Intrauterine exposure to hyperglycaemia in pregnancy and risk of adiposity in the offspring at 10 years of age – A community based retrospective cohort study in Sri Lanka

Background Intrauterine exposure to a hyperglycemic environment can cause long term changes in body composition resulting in increased adiposity and cardio metabolic risk in the offspring. The aim of this study was to determine the association between hyperglycaemia in pregnancy (HIP) and risk of adiposity in the offspring at 10-11 years of age. Methods A retrospective cohort study was conducted in the Colombo district, Sri Lanka. 7205 children who were born in 2005 were identified through schools and Public Health Midwives in the community. Mothers of these children still possessing antenatal records were interviewed and relevant data were extracted from medical records to identify eligible participants. Exposure status (hyperglycaemia in pregnancy) was ascertained based on client held antenatal records. 159 children of mothers with HIP (exposed) and 253 children of mothers with no HIP (non-exposed) were recruited. Height, weight, waist circumference and triceps skin fold thickness (TSFT) of participants were measured to ascertain outcome status. Results The mean ages (SD) of exposed and non-exposed groups were 10.9 (0.3) and 10.8 (0.3) years respectively. The median BMI (17.6 vs 16.1, p<0.001), waist circumference (63cm vs 59.3 cm, p<0.001) and triceps skinfold thickness (13.7mm vs 11.2mm, p< 0.001) were significantly higher in the exposed group than in the non-exposed group. Children who were exposed to intrauterine hyperglycaemia were more likely to be overweight (aOR=2.5, 95% CI 1.3-4.7), have abdominal obesity (aOR=2.9, 95% CI 1.2-6.8) and high TSFT > 70th centile (aOR=2.1, 95% CI 1.2-3.9) at 10-11 years of age than children who were not exposed after adjusting for maternal BMI, birth weight and birth order. Conclusions Intrauterine exposure to HIP is associated with significantly higher risk of adiposity in the offspring at 10 years of age.

cases are due to diabetes in pregnancy (DIP) which is either pre-existing type 1 or type 2 diabetes 36 or diabetes first detected at any time during the index pregnancy [1,2]. The number of women 37 having hyperglycaemia in pregnancy is increasing as a result of the increasing prevalence of 38 obesity and diabetes in women and higher age at childbirth [3]. 39 Pederson's hyperglycemia-hyperinsulinism hypothesis, as proven by several studies, is still the 40 basis of research on feto-maternal metabolism [4,5]. This hypothesis postulates that deficiency of 41 maternal insulin causes a rise in maternal glucose, which in turn increases fetal glucose levels. 42 This results in fetal hyperinsulinaemia which stimulates fetal growth and adiposity. Frienkel and 43 Metzger stated that deficiency of maternal insulin causes an increased influx of mixed nutrients or 44 fuels (glucose, amino acids, lipids, ketones) into fetal circulation resulting in hyperinsulinaemia 4 45 [4]. Frienkel presented the concept "fuel-mediated teratogenesis" to describe alterations that goes 46 beyond organogenesis causing long-range effects on anthropometric, metabolic and behavioral 47 functions in the offspring due to abnormal fuel mixtures in maternal metabolism due to in pregnancy have had a small number of exposed offspring thus limiting the power of such 57 studies [10,11]. A large number of studies have reported a positive association between HIP and 58 overweight and obesity [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27], while few studies have not shown such an association [11, 30]. Given the limited evidence from South Asian populations for risk estimates for childhood 60 obesity that are attributable to maternal diabetes in utero, further studies in these populations were 61 identified as an important research need [31]. South Asians present with greater metabolic risk at 62 lower levels of BMI compared with other ethnic groups, with type 2 diabetes developing at a 63 younger age and rapidly progressing to other complications [32][33][34]. Many studies have shown 64 that being obese in childhood and adolescence is associated with obesity in the adult life, and 65 overweight in adolescence is considered an important predictor of long-term morbidity and 66 mortality [28,[35][36][37][38]. Given the high risk of diabetes and cardiovascular diseases and rising trend 5 67 of obesity among South Asians, it is imperative that we identify risk groups and target interventions 68 from early life to mitigate the escalating epidemic of non-communicable diseases.

69
The aim of this study was to determine the association between the intrauterine exposure to 70 hyperglycaemia and anthropometric measurements in offspring at 10 -11 years of age in Sri Lanka 71 and to determine whether the association was independent of child's birth weight, parity and 72 mother's pre-pregnancy BMI. Lanka. The PHM delivers maternal and child care services as the grass roots level healthcare 87 worker. The PHM maintains a paper-based record keeping system for maternal and child care 88 services and all live births in a given PHM area are recorded in the "Birth and Immunization Register" (BI Register) by the PHM. In the current study, we identified children born in 2005 90 through the BI registers and through schools in the selected MOH areas.

91
There was no universal screening programme to screen for HIP in Sri Lanka in 2005. During this 92 period, GDM screening in the antenatal clinics, as per national guidelines at that time, was based 93 on assessment of risk factors (41). These women underwent 75g oral glucose tolerance testing 94 mainly at gestation weeks 24-28. WHO (1999) criteria for 2-hour post 75g oral glucose load 95 (≥140mg/dl) was taken as the criterion for diagnosis of GDM (42).

96
Since Sri Lanka does not have an electronic database system for keeping patient records and paper-97 based records are stored only for 5 years in the health institutions, tracing patient held antenatal 98 records to verify exposure status (hyperglycaemia in pregnancy) was the best possible option 99 available. A feasibility study conducted beforehand to verify the availability of patient held 100 antenatal records revealed that approximately 70% of women had antenatal records 10 years after 101 the delivery.

102
The study was conducted in three stages. In the first stage of the study, a self-administered 103 questionnaire to obtain information on history of hyperglycaemia in the index pregnancy, 104 availability of antenatal records and blood sugar assessment reports of the index pregnancy was 105 sent to all mothers of 2005 born children identified through the BI registers in the community and 106 through schools in the selected MOH areas. 107 We defined occurrence of hyperglycaemia in the index pregnancy as a positive answer (yes) to the 108 question 'Did you have high blood sugar / diabetes during the index pregnancy'. Given the high 109 literacy level among women in Sri Lanka, most women were aware of whether they had diabetes 110 during pregnancy.

111
A total of 7205 children who were born in 2005 were identified in stage 1. The prevalence of self-112 reported hyperglyceamia in the index pregnancy was 3.5% (N=257). Eighty eight percent (n=226) 113 of mothers of children exposed to HIP still had antenatal records of index pregnancy compared to 114 69% (n=4811) of mothers of children not exposed to HIP. Potential participants for the main study 115 were identified at the end of the first stage. All children whose mothers had antenatal records and 116 gave a history of HIP during the index pregnancy were considered as "potential participants" to be 117 included in the "exposed group". For each potential participant in the exposed group, two children 118 of mothers with antenatal records and no history of HIP during the index pregnancy were selected 119 from the same PHM area as "potential participants" to be included in the "non-exposed group".

120
During the second stage, the mothers of all potential participants of "exposed" and "non-exposed" 121 groups were invited to participate in the "eligibility assessment sessions". These eligibility 122 assessment sessions were conducted at PHM area level as it was easily accessible to all mothers 123 thus maximizing participation.

124
The research team interviewed the mothers of potential participants and scrutinized their antenatal 125 and medical records to identify participants meeting the inclusion criteria (born in 2005, 126 availability of antenatal records, singleton pregnancy) which were previously decided by a group 127 of experts comprising of specialists in obstetrics, obstetric medicine and public health. Having 128 received antenatal care in a unit lead by a Consultant Obstetrician was one of the eligibility criteria 129 for both "exposed" and "non exposed" groups to counter the possibility of misclassification due 130 to limiting the GDM screening to high risk pregnancies in 2005. 131 170 children exposed to HIP and 291 children not exposed to HIP were identified as eligible and 132 were invited for the study. A sample size of 161 in each group was required to detect a 15% 133 difference in the risk of being overweight with 90% power, an alpha error of 0.05 and a 1:1 ratio 8 134 between children exposed and not exposed to hyperglcyaemia in utero (23). In the third stage, 159 135 offspring of women with HIP (OHIP) and 253 offspring of women with no HIP (ONHIP) in the 136 index pregnancy participated in the study. Among the OHIP, 86.8% (n=138) were exposed to   Anthropometric measurements of the participants were obtained early in the morning following 154 standard operating procedures to ascertain outcome status. Weight and height were measured in 155 light clothing and without shoes. Weight was measured to the nearest 0.1 kg using a calibrated 9 156 digital scale (SECA 876). Height was measured to the nearest 0.1 cm using a SECA stadiometer.

157
Waist circumference was measured to the nearest 0.1 cm at the mid-point between the lowest rib 158 and the top of the iliac crest with a non-elastic tape. Triceps skinfold thickness was measured to 159 the nearest 0.2mm using a Harpenden skinfold caliper. Two measurements were taken and the 160 mean was used for analysis. The same instruments were calibrated regularly and used throughout 161 the study. Overweight was defined as a BMI for age > +1 SD (equivalent to BMI 25kg/m 2 at 19 years)(44). High triceps skinfold thickness was defined as TSFT above the 70 th percentile for age and sex.

186
Since there are racial differences in skinfold thickness (50) Table 1. Characteristics of the exposed (OHIP) and non-exposed (ONHIP) groups At the time of the outcome assessment, the age of all participants ranged between 10.3 years to 231 11.6 years with a mean of 10.85years (SD=0.39). Mothers of children exposed to HIP were older 232 and had significantly higher BMI at the booking visit in the first trimester compared to mothers of 233 non-exposed children (p<0.001). Exposed children were heavier at birth and had a shorter 234 gestational age compared to non-exposed children (p<0.001). About half of the children in ONHIP 235 group were firstborns compared to only one third of children in the OHIP group (p=0.002).

236
Sociodemographic characteristics, breast feeding practices, dietary energy intake and physical 237 activity level were not significantly different between the two groups.

248
1 Data were available for 152 exposed and 236 non-exposed children. The mean BMI-for-age z-score of exposed children was significantly higher than that of non-253 exposed children (P<0.001). Exposed children were significantly heavier and had significantly 254 higher median BMI, WC and TSFT than the non-exposed children (p< 0.001).   The prevalence of overweight, abdominal obesity and high TSFT were significantly higher among 265 the offspring of mothers who had HIP. The high prevalence of abdominal obesity (7.1%) and high 266 TSFT (20.8%) even among the children not exposed to HIP is a concern. Children exposed to HIP 267 were 2 times more likely to be overweight and have abdominal obesity and have a TSFT > 70 th 268 percentile than non-exposed children (p < 0.01). Prevalence of obesity was similar in both groups.  Table 4.  Even after adjustment for maternal BMI, birth weight and birth order, exposure to HIP was a 284 significant predictor of overweight, abdominal obesity and high TSFT in the offspring at 10 years 285 of age. Maternal overweight in the first trimester, a proxy for pre-pregnancy overweight, is an 286 independent risk factor for offspring overweight and high TSFT at 10-11 years. Similarly, being 287 the first-born child carries a more than two-fold increased risk of overweight and high TSFT 288 independent of maternal BMI, birth weight and exposure to HIP. (BMI-z-score > +1SD) was significantly higher among OHIP compared to ONHIP (30.8% vs 302 16.2%). Our results are similar to findings of other studies that have reported a higher risk of 303 overweight and obesity among offspring of mothers who had HIP (11,14,17,18,(21)(22)(23)(24)(25)(57)(58)(59)(60).

304
However, in contrast to other studies, the prevalence of obesity (BMI-z-score > +2SD) was similar 305 in the exposed and non-exposed groups in our study.

306
In our study, children exposed to intrauterine hyperglycaemia had a significantly higher waist 307 circumference at 10 years compared to non-exposed children. Previous studies have reported 308 similar findings of significantly higher waist circumference among offspring exposed to 309 hyperglycaemia in utero including a multinational study involving 206 offspring of GDM mothers 310 and 4534 offspring of non-GDM mothers from 12 countries (24,61,62).

311
In our study, children exposed to HIP had significantly higher TSFT than children not exposed to 312 HIP (13.3mm vs 9.9 mm; p< 0.001). Wright et al, observed that children exposed to GDM had 313 significantly higher sum of skinfold thicknesses (Subscapular and Triceps) than non-exposed 314 children (63). Cumme et al, reported increased subscapular to triceps skinfold thickness ratio in 315 children exposed to HIP (62). Krishnaveni et al. from India, observed significantly higher TSFT 316 among the offspring of diabetic mothers compared to offspring of non-diabetic mothers at 5 years 317 of age (26). When the same cohort was assessed at 9.5 years of age, they observed a significantly 318 higher BMI and TSFT among girls exposed to intrauterine hyperglycaemia but not among boys 319 (14). No significant difference between the growth of the boys and girls was observed in our study 320 (results not shown).

321
In contrast to the many studies where the association between maternal HIP and child overweight 322 attenuated towards the null after adjusting for maternal BMI (11,56,57)(62), our results were 323 statistically significant even after adjusting for maternal BMI, child's birth weight and birth order. 324 We included offspring of women with any type of HIP (gestational diabetes, pre-existing diabetes 325 or overt diabetes first detected in pregnancy) in the "exposed" group without stratification by type 326 of diabetes based on previous research which showed that long-term consequences of HIP on 327 offspring overweight are independent of mother's diabetes type (25,64,65). A sub-group analysis Having a large number of offspring exposed to HIP is a major strength of our study. Selecting both 349 "exposed" and "non-exposed" children from the same source population in the community based 350 on antenatal records reduced recall bias and misclassification. Since exposure was assigned on an 20 351 earlier date than the outcome was measured in the child, it is unlikely that the outcomes of interest 352 would have influenced the classification of exposure status. Children whose mothers received 353 antenatal care from a consultant obstetrician were selected in both exposed and non-exposed pregnancy. But this data was not available for the majority of the participants. We adjusted for the 375 birth weight of the child which can be taken as a proxy measure for weight gain in pregnancy.

376
Based on the national guidelines on antenatal care in Sri Lanka, birth weight ≥ 3.5kg was taken as 377 macrosomia (41).

378
The results of this study have several important public health implications. Locally generated 379 evidence in this study would be an eye opener for clinicians, field health care workers and health 380 policy makers to take necessary actions to follow up exposed children closely during the critical Conclusions 391 Children exposed to intrauterine hyperglycaemia have higher BMI, waist circumference and TSFT 392 at 10-11 years compared to children who were not exposed independent of maternal pre-pregnancy 393 overweight, birth weight and birth order. It is imperative to implement long term follow up for 394 children exposed to hyperglycaemia in pregnancy with anthropometric assessment and life style 395 modification advice to reduce the risk of developing overweight and associated metabolic and 396 cardiovascular disturbances. waist for height percentiles in urban South Indian children aged 3-16 years. Indian Pediatr.