This was a prospective follow-up study of children in the Shanghai Birth Cohort (SBC) (34). The SBC is a prospective cohort involving 4127 pregnancies in six tertiary obstetric care hospitals in Shanghai between 2013 and 2016. The study was approved by the research ethics boards of Shanghai Xinhua Hospital (the coordination center) and all participating hospitals. Written informed consent was obtained from all study participants.

There was a total of 3692 live births (3331 singletons) in the SBC, and 2207 singleton children remained in follow-ups at age 4 years. We excluded children whose mothers had pre-gestational diabetes (n=7) or preeclampsia/eclampsia (n=27), and children with birth defects (n=16), born at gestational age <34 weeks (n=13), or missing information on maternal GDM (n=34), child’s sex or gestational age at delivery (n=55). The final study sample included 2055 children with at least one follow-up length/height measurement between ages 6 weeks and 4 years. The numbers of children with data available on length/height were 1772 (86%) at birth, 1801 (88%) at 6 weeks, 1739 (85%) at 6-months, 1805 (88%) at 12 months, 1874 (91%) at 2 years, and 1951 (95%) at 4 years of age, respectively. There were 1139 children with complete data on length/height measurements at all the six time points. Figure 1 illustrates the selection of study subjects and follow-up length/height measurements from birth to 4 years of age.

GDM was diagnosed by a 75g oral glucose tolerance test at 24-28 weeks of gestation according to International Association of Diabetes and Pregnancy Study Groups (IADPSG)’ criteria (35): if any blood glucose value reaches or exceeds the following thresholds: fasting 5.1 mmol/L, 1- hour 10.0 mmol/L, and 2-hour 8.5 mmol/L. There were 369 mothers with GDM in the study cohort.

Available maternal/pregnancy characteristics included age, ethnicity, parity, education, smoking or alcohol use during pregnancy, hypertensive disorders in pregnancy (chronic or gestational, excluding preeclampsia or eclampsia), height, pre-pregnancy weight, gestational weight gain, mode of delivery, infant’s sex, birth weight and gestational age. Pre-pregnancy body mass index (BMI) was calculated as weight (kg)/(height [m]) ^2. Gestational weight gain was calculated as the difference between the weight measured within one week before delivery and pre-pregnancy weight, and was classified as inadequate, adequate, and excessive if it was below, within, or above the recommendations of the 2009 Institute of Medicine guidelines (36): 12.5-18.0 kg for underweight; 11.5-16.0 kg for normal weight; 7.0-11.5 kg for overweight; and 5.0-9.0 kg for obese women. Women were classified by pre-pregnancy BMI as underweight (<18.5 kg/m²), normal weight (18.5-24.9 kg/m²), overweight (25.0-29.9 kg/m²) or obese (30.0+ kg/m²). Gestational age was determined by the date of last menstruation period and confirmed by first-trimester ultrasound dating. If the difference between the 2 estimates was more than 2 weeks, the ultrasound dating-based estimate was used.

Length/height was measured at birth and postnatal follow-ups at ages 6 weeks (42 days), 6 months, 1, 2 and 4 years by trained research staff. Length (birth to age 2 years) and standing height (age 4 years) were measured following standardized operating protocols. Length (birth to 2 years of age) were measured in supine position using a Seca 416 Infantometer (Seca Netherlands, Hamburg). At age 4 years, standing height was measured by a wall-mounted stadiometer (Seca Netherlands, Hamburg). Length and height measurements were measured twice to the closest 0.1 cm, and the averages were taken as the final values.

Length/height-for-age Z scores were calculated according to the World Health Organization (WHO) growth references (37) using the ‘anthro’ and ‘anthroplus’ package in R.

Data are presented as mean ± SD for continuous variables or frequency (percentage) for categorical variables. The independent t-test and Pearson’s chi-square test were used to compare differences in continuous and categorical variables between GDM and euglycemic pregnancy groups, respectively. Generalized linear models were employed to evaluate the association between GDM and LAZ/HAZ at each follow-up time point. Generalized estimating equation (GEE) models were used to assess the association between GDM and linear growth trajectory from birth to 4 years of age overall accounting for the correlations in length/height measurements over ages within the same subjects. In the adjusted models, we included pre-specified known factors affecting length/height (maternal height, infant’s sex, gestational age at birth). Other co-variables, including maternal age, ethnicity, parity, smoking in pregnancy, alcohol drinking in pregnancy, pre-pregnancy BMI, gestational weight gain, hypertensive disorders in pregnancy, mode of delivery and breastfeeding were subject to a stepwise regression selection process. Only co-variables with P<=0.20 would be retained in the final parsimonious models.

Since pre-pregnancy BMI may be considered an upstream risk factor of GDM, gestational weight gain may be influenced by maternal GDM, and GDM may affect postnatal linear growth through affecting gestational age at birth, we repeated the stepwise regression models in sensitivity analyses excluding gestational age, or gestational weight gain, or both pre-pregnancy BMI and gestational weight gain. To disentangle the potential confounding effects of maternal hypertension from GDM in the associations with offspring’s linear growth, we excluded those exposed to maternal hypertensive disorders in a sensitivity analysis. Considering the potential impact of paternal height, we also conducted a sensitivity analysis to include paternal height (with a high frequency of missing values) in the adjusted models in a sensitivity analysis.

Effect sizes in regression coefficients (β) with 95% confidence intervals (CIs) were presented. Missing values in continuous co-variables were not imputed. Missing values in categorical co-variables were taken as a valid category in regression models. P<0.05 was considered statistically significant in assessing the primary association of interest between GDM and linear growth from birth to age 4 years in a GEE model. P values in other comparisons were for exploratory information only, and multiple comparisons were not accounted for. All statistical analyses were conducted using R version 4.2.3 and STATA V.15.

Table 1 presents the characteristics of subjects in the final study cohort. The average age of mothers at delivery was 29 years, and all mothers were over 20 years of age. There were 369 mothers (18%) with GDM. Compared to mothers with a euglycemic pregnancy, mothers with GDM were older (mean: 30.2 vs. 28.7 years), less likely to have completed college or higher education, had higher pre-pregnancy BMI (mean: 24.1 vs. 22.2 kg/m2), but were less likely to have excessive weight gain during pregnancy (12.7% vs. 24.7%). GDM mothers were more likely to be affected by hypertensive disorders in pregnancy (7.9% vs. 3.3%). There were no significant differences in maternal height, smoking or alcohol consumption during pregnancy between the two groups. Among mothers with GDM and information available on treatments (n=243), 24 (9.9%) mothers required the use of insulin in the management of hyperglycemia, the rest took dietary and life style interventions only. The infants of mothers with GDM had a slightly shorter average gestational age at delivery (mean: 38.8 vs. 39.0 weeks), and were more likely to be delivered via cesarean section (59.5% vs. 50.5%), while birth weights were comparable. A slightly higher proportion of infants of GDM mothers had breastfeeding less than 6 months (29.6% vs. 23.7%).

Table 2 presents the data on length/height in centimeter and for age z scores (LAZ/HAZ) comparing the offspring of GDM vs. euglycemic mothers. Average birth length was similar, but average LAZ/HAZ scores were consistently lower in the offspring of GDM mothers in all postnatal follow-up age points (6 weeks, 6 months, 1, 2 and 4 years), and statistically significantly so at 6 weeks, 6 months and 2 years. On average, crude LAZ/HAZ values were 0.08 to 0.22 lower in the offspring of GDM vs. euglycemic mothers at 6 weeks to 4 years of age. The mean differences in Z scores between the two groups showed a narrowing trend over increasing ages (Z score differences: 6 weeks -0.22, 6 months -0.20, 1 year -0.11, 2 years -0.17, 4 years -0.08).

Adjusting for maternal and child characteristics, the offspring of mothers with GDM had consistently lower LAZ/HAZ than the offspring of euglycemic mothers, with mean differences (regression coefficients) in the range of -0.11 to -0.18 between ages 6 weeks and 4 years (adjusted P<0.05 at all the five age points: 6 weeks, 6 months, 1, 2 and 4 years) ( Table 3 ). Accounting for within-subject correlations in length//height, the children born to mothers with GDM exhibited a significantly reduced linear growth from birth to age 4 years overall, with a mean linear growth deficit of 0.12 (95% CI: 0.04, 0.21) in LAZ/HAZ after adjusting for maternal height, pre-pregnancy BMI, gestational weight gain, alcohol drinking status in pregnancy, gestational age at delivery and child’s sex (other maternal and child factors did not affect the comparisons and were excluded at P>0.2).

Sensitivity analyses excluding gestational age, gestational weight gain, or pre-pregnancy BMI and gestational weight gain as covariates, or excluding children born to mothers with hypertensive disorders in pregnancy in the analyses showed similar results as in the main analysis ( Supplementary Tables S2 , S3 ). With further adjustment for paternal height, the associations showed a similar pattern, but could not detect a statistically significant effect of GDM on linear growth due to the relatively small sample size (data available on paternal height: n=36 for children of GDM mothers) ( Supplementary Table S4 ).

Among children of mothers with GDM, there were no significant differences comparing children of mothers who received insulin treatment versus those who received dietary and lifestyle interventions only in LAZ/HAZ scores at birth (mean ± SD: -0.33 ± 0.93 vs. -0.06 ± 0.80, P = 0.20) or age 4 years (0.56 ± 1.02 vs. 0.44 ± 0.93, P = 0.60). There was no significant difference in linear growth from birth to age 4 years overall in children of GDM mothers who received insulin treatment or not ( Supplementary Table S5 ).