Deciphering the pre-pregnancy BMI-GDM link: the mediating and interacting role of early-pregnancy fatty liver index

Gestational diabetes mellitus (GDM) is a common pregnancy complication with complex metabolic etiology, increasing risks for adverse perinatal outcomes and long-term maternal and offspring health. Pre-pregnancy body mass index (pre_BMI) and hepatic fat accumulation have been implicated in GDM, but their nonlinear associations, threshold effects, and mediation roles remain unclear in Asian populations. In this secondary analysis of a prospective cohort, 586 Korean singleton pregnancies were assessed. Pre_BMI was self-reported by participants before pregnancy, and fatty liver index (FLI) was calculated using blood and anthropometric data collected at 10–14 weeks gestation. GDM was diagnosed via a two-step oral glucose tolerance test at 24–28 weeks. Generalized additive models, segmented regression, mediation, and additive interaction analyses were applied, adjusting for maternal age and parity. Subgroup, sensitivity, and propensity score-matched analyses evaluated robustness. Pre_BMI was positively associated with GDM risk. FLI_log exhibited a nonlinear dose–response relationship with GDM risk, with a threshold identified at 1.12. Above this threshold, GDM risk increased substantially (OR = 165.454). Mediation analysis revealed that FLI_log accounted for 45.2% of the total association between pre_BMI and GDM risk, suggesting that hepatic fat accumulation may represent an important intermediate pathway linking pre-pregnancy obesity to GDM development. Additive interaction analysis further demonstrated a significant synergistic interaction between high pre_BMI and elevated FLI_log (RERI = 7.461; AP = 0.785), indicating an excess risk of GDM when both factors coexisted. These findings remained consistent across subgroup and sensitivity analyses. Early-pregnancy hepatic fat accumulation may partially mediate the association between pre-pregnancy obesity and GDM risk and may further strengthen their combined effect. Identification of the FLI_log threshold provides a potential quantitative indicator for early GDM risk stratification and targeted intervention, offering new insights into the mechanisms underlying obesity-related GDM and supporting precision metabolic management among high-risk Asian pregnant women.