# Gestational diabetes mellitus and the risk of autism spectrum disorder in offspring: a population-based retrospective cohort study

## TL;DR

This study found no significant link between gestational diabetes and autism in children after adjusting for other factors.

## Contribution

The study is novel in using a large cohort and adjusting for confounders to assess gestational diabetes and autism risk.

## Key findings

- Unadjusted analysis showed higher ASD rates in children of mothers with gestational diabetes.
- After adjusting for confounders, gestational diabetes was not significantly linked to autism risk.
- Kaplan-Meier analysis showed differences in cumulative ASD incidence across gestational diabetes subtypes.

## Abstract

While several environmental and perinatal factors have been associated with the development of autism spectrum disorder (ASD), there is still much to uncover. In this study, we investigated the possible association between gestational diabetes mellitus (GDM), a condition that is becoming more widespread worldwide, and the risk of ASD.

A population-based retrospective cohort study was conducted using data from a tertiary referral hospital and affiliated community clinics. ASD diagnoses were identified through centralized outpatient and hospital records and were established by qualified specialists in accordance with DSM-5 criteria during long-term childhood follow-up. The incidence of ASD in offspring was compared between pregnancies complicated by GDM, categorized as A1 (diet-controlled) or A2 (requiring pharmacologic treatment), and pregnancies without GDM. Cumulative incidence of ASD was estimated using Kaplan-Meier survival analysis, and a Cox proportional hazards model was applied to adjust for potential confounders.

Among 115,063 deliveries included in the study, 3,461 (3.0%) were complicated by GDM A1 and 1,164 (1.0%) by GDM A2. Overall, 767 offspring were diagnosed with ASD during childhood. Univariate analysis demonstrated a statistically significant association between GDM severity and ASD incidence (1.5% for GDM A2, 1.0% for GDM A1, and 0.6% for no GDM; p<0.001). Kaplan-Meier analysis demonstrated a significant difference in cumulative ASD incidence across GDM subtypes (log-rank p<0.001). However, after adjustment for confounders in a multivariable Cox model, neither GDM A1 nor GDM A2 was statistically significantly associated with ASD risk (aHR 1.18, 95% CI 0.83–1.66; aHR 1.56, 95% CI 0.95–2.56, respectively).

Our findings suggest no statistically significant association between GDM and ASD in offspring after adjustment.