exposure to estimate risk, though differences in use may
positive airway pressure. NICHD BPD definitions2 were applied modify BPD evolution.
at 36 weeks’ PMA. Infants receiving nHF at 36 weeks’ PMA were Multiple factors may contribute to the high incidence of sBPD or
classified as having sBPD only if they required an FiO2 >0.21. If death (48.1%) observed in our population. Intensive care is provided
to infants born at 23 weeks’ GA in centres included in this study.
A subset of this data was presented in poster format at the 2019 Annual Congress of the Of the 10 infants born at 23 weeks’ GA included in this cohort, five
Perinatal Society of Australia and New Zealand.
© 2020 Foundation Acta Pædiatrica. Published by John Wiley & Sons Ltd
|
166
wileyonlinelibrary.com/journal/apa 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
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)
1
died prior to 36 weeks’ PMA and the five survivors had sBPD. The Newborn Research, The Royal Women’s Hospital, Melbourne,
lack of a physiological test of BPD severity and nHF use at 36 weeks’ Vic., Australia
2
PMA may have increased the diagnosis of sBPD. Sixty-two infants Department of Obstetrics and Gynaecology, The University of
received nHF at 36 weeks’ PMA; how these infants’ BPD is best cat- Melbourne, Melbourne, Vic., Australia
3
egorised is currently unknown. Finally, clinical practices in our nurs- Monash Newborn, Monash Children’s Hospital, Melbourne,
eries may increase BPD rates. Vic., Australia
The BPD estimator has limited ethnicity options. Those available,
White, Black or Hispanic, do not reflect our community with a sub- Correspondence
stantial Asian population. Elizabeth K. Baker, Newborn Research, The Royal Women’s
In the original population, equal numbers of infants died before Hospital, 20 Flemington Road, Parkville, Vic. 3052, Australia.
36 weeks’ PMA or had sBPD (13% for each outcome). In our popu- Email: elizabeth.baker2@thewomens.org.au
lation, sBPD incidence was four-fold higher than that of death. This
differential may reflect recent increases in preterm infant survival ORCID
rates and an associated BPD increase.5 Elizabeth K. Baker https://orcid.org/0000-0001-8523-951X
Infants ≤28 weeks’ GA were included in our cohort, whereas
the estimator was based on a population of infants <30 weeks’ GA. REFERENCES
Whilst limiting comparison at population level, this difference should 1. Laughon MM, Langer JC, Bose CL, et al. Prediction of bronchopulmo-
not influence individuals’ estimated risk. nary dysplasia by postnatal age in extremely premature infants. Am J
Respir Crit Care Med. 2011;183(12):1715-1722.
Despite limitations, the BPD estimator distinguishes infants with
2. Jobe A, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit
sBPD/death from those without in our population. However, the Care Med. 2001;163:1723-1729.
BPD estimator needs recalibrating for local populations. 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.
C O N FL I C T O F I N T E R E S T
4. Fanaroff AA, Stoll BJ, Wright LL, et al. Trends in neonatal morbidity
The authors have no conflicts of interest to declare. 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.
F U N D I N G I N F O R M AT I O N
2015;314(10):1039-1051.
National Health and Medical Research Council, Australia (Practitioner
Fellowship No. 1157782 to PGD).
S U P P O R T I N G I N FO R M AT I O N
Elizabeth K. Baker1,2,3 Additional supporting information may be found online in the
1,2
Peter G. Davis Supporting Information section.
| |
Received: 13 April 2020 Revised: 20 July 2020 Accepted: 28 September 2020
DOI: 10.1111/apa.15604
BRIEF REPORT
1
Neonatal Intensive Care Unit, IRCCS
Istituto Giannina Gaslini, Genoa, Italy Abstract
2
Epidemiology and Biostatistics Unit, IRCCS Background: Post-haemorrhagic ventricular dilatation (PHVD) still represents an
Istituto Giannina Gaslini, Genoa, Italy
important cause of brain injury in premature infants. Intervention for PHVD is rec-
3
Paediatric Neurosurgery Unit, IRCCS
Istituto Giannina Gaslini, Genoa, Italy
ommended once Ventricular Index (VI) crosses the 97th percentile + 4 mm line ac-
4
Department of Neurosciences, cording to Levene.
Rehabilitation, Ophthalmology, Genetics, Objectives: We aimed to compare PHVD severity, timing of intervention, and out-
Maternal and Child Health (DINOGMI),
University of Genoa, Genoa, Italy come between outborn infants transferred to a level IV NICU in order to be treated
for PHVD and a control population of inborn infants.
Correspondence
Alessandro Parodi, Neonatal Intensive Care Methods: Preterm infants with PHVD requiring treatment were divided into: out-
Unit, IRCCS Istituto Giannina Gaslini, Via born infants (transferred to our NICU in order to be treated for PHVD) and inborn
Gaslini 5, 16148 Genoa, Italy.
Email: alessandroparodimail@gmail.com 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. |
169
KEYWORDS
1 | I NTRO D U C TI O N Demographic, clinical and ultrasound data were compared be-
tween the two groups. Measurements of non-dilated lateral ventri-
Post-haemorrhagic ventricular dilatation (PHVD) remains a major cles were excluded if there was unilateral hydrocephalus.
cause of brain injury in preterm infants. Cerebrospinal fluid Parameters were compared using the Mann-Whitney test for
drainage is often needed to decrease dilatation and prevent fur- continuous variables and the chi-square or Fisher's exact test for
ther damage.1 Like most European neonatal intensive care units categorical variables. The statistical analysis was performed using
(NICUs), we treat infants once the ventricular index (VI) has SPSS for Windows, version 20 (IMB Corp).
crossed the 97th percentile (p97) and 4mm line, 2,3 as recently Of the 27 preterm infants treated for postnatal-onset PHVD, 12
recommended.1,4 were outborn and 15 were inborn. Decisions about EVD treatment
We had noticed that neonates transferred to our level-four NICU were made the same day that outborn infants were admitted. They
had extremely severe PHVD frequently surpassing the treatment were significantly older than inborn infants at the time of the deci-
threshold. Our aim was to assess the degree of ventricular dilatation, sion and actual surgery (Table 1).
the timing of surgery and neurodevelopmental outcome in inborn We included 47/54 lateral ventricle measurements for 21 out-
and outborn patients. The hypothesis was that infants were being born and 26 inborn patients: six measurements were missing due
referred to our centre later than recommended. to porencephaly involving the frontal horn, following periventricular
The clinical database at the Gaslini Institute of Genoa, Italy, was haemorrhagic infarction, and we excluded a non-dilated lateral ven-
retrospectively searched for all preterm infants treated for postna- tricle in one patient with unilateral PHVD. The VI-p97 values were
tal-onset PHVD from January 2012 to December 2018. Our NICU significantly higher in the outborn group (Table 1).
has a longstanding practice to treat PHVD once the VI has crossed A ventriculoperitoneal shunt was inserted in 66.7% of outborn
3
the p97 + 4 mm line according to Levene by external ventricular and 40.0% of inborn patients (P = .18). A higher rate of developmen-
drain (EVD) placement. This temporary drain progressively reduces tal delay in all domains was observed in outborn patients compared
ventricular dilatation by continuously draining 10–15 ml/kg of cere- to inborn patients, although statistically not significant (Table 1). A
brospinal fluid daily, according to ventricular size changes and aim- similar trend was observed after excluding subjects with parenchy-
ing for VI < p97 over the next two weeks. Once the cerebrospinal mal lesions (3 patients with infarction in each group and 1 inborn
fluid protein concentration is below 1 g/L, EVD is removed and ven- patient with cystic periventricular leukomalacia), reaching statisti-
tricular size is closely monitored to identify patients with recurrent cal significance for overall adaptive functioning (66.7% of outborn
progressive dilatation who require a permanent ventriculoperitoneal vs 18.2% of inborn patients with VABS-II composite score < 70,
shunt. P = .04).
Study patients were divided into outborn infants transferred to Infants transferred from other centres with progressive PHVD
the unit for PHVD treatment and inborn infants admitted before were treated later and reached more severe ventricular dilatation
PHVD onset. Outborn infants who had already received PHVD than inborn infants. Their ventricular measurements were sig-
treatment were excluded. nificantly higher before surgery and most widely exceeded the
The clinical data collected included the admission age of outborn p97 + 4 mm cut-off commonly used in Europe. 2 Interestingly, our
patients, the age when EVD placement was decided and provided follow-up data showed a trend towards a higher rate of permanent
and any need for a ventriculoperitoneal shunt. The ultrasound data shunt and developmental delay in outborn than in inborn patients,
included the prospectively measured VI in each lateral ventricle on although the sample was too small to reach statistical significance
the day that we decided to treat inborn and outborn infants. VI was for most outcomes. These results agree with data showing impaired
retrospectively plotted on Levene's graph to calculate the difference neurodevelopmental outcomes for EVD treatment after 25 days of
between the VI and the p97 (VI-p97), according to postmenstrual life compared to earlier treatment.5 Furthermore, the p97 + 4 mm
3
age. treatment threshold is supported by evidence suggesting that early
Neurodevelopmental outcome was assessed using the Vineland PHVD management (ie based on ventricular measurements) results
Adaptive Behavior Scales II (VABS-II), a parental questionnaire in- in improved neurodevelopmental outcome and in reduced shunt rate
vestigating four domains of adaptive behaviour (communication, when compared to late treatment (ie based on clinical signs of in-
daily living skills, socialisation and motor skills) and overall adaptive creased intracranial pressure).1,4
functioning. Developmental delay was defined as a score <70 (−2 Delays in transferring infants to our centre may have been due to
SD or less). 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
BRIEF REPORT
The human brain experiences intensive connectivity and size trans- were over 35. We found that 9% of the infants had experienced
formations after birth.1 Head circumference (HC) shows important dengue, malaria or pneumonia at least once in their first year, and
increments in size, specifically in the first year; then, this growth rate these were significantly associated with lower HC z-score (mean
decelerates over time. 2 Identifying the different elements required −0.18 ± 0.91, P < 0.01). Overall, 2% of the infants were stunted
for proper physical growth in children during critical periods is rel- (height for age ≤ 2 z-score), 8% were overweight (weight for age ≥ 2
evant for primary health care. z-score) and 1% had an HC measure of ≤2 z-score (data not shown).
This study investigated the HC predictors in the first year using After multiple adjustments, there were number of variables that
data from the Maternal and Child Health and Nutrition in Acre were positive predictors of the HC z-score in the first year of life, in
(MINA-Brazil) study carried out in Cruzeiro do Sul, Acre State, addition to the child's age, primiparity, type of delivery and birth-
Western Brazilian Amazon. We enrolled 1246 mother-infant pairs at weight. These were as follows: higher maternal education, mother
3
birth, as previously described, and 774 (52% female) took part in the living with a partner, exclusive breastfeeding for at least 90 days
1-year follow-up study from August 2016 to July 2017, when they of life and conditional weight. The final model had a HC z-score
were 10-15 months of age. The baseline survey at birth provided variation value of 30% (Table 1).
data on the mother's and infant's health conditions. The follow-up We believe this was the first birth cohort study to determine
visit detailed the child's health profile and anthropometric status. whether conditional weight was a predictor of HC and to address
Structured questionnaires, administered by trained researchers, social and nutritional factors. Our results also showed that exclusive
collected data on the participants’ sociodemographic information, breastfeeding in the first 90 days of life was an independent pre-
the child's health conditions and morbidity and feeding practices in dictor of HC in the first year. Although HC may be a significant way
early childhood. Perinatal and birth data were collected from hospi- of estimating brain volume, other studies using different methods
tal records. The participants, and those lost to follow-up at 1 year, may reinforce the need for deeper investigation of the role of diet
had similar profiles with regard to maternal education, primiparous on brain development. An American cross-sectional study examined
mothers, delivery methods, low birthweight and the children's sex.3 magnetic resonance imaging scans and the dietary feeding patterns
Body weight change in the first year of life was estimated using of 133 infants aged 10 months to 4 years. The results showed that
a conditional weight (variable that expressed the weight change, but longer breastfeeding duration significantly improved brain develop-
was not correlated with the birthweight values).4 Normality assump- ment and that receiving breast milk exclusively for at least the first
tions for the HC standard deviation scores (z-scores) were confirmed 90 days had a positive impact on white matter microstructure.5
using the Shapiro-Wilk test. To report the distribution of HC z-score The study limitations included the high dropout rate and the
values, according to the independent variables, we used the Student possible unreliability of self-reported data. In addition, the in-
t test for dichotomous variables. The regression coefficients (β) and formation on other possible predictors of HC was not extensive,
their 95% confidence intervals (95% CIs) were calculated by multiple limiting the generalisability of our findings, due to potential selec-
linear regression, to assess the independent associations with HC tion bias. Another limitation was using just one timepoint for HC,
z-score values at the follow-up visit. All reported P values are two- which precluded estimating long-term predictors and trajectories.
tailed. The statistical analyses were performed using Stata version However, the anthropometric data were collected by trained re-
16.0 (StataCorp). searchers, which improved reliability, and this was the first popula-
At the 1-year follow-up visit, we observed lower HC z-score tion-based cohort study in the Western Brazilian Amazon to assess
values among children from socially vulnerable mothers. Maternal the predictors of HC during a critical period of the children's lives.
age ranged from 14 to 43 years: 22% were under 20, and 11% 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
A full list of members of the MINA-Brazil Study Group is presented in the effective strategies for childcare.
Acknowledgements.
© 2020 Foundation Acta Paediatrica. Published by John Wiley & Sons Ltd
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
2. Sacco R, Gabriele S, Persico AM. Head circumference and brain size capital in countries of low and middle income: findings from five birth
in autism spectrum disorder: a systematic review and meta-analysis. cohort studies. Lancet. 2013;382:525-534.
Psychiatry Res. 2015;234:239-251. 5. Deoni SCL, Dean DC, Piryatinsky I, et al. Breastfeeding and early
3. Cardoso MA, Matijasevich A, Malta MB, et al. Cohort profile: the white matter development: a cross-sectional study. NeuroImage.
Maternal and Child Health and Nutrition in Acre, Brazil, birth cohort 2013;82:77-86.
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
DOI: 10.1111/apa.15518
BRIEF REPORT
Feeding with human milk has been recognised as an essential com- unit questionnaire recorded information on the units’ characteristics
ponent of newborn care and is especially important for infants born and policies, including procedures for managing MOM. Of the 135
very preterm (VPT, below 32 weeks of gestation) who face higher eligible NICUs, 134 (99.3%) replied.
risks of adverse outcomes. WHO recommends that infants who can- Descriptive statistics are presented as numbers and valid
not be fed mother's own milk (MOM) should receive donor human proportions.
milk. Direct feeding at the breast takes time to establish after VPT Ethical approval was obtained in each region from appropriate
birth and procedures are required for expressing, collecting, stor- ethics committees, as required by national legislation, and for the
ing and administering breast milk that respect microbiological safety European database by the French data protection committee.
rules and ensure nutritional and immunological quality. However, as As shown in Table 1, 72% of the 134 neonatal units were level III
recommendations are scarce, these procedures appear to be depen- and 91% had a written protocol (developed at unit or regional level)
dent on organisational structures and policies of the units.1-3 for breastfeeding/human milk use; when MOM was not available,
Breast milk is one of the most common modes of postnatal 34% used human banked milk for all VPT infants (range: 0%-100% in
human cytomegalovirus (HCMV) transmission in preterm infants Sweden) and 13% for some groups only (range: 0%-50%).
with rates for fresh or frozen MOM ranging from 8% to 37% and Management of MOM varied widely across the countries.
about 10% probability of infant infection. 2 The debate on the risk of Overall, 56% of units reported using fresh MOM (not frozen and
neonatal morbidity and long-term neurodevelopmental impairment unpasteurised milk) without restrictions regarding gestational age,
in VPT infants with HCMV infection is ongoing,4 raising additional birthweight or risk of HCMV transmission (range: 0% in Germany to
challenges for establishing best practices for managing MOM.1,2 In 100% in Netherlands and Denmark). Different practices were also
2012, the American Academy of Paediatrics stated that the benefits observed between units in the same country, particularly in France
of fresh MOM from HCMV-seropositive mothers outweighed the and Italy. In most units (70%), MOM was not pasteurised (range:
risk of HCMV infection. In 2018, the working group of the French 23%-100%); and in 29 (22%) units, all VPT infants received pasteur-
Neonatal Society on fresh human milk use in preterm infants ad- ised MOM (range: 0%-73% in France).
vised pasteurisation of MOM from HCMV-seropositive mothers HCMV serology on all mothers who express their milk (or
2
for infants born <28 weeks or <1000 grams. In addition to HCMV HCMV polymerase chain reaction of the milk) was required in 29%
transmission, also bacterial contamination of expressed milk is sus- of units (range: 0% in 6 countries to 93% in Germany). Wide differ-
pected to lead to significant problems in VPT infants. 2 ences were found between NICUs in France and Italy. Among these
Published reports have revealed differences in MOM manage- units (n = 36), 47% used human bank/MOM pasteurised, 26% fro-
1,3
ment among neonatal intensive care units (NICUs) and countries. zen-thawed MOM, 3% untreated fresh MOM and 24% formula in the
We aimed to compare practices for handling MOM for VPT infants case of HCMV-seropositive mothers.
using standardised questions in European NICUs. Systematic bacteriological analyses of the fresh or fro-
Data were collected as part of the area-based EPICE (Effective zen-thawed MOM were not performed in 76% of units (range: 29%-
Perinatal Intensive Care in Europe) study set-up to investigate the 100%), performed in <10% for the first milk feeding only, and every
use of evidence-based practices for the care of VPT infants in 19 week in 7%.
regions in 11 countries.5 These large variations in practice between countries could reflect
In 2012, structured questionnaires were sent to the head of all differences in local regulations or guidelines, as well as lack of strong
NICUs with at least 10 VPT admissions during the study period. The 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
TA B L E 1 Management of mother's own milk for very preterm infants in the neonatal unit, by countryb
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
1
was also observed between units within the same country, revealing EPIUnit–Instituto de Saúde Pública, Universidade do Porto,
that different options can operate locally, such as milk bank availabil- Porto, Portugal
2
ity which may differ between regions within countries. Study limita- Obstetrical, Perinatal and Pediatric Epidemiology Research
tions include data collected in 2012, which may underestimate current Team (EPOPé), CRESS, INSERM, INRA, Université de Paris, Paris,
human milk bank availability, and a smaller number of units in par- France
3
ticipating regions in some countries; however, no more recent com- Department of Women’s and Children’s Health, Karolinska
parable information on practices is available from so many European Institutet, Stockholm, Sweden
4
countries. Department of Neonatal and Infant Medicine, Tallinn Children's
While it is recognised that MOM should be encouraged as the Hospital, Tallinn, Estonia
5
primary feeding method for VPT infants, this variation indicates Department of Pediatrics, University of Tartu, Tartu, Estonia
6
substantial differences in attitudes about what constitutes best Clinical Care and Management Innovation Research Area,
practices for the management of MOM among European neonatol- Bambino Gesù Children’s Hospital, IRCCS, Rome, Italy
7
ogists. To guide practice, further studies are needed to reinforce the Children´s Hospital, University Hospital, Philipps University,
evidence base on ensuing outcomes of HCMV infection as well as Marburg, Germany
8
practices for managing these risks while ensuring that VPT infants Departamento de Ciências da Saúde Pública e Forenses e
can benefit from MOM. Educação Médica, Faculdade de Medicina, Universidade do
Porto, Porto, Portugal
AC K N OW L E D G E M E N T
The authors are grateful to the staff from the participating neonatal Correspondence
units. Carina Rodrigues, Instituto de Saúde Pública da Universidade
do Porto, Rua das Taipas, 135, 4050-600 Porto, Portugal.
C O N FL I C T O F I N T E R E S T Email: carina.rodrigues@ispup.up.pt
None to declare.
ORCID
F U N D I N G I N FO R M AT I O N Carina Rodrigues https://orcid.org/0000-0003-1720-2808
The research received funding from the European Union's Seventh Jennifer Zeitlin https://orcid.org/0000-0002-9568-2969
Framework Programme (grant 259882) and from the Foundation Marina Cuttini https://orcid.org/0000-0002-3284-6874
for Science and Technology - FCT (Portuguese Ministry of Science, Henrique Barros https://orcid.org/0000-0003-4699-6571
Technology and Higher Education), the Operational Programmes
Competitiveness and Internationalization (COMPETE 2020) and REFERENCES
Human Capital (POCH), Portugal 2020, under the EPIUnit (POCI-01- 1. Klotz D, Jansen S, Gebauer C, Fuchs H. Handling of breast milk by
0145-FEDER-006862; UID/DTP/04750/2019) and the PhD grant neonatal units: large differences in current practices and beliefs.
Front Pediatr. 2018;6:235.
SFRH/BD/111794/2015 (CR).
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.
Carina Rodrigues1 Acta Paediatr. 2018;107(7):1145-1155.
3. Bonet M, Forcella E, Blondel B, et al. Approaches to supporting lacta-
Jennifer Zeitlin2
tion and breastfeeding for very preterm infants in the NICU: a qualita-
Emilija Wilson3 tive study in three European regions. BMJ Open. 2015;5(6):e006973.
Liis Toome 4,5 4. Weimer KED, Kelly MS, Permar SR, Clark RH, Greenberg RG.
Marina Cuttini6 Association of adverse hearing, growth, and discharge age outcomes
with postnatal cytomegalovirus infection in infants with very low
Rolf F. Maier7
birth weight. JAMA Pediatr. 2020;174(2):133-140.
Véronique Pierrat 2
5. Zeitlin J, Maier RF, Cuttini M, et al. Cohort profile: Effective Perinatal
Henrique Barros1,8 Intensive Care in Europe (EPICE) very preterm birth cohort. Int J
the EPICE Research Group Epidemiol. 2020;49(2):372-386.