Universal Thermal Climate Index and Gestational Diabetes Mellitus Risk: Effect Modification by Green Space in a Chinese Prospective Cohort

BACKGROUND: Thermal stress has been associated with gestational diabetes mellitus (GDM), but ambient temperature alone inadequately captures integrated thermal stress. The Universal Thermal Climate Index (UTCI) is comprehensive, while residential greenness's role remains unclear. This study aimed to investigate the nonlinear and lagged associations between maternal thermal stress, assessed using the UTCI, and GDM risk, and to examine whether residential greenness modified these associations. METHODS: This hospital-based prospective birth cohort included 2,984 pregnant women recruited in Fuzhou, China, between January 2020 and October 2024. GDM was defined by a 75-g oral glucose tolerance test (OGTT) or a documented clinical diagnosis during pregnancy. Maternal thermal exposure during the 12-week preconception period and the first 24 weeks of pregnancy was assessed using UTCI based on spatiotemporally matched reanalysis data. Residential greenness was characterized using the annual mean Enhanced Vegetation Index (EVI). Distributed lag nonlinear models were applied to examine the nonlinear and lagged associations between UTCI and GDM and to evaluate whether residential greenness modified these associations. RESULTS: Compared with the median UTCI, lower UTCI during the 12-week preconception period was associated with a lower GDM risk (RR = 0.67, 95% CI: 0.53-0.84), whereas it was associated with an increased risk during the first trimester (RR = 1.15, 95% CI: 1.02-1.28). Higher UTCI showed delayed associations with GDM, with the strongest risk observed during the second trimester (RR = 1.17, 95% CI: 1.01-1.36). Women with junior college or undergraduate education were more susceptible to thermal stress. Residential greenness modified the UTCI-GDM association during the preconception period, with lower risks observed under combinations of lower UTCI with moderate-to-low EVI and higher UTCI with extremely high EVI. CONCLUSION: The association between thermal stress and GDM varied across exposure windows. Preconception cold exposure was associated with a lower GDM risk, whereas early-pregnancy cold exposure was associated with a higher risk. The second trimester appeared to be a critical susceptibility window for heat-related GDM risk. Residential greenness modified the UTCI-GDM association during the preconception period. These findings support time-specific thermal exposure interventions and urban greening strategies to reduce maternal thermal vulnerability and improve GDM prevention.