Bentley et al.

BMC Pediatrics (2016) 16:55

DOI 10.1186/s12887-016-0591-0

RESEARCH ARTICLE Open Access

Gestational age, mode of birth and

breastmilk feeding all influence acute early
childhood gastroenteritis: a record-linkage
cohort study
Jason P. Bentley1,4*, Judy M. Simpson2, Jenny R. Bowen1,3, Jonathan M. Morris1, Christine L. Roberts1
and Natasha Nassar1

Abstract
Background: Acute gastroenteritis (AGE) is a leading cause of infectious morbidity in childhood. Clinical studies
have implicated caesarean section, early birth and formula feeding in modifying normal gut microbiota development
and immune system homeostasis in early life. Rates of early birth and cesarean delivery are also increasing worldwide.
This study aimed to investigate the independent and combined associations of the mode and timing of birth and
breastmilk feeding with AGE hospitalisations in early childhood.
Methods: Population-based record-linkage study of 893,360 singleton livebirths of at least 33 weeks gestation without
major congenital conditions born in hospital, New South Wales, Australia, 20012011. Using age at first AGE hospital
admission, Cox-regression was used to estimate the associations for gestational age, vaginal birth or caesarean delivery
by labour onset and formula-only feeding while adjusting for confounders.
Results: There were 41,274 (4.6 %) children admitted to hospital at least once for AGE and the median age at first
admission was 1.4 years. Risk of AGE admission increased with decreasing gestational age (3738 weeks: 15 %
increased risk, 3336 weeks: 25 %), caesarean section (20 %), planned birth (17 %) and formula-only feeding (18 %). The
rate of AGE admission was highest for children who were born preterm by modes of birth other than vaginal birth
following the spontaneous onset of labour and who received formula-only at discharge from birth care (6278 %).
Conclusions: Vaginal birth following spontaneous onset of labour at 39+ weeks gestation with any breastfeeding
minimised the risk of gastroenteritis hospitalisation in early childhood. Given increasing trends in early planned birth
and caesarean section worldwide, these results provide important information about the impact obstetric interventions
may have on the development of the infant gut microbiota and immunity.
Keywords: Acute gastroenteritis, Early term birth, Caesarean section, Child, Healthy start to life, Breastfeeding

Background [1, 2]. Many factors in early childhood are associated

Acute gastroenteritis is characterised by viral or bacterial with an increased susceptibility to gastroenteritis, such
infection causing diarrhea and vomiting and is a leading as poor social conditions, diet and antibiotic use [3, 4].
cause of infectious morbidity in infants and children world- Additionally, gut microbial composition and immuno-
wide even in developed countries including Australia, logical immaturity in the newborn may also play an
where the incidence is highest in the first two years of life important role [58].
Bacterial exposures from the birth canal and perianal
* Correspondence: jben9630@uni.sydney.edu.au region during vaginal birth are important precursors for
1
Clinical and Population Perinatal Health Research, Kolling Institute, University
of Sydney, Sydney, NSW, Australia the colonisation of the gut in the first few days of life.
4
University Department of Obstetrics, Building 52, Royal North Shore To prepare for this, cells of the adaptive immune system
Hospital, St Leonards, NSW 2065, Australia are recruited to the fetal intestinal tissue with a
Full list of author information is available at the end of the article

2016 Bentley et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Bentley et al. BMC Pediatrics (2016) 16:55 Page 2 of 10

transition to adult T-cells occurring in the third trimes- from birth until the age of 6 years, death or the end
ter [5]. Once born, a multitude of pathways activate to of the study period (30 June 2012), whichever oc-
prepare the immune system and intestinal epithelial cells curred first.
to manage the high density of bacteria in the gut, estab- This study used linked birth, hospital and death records
lishing a homeostasis between the immune system and from the NSW Perinatal Data Collection (PDC), NSW
gut microbiota [6]. The later the gestational age at birth, Admitted Patient Data Collection (APDC) and Registry of
the better prepared the newborns immune system is for Births, Deaths and Marriages Death Registrations (fact of
establishing homeostasis. Bacterial colonisation and the death) respectively. The PDC is a population-based statu-
immune response in the gut are further supported by tory collection covering all live births and stillbirths of at
exposure to the nutritional, growth and immunological least 20 weeks gestation or, if gestational age is unknown,
factors contained in breastmilk [7]. Clinical studies at least 400 grams birthweight. It contains information on
have shown gut colonisation is typically imbalanced maternal characteristics, pregnancy, labour and delivery
towards bacterial species such as E. Coli. in infants factors, and infant outcomes. The APDC includes demo-
delivered by caesarean section or fed formula rather graphic and hospitalisation related data for every inpatient
than breastmilk [8]. admitted to any public or private hospital in NSW. Diag-
This suggests potential common biological mechanisms noses for each admission are coded according to the 10th
by which shortened gestation, delivery by caesarean sec- revision of the International Classification of Disease,
tion and a lack of breastmilk exposure may increase sus- Australian Modification (ICD-10-AM) [20]. Probabilistic
ceptibility to gut infections by disturbing or modifying gut record linkage of these data was performed by the NSW
microbiota and immune system homeostasis in early life. Centre for Health Record Linkage using methods de-
Previous population-based studies have investigated the scribed previously and only de-identified information
independent associations of vaginal birth and breastmilk was provided to the researchers [21]. The data sources
feeding with childhood gastroenteritis [911]. Few have used for this study require ethical and data custodian
examined the association with gestational age, especially approval to access, link (by an independent and ap-
those born around term, either preterm or early term proved authority) and release for research. Approval for
(3738 weeks gestation) and there is evidence these in- the record linkage and use of the data for research was
fants and children are at an increased risk of morbidity obtained from the NSW Population and Health Ser-
generally [12]. The combined risk of gastroenteritis associ- vices Research Ethics Committee and the appropriate
ated with these birth characteristics is currently unknown, data custodians.
but such information is important given worldwide
increasing rates of early planned birth and delivery by
caesarean section [1316], which are also associated with Mode of birth, timing of birth and infant feeding at
reduced rates of breastmilk feeding [17, 18]. Record discharge
linkage of large routinely collected population-based The study factors of interest were mode of birth, gesta-
data with standardised clinical information provides a tional age (timing of birth) and infant feeding status at dis-
powerful approach to investigate the combined risk of charge from birth care. Mode of birth was defined using
gastroenteritis for multiple birth characteristics. the combination of labour onset and type of birth (vaginal
The aim of this study was to investigate the inde- birth or caesarean section) and categorised as vaginal birth
pendent and combined associations of the mode and following spontaneous onset of labour, caesarean section
timing of birth and breastmilk feeding with gastro- following spontaneous onset of labour, vaginal birth follow-
enteritis hospitalisations in early childhood. ing labour induction, caesarean section following labour in-
duction, or pre-labour caesarean section. Gestational age is
Methods reported in completed weeks of gestation, as determined
Study population by the best clinical estimate including early ultrasound and
The study population included all singleton live births last menstrual period. This was categorised as preterm
of 33 weeks gestation from 2001 to 2011 in New South (3336 weeks), early-term (3738 weeks) or term (3942
Wales (NSW), Australia. Stillbirths and births to non- weeks) birth. Infant feeding status at discharge from birth
NSW resident mothers were excluded as these births have care is recorded using one or more of the following three
no opportunity for follow-up through record linkage with categories: breastfeeding, expressed breastmilk or in-
hospital admissions in NSW. Infants with major congeni- fant formula. These categories were used to create two in-
tal conditions, born before 33 weeks gestation, or twins dependent groups: any breastmilk feeding (breastfeeding
and higher-order births were excluded as they have or expressed breastmilk feeding without infant formula)
different risk profiles, outcomes and models of care and formula-only feeding (infant formula without breast-
[19]. Each child in the study population was followed feeding or expressed breastmilk feeding).
Bentley et al. BMC Pediatrics (2016) 16:55 Page 3 of 10

Study outcome hospitalisations. The variations investigated were: a sub-

The study outcome was hospital admissions for gastro- population of low risk liveborn singleton pregnancies
enteritis, which we refer to as acute gastroenteritis (infants born at 37 weeks gestation, cephalic present-
(AGE). Primary or additional diagnoses for gastroenter- ing, and a birthweight between the 10th and 90th percen-
itis (ICD-10-AM: A00-A09 or K52) were used to identify tiles for gestational age and sex, to women aged 2034
admissions. Inter-hospital transfers were treated as a years without medical conditions), using children with
continuation of an admission and not a new admission. no hospital admissions only as the controls, using only a
AGE admissions within 7 days of a previous AGE admis- primary diagnosis of gastroenteritis, restricting to AGE
sion were also treated as a single event. hospitalisations within the first year of life or within the
first two years of life (rather than six), and restricting
Statistical analysis to births from July 2007 onwards (universal rotavirus
The proportion and number of children with none, one, vaccination).
or more than one AGE admission in the study period by The level of missing information was minimal for all
maternal and perinatal characteristics were calculated. variables (<0.01 to 0.10 %), with the exception of infant
Cox proportional hazards regression was used to estimate feeding at discharge. Collection of infant feeding status
the adjusted Hazard Ratios and 95 % confidence intervals began in mid-2006 and was only available for 51.8 % of
for the independent and combined associations between births, within which 0.95 % were missing. As the per
the study exposures and first AGE hospitalisation with cent missing across all variables except infant feeding
child age as the timescale and age at discharge from birth affected only 0.23 % of records, these were excluded.
care as entry into observation. For censored individuals, However, as infant feeding was an exposure of interest,
age was recorded as the earliest of death, sixth birthday, it was imputed using a logistic model following recom-
or end of the study period (30 June 2012). mendations in the literature (see Additional file 1). All
The covariates used in the study reflect known risk fac- analyses were performed using Stata 13.0 (StataCorp
tors identified in the literature [4, 9, 12, 22]. Covariates in- LP, TX, USA).
cluded were: parity (primiparae or multiparae), maternal
age (<20, 2024, 2529, 3034, 3539, 40+ years), country Results
of birth (Australia-born or other), socio-economic status Of the 893,360 children included in the study, 28 % were
quintile (Australian Bureau of Statistics Socio-Economic delivered by caesarean section, 41 % were planned births
Index For Areas Index of Relative Socio-economic (pre-labour caesarean or following labour induction), 27 %
Advantage) [23], smoking during pregnancy, hypertensive were born before 39 weeks gestation and 12 % were fed
disorders of pregnancy (gestational hypertension, pre- only formula in the birth admission (Table 1). There were
eclampsia or eclampsia), diabetes mellitus in pregnancy 38,085 (4.3 %) children admitted to hospital for AGE once
(gestational or pre-existing) [24, 25], babys sex, 5-minute and 3,189 (0.4 %) more than once (7.7 % recurrence rate).
Apgar score < 7, birthweight (standardised by gestational The proportion of children admitted for AGE was higher
age and sex) [26], presence of AGE or other infection for those delivered by caesarean section, born before 39
specific to the perinatal period (ICD-10-AM: P35-P39), weeks gestation and fed only formula. Average follow-
admission to a Special Care Nursery (SCN) or Neonatal up per child was 4.37 years (standard deviation: 1.88),
Intensive Care Unit (NICU) and infant birth admission with a total follow-up time of 3,907,163 years. For the
length of stay (standardised by gestational age and type of 41,274 children admitted, the median age at first AGE
birth using the study population). To account for the hospital admission was 1.43 years (Inter-quartile range
inclusion of rotavirus vaccination in the national immun- 0.77 to 2.48 years).
isation program from 1 July 2007, year of birth was cate- Compared to vaginal birth following spontaneous on-
gorised as before July 2007 or July 2007 onwards. All set of labour, all other modes of birth were independ-
covariates were adjusted for in the analysis except for ently associated with a 1223 % increased rate of AGE
admission to SCN or NICU and 5-minute Apgar score < 7 admission (Table 2). In general those modes of birth in-
which were omitted from the model due to high co- cluding delivery by caesarean section were similar with
linearity with birth at 3336 weeks gestation. The assump- largely overlapping confidence intervals (1923 % in-
tion of proportional hazards was assessed using standard creased rate of admission), while vaginal birth following
diagnostics and found to be reasonable. labour induction was intermediary to these modes of
Cox-regression was used to estimate the associations birth and vaginal birth following spontaneous onset of
for the study exposures under variations of the study labour (12 % increased rate of admission). Adjusted
population and design, to assess the robustness of the associations for all variables are presented in the
main findings and for comparability with other studies Additional file 2. The rate of AGE admission increased
of associations between birth factors and childhood with decreasing gestational age. Birth at 3738 weeks
Bentley et al. BMC Pediatrics (2016) 16:55 Page 4 of 10

Table 1 Maternal and perinatal characteristics for children admitted to hospital once and more than once for acute gastroenteritis,
NSW 20012011
Variable Number of AGE hospital admissions Total (N = 893,360)
None (N = 852,086) One (N = 38,085) Two or more (N = 3189)
N Row %a N Row %a N Row %a N Col. %b
Mode of birth
Vaginal birth spontaneous onset of labour 445,535 95.6 18,929 4.1 1469 0.3 465,933 52.2
Vaginal birth labour induction 173,429 95.1 8220 4.5 722 0.4 182,371 20.4
Caesarean section pre-labour 132,892 95.3 6017 4.3 561 0.4 139,470 15.6
Caesarean section spontaneous onset of labour 57,537 94.9 2820 4.7 252 0.4 60,609 6.8
Caesarean section labour induction 42,693 94.9 2099 4.7 185 0.4 44,977 5.0
Gestational age (weeks)
3336 36,508 94.4 1952 5.0 225 0.6 38,685 4.3
3738 188,164 95.0 8985 4.5 834 0.4 197,983 22.2
3942 627,414 95.5 27,148 4.1 2130 0.3 656,692 73.5
Maternal age (years)
<20 31,020 92.8 2154 6.4 247 0.7 33,421 3.7
2024 118,300 94.0 6902 5.5 676 0.5 125,878 14.1
2529 235,736 95.1 11,235 4.5 926 0.4 247,897 27.8
3034 282,186 95.8 11,464 3.9 853 0.3 294,503 33.0
3539 153,271 96.4 5266 3.3 406 0.3 158,943 17.8
40+ 31,573 96.5 1064 3.3 81 0.2 32,718 3.7
Primiparae 353,744 95.0 17,306 4.6 1468 0.4 372,518 41.7
Australian Born 598,374 95.0 29,024 4.6 2566 0.4 629,964 70.5
Socio-economic advantage
1st Quintile (Highest) 173,516 96.1 6605 3.7 457 0.3 180,578 20.2
2nd Quintile 162,516 95.9 6420 3.8 501 0.3 169,437 19.0
3rd Quintile 178,752 95.5 7823 4.2 604 0.3 187,179 21.0
4th Quintile 175,417 94.9 8557 4.6 775 0.4 184,749 20.7
5th Quintile (Lowest) 161,885 94.4 8680 5.1 852 0.5 171,417 19.2
Smoking during pregnancy 126,685 94.2 7101 5.3 686 0.5 134,472 15.1
Diabetes 53,822 95.3 2420 4.3 220 0.4 56,462 6.3
Hypertension 74,126 94.5 3956 5.0 364 0.5 78,446 8.8
Babys sex - Male 436,304 95.2 20,085 4.4 1724 0.4 458,113 51.3
Year of birth
Prior to 1 July 2007 473,607 93.6 29,552 5.8 2662 0.5 505,821 56.6
From 1 July 2007 378,479 97.7 8533 2.2 527 0.1 387,539 43.4
Birthweight z-score
< -2 16,064 94.1 897 5.3 117 0.7 17,078 1.9
[-2,-1) 113,474 94.9 5570 4.7 482 0.4 119,526 13.4
[-1,0) 305,914 95.3 13,855 4.3 1154 0.4 320,923 35.9
[0,1) 281,683 95.5 12,173 4.1 994 0.3 294,850 33.0
[1,2) 108,087 95.7 4467 4.0 374 0.3 112,928 12.6
2 26,864 95.8 1123 4.0 68 0.2 28,055 3.1
Bentley et al. BMC Pediatrics (2016) 16:55 Page 5 of 10

Table 1 Maternal and perinatal characteristics for children admitted to hospital once and more than once for acute gastroenteritis,
NSW 20012011 (Continued)
Admission to SCN/NICU 111,483 94.2 6205 5.2 663 0.6 118,351 13.3
5 Minute Apgar Score < 7 10,947 94.5 584 5.0 59 0.5 11,590 1.3
Infection in birth admissionc 17,475 94.1 991 5.3 109 0.6 18,575 2.1
Birth admission length of stay z-score
< -1 81,018 96.3 2893 3.4 204 0.2 84,115 9.4
[-1,0) 346,181 95.7 14,527 4.0 1171 0.3 361,879 40.5
[0,1) 351,360 95.2 16,317 4.4 1371 0.4 369,048 41.3
1 73,527 93.9 4348 5.6 443 0.6 78,318 8.8
Formula-only feeding*
Yes 52,360 96.4 1840 3.4 127 0.2 54,327 11.9
No 393,156 97.4 9936 2.5 660 0.2 403,752 88.1
*Complete cases only (n = 458,079), SCN Special Care Nursery, NICU Neonatal Intensive Care Unit, Col. Column
a
Per cent of all children in the row. bPer cent of all children in the study population. cIncludes AGE or ICD-10-AM: P35-P39

was associated with a 15 % increase in the rate of AGE Discussion

admission (adjusted hazard ratio [aHR], 1.15; 95 % This is the first population-based study to specifically in-
Confidence Interval [CI], 1.121.17), and for births at vestigate the combined effects of mode and timing of
3336 weeks gestation was 23 % (aHR, 1.23; 95 % CI, birth and breastmilk feeding at discharge from birth care
1.181.29). Infants fed only formula (aHR, 1.18; 95 % CI, on early childhood gastroenteritis hospital admissions.
1.111.24) were also more likely to be admitted for AGE. The results show an increased rate of admission in early
Results for the combined associations are presented in childhood for being born before 39 weeks gestation, by
Fig. 1. Compared with vaginal birth following spontan- modes of birth other than vaginal birth following
eous onset of labour at 39+ weeks gestation with any spontaneous onset of labour and formula-only feeding
breastmilk feeding at discharge, early term births with at discharge from birth care. The combined effects
formula-only feeding had an increased rate of AGE of highlight the benefit of normal birth and early breastmilk
35 % and preterm birth a 45 % increased rate. Children exposure, and are also suggestive of their impact on subse-
born at early term with formula-only feeding had a higher quent gastrointestinal health by possibly modifying or dis-
rate of admission compared to preterm births that turbing gut microbiota and immune system homeostasis.
had some breastmilk feeding (35 % versus 23 %). For the These findings are also pertinent in the context of increas-
other modes of birth with formula-only feeding, early ing rates of caesarean section and early planned birth.
term births had increased admission rates of 5166 %, To our knowledge this is the first study to identify that
and preterm births had the highest rates ranging from children born preterm and early term, compared with
6278 %. Within births 39+ weeks gestation, all modes of children born at full term (39+ weeks gestation), had a
birth with formula-only feeding compared with vaginal 3852 % and 2841 % increased rate of AGE admission
birth following spontaneous onset of labour, had increased respectively, regardless of the mode of birth. Previous
rates of admission by 3145 %. studies have demonstrated an increased rate of overall
The results were mostly robust to changes in study pediatric or respiratory hospitalisations for preterm and
population or design. Restriction to the first two years of early term births [12, 2729]. We also found for preterm
life, low risk pregnancies, the period of rotavirus vaccin- and early term births, those with modes of birth other
ation or a primary diagnosis of AGE provided generally than vaginal birth following spontaneous onset of labour
similar adjusted hazard ratios to those for the selected and formula-only feeding at discharge had even higher
study population (Table 2). The impact of formula-only rates of AGE admission, 5166 % and 6278 % respe-
feeding was stronger (34 % versus 18 %) when restricting ctively. Interestingly, children born at early term had
to the first year of life. Restricting the control group to higher AGE admission rates than those born preterm
children with no hospital admissions provided a stronger who received any breastmilk feeding in the birth admis-
association for gestational age (3336 weeks, 44 % ver- sion. Nevertheless, the most vulnerable group of chil-
sus 23 %; 3738 weeks, 21 % versus 15 %). For the ad- dren identified were those born preterm by modes of
justed final models and combined associations for the birth other than vaginal birth following the spontaneous
investigated changes in study population or design see onset of labour and who received formula-only at dis-
Additional files 2 and 3. charge from birth care.
 Bentley et al. BMC Pediatrics (2016) 16:55
Table 2 Associations for age at first hospital admission for acute gastroenteritis by mode of birth, timing of birth and infant formula only at birth for the overall study and
selected sub-populations, NSW 20012011a
Study Population Low-riskc Healthy controlsd From 1 July 2007e Primaryf 1-yearg 2-yearh
Study population (N) 893,360 435,417 610,538 387,539 893,360 893,360 893,360
Children admitted (%) 4.6 4.6 6.8 2.3 4.1 1.6 3.0
aHR (95 % CI) aHR (95 % CI) aHR (95 % CI) aHR (95 % CI) aHR (95 % CI) aHR (95 % CI) aHR (95 % CI)
Mode of birth
Vaginal birth spontaneous onset of labour 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference]
Vaginal birth labour induction 1.12 (1.091.15) 1.12 (1.081.16) 1.15 (1.121.18) 1.10 (1.041.16) 1.13 (1.101.16) 1.18 (1.131.23) 1.14 (1.101.17)
Caesarean section pre-labour 1.19 (1.161.23) 1.17 (1.111.22) 1.24 (1.201.27) 1.20 (1.131.27) 1.20 (1.161.23) 1.24 (1.181.31) 1.22 (1.171.26)
Caesarean section spontaneous onset of labour 1.20 (1.161.25) 1.15 (1.081.22) 1.24 (1.201.29) 1.22 (1.131.33) 1.21 (1.161.26) 1.25 (1.171.34) 1.24 (1.181.30)
Caesarean section labour induction 1.23 (1.181.29) 1.31 (1.221.41) 1.28 (1.221.34) 1.18 (1.071.29) 1.23 (1.171.29) 1.33 (1.231.43) 1.26 (1.201.33)
Gestational age (weeks)
3336 1.23 (1.181.29) - 1.44 (1.371.50) 1.28 (1.161.40) 1.23 (1.171.29) 1.28 (1.191.38) 1.24 (1.171.31)
3738 1.15 (1.121.17) 1.14 (1.101.19) 1.21 (1.181.24) 1.17 (1.111.23) 1.14 (1.111.17) 1.19 (1.151.24) 1.15 (1.121.19)
3942 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference] 1.00 [Reference]
Formula-only feedingb 1.18 (1.111.24) 1.17 (1.081.26) 1.19 (1.141.25) 1.20 (1.141.28) 1.17 (1.111.24) 1.34 (1.251.44) 1.24 (1.171.31)
CI Confidence Interval, aHR Hazard Ratio. aAll models were adjusted for maternal country of birth, maternal smoking during pregnancy, socio-economic advantage, parity, diabetes, hypertension, babys sex, year of
birth, birthweight, and length of stay and infections in the birth admission (AGE or ICD-10-AM: P35-P39), with the following exceptions; for the low-risk group diabetes and hypertension were not applicable and admis-
sion to a special care nursery or neonatal intensive care was able be included as preterm was not in the sub-group, for births occurring after 1 July 2007, the indicator for pre/post introduction of universal rotavirus
vaccination was not applicable
b
Reference category is absence of risk factor and the reported aHR includes uses imputed values
c
Population restricted to low risk pregnancies: 10th90th percentile birthweight for gestational age and sex, cephalic presenting, term births (37 weeks) to mothers aged 2034 years without medical conditions
d
Population restricted to children with one or more AGE hospital admissions or no hospital admissions
e
Population restricted to children born after the inclusion of rotavirus vaccination in the Australian National Immunisation Program (1 July 2007)
f
Age at first hospital admission with a primary diagnosis of AGE was used to define the event
g
The age at first AGE hospital admission within the first year of life was used to define the event. For censored individuals, age was recorded as the earliest of death, first birthday, or end of the study period (30 June 2012)
h
The age at first AGE hospital admission within the first two years of life was used to define the event. For censored individuals, age was recorded as the earliest of death, second birthday, or end of the study period
(30 June 2012)

Page 6 of 10
Bentley et al. BMC Pediatrics (2016) 16:55 Page 7 of 10

The increased rate of AGE admission with decreasing exposed to breastmilk have less diverse gut microbiota
gestational age may be explained in part by the under- dominated by bad bacteria [8, 31, 32]. The similarity of
preparedness of the newborns immune system, particu- the adjusted associations for caesarean section regardless
larly in the gut epithelium, to respond to the initial of the onset of labour is consistent with the theory of
microbial colonisation at birth. Differences in markers beneficial exposure to bacteria at the time of vaginal
of immune function between infants born before and after birth [9, 33]. The higher rate of admission for infants fed
37 weeks gestation have been reported previously [30]. only formula is consistent with the idea that with min-
This may explain why vaginal births following labour imal or no breastmilk exposure, there is a loss of the
induction had an increased rate of admission compared to associated microbial and immunological benefits in early
those following the spontaneous onset of labour, as a deci- childhood. While our estimate (aHR, 1.18) was lower
sion has been made to deliver. Relative to the developmen- than other studies that have examined breastmilk feed-
tal trajectory of the infant, it may be that the necessary ing and AGE, these generally followed infants for the
time required for the infants innate immune response first 612 months of life, where a significant proportion
to sufficiently mature has been circumvented. of AGE admissions occur [34, 35]. The aHR of 1.34 from
Variation in gut microbial composition in infants by our analysis restricted to AGE admissions in the first
mode of birth (vaginal birth or caesarean section) and year of life is similar to these studies, suggesting a poten-
feeding status (breastmilk or formula) is well supported tially stronger association earlier in life.
by clinical evidence [8, 31, 32]. These studies highlight Previous population-based record-linkage studies of
that infants delivered by caesarean section or not term births and without breastmilk feeding information

Fig. 1 Combined associations for age at first hospital admission for acute gastroenteritis by mode of birth, timing of birth and infant formula only
at birth, NSW 20012011. The reference category is vaginal, birth following spontaneous onset of labour at 39+ weeks gestation with breastmilk
feeding at discharge. Associations adjusted for maternal country of birth, maternal smoking during pregnancy, socio-economic advantage, parity,
diabetes, hypertension, babys sex, year of birth, birthweight, and length of stay and infections in the birth admission (AGE or ICD-10-AM: P35-P39)
Bentley et al. BMC Pediatrics (2016) 16:55 Page 8 of 10

found, as we did, an increased risk of AGE for children findings are due to chance. However, some caution is
delivered by caesarean section [9, 10]. However, with the warranted as some of the associations for the study fac-
known association between caesarean section and diffi- tors are small and by chance a statistically significant
culty initiating breastfeeding [18], the combined associ- association may be found when performing many com-
ation for both factors is of particular interest. We found parisons. Using available administrative data has some
that despite term birth, children delivered by caesarean limitations, as not all potentially relevant characteristics
section and formula-only feeding at discharge from birth such as diet, environment or antibiotic use during preg-
care had a 4045 % increased rate of admission. Even nancy and early childhood could be investigated. Des-
for infants with breastmilk exposure, caesarean section pite the lack of information on long term breastfeeding
was still associated with increased rates of AGE admis- outcomes such as the duration of exclusive breastfeeding,
sion (1923 %). This suggests that for infants delivered previous studies have found in-hospital formula supple-
by caesarean section, breastfeeding initiation and dur- mentation is associated with early cessation of exclusive
ation are important factors for reducing the risk of AGE or any breastmilk feeding post-discharge [4648]. Id-
in early childhood and alternative methods of exposure entified cases of acute gastroenteritis were based on
to beneficial bacteria at the time of birth are required. hospital admissions only which represent the severe
Changes in clinical obstetric practice have seen an in- end of the clinical spectrum and do not include mild
crease in rates of planned birth before 39 weeks gesta- cases that may be treated through out-patient facilities
tion and caesarean section worldwide [1316], and the or primary care services.
adverse impact of caesarean section on breastfeeding is
well-established [18]. As these trends relate to factors Conclusions
hypothesised to have a common biological basis for af- We have shown using a large population-based record-
fecting the risk of acute gastroenteritis, the impact of a linkage study that the rate of acute gastroenteritis hospita-
continuation of these patterns on AGE should not be lisations in early childhood are increased for births by
underestimated. Although, the increasing recognition of caesarean section or induction of labour, before 39 weeks
the potential harms of early elective births and subse- gestation and for infants fed only formula at discharge
quent introduction of clinical guidelines, policies and from birth care. The combined effects of these factors
interventions to reduce labour induction or pre-labour highlight the benefit of normal birth and early breastmilk
caesarean section for non-medical reasons before 3940 exposure, for reducing the risk of gastroenteritis hospital-
weeks gestation may counter these trends [3640]. isation in early childhood. These previously unknown
As planned birth before 39 weeks increases the per- combined effects represent useful information against a
ceptions of women about what constitutes normal birth backdrop of increasing rates of caesarean section and early
is also likely to be altered. Recent studies have demon- planned birth, and their potential impact on gut micro-
strated that almost one in four women believed that a biota and immune system homeostasis.
baby was full-term at 3436 weeks gestation, that one
in two believed full term was 3738 weeks and more Availability of data and materials
than 90 % believed it was safe to deliver before 39 weeks The data used in this study cannot be shared by the
[41, 42]. Another study demonstrated that many women Authors due to the use and release of the data being sub-
had little knowledge of the benefits or risks of a caesarean ject to data custodian and ethics approval and conditions
section, yet almost half indicated that a caesarean section that require the data only be used for approved research,
without medical indication should be given on request by approved persons directly involved in the project and
[43]. Given the multi-faceted changes in practice and following the completion of a confidentiality undertaking
attitudes towards earlier births and interventions, similar prior to the information being released.
studies to ours are vital to assess the impact on long term
outcomes to inform clinicians, and women and their Additional files
families.
The strengths of this study are that it is a large Additional file 1: Imputation.pdf summarises the imputation approach
population-based cohort, using data and variables with for formula-only feeding at discharge from birth care, and compares the
demonstrated accuracy and validity [44, 45]. Using a associations for the study factors and all covariates between the imputation
and complete case analysis. (DOC 92 kb)
large population-based cohort also reduces the impact
Additional file 2: All adjusted associations.pdf summarises the
of genetic diversity and enables complete ascertainment adjusted associations for the study factors and all covariates used in the
of hospital admissions. The mode and timing of birth adjustment for the main and additional study populations. (DOC 84 kb)
and infant feeding at discharge were all statistically sig- Additional file 3: Additional combined adjusted associations.pdf
nificant and consistently so for all analyses (coefficient summarises the combined associations for the study factors for all
additionally investigated study populations. (DOC 443 kb)
p-values were typically < 0.001), so it is unlikely our
Bentley et al. BMC Pediatrics (2016) 16:55 Page 9 of 10

Abbreviations 12. Lain SJ, Nassar N, Bowen JR, Roberts CL. Risk factors and costs of hospital
AGE: acute gastroenteritis; aHR: adjusted Hazard Ratio; APDC: admitted patient admissions in first year of life: a population-based study. J Pediatr. 2013;
data collection; CI: confidence interval; ICD-10-AM: International Classification of 163(4):10149.
Disease, 10th revision, Australian Modification; NICU: Neonatal Intensive Care Unit; 13. Morris JM, Algert CS, Falster MO, Ford JB, Kinnear A, Nicholl MC, Roberts CL.
NSW: New South Wales; PDC: perinatal data collection; SCN: special care nursery. Trends in planned early birth: a population-based study. Am J Obstet Gynecol.
2012;207(3):186.e181e188.
Competing interests 14. Menacker F, Hamilton BE. Recent trends in cesarean delivery in the United
The author(s) declare that they have no competing interests. States. NCHS Data Brief. 2010;35:18.
15. Betrn AP, Merialdi M, Lauer JA, Bing-Shun W, Thomas J, Van Look P,
Authors contributions Wagner M. Rates of caesarean section: analysis of global, regional and
JPB and NN conceived the project and developed the idea in collaboration with national estimates. Paediatr Perinat Epidemiol. 2007;21(2):98113.
JMS, JRB, JMM and CLR. All authors (JPB, JMS, JRB, JMM, CLR, NN) contributed to 16. Nassar N, Schiff M, Roberts CL. Trends in the distribution of gestational age
the study design, CLR and NN were responsible for data acquisition and JPB and contribution of planned births in New South Wales, Australia. PLoS
analysed the data with the support of JMS. All authors (JPB, JMS, JRB, JMM, CLR, ONE. 2013;8(2), e56238.
NN) were involved in the interpretation of results. JPB and NN initially drafted 17. Lutsiv O, Giglia L, Pullenayegum E, Foster G, Vera C, Chapman B, Fusch C,
the manuscript and all authors (JPB, JMS, JRB, JMM, CLR, NN) were involved in McDonald SD. A population-based cohort study of breastfeeding according
critical revision of the intellectual content. All authors (JPB, JMS, JRB, JMM, CLR, to gestational age at term delivery. J Pediatr. 2013;163(5):12838.
NN) approved the final version of the manuscript. 18. Prior E, Santhakumaran S, Gale C, Philipps LH, Modi N, Hyde MJ.
Breastfeeding after cesarean delivery: a systematic review and meta-analysis
Acknowledgements of world literature. Am J Clin Nutr. 2012;95(5):111335.
We would like to acknowledge the NSW Ministry of Health for providing access 19. NSW Department of Health. In: Statewide Services Development Branch,
to population health data and the NSW Centre for Health Record Linkage for editor. Guide to the Role Delineation of Health Services (3rd edition).
linking the data sets. JPB was supported by an Australian Postgraduate Award Sydney: NSW Health Department; 2002.
Scholarship, Sydney University Merit Award and a Northern Clinical School 20. The International Statistical Classification of Diseases and Related Health
Scholarship Award, CLR was supported by an Australian National Health and Problems, Australian Modification Tabular List of Diseases and Alphabetic
Medical Research Council Senior Research Fellowship (#APP1021025) and NN Index of Diseases. http://nccc.uow.edu.au/CasemixTools/icd10am/index.html
an Australian National Health and Medical Research Career Development Accessed 6 Nov 2014.
Fellowship (#APP1067066). 21. Bentley JP, Ford JB, Taylor LK, Irvine KA, Roberts CL. Investigating linkage
rates among probabilistically linked birth and hospitalization records. BMC
Author details Med Res Methodol. 2012;12:149.
1
Clinical and Population Perinatal Health Research, Kolling Institute, University 22. Newman RD, Grupp-Phelan J, Shay DK, Davis RL. Perinatal risk factors for
of Sydney, Sydney, NSW, Australia. 2Sydney School of Public Health, infant hospitalization with viral gastroenteritis. Pediatrics. 1999;103(1), e3.
University of Sydney, Sydney, NSW, Australia. 3Department of Neonatology, 23. Australian Bureau of Statistics. Socio-economic Indexes for Areas (SEIFA), Data
Royal North Shore Hospital, Sydney, NSW, Australia. 4University Department only, 2006. Catalogue 2033.0.55.001 http://www.abs.gov.au/AUSSTATS/abs@.
of Obstetrics, Building 52, Royal North Shore Hospital, St Leonards, NSW nsf/DetailsPage/2033.0.55.0012006?OpenDocument Accessed 15 Jul 2014.
2065, Australia. 24. Chen JS, Roberts CL, Simpson JM, Ford JB. Prevalence of pre-eclampsia,
pregnancy hypertension and gestational diabetes in population-based data:
Received: 6 February 2015 Accepted: 20 April 2016 impact of different ascertainment methods on outcomes. Aust N Z J Obstet
Gynaecol. 2012;52(1):915.
25. Tran DT, Roberts CL, Havard A, Jorm LR. Linking birth records to hospital
References admission records enhances the identification of women who smoke
1. Parashar UD, Gibson CJ, Bresee JS, Glass RI. Rotavirus and severe childhood during pregnancy. Aust N Z J Public Health. 2014;38(3):25864.
diarrhea. Emerg Infect Dis. 2006;12(2):3046. 26. Dobbins TA, Sullivan EA, Roberts CL, Simpson JM. Australian national
2. Dey A, Wang H, Menzies R, Macartney K. Changes in hospitalisations for birthweight percentiles by sex and gestational age, 19982007. Med J Aust.
acute gastroenteritis in Australia after the national rotavirus vaccination 2012;197(5):2914.
program. Med J Aust. 2012;197(8):4537. 27. Boyle EM, Poulsen G, Field DJ, Kurinczuk JJ, Wolke D, Alfirevic Z, Quigley MA.
3. Wilking H, Hohle M, Velasco E, Suckau M, Eckmanns T. Ecological analysis of Effects of gestational age at birth on health outcomes at 3 and 5 years of age:
social risk factors for Rotavirus infections in Berlin, Germany, 20072009. Int population based cohort study. BMJ. 2012;344, e896.
J Health Geogr. 2012;11(1):37.
28. Brown HK, Speechley KN, Macnab J, Natale R, Campbell MK. Neonatal
4. Dennehy PH, Cortese MM, Bgu RE, Jaeger JL, Roberts NE, Zhang R,
morbidity associated with late preterm and early term birth: the roles of
Rhodes P, Gentsch J, Ward R, Bernstein DI, et al. A case-control study to
gestational age and biological determinants of preterm birth. Int J
determine risk factors for hospitalization for rotavirus gastroenteritis in U.S.
Epidemiol. 2014;43(3):80214.
children. Pediatr Infect Dis J. 2006;25(12):112331.
29. Paranjothy S, Dunstan F, Watkins WJ, Hyatt M, Demmler JC, Lyons RA, Fone D.
5. Burt TD. Fetal regulatory T cells and peripheral immune tolerance in utero:
Gestational age, birth weight, and risk of respiratory hospital admission in
implications for development and disease. Am J Reprod Immunol. 2013;
childhood. Pediatrics. 2013;132(6):e15629.
69(4):34658.
6. Maynard CL, Elson CO, Hatton RD, Weaver CT. Reciprocal interactions of the 30. Skogstrand K, Hougaard DM, Schendel DE, Bent N-P, Svrke C, Thorsen P.
intestinal microbiota and immune system. Nature. 2012;489(7415):23141. Association of preterm birth with sustained postnatal inflammatory
7. Ballard O, Morrow AL. Human milk composition: nutrients and bioactive response. Obstet Gynecol. 2008;111(5):111828.
factors. Pediatr Clin North Am. 2013;60(1):4974. 31. Biasucci G, Rubini M, Riboni S, Morelli L, Bessi E, Retetangos C. Mode of
8. Azad MB, Konya T, Maughan H, Guttman DS, Field CJ, Chari RS, Sears MR, delivery affects the bacterial community in the newborn gut. Early Hum
Becker AB, Scott JA, Kozyrskyj AL. Gut microbiota of healthy Canadian Dev. 2010;86 Suppl 1:135.
infants: profiles by mode of delivery and infant diet at 4 months. CMAJ. 32. Penders J, Thijs C, Vink C, Stelma FF, Snijders B, Kummeling I, van den
2013;185(5):38594. Brandt PA, Stobberingh EE. Factors influencing the composition of the
9. Roberts CL, Algert CS, Ford JB, Nassar N. Mode of delivery may be the risk intestinal microbiota in early infancy. Pediatrics. 2006;118(2):51121.
factor for infant infectious morbidity. Arch Dis Child. 2012;97(8):759. 33. Romero R, Korzeniewski SJ. Are infants born by elective cesarean delivery
10. Hkansson S, Klln K. Caesarean section increases the risk of hospital without labor at risk for developing immune disorders later in life? Am J
care in childhood for asthma and gastroenteritis. Clin Exp Allergy. 2003; Obstet Gynecol. 2013;208(4):2436.
33(6):75764. 34. Duijts L, Ramadhani MK, Moll HA. Breastfeeding protects against infectious
11. Bager P, Simonsen J, Ethelberg S, Frisch M. Cesarean delivery and risk of diseases during infancy in industrialized countries. A systematic review.
intestinal bacterial infection. J Infect Dis. 2010;201(6):898902. Matern Child Nutr. 2009;5(3):199210.
Bentley et al. BMC Pediatrics (2016) 16:55 Page 10 of 10

35. Ladomenou F, Moschandreas J, Kafatos A, Tselentis Y, Galanakis E. Protective

effect of exclusive breastfeeding against infections during infancy: a
prospective study. Arch Dis Child. 2010;95(12):10048.
36. Clark SL, Frye DR, Meyers JA, Belfort MA, Dildy GA, Kofford S, Englebright J,
Perlin JA. Reduction in elective delivery at <39 weeks of gestation: comparative
effectiveness of 3 approaches to change and the impact on neonatal intensive
care admission and stillbirth. Am J Obstet Gynecol. 2010;203(5):449.e4416.
37. MacDorman MF, Reddy UM, Silver RM. Trends in stillbirth by gestational age
in the United States, 20062012. Obstet Gynecol. 2015;126(6):114650.
38. Little SE, Zera CA, Clapp MA, Wilkins-Haug L, Robinson JN. A Multi-State
Analysis of Early-Term Delivery Trends and the Association With Term
Stillbirth. Obstet Gynecol. 2015;126(6):113845.
39. March of Dimes. Prematurity Campaign 10-year anniversary report. 2013.
Available at:www.marchofdimes.com/materials/prematurity-campaign-
progress-report-2012.pdf. Accessed July 6, 2014.
40. Chang HH, Larson J, Blencowe H, Spong CY, Howson CP, Cairns-Smith S,
Lackritz EM, Lee SK, Mason E, Serazin AC, et al. Preventing preterm births:
analysis of trends and potential reductions with interventions in 39
countries with very high human development index. Lancet. 2013;
381(9862):22334.
41. Goldenberg RL, McClure EM, Bhattacharya A, Groat TD, Stahl PJ. Womens
perceptions regarding the safety of births at various gestational ages.
Obstet Gynecol. 2009;114(6):12548.
42. Zhang LY, Todd AL, Khambalia A, Roberts CL. Womens beliefs about the
duration of pregnancy and the earliest gestational age to safely give birth.
Aust N Z J Obstet Gynaecol. 2015;55(2):15662.
43. Dursun P, Yanik FB, Zeyneloglu HB, Baser E, Kuscu E, Ayhan A. Why women
request cesarean section without medical indication? J Matern Fetal
Neonatal Med. 2011;24(9):11337.
44. Roberts CL, Bell JC, Ford JB, Morris JM. Monitoring the quality of maternity
care: how well are labour and delivery events reported in population health
data? Paediatr Perinat Epidemiol. 2009;23(2):14452.
45. NSW Health Department: Validation Study: NSW Midwives Data Collection
1998. NSW Public Health Bull. 2000;9:979.
46. Chantry CJ, Dewey KG, Peerson JM, Wagner EA, Nommsen-Rivers LA. In-
hospital formula use increases early breastfeeding cessation among first-time
mothers intending to exclusively breastfeed. J Pediatr. 2014;164(6):133945.
47. Semenic S, Loiselle C, Gottlieb L. Predictors of the duration of exclusive
breastfeeding among first-time mothers. Res Nurs Health. 2008;31(5):42841.
48. Forster D, McLachlan H, Lumley J. Factors associated with breastfeeding at
six months postpartum in a group of Australian women. Int Breastfeed J.
2006;1(1):18.

Submit your next manuscript to BioMed Central

and we will help you at every step:
We accept pre-submission inquiries
Our selector tool helps you to find the most relevant journal
We provide round the clock customer support
Convenient online submission
Thorough peer review
Inclusion in PubMed and all major indexing services
Maximum visibility for your research

Submit your manuscript at

www.biomedcentral.com/submit