Association between maternal diabetes and neurodevelopmental outcomes in children: a systematic review and meta-analysis of 202 observational studies comprising 56·1 million pregnancies.
Level 3 - non-randomized controlled study
Systematic review and meta-analysis of observational cohort and case-control studies
PubMed 40209722 · doi:10.1016/S2213-8587(25)00036-1
What was done
A systematic review and meta-analysis (PROSPERO CRD42023395464) searched PubMed, Web of Science, Embase, and EBSCO through Dec 1, 2024, for observational studies evaluating the association between maternal diabetes (gestational and pre-gestational) and offspring neurodevelopmental outcomes classified by DSM-5 and ICD-11. Risk ratios (RRs) or standardized mean differences with 95% CIs were synthesized using random-effects models.
What was found
Across 202 studies encompassing 56,082,462 mother-child pairs (110 on gestational diabetes, 80 on pre-gestational diabetes), maternal diabetes was associated with increased risks of neurodevelopmental disorders and lower intelligence and psychomotor scores (numerical score changes not reported in abstract). In the 98 studies (49%) adjusting for multiple confounders, children exposed to maternal diabetes had increased risks of: - Any neurodevelopmental disorder: RR 1.28 (95% CI 1.24–1.31) - ADHD: RR 1.30 (95% CI 1.24–1.37) - Autism spectrum disorder: RR 1.25 (95% CI 1.20–1.31) - Intellectual disability: RR 1.32 (95% CI 1.18–1.47) - Specific developmental disorders: RR 1.27 (95% CI 1.17–1.37) - Communication disorder: RR 1.20 (95% CI 1.11–1.28) - Motor disorder: RR 1.17 (95% CI 1.10–1.26) - Learning disorder: RR 1.16 (95% CI 1.06–1.26) Pre-gestational diabetes showed a significantly stronger association than gestational diabetes (RR 1.39 [95% CI 1.34–1.44] vs 1.18 [95% CI 1.14–1.23]; subgroup difference p < 0.0001).
Why it matters
Synthesizing 56 million pregnancies, this study confirms a consistent 16% to 39% increased relative risk across multiple childhood neurodevelopmental disorders following maternal diabetes, with risks highest for pre-existing diabetes.
Limits
All included studies were observational, preventing determination of causality. Only 49% of studies adjusted for multiple confounders, leaving residual confounding from maternal obesity, genetics, or glycemic control unaddressed; point estimates for cognitive/psychomotor test scores were not provided in the abstract.