Introduction
Offspring of gestational diabetic mothers (OGDM) are at increased risk of macrosomia [
1] and higher newborn adiposity [
2]. These physical traits are associated with obstetric and neonatal complications, including prematurity, shoulder dystocia, hypoglycaemia and jaundice [
3]. Subsequently, OGDM demonstrate a relatively slow weight gain or ‘catch-down’ growth until 2 years of age [
4,
5]. Weight gain is then accelerated after age 5 years [
6] and linked to long-term increased metabolic risks, including obesity and type 2 diabetes [
7‐
9]. However, in recent years, some studies suggest that birth size of OGDM may be normalising [
10,
11].
This could be attributed to the changing diagnostic criteria for gestational diabetes mellitus (GDM) and subsequent more intensive management of gestational hyperglycaemia. In 2010, the International Association of Diabetes and Pregnancy Study Groups (IADPSG) suggested more stringent GDM diagnostic criteria than those used previously [
12]. This followed results from the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study, involving 23,000 non-GDM pregnant women across nine countries. The study found that maternal blood glucose was positively correlated with increased birthweight, cord blood C-peptide level, neonatal hypoglycaemia and Caesarean section delivery rates [
13,
14]. Even before the implementation of these guidelines, several recent randomised controlled trials had shown decreased perinatal complications and more normal birthweights after intensive treatment of glucose intolerance in pregnancy (dietary advice, glucose monitoring, and insulin as needed), compared with routine pregnancy care [
15‐
17]. The introduction of the IADPSG guidelines may lead to more mothers being treated and perhaps more intensively. The impact of these temporal changes in management on early infancy growth has not been documented. As well as large birthweight, low birthweight and rapid postnatal ‘catch-up’ growth may also have implications for future health [
18].
We hypothesised that recent changes in the detection and management of GDM could have an impact on classical growth differences long-observed between OGDM and other infants. Within the context of an established prospective birth cohort, we retrospectively applied the 2010 IADPSG criteria to two groups of OGDM born in non-overlapping years and compared birth size and early infancy anthropometry in each group, against a control group unaffected by GDM.
Methods
Statistical analyses
Infancy age- and sex-appropriate SD scores (SDS) were calculated for weight and length measurements (with adjustment for gestational age at birth and 3 months), by comparison with the UK 1990 growth reference using LMS growth software [
21]. For each of the four skinfold thickness measurements an internal SDS was calculated, using residuals from a linear regression model, adjusting for infancy age, (gestational age at birth and 3 months) and sex. Mean skinfold thickness SDS was used in analyses. Maternal BMI was derived from self-reported pre-pregnancy weight divided by the square of measured height (kg/m
2). Birth ponderal index was calculated by dividing the infant’s birthweight by its birth length cubed (kg/m
3). Deprivation was assessed using an integrated index based on residential postcodes [
22].
Maternal and birth characteristics were compared between groups using ANOVA with Bonferroni post hoc analysis for continuous variables, and χ2 tests for categorical outcomes. Unless otherwise stated, all data are presented as means ± SDs.
Multiple linear regression was used to investigate the effect of GDM on birth outcomes, allowing adjustment for potential confounders, including infant sex, postnatal age, gestational age, pre-pregnancy maternal BMI, maternal height, parity, breastfeeding history at 3 months, delivery method, maternal ethnicity, socioeconomic status reflected by Index of Multiple Deprivation (IMD) and pregnancy smoking history. All confounders were chosen a priori through the extensive work of CBGS and the Avon Longitudinal Study of Parents and Children (ALSPAC) [
18].
Under the traditional listwise deletion method, only 68% of the control group and 64% of both recent and earlier OGDM had complete data on all covariates. Covariates with most missing values were maternal pre-pregnancy BMI for control infants and ‘earlier OGDM’, and smoking history for recent OGDM. Data were primarily missing due to incomplete perinatal questionnaire responses. Missing covariates including IMD (
n = 3), parity (
n = 4), maternal ethnicity (
n = 8), smoking history during pregnancy (
n = 39), maternal pre-pregnancy BMI (
n = 185), maternal height (
n = 148), delivery method (
n = 27) and infant feeding history (
n = 189) were imputed under the assumption that they are missing at random. The R package ‘Multiple Imputations via Chained Equations (MICE)’ was used to generate 20 imputed datasets, using normal linear regression for continuous variables and logistic linear regression for binary variables. Analyses run on each dataset were pooled according to Rubin’s rules [
23]. Imputed values compared reasonably to observed values, and the results (i.e. linear regression model on birth data, Table
2) using listwise deletion were similar to imputed values, therefore imputed values were presented in the subsequent analyses.
In the visit measurements, missing data were commonly due to loss-to-follow-up or drop outs. In order to capitalise the longitudinal growth data with good handling of missing values, linear mixed-effects models were used to relate the continuous growth outcome variables (weight, height and skinfold thickness) to visit time point, cohort group, and their interaction with infant age, taking into account the same confounders as in the linear regression models for birth measurements. Due to non-linear relationships with age (indicated by significant estimates for age-squared), time was modelled using linear splines with knots at ages 3 and 12 months. Models were fitted to the data by restricted maximum likelihood (REML).
Statistical analyses were conducted using SPSS (IBM SPSS Statistics for Windows, version 25.0; IBM, Armonk, New York, USA) and R (version 3.3.2; R Foundation for Statistical Computing, Vienna, Austria). p < 0.05 was considered statistically significant.
Discussion
This observational study demonstrates significant differences in both birth size and subsequent infancy growth trajectories between recent and earlier OGDM compared with control infants. Recent OGDM had comparable birthweight and length SDS, but unexpectedly reduced skinfold thickness, indicating lower adiposity, compared with control infants. Conversely, earlier OGDM were heavier at birth, consistent with the traditional description of OGDM [
1,
2].
Some recent studies concur with our findings. An Australian study (2013) [
10] found no significant difference between recent OGDM and control infants’ birthweights; and a UK study (2016) [
11] reported lower weights and lengths for OGDM vs control infants at 2 weeks of life. Recent GDM trials (no treatment vs lifestyle advice +/− insulin) also suggest a shift towards the normal population distribution of birthweights in OGDM [
16,
17].
Reduced subcutaneous fat at birth in our recent OGDM may be a novel finding. Au et al [
10] demonstrated lower body fat percentage in OGDM at birth, (7.9 ± 4.5% vs 9.5 ± 4.3% in the control group), but this was not statistically significant in their relatively small study (
N = 67) [
24]. Conversely, a recent systematic review and meta-analysis published in 2017 concluded that newborn adiposity is still increased in OGDM [
25]. However, although the overall numbers in that analysis were large, individual studies were small, and many combined type 1 diabetes, type 2 diabetes and GDM. Two of the most recent studies showed no body fat percentage difference between OGDM and control infants [
26,
27]. Logan et al [
11] found no difference in total adipose tissue mass on MRI in OGDM compared with control infants at 11 days of age. Therefore, our study and recent literature suggest that GDM diagnosed and treated over the last decade may result in offspring birthweight and length comparable with the general population, and even reduced adiposity.
Our recent OGDM cohort showed significantly increased weight and skinfold thickness gains compared with the control group from birth to 3 months, despite similar breastfeeding rates. A comparable UK cohort (2011–2014) also found greater weight and adiposity gains from birth to 2.5 months in OGDM [
11]. However, in contrast with our findings, their OGDM cohort still had greater total adipose tissue at 2.5 months, adjusting for sex and maternal pre-pregnancy BMI [
11]. Our recent OGDM then showed reduced gains in weight and skinfold thickness from 3 to 12 months, resulting in significantly reduced weight and adiposity at 12 months compared with control infants. Subcutaneous adiposity in this group remained lower than in the control group until 24 months of age. In contrast, earlier macrosomic OGDM showed expected catch-down growth, with slightly reduced weight and skinfold thickness gains from birth to 3 months, and significantly decreased gains in weight and length between 3 and 12 months, compared with control infants. However, they still had higher adiposity than control infants at all time points.
The smaller birth size seen in recent OGDM was evident despite higher maternal BMI and higher OGTT 60 min glucose concentrations compared with earlier OGDM. Both groups were retrospectively defined using the IADPSG criteria. Therefore the normalisation of birth anthropometry seen in recent OGDM is probably due to intensification of GDM monitoring and treatment, rather than inclusion of individuals with ‘milder’ GDM. We hypothesise that this could result from tighter glycaemic control per se, direct effects of medication transported across the placenta, or interactions between these environmental factors, genetic predisposition and epigenetic modulation. Conversely, earlier OGDM showed predicted increased birthweight, due presumably to greater nutrient supply in pregnancy and fetal hyperinsulinism. A weakness of our study is that we do not have uniform data on glucose variability, maternal treatments and HbA1c to further inform this debate.
Amelioration of the classic macrosomic phenotype is likely to be associated with fewer adverse outcomes; however, the long-term effect of early reduced infancy weight and subcutaneous fat could be associated with risk itself, particularly if leading to catch-up growth, which has been associated with risk for childhood obesity and adult type 2 diabetes [
18,
28]. The finding that tight glucose control can not only normalise birthweight but also be associated with reduced body size has been previously reported. Langer et al investigated three GDM groups and showed that the group with lowest maternal glucose values had a higher proportion of small for gestational age (SGA) infants [
29]. The timing of treatment could also play a role as a recent study reported that early GDM treatment was associated with a higher rate of SGA-related neonatal intensive care unit admissions, whereas later treatment resulted in more large for gestational age infants [
30]. The recent HAPO data relating to the follow-up of infants born to mothers with a wide range of glucose values at 28 weeks gestation confirm a positive relationship between those levels and adiposity at 10–14 years (skinfold thickness and air displacement plethysmography) [
31]. As well as linking high antenatal glucose exposures to childhood overweight/obesity, we could also infer from these data that lower glucose exposures might result in persisting reduced adiposity. Therefore, while there are clear advantages of intensive multidisciplinary GDM management, there may also be negative implications for some OGDM.
In addition to more extensive diet and lifestyle advice, medical treatment of GDM has changed significantly over recent years [
11]. Metformin is now commonly used worldwide, often as first-line medication, and crosses the placenta in significant amounts. While we were unable to include medication use in our analyses, 20% of women were treated with metformin (+/− insulin) during recent GDM recruitment, compared with near zero for the earlier GDM (clinic data). We therefore postulate that metformin itself may at least partly explain the differences seen between recent and earlier OGDM anthropometry and growth trajectories. The Metformin in Gestational Diabetes (MiG) trial, randomising women with GDM to metformin (+/− insulin if needed) or insulin, suggested that metformin might affect infancy fat deposition patterns. There were no differences at birth [
32]. However, at 2 years of age, children from the metformin group had increased subscapular and biceps skinfold thickness, despite no difference in overall fat, suggesting a more favourable fat distribution [
32]. Our recent OGDM cohort with greater metformin exposure, compared with the earlier group, shows preferentially increased gains in 3 month subscapular skinfold thickness and then reduced gains until 2 years of age (ESM Table
1).
At 7–9 years, OGDM randomised to metformin in the MiG trial had similar total body fat and metabolic measures, although the 9 year olds were larger [
33]. A study in polycystic ovary syndrome (PCOS) suggested a growth restriction effect of metformin in infants of normal-weight mothers [
34]. It is therefore hard to interpret whether metformin may confer a beneficial fat distribution or an increased long-term risk of obesity. Metformin effects may also differ depending on maternal weight gain, glycaemic control and other environmental factors. Further studies are needed to elucidate the effect of metformin itself, effects on maternal energy intake and weight gain, and interactions with other environmental factors, on adiposity distribution.
To our knowledge, no previous study has investigated growth trajectories of recent OGDM, in comparison with a control group, up to 24 months of age. Strengths of this study include measures of length and skinfold thickness adiposity, in addition to weight, in a large cohort. Collection of detailed maternal and demographic data allowed adjustment for potential confounding factors. Use of the IADPSG diagnostic criteria [
12] means that our results are relevant to populations worldwide, where GDM is now diagnosed using lower thresholds and is more aggressively treated. The ‘recent’ OGDM cohort showed a slightly increased prevalence of ethnic minority groups. However, ethnicity was not a significant covariate in our growth models. Local hospital demographic data (not shown) for all GDM mother–infant dyads born at the same time as the recent group were similar to the study population, and therefore anthropometric findings of recent OGDM at birth were unlikely to result from study participation bias. Furthermore, the anthropometric measures for more recent control participants recruited in Cambridge have not shown any evidence of a secular trend in infancy growth (data not shown).
Limitations of the study include that the two OGDM groups are not fully comparable, and no details are available of glycaemic control after GDM diagnosis, although it is likely that more intensive treatment of recent GDM women led to tighter glycaemic control. Since data regarding weight gain during pregnancy were incomplete, it cannot be confirmed that recent GDM women had adequate pregnancy weight gain. The study population was large compared with most previous studies; however, there were insufficient numbers to investigate the individual effects of metformin and insulin on anthropometric outcomes. A further limitation is that 19% of women in the earlier cohort were identified retrospectively and did not receive GDM treatment. However, excluding these women would still give similar outcomes in the regression models (ESM Table
2). Studies are needed to further understand the mechanisms responsible for anthropometric outcomes in OGDM, and ideal GDM treatment going forward. We believe that the trend has been towards stricter application of ‘targets’ blood testing and the greater use of metformin. A weakness of our study is that we do not have accurate details of glucose concentrations, HbA
1c or metformin use in these populations. Going forward, continuous glucose monitoring data defining individual glucose exposures may clarify these factors. Our work and that of others suggests that it will be informative to further study adiposity distribution in OGDM, including subcutaneous and visceral fat deposits, to investigate beneficial vs undesirable adiposity gains.
It is debatable whether reduced early anthropometric measures in OGDM will have positive or negative implications, particularly for longer-term health. In the neonatal period it may be advantageous, allowing normal birthweights and reduced pregnancy complications. However, this could result in increased numbers of SGA infants, and associated comorbidities. Reduced size at birth, leading to subsequent early catch-up growth, may also lead to later increased metabolic disease risk [
35].
Acknowledgements
The authors acknowledge the CBGS research nurses S. Smith, A-M. Wardell and K. Forbes, Addenbrooke’s Hospital, Cambridge, UK. The authors are also very grateful to the diabetes in pregnancy team, Rosie Maternity Hospital, Cambridge, UK, particularly R. Harding and K. Stubbington, for their help with recruitment. The authors thank all the families who contributed to the study, the staff at the NIHR-Wellcome Trust Clinical Research Facility, Cambridge, the NIHR Cambridge Comprehensive Biomedical Research Centre, and the midwives at the Rosie Maternity Hospital, Cambridge, UK.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.