International Journal of Obesity (2014) 38, 1268–1274 © 2014 Macmillan Publishers Limited All rights reserved 0307-0565/14 www.nature.com/ijo

PEDIATRIC ORIGINAL ARTICLE

Maternal pregravid body mass index and child hospital admissions in the first 5 years of life: results from an Australian birth cohort CM Cameron1,2, R Shibl3, RJ McClure4, S-K Ng2,5 and AP Hills2,6 OBJECTIVES: To examine the association of maternal pregravid body mass index (BMI) and child offspring, all-cause hospitalisations in the first 5 years of life. METHODS: Prospective birth cohort study. From 2006 to 2011, 2779 pregnant women (2807 children) were enrolled in the Environments for Healthy Living: Griffith birth cohort study in South-East Queensland, Australia. Hospital delivery record and selfreport baseline survey of maternal, household and demographic factors during pregnancy were linked to the Queensland Hospital Admitted Patients Data Collection from 1 November 2006 to 30 June 2012, for child admissions. Maternal pregravid BMI was classified as underweight ( o18.5 kg m−2), normal weight (18.5–24.9 kg m−2), overweight (25.0–29.9 kg m−2) or obese (⩾30 kg m−2). Main outcomes were the total number of child hospital admissions and ICD-10-AM diagnostic groupings in the first 5 years of life. Negative binomial regression models were calculated, adjusting for follow-up duration, demographic and health factors. The cohort comprised 8397.9 person years (PYs) follow-up. RESULTS: Children of mothers who were classified as obese had an increased risk of all-cause hospital admissions in the first 5 years of life than the children of mothers with a normal BMI (adjusted rate ratio (RR) = 1.48, 95% confidence interval 1.10–1.98). Conditions of the nervous system, infections, metabolic conditions, perinatal conditions, injuries and respiratory conditions were excessive, in both absolute and relative terms, for children of obese mothers, with RRs ranging from 1.3–4.0 (PYs adjusted). Children of mothers who were underweight were 1.8 times more likely to sustain an injury or poisoning than children of normal-weight mothers (PYs adjusted). CONCLUSION: Results suggest that if the intergenerational impact of maternal obesity (and similarly issues related to underweight) could be addressed, a significant reduction in child health care use, costs and public health burden would be likely. International Journal of Obesity (2014) 38, 1268–1274; doi:10.1038/ijo.2014.148

INTRODUCTION Consistent with population-level changes in body mass index (BMI) and epidemic proportions of overweight and obesity in the developed world, significant mortality, morbidity and rising health care costs are evident.1,2 Obesity is now also a burgeoning problem in the developing world, particularly in women,3 in addition to the perpetuation of high levels of underweight and malnourishment in low- and middle-income countries.3,4 Current evidence demonstrates that an abnormal pregravid BMI, as a result of malnourishment or obesity, is associated with problems with conception, pregnancy complications as well as adverse birth and early neonatal outcomes.5,6 Compared with healthy weight mothers, pregnant women with very low pregravid BMI’s have an increased incidence of preterm deliveries, small-forgestational-age babies, as well as increased neonatal mortality.7,8 It is also well established that maternal prepregnancy obesity increases maternal morbidity with conditions such as gestational diabetes mellitus9 and hypertensive complications,10 as well as higher rates of caesarean sections11 and postoperative

complications.6 One recent population-based study found associations between maternal BMI during pregnancy, short-term morbidity and increased maternal and neonatal health service costs.12 Maternal obesity has been associated with other adverse neonatal outcomes including increased risk of birth injury,13,14 low Apgar scores,15,16 respiratory distress syndrome,13 hypoglycemia,13 neural tube defects6 and macrosomia.17,18 While pregnancy and early neonatal outcomes related to extreme maternal pregravid BMIs are well understood, there is less evidence regarding long-term intergenerational impact.6,19 Recent studies have begun to investigate the role of prepregnancy obesity and childhood respiratory conditions,5,20 child obesity,21 asthma,22 autism,23 attention deficit hyperactivity disorder24 and offspring adult cardiovascular events.25 However, no studies to date have compared maternal pregravid BMIs and patterns of offspring hospitalisations in early childhood. Therefore, the aim of this study was to examine the association of maternal pregravid BMI and child offspring all-cause hospitalisations in the first 5 years of life.

1 Centre of National Research on Disability and Rehabilitation, School of Human Services and Social Work, Griffith University, Griffith, Queensland, Australia; 2Griffith Health Institute, Griffith University, Griffith, Queensland, Australia; 3Faculty of Business, Queensland University of Technology, Brisbane, Queensland, Australia; 4Monash Injury Research Institute, Monash University, Monash, Victoria, Australia; 5School of Medicine, Griffith University, Griffith, Queensland, Australia and 6Mater Mothers’ Hospital, Mater Research Institute - University of Queensland, Brisbane, Queensland, Australia. Correspondence: Dr CM Cameron, Centre of National Research on Disability and Rehabilitation, School of Human Services and Social Work, Griffith University, Logan campus L05 4.57, University Drive, Meadowbrook, Griffith, Queensland 4131, Australia. E-mail: cate.cameron@griffith.edu.au Received 1 April 2014; revised 27 June 2014; accepted 18 July 2014; accepted article preview online 25 July 2014; advance online publication, 19 August 2014

Maternal body mass index and child hospitalisations CM Cameron et al

MATERIALS AND METHODS Study design Stratified cohort study.

Participants and setting Participants were child offspring of mothers in the Environments for Healthy Living (EFHL) study, who were residing in Queensland at the time of enrollment. EFHL is a longitudinal birth cohort study in Southeast Queensland and Northern New South Wales, Australia, examining a number of factors hypothesised to affect child health and development, including obesity.26 Women were enrolled over a 6-year period (2006– 2011) for 4 months each year, from three public maternity hospitals, while attending routine antenatal appointments during the third trimester of pregnancy.

Theoretical framework This analysis, and the wider EFHL study from which the data were obtained, are based on an ecoepidemiological model of health.26 This model highlights the multilevel nature of factors and interactions, including the environment, economic and social context, individual behaviours and genetic factors, which may influence health outcomes.27,28 These multilevel influences operate throughout the life course, with prenatal and early years of life having a crucial role in later outcomes.29,30 The fetal environment is considered a key factor in the aetiology of some diseases later in life, including asthma and respiratory functioning with health behaviours (for example, smoking and alcohol consumption) adopted by parents, important determinants of intrauterine growth.31 There is increasing interest in examining the role of nutrition, and extreme pregravid parental BMIs, in shaping health and development during the prenatal, infant, early childhood and adolescent periods.32 It is in this context that the current analysis was derived.

Ethics approval EFHL was approved by the Human Research Ethics Committees of the three participating hospitals (Logan Hospital, Gold Coast Hospital and The Tweed Hospital) and Griffith University. Each participant gave written informed consent for completion of a maternal baseline survey, the release of hospital perinatal data related to the birth of their child and linkage of their child’s inpatient state hospital records. All research data were de-identified and stored for analysis.

Data sources The baseline survey consisted of 48 self-report items including maternal, household and demographic factors during pregnancy.26 Hospital perinatal records were extracted for information on the pregnancy, delivery and newborn status. Child hospitalisations were extracted from Queensland Hospital Admitted Patients Data Collection from 1 November 2006 to 30 June 2012, inclusive. Child and maternal details and demographic data were linked by the Queensland Health Department using a linkage software applying deterministic and probabilistic methodologies, as well as manual clerical reviews where required.

Calculation of person years Children in the cohort were born between 2006 and 2011, with hospital data available up to mid 2012, and therefore, the length of follow-up differed considerably. In addition, if a participant’s family had moved out of Queensland, the child would not be utilising Queensland hospital services, and similarly, if the child was deceased. Using dates of birth and coverage dates, individual person years (PYs) were calculated for each child with the total time he or she was residing in the state of Queensland, alive and eligible for health care, from birth to 5 years.

Variables Cohort stratification. Child participants were stratified by their mother’s pregravid BMI determined by maternal self-report of weight and height. Maternal pregravid BMI was classified as underweight ( o18.5 kg m−2), normal weight (18.5–24.9 kg m−2), overweight (25.0–29.9 kg m−2) or obese © 2014 Macmillan Publishers Limited

(⩾30 kg m−2).33 Participants with missing maternal BMI data were excluded from the analysis. Other factors. Demographic, maternal and household variables were derived from maternal self-report survey data. Maternal factors included highest level of education obtained (tertiary degree, trade/apprenticeship, secondary school, not complete school), marital status (single/two-parent family), age at the time of the birth of the child (o 25, 25–29, 30–34 and 35 + years), diabetes—preexisting or gestational (yes/no) and adverse health behaviours during pregnancy. Adverse behaviours included cigarette smoking (yes/no), recreational drug use and/or high-level alcohol consumption (yes/no). High-level alcohol consumption was defined as five or more standard drinks on any one occasion after the first trimester of pregnancy. Alcohol consumption after the first trimester of pregnancy was used as a marker of the mother being aware they were pregnant and therefore conscious of health behaviour choices. Household variables included the number of other children living in the household (0, 1–2, 3+), their partner’s employment status (employed/not working) and income quintile. Gross annual household income was reported in $10,000 increments and standardised in AUD$2010 using published Consumer Price Index34,35 then divided into income quintiles. Child and birth data manually extracted from hospital perinatal records, included child gender (male/female), plurality (singleton/twins), gestational age ( o36 weeks/36+ weeks), birth mode (vaginal/caesarean), newborn admission to intensive or special care (yes/no) and birth weight. Using Australian National birth weights for full-term singletons, birth weight was classified as low (o 2500 g), normal (2500–4000 g) and macrosomia (44000 g).36 Outcome measures. Total number of hospitalisations and diagnostic groupings in the first 5 years of life. The hospitalisation for the birth itself was not included in the number of hospitalisations. Hospital transfers were kept as separate records. Social admissions were excluded. Diagnostic groupings were determined from the extracted hospital records during the follow-up period. Using the primary diagnostic field, admissions were broadly categorised according to the 21 chapters of ICD-10-AM.

Analysis Data cleaning and analyses were undertaken using SAS 9.3 software (SAS Institute). The statistical significance of differences between groups was assessed by χ2 test for categorical data, Gamma statistic for ordinal data and the Kruskal–Wallis test for continuous data because of non-normal distributions and the number of maternal pregravid BMI groups. All tests were two-sided with a 5% level of significance. Rates of all-cause admissions and ICD-10 diagnostic groups were calculated for the four maternal pregravid BMI groups taking into account PYs exposure data. Negative binomial regression was used to estimate crude and adjusted RR between exposure (maternal pregravid BMI) and outcome (number of hospitalisations) for child participants, accounting for individual PYs exposure time. Negative binomial regression can be considered as a generalisation of Poisson regression for modelling count data, with an extra dispersion parameter (alpha) to model the overdispersion. A likelihood ratio test of alpha equal to zero was adopted to compare this negative binomial regression model with a Poisson regression model. A Vuong test SAS macro37,38 was used to check whether a zero-inflated negative binomial regression model was required to account for the presence of 'excess zeroes' in the outcome data. On the basis of likelihood ratio test, the Vuong test and the goodness-of-fit criteria, the negative binomial regression model was the best model for the hospital count data.39,40 Weinberg states that any factor which is thought to be intermediate on the causal pathway between the exposure and the disease, or outcome of interest, should not be adjusted for in the analysis.41 On the basis of literature and clinical evidence, three factors (birth weight, gestational age and caesarean delivery) were hypothesised to be part of the causal pathway and were therefore not adjusted for in the analysis. Extremes in maternal pregravid BMI have been associated with low birth weight and preterm delivery for underweight groups, and macrosomia and an increased rate of caesarean delivery for overweight and obese groups.42–44 In turn, low birth weight, preterm births and caesarean deliveries have been shown to increase the risk of hospitalisation and length of stay during the neonate period.45–49 Factors shown to be associated with both the exposure and the outcomes in univariate analysis where Po 0.1, were included in the model as potential confounders. The final model included maternal diabetes International Journal of Obesity (2014) 1268 – 1274

1269

Maternal body mass index and child hospitalisations CM Cameron et al

1270 (preexisting and gestational), smoking during pregnancy, high-risk alcohol or drug use during pregnancy, marital status, partner employment status and newborn admission to intensive or special care. Owing to the small number of events within single years, regression analysis of the total number of admissions between ‘0–5 years’ was conducted.

RESULTS In total, 3105 mothers in the EFHL study resided in Queensland at the time of enrollment. Of those, 2779 women had complete pregravid BMI data (89.5%) and were included in the analysis, with 2807 children (including 28 sets of twins). Automated linkage and manual searching found records in Queensland Hospital Admitted Patients Data Collection for 97.1% of the child participants. Almost half of the maternal cohort was classified by their pregravid BMI as either overweight/obese (39.9%) or underweight (8.0%). Sociodemographic and birth characteristics for maternal BMI groups Women in the underweight BMI group were more likely to be under 25 years of age, be a single parent, have a partner who was not working, have no other children living in the household and reported adverse health behaviours during pregnancy (smoking, drug use/high-level alcohol consumption), compared with women with a normal BMI classification (all Po0.001) (Table 1). Women in the overweight and obese groups were more likely to be older and have other children living in the household, compared with women with a normal BMI (Po0.001). Women in the obese category also had a higher proportion of partners who were not working, singleparent families and were least likely to have completed tertiary education compared with women with a normal BMI (all Po0.001). Almost double the number of women with diabetes (preexisting and gestational) was found in the obese group compared with mothers with a normal BMI (5.9% vs 3.2%) (P = 0.0074). As maternal pregravid BMI increased, the number of women who had caesarean sections also increased from 23.5% in the underweight group to 35.6% in the obese group (Po0.0001). As maternal pregravid BMI increased, the number of newborn children who required admission to intensive or special care also increased, from 11.8% in the underweight group to 20.1% in the obese group (P = 0.001). There was no difference in the gender of the children (P = 0.6554) or the proportion of babies born less than 36 weeks gestation (P = 0.4662) between the maternal BMI groups (Table 1). Follow-up All children in the study were born between 2006 and 2011, with hospital admissions data available up to and including 30 June 2012, giving a total 8397.9 PYs follow-up from birth to 5 years of age. During the study period, there was a known 4% ‘out of state’ change in residence and 0.2% child deaths, which were censored from the PYs on the dates of departure and dates of death. There was a significant difference in the length of follow-up time for the pregravid BMI groups with less follow-up time of the children of obese women (P = 0.006). Consistent with this, a previous analysis indicated that across the years of enrollment an increasing proportion of women were classified with pregravid obesity.50 Hospital admissions Of the total 2807 children, 638 had one or more hospital admissions in the follow-up period (22.7%). With the exception of year 3, the children of obese mothers had higher rates of allcause hospital admissions each year, compared with the children of mothers with a normal BMI (Table 2). Children of obese mothers had a 5-year overall rate of 2.3 vs 1.24 adms per 10 PYs in the children with mothers with a normal BMI. This was followed by the children of underweight mothers with a rate of 1.39 adms per 10 International Journal of Obesity (2014) 1268 – 1274

PYs. There was some instability in the rates of admissions per year, most likely due to small numbers, and in the maternal obese category, possibly caused by one outlier case with a large number of admissions in years 4 and 5. However, the first year of life had consistently high rates of child admissions for all maternal BMI groups compared with the other four years. After adjusting for known confounders, including marital status, partner’s employment status, maternal diabetes, adverse health behaviours during pregnancy and newborn admission to special care unit, children of mothers who were classified as obese had significantly more all-cause hospital admissions in the first 5 years of life than the children of mothers with a normal BMI (Adj RR = 1.48, 95% confidence interval 1.10–1.98) (Table 3). There was no difference in the adjusted rates of hospital admissions for the children of mothers who were classified as underweight or overweight. The only covariate that remained significant in the adjusted model was whether the newborn had been admitted to intensive or special care after delivery (P = 0.0002). Admissions by ICD-10 diagnostic groupings To determine if there were particular patterns of service use or specific conditions that affected children of mothers with different pregravid BMIs, child hospital admissions were categorised using the primary diagnostic field, grouped according to the broad diagnostic chapters of ICD-10-AM. For the total cohort, the most common reason for admission to hospital in the first years of life was for respiratory conditions (25.1%), followed by ill-defined conditions (8.8%), injury and poisonings (8.5%) and infectious diseases (8.4%). Children of mothers classified as obese experienced higher rates of admissions for a number of diagnostic types compared with children of mothers with a normal BMI (Figure 1). Although neoplasms had the greatest relative difference (RR = 9.5), in absolute terms, neoplasm accounted for less than 1% of all admissions. Children of mothers in the obese group also had 4.0 times the rate of admission for nervous system diseases compared with those with mothers with a normal BMI (26.0 vs 6.4 adms/1000 PYs), 2.3 times the rate of admissions for infectious diseases (19.6 vs 8.6 adms/1000 PYs), 2.1 times the rate of endocrine and metabolic disorders (5.6 vs 2.7 adms/1000 PYs), 1.7 times the rates of conditions from the perinatal period (14.7 vs 8.9 adms/1000 PYs) and 1.6 times the rate of injuries and poisonings (16.1 vs 10.2 adms/1000 PYs). However, the greatest absolute burden was accounted for by respiratory conditions with a rate 1.3 times that of children of normal BMI mothers (40.7 vs 31.0 adms/1000 PYs). Similarly, children of mothers in the underweight group had the greatest relative difference in admissions for neoplasms (RR = 9.1), but accounted for few admissions overall (Figure 2). The diagnostic group that accounted for the greatest absolute and relative difference for the children of underweight mothers (compared with children of mothers with a normal BMI) was injuries and poisonings (RR = 1.8; 18.1 vs 10.2 adms/1000 PY). Rates of admissions for children in the maternally overweight group compared with the normal BMI group were elevated in similar diagnostic types as the children of obese mothers but to a lesser extent, including diseases of the nervous system (RR = 1.4; 8.8 vs 6.4 adms/1000 PYs), injury and poisonings (RR = 1.3; 13.3 vs 10.2 adms/1000 PYs) and infectious diseases (RR = 1.2; 10.5 vs 8.6 adms/1000 PYs) (Figure 3). DISCUSSION The present study is the first to identify a significant association between maternal pregravid obesity and all-cause child hospital admissions and diagnostic patterns, in the early years of life. Considerable research evidence suggests that maternal overweight and obesity are causally related to pregnancy © 2014 Macmillan Publishers Limited

Maternal body mass index and child hospitalisations CM Cameron et al

1271 Table 1.

Maternal, household characterisitics and child birth characteristics by pregravid maternal BMI status

Maternal cohort (n = 2779)

Maternal and delivery characteristics (n = 2779) Maternal age (in years) o 25 years 25–29 years 30–34 years 35+ years Missing = 10 Maternal education Not complete school Secondary school Trade/apprenticeship Tertiary degree Missing = 12 Diabetesa No Yes Missing = 104 Cigarette smoking No Yes Missing = 11 High-risk alcohol or drug use No Yes Missing = 537 Birth mode Vaginal Caesarean Missing = 95 Household characteristics (n = 2779) Marital status Two-parent family Single-parent family Missing = 9 Other children in household None 1–2 children 3+ children Missing = 25 Partner employment status Employed Not working Missing = 148 Annual household incomeb Lowest quintile 2nd quintile 3rd quintile 4th quintile Highest quintile Missing = 407 Child characteristics (n = 2807) Gender Male Female Missing = 88 Plurality Singleton Twin Missing = 0 Birth weight Low ( o 2500 g) Normal (2500–3999 g) Macrosomia (44000 g) Missing = 94 Gestational age o 36 weeks 36+ weeks Missing = 92 Intensive or special care No Yes Missing = 90

Underweight n (%) 223 (8.0)

Normal weight n (%) 1447 (52.1)

Overweight n (%) 610 (21.9)

Obese n (%) 499 (18.0)

78 62 44 38

(35.2) (27.9) (19.8) (17.1)

330 423 408 280

(22.9) (29.3) (28.3) (19.4)

113 195 165 136

(18.6) (32.0) (27.1) (22.3)

106 149 125 117

(21.3) (30.0) (25.2) (23.5)

49 73 57 41

(22.3) (33.2) (25.9) (18.6)

264 414 413 351

(18.3) (28.7) (28.6) (24.4)

114 196 190 107

(18.8) (32.3) (31.3) (17.6)

112 (22.5) 174 (34.9) 152 (30.5) 60 (12.1)

210 (99.1) 2 (0.9)

1354 (96.8) 45 (3.2)

566 (96.3) 22 (3.7)

448 (94.1) 28 (5.9)

146 (65.5) 77 (34.5)

1121 (77.7) 322 (22.3)

461 (76.0) 146 (24.0)

356 (71.9) 139 (28.1)

148 (80.9) 35 (19.1)

1040 (91.3) 99 (8.7)

468 (93.8) 31 (6.2)

389 (92.4) 32 (7.6)

163 (76.5) 50 (23.5)

1049 (74.7) 355 (25.3)

391 (66.4) 198 (33.6)

309 (64.6) 169 (35.6)

182 (81.6) 41 (18.4)

1281 (88.8) 161 (11.2)

540 (89.1) 66 (10.9)

419 (84.0) 80 (16.0)

117 (52.7) 89 (40.1) 16 (7.2)

603 (42.2) 710 (49.6) 117 (8.2)

196 (32.5) 340 (56.4) 67 (11.1)

162 (32.5) 246 (49.3) 91 (18.2)

185 (88.5) 24 (11.5)

1295 (94.0) 82 (6.0)

547 (94.8) 30 (5.2)

423 (90.4) 45 (9.6)

50 43 34 22 32

219 210 240 272 286

(17.8) (17.1) (19.6) (22.2) (23.3)

97 (18.2) 116 (21.8) 103 (19.4) 110 (20.7) 106 (19.9)

75 (17.4) 112 (25.9) 96 (22.2) 80 (18.5) 69 (16.0)

111 (52.1) 102 (47.9)

739 (51.9) 685 (48.1)

293 (48.9) 306 (51.1)

245 (50.7) 238 (49.3)

223 (100.0) 0 (0.0)

1430 (97.7) 34 (2.3)

602 (97.4) 16 (2.6)

496 (98.8) 6 (1.2)

9 (4.3) 191 (90.5) 11 (5.2)

40 (2.8) 1201 (84.5) 180 (12.7)

21 (3.5) 474 (79.1) 104 (17.4)

9 (1.9) 364 (75.5) 109 (22.6)

8 (3.8) 205 (96.2)

51 (3.6) 1372 (96.4)

28 (4.7) 570 (95.3)

19 (4.0) 462 (96.0)

187 (88.2) 25 (11.8)

1229 (86.3) 195 (13.7)

490 (81.9) 108 (18.1)

386 (79.9) 97 (20.1)

(27.6) (23.8) (18.8) (12.1) (17.7)

Abbreviation: BMI, body mass index. aPreexisting and gestational diabetes. bAUD$2010.

© 2014 Macmillan Publishers Limited

International Journal of Obesity (2014) 1268 – 1274

Maternal body mass index and child hospitalisations CM Cameron et al

1272 Table 2. Child hospital admissions from birth to 5 years of age by maternal pregravid BMI, overall and by 1-year time intervals: number of hospitalisations (excluding birth admission), rates reported per 10 PYs (n = 2807) Underweight

Year 1 Year 2 Year 3 Year 4 Year 5 Total 5 years

Normal weight

Overweight

n

Rate

n

Rate

n

38 22 15 7 10 92

1.83 1.25 1.13 0.80 1.73 1.39

275 120 74 55 35 557

1.98 1.03 0.83 0.86 0.82 1.24

95 48 38 27 13 220

Rate

Obese n

Rate

1.63 128 2.73 1.00 58 1.49 1.08 14 0.52 1.14 52 2.96 0.85 88 7.34 1.22 327 2.30

Abbreviations: BMI, body mass index; PYs, person years.

complications, increased birth weight, difficulties associated with the delivery of large-for-gestational-age babies and neonatal admissions to care.10,12 Similarly, there is some evidence that maternal morbid obesity (BMI ⩾ 40 kg m − 2) is causally related to higher levels of long-term adiposity in offspring; however, further study of this relationship is warranted. There is also a need to better understand the history and timing of prepregnancy weight gain in addition to gestational weight gain, to appreciate the factors contributing to risk of short- and long-term unhealthy body composition in offspring and cumulative impact across generations.32 Research into long-term outcomes for offspring of mothers with extreme pregravid BMIs is increasing with some evidence of adverse learning, development and poorer child health outcomes.6,10,19,51 Although the link between maternal obesity and an increased risk of neonate care has been shown,12 a few recent studies have begun to examine hospital admissions for particular conditions in children of obese mothers, including childhood respiratory problems,20 and adult offspring cardiovascular events.25 This study demonstrated a strong association between maternal obesity and child offspring hospitalisations in the first 5 years of life. After adjusting for maternal, sociodemographic and birth factors, children of obese mothers had almost 1.5 times the number of admissions to hospital than children of mothers with a normal BMI. This significance was sustained when one child in the maternally obese category with an excess number of admissions was excluded from the analysis. Different diagnostic patterns of service use were evident for children of mothers with extreme pregravid BMIs. Conditions of the nervous system, infections, metabolic conditions, perinatal conditions, injuries and respiratory conditions were excessive, in both absolute and relative terms, for children of obese mothers, with RRs ranging from 1.3–4.0. Although smaller, a similar diagnostic pattern of increased admissions was found for children of mothers who were overweight. These findings support one case–control study that examined hospital admissions for respiratory conditions in children (0–5 years) and maternal pregravid BMI. Similarly, Parsons et al.20 found an elevated maternal pregravid BMI was associated with higher risk of early childhood respiratory hospitalisations. In the current study, the most notable difference for children of mothers who were underweight was that they were 1.8 times more likely to sustain an injury or poisoning than children of normalweight mothers. This is likely, in part, to reflect the sociodemographic differences in the underweight maternal group with a predominance of younger mothers, lowest income households, increased partner unemployment, increased prevalence of smoking, high-risk alcohol and drug use and lower education levels. Consistent with the literature, this study found sociodemographic and birth differences between the maternal BMI groups.52,53 Low levels of maternal International Journal of Obesity (2014) 1268 – 1274

Table 3. Unadjusted and adjusted RRs for all-cause child hospital admissions in the first 5 years of life by maternal pregravid BMI status

Maternal BMI Underweight Normal weight Overweight Obese

Crude RR (95% CI)

Adjusted RR (95% CI)a

1.06 (0.75–1.52) Reference 0.94 (0.74–1.20) 1.69 (1.32–2.15)*

1.00 (0.65–1.54) Reference 0.95 (0.71–1.28) 1.48 (1.10–1.98)*

Maternal characteristics Diabetesb No Yes Smoked during pregnancy No Yes High-risk alcohol or drug use No Yes Household characteristics Marital status Two-parent family Single-parent family Partners employment status Employed Not working Child characteristics Intensive or special care No Yes

Reference 0.74 (0.40–1.39) Reference 1.06 (0.81–1.40) Reference 0.85 (0.56–1.29)

Reference 1.27 (0.85–1.89) Reference 1.41 (0.92–2.16)

Reference 1.79 (1.32–2.43)*

Abbreviations: BMI, body mass index; CI, confidence interval; RR, rate ratio. *Po0.01. aAdjusted for factors shown in univariate analysis with a P-value o0.1. bPreexisting and gestational diabetes.

education, partner unemployment and single-parent families were associated with both underweight and obese mothers. Gestational diabetes, caesarean sections, child macrosomia and neonate admissions were more likely for mothers in the obese group and their child offspring. Study strengths and weaknesses The present study has numerous strengths. It is the first study to examine the relationship between maternal BMI and all-cause child hospital admissions and diagnostic patterns of admissions in children 0–5 years of age. The prospective longitudinal data with linkage to administrative health data minimised loss to follow-up common in many cohort studies. The study had a sample size of 2807 children with almost 8400 PYs of follow-up time. Over 97.1% of child participants were identified in the state-based hospital data, which included the birth record. An optimal feature of maternal and offspring health studies is the incorporation of objective or measured as opposed to selfreport estimates of height and weight and subsequent BMI classification. A weakness of the current study was the reliance on self-report of maternal height and weight measures. Further, women were recruited to the study in the latter weeks of pregnancy through routine antenatal hospital clinics so those with high risk and/or complications during pregnancy, plus infants born significantly preterm, were not included.50 This is likely to account for why no difference was found for the gestational ages across maternal BMI groups that is in contrast to other research.6 However, previous analysis of the baseline data indicated maternal characteristics were broadly representative of the regional study and national reference population, and showed little change across the recruitment years.50 © 2014 Macmillan Publishers Limited

Maternal body mass index and child hospitalisations CM Cameron et al

1273 Normal

Obese

Diagnostic groups

Respiratory diseases Congenital Anomalies Injury and Poisoning Ill defined conditions Perinatal conditions Infectious diseases Digestive diseases Ear and mastoid Nervous system Genitourinary Skin diseases Endocrine/Metabolic Eye and adnexa Musculoskeletal Circulatory Neoplasms 0

5

10 15 20 25 30 35 Admission rates /1000 PYs

40

45

Figure 1. Five year rates of ICD-10 diagnostic grouping of hospital admissions per 1000 PYs, for children of maternal pregravid obese BMI group compared with normal BMI group.

Diagnostic groups

Normal

Underweight

Respiratory diseases Congenital Anomalies Injury and Poisoning Ill defined conditions Perinatal conditions Infectious diseases Digestive diseases Ear and mastoid Nervous system Genitourinary Skin diseases Endocrine/Metabolic Eye and adnexa Musculoskeletal Circulatory Neoplasms

CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS 0

5

10

15 20 25 30 35 Admission rates /1000 PYs

40

45

Figure 2. Five year rates of ICD-10 diagnostic grouping of hospital admissions per 1000 PYs, for children of maternal pregravid underweight BMI group compared with normal BMI group. Normal

Diagnostic groups

front of mind for many women as they enter the childbearing years. Evidence also suggests that the preconception period (for those women planning to fall pregnant) and the duration of pregnancy are ‘teachable moments’ and therefore, windows of opportunity to influence the health practices of women. However, given the extent of maternal obesity in the Australian context and across the developed world, knowledge and awareness of health benefits has not translated into behaviour change in a sufficient number of women. Importantly, maternal weight is modifiable with support and appropriate engagement in physical activity throughout pregnancy and attendant consideration of energy intake. For the vast majority of women, regular physical activity sufficient to provide health benefits, is not only possible but potentially of critical importance to maternal and offspring health. This study adds support to a growing body of evidence of the health consequences of maternal overweight and obesity on pregnancy and delivery outcomes plus offspring health and development. Children of women classified as obese were found to have an increased rate of hospitalisations, particularly for conditions of the nervous system, infections, metabolic and respiratory conditions. Although it was not possible for this study to provide an insight into underlying causal mechanisms between maternal pregravid weight and child offspring hospital admissions,54 results suggest that if the intergenerational impact of maternal overweight and obesity (and similarly issues related to underweight) could be addressed, a significant reduction in child health care use, costs and public health burden would be likely.

Overweight

Respiratory diseases Congenital Anomalies Injury and Poisoning Ill defined conditions Perinatal conditions Infectious diseases Digestive diseases Ear and mastoid Nervous system Genitourinary Skin diseases Endocrine/Metabolic Eye and adnexa Musculoskeletal Circulatory Neoplasms

The research reported in this publication is part of the Griffith Study of Population Health: Environments for Healthy Living (EFHL) (Australian and New Zealand Clinical Trials Registry: ACTRN12610000931077). Core funding to support EFHL is provided by Griffith University. The EFHL project was conceived by RJM, CMC, Professor Judy Searle and Professor Ronan Lyons. We are thankful for the contributions of the Project Manager, Rani Scott, and the current and past Database Managers. We gratefully acknowledge the administrative staff, research staff and the hospital antenatal and birth suite midwives of the participating hospitals for their valuable contributions to the study, in addition to the expert advice provided by Research Investigators throughout the project. CMC was supported by a Public Health Fellowship (ID 428254) from the National Health and Medical Research Council (NHMRC) Australia.

AUTHOR CONTRIBUTIONS CMC originated the research, conducted the statistical analyses and led the writing of the article. APH contributed to the writing of literature review and discussion in the article. CMC and RJM originated the EFHL study and supervised all aspects of its implementation. All authors assisted with conceptualising ideas, interpreted research findings and contributed to the writing of the article. All authors read and approved the final manuscript.

REFERENCES 0

5

10

15 20 25 30 35 Admission rates /1000 PYs

40

45

Figure 3. Five year rates of ICD-10 diagnostic grouping of hospital admissions per 1000 PYs, for children of maternal pregravid overweight BMI group compared with normal BMI group.

Public health implications and conclusion Knowledge and understanding of the importance of healthy eating and activity behaviours for maternal and offspring health is © 2014 Macmillan Publishers Limited

1 Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012; 380: 2224–2260. 2 World Health Organization. Global Status Report on Noncommunicable Diseases 2010. World Health Organization: Geneva, 2011. 3 Kanter R, Caballero B. Global gender disparities in obesity: a review. Adv Nutr 2012; 3: 491–498. 4 Keats S, Wiggins S. Future diets: implications for agriculture and food prices. Overseas Dev Inst 2014; http://www.odi.org/sites/odi.org.uk/files/odi-assets/pub lications-opinion-files/8776.pdf.

International Journal of Obesity (2014) 1268 – 1274

Maternal body mass index and child hospitalisations CM Cameron et al

1274 5 Haberg SE, Stigum H, London SJ, Nystad W, Nafstad P. Maternal obesity in pregnancy and respiratory health in early childhood. Paediatr Perinat Epidemiol 2009; 23: 352–362. 6 Ruager-Martin R, Hyde MJ, Modi N. Maternal obesity and infant outcomes. Early Hum Dev 2010; 86: 715–722. 7 Kalk P, Guthmann F, Krause K, Relle K, Godes M, Gossing G et al. Impact of maternal body mass index on neonatal outcome. Eur J Med Res 2009; 14: 216–222. 8 Sebastian Manzanares G, Angel Santalla H, Irene Vico Z, Lopez Criado MS, Alicia Pineda L, Jose Luis Gallo V. Abnormal maternal body mass index and obstetric and neonatal outcome. J Matern Fetal Neonatal Med 2012; 25: 308–312. 9 Scott-Pillai R, Spence D, Cardwell C, Hunter A, Holmes V. The impact of body mass index on maternal and neonatal outcomes: a retrospective study in a UK obstetric population, 2004-2011. BJOG 2013; 120: 932–939. 10 Lawlor DA, Relton C, Sattar N, Nelson SM. Maternal adiposity--a determinant of perinatal and offspring outcomes? Nat Rev Endocrinol 2012; 8: 679–688. 11 Marshall NE, Guild C, Cheng YW, Caughey AB, Halloran DR. The effect of maternal body mass index on perinatal outcomes in women with diabetes. Am J Perinatol 2014; 31: 249–256. 12 Denison FC, Norwood P, Bhattacharya S, Duffy A, Mahmood T, Morris C et al. Association between maternal body mass index during pregnancy, short-term morbidity, and increased health service costs: a population-based study. BJOG 2014; 121: 72–81; discussion 82. 13 Blomberg M. Maternal obesity, mode of delivery, and neonatal outcome. Obstet Gynecol 2013; 122: 50–55. 14 Kaplan-Sturk R, Akerud H, Volgsten H, Hellstrom-Westas L, Wiberg-Itzel E. Outcome of deliveries in healthy but obese women: obesity and delivery outcome. BMC Res Notes 2013; 6: 50. 15 Galtier-Dereure F, Boegner C, Bringer J. Obesity and pregnancy: complications and cost. Am J Clin Nutr 2000; 71: 1242S–1248SS. 16 Minsart AF, Buekens P, De Spiegelaere M, Englert Y. Neonatal outcomes in obese mothers: a population-based analysis. BMC Pregnancy Childbirth 2013; 13: 36. 17 Yu Z, Han S, Zhu J, Sun X, Ji C, Guo X. Pre-pregnancy body mass index in relation to infant birth weight and offspring overweight/obesity: a systematic review and meta-analysis. PLoS One 2013; 8: e61627. 18 Ovesen P, Rasmussen S, Kesmodel U. Effect of prepregnancy maternal overweight and obesity on pregnancy outcome. Obstet Gynecol 2011; 118: 305–312. 19 O'Reilly JR, Reynolds RM. The risk of maternal obesity to the long-term health of the offspring. Clin Endocrinol (Oxf) 2013; 78: 9–16. 20 Parsons EC, Patel K, Tran BT, Littman AJ. Maternal pre-gravid obesity and early childhood respiratory hospitalization: a population-based case-control study. Matern Child Health J 2013; 17: 1095–1102. 21 Reynolds RM, Osmond C, Phillips DI, Godfrey KM. Maternal BMI, parity, and pregnancy weight gain: influences on offspring adiposity in young adulthood. J Clin Endocrinol Metab 2010; 95: 5365–5369. 22 Patel SP, Rodriguez A, Little MP, Elliott P, Pekkanen J, Hartikainen AL et al. Associations between prepregnancy obesity and asthma symptoms in adolescents. J Epidemiol Community Health 2012; 66: 809–814. 23 Tanne JH. Maternal obesity and diabetes are linked to children's autism and similar disorders. BMJ 2012; 344: e2768. 24 Chen Q, Sjolander A, Langstrom N, Rodriguez A, Serlachius E, D'Onofrio BM et al. Maternal pre-pregnancy body mass index and offspring attention deficit hyperactivity disorder: a population-based cohort study using a siblingcomparison design. Int J Epidemiol 2014; 43: 83–90. 25 Reynolds RM, Allan KM, Raja EA, Bhattacharya S, McNeill G, Hannaford PC et al. Maternal obesity during pregnancy and premature mortality from cardiovascular event in adult offspring: follow-up of 1 323 275 person years. BMJ 2013; 347: f4539. 26 Cameron CM, Scuffham PA, Spinks A, Scott R, Sipe N, Ng S et al. Environments for Healthy Living (EFHL) Griffith birth cohort study: background and methods. Matern Child Health J 2012; 16: 1896–1905. 27 Nicholson JM, Sanson A. A new longitudinal study of the health and wellbeing of Australian children: how will it help? Med J Aust 2003; 178: 282–284. 28 Lynch JW. Social epidemiology: some observations on the past, present and future.Australas Epidomol 2000; 7: 7–15.

International Journal of Obesity (2014) 1268 – 1274

29 Eriksson JG. Epidemiology, genes and the environment: lessons learned from the Helsinki Birth Cohort Study. J Intern Med 2007; 261: 418–425. 30 Stettler N, Iotova V. Early growth patterns and long-term obesity risk. Curr Opin Clin Nutr Metab Care 2010; 13: 294–299. 31 Alati R, Al Mamun A, O'Callaghan M, Najman JM, Williams GM. In utero and postnatal maternal smoking and asthma in adolescence. Epidemiology 2006; 17: 138–144. 32 Herman DR, Taylor Baer M, Adams E, Cunningham-Sabo L, Duran N, Johnson DB et al. Life course perspective: evidence for the role of nutrition. Matern Child Health J 2014; 18: 450–461. 33 World Health Organization. Obesity: Preventing and Managing the Global Epidemic: Report of a WHO Consultation. World Health Organization: Geneva, 2000. 34 Australian Bureau of Statistics. Comsumer Price Index, Australia 2006-10. 35 Australian Bureau of Statistics 2011. Household Income and Income Distribution, Australia, 2009-10. 36 Roberts CL, Lancaster PA. Australian national birthweight percentiles by gestational age. Med J Aust 1999; 170: 114–118. 37 Vuong QH. Likelihood ratio tests for model selection and non-nested hypotheses. Econometrica 1989; 57: 307–333. 38 SAS Sample 42514 Tests for Comparing Nested and Nonnested Models (Program). SAS Institute Inc.: Cary, NC, USA. http://support.sas.com/kb/42/514.html. 39 McCullagh P, Nelder JA. Generalized linear models. 2nd edn. Chapman and Hall, 1989. 40 Gardner W, Mulvey EP, Shaw EC. Regression analyses of counts and rates: poisson, overdispersed Poisson, and negative binomial models. Psychol Bull 1995; 118: 392–404. 41 Weinberg CR. Toward a clearer definition of confounding. Am J Epidemiol 1993; 137: 1–8. 42 Vinayagam D, Chandraharan E. The adverse impact of maternal obesity on intrapartum and perinatal outcomes. ISRN Obstetr Gynecol 2012; 2012: 939762. 43 Lynch CM, Sexton DJ, Hession M, Morrison JJ. Obesity and mode of delivery in primigravid and multigravid women. Am J Perinatol 2008; 25: 163–167. 44 Leddy MA, Power ML, Schulkin J. The impact of maternal obesity on maternal and fetal health. Rev Obstet Gynecol 2008; 1: 170–178. 45 Petrou S, Mehta Z, Hockley C, Cook-Mozaffari P, Henderson J, Goldacre M. The impact of preterm birth on hospital inpatient admissions and costs during the first 5 years of life. Pediatrics 2003; 112: 1290–1297. 46 Yuksel B, Greenough A. Birth weight and hospital readmission of infants born prematurely. Arch Pediatr Adolesc Med 1994; 148: 384–388. 47 Soskolne EI, Schumacher R, Fyock C, Young ML, Schork A. The effect of early discharge and other factors on readmission rates of newborns. Arch Pediatr Adolesc Med 1996; 150: 373–379. 48 Goer H, Romano A, Sakala C. Vaginal or Cesarean Birth: What Is at Stake for Women and Babies? 2012, New York, USA. 49 Martens PJ, Derksen S, Gupta S. Predictors of hospital readmission of Manitoba newborns within six weeks postbirth discharge: a population-based study. Pediatrics 2004; 114: 708–713. 50 Cameron CM, Scuffham PA, Shibl R, Ng S, Scott R, Spinks A et al. Environments For Healthy Living (EFHL) Griffith birth cohort study: characteristics of sample and profile of antenatal exposures. BMC Pub Health 2012; 12: 1080. 51 Hinkle SN, Sharma AJ, Kim SY, Schieve LA. Maternal prepregnancy weight status and associations with children's development and disabilities at kindergarten. Int J Obes (Lond) 2013; 37: 1344–1351. 52 Heslehurst N, Ells LJ, Simpson H, Batterham A, Wilkinson J, Summerbell CD. Trends in maternal obesity incidence rates, demographic predictors, and health inequalities in 36,821 women over a 15-year period. BJOG 2007; 114: 187–194. 53 Heslehurst N, Rankin J, Wilkinson JR, Summerbell CD. A nationally representative study of maternal obesity in England, UK: trends in incidence and demographic inequalities in 619 323 births, 1989-2007. Int J Obes (Lond) 2010; 34: 420–428. 54 Langley K, Thapar A. Commentary: Maternal pre-pregnancy BMI and offspring ADHD: a lesson in the importance of testing causal pathways. Int J Epidemiol 2014; 43: 91–93.

© 2014 Macmillan Publishers Limited