Abstract
Gestational diabetes mellitus (GDM) is a prevalent pregnancy complication with significant short- and long-term consequences for mothers and offspring. While environmental factors, such as obesity and diet, contribute to the risk, genetic predisposition also plays a role in the pathogenesis of GDM. Genome-wide association studies have identified multiple susceptibility loci, including MTNR1B, CDKAL1, and IRS1, which represent mechanistically distinct pathways affecting β-cell function, insulin secretion, and peripheral insulin signaling. This review provides a unified mechanistic framework explaining why these three genes, despite individually modest effect sizes, offer complementary insights into GDM pathophysiology that extend beyond other established loci such as TCF7L2. We critically evaluate the current evidence for genetic risk scores in GDM prediction, acknowledging that their incremental predictive value beyond traditional clinical factors remains modest AUC improvement typically <0.05). The integration of genetic variants with epigenetic modifications is discussed, with careful attention to distinguishing causal mechanisms from correlative findings. We emphasize significant limitations in current research, including population stratification, winner's curse effects, and the predominance of East Asian cohorts. While genetic insights may eventually inform risk stratification, substantial barriers remain before clinical implementation, including insufficient predictive accuracy, lack of cost-effectiveness data, and limited generalizability across diverse populations. Future directions include integrating multi-omics data, developing ethnically validated polygenic risk scores, and conducting pragmatic randomized controlled trials to establish the clinical utility of precision prevention strategies.