Residential greenness and incident type 2 diabetes: A large prospective cohort study in China integrating evidence from 17 global cohorts.

The association of residential greenness on type 2 diabetes (T2D) remain unconfirmed. We aimed to validate this using a large Chinese prospective cohort and pooled global cohorts. The Prospective Environment And Health Epidemiological Study in Northwestern China (PEAHEN-China) included 5,869,738 adults in northwest China followed from 2017 to 2021. T2D was diagnosed through fasting blood glucose (FBG) levels, glycated hemoglobin (HbA1c) levels, and self-report. Residential greenness exposure was estimated using the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), and percentage of green spaces (GS%). Cox regression models and generalized estimating equation models were used to assess the associations of residential greenness with incidence of T2D as well as levels of FBG and HbA1c. Subgroup, mediation, sensitivity analyses, systematic review, and meta-analysis were also conducted. During a median 3.6-year follow-up, 448,948 new T2D cases occurred. In adjusted model, highest vs. lowest quartile of NDVI, EVI, and GS% showed 6.1% (95% CI: 5.2-7.0%), 3.0% (95% CI: 2.0-4.0%), and 2.2% (95% CI: 1.2-3.2%) lower risk of T2D. Stronger associations were observed in younger, women, urban, and educated participants. In exploratory mediation analyses, PM2.5, air temperature, and physical activity might explain part of the greenness-T2D association. Similar associations were found in the exploratory FBG and HbA1c analysis. Systematic review of 17 global cohorts found 13 reporting lower T2D incidence with greater greenness levels, supported by meta-analysis. Our study combined with global cohorts suggest that people living in greener areas have lower T2D risk, offering insights for T2D primary care.