Objective We explored associations between long-term residential surrounding greenness and incident type 2 diabetes (T2DM), hypertension, and cardiovascular diseases (CVDs)—coronary heart disease (CHD), stroke, and heart failure (HF)—in a large French urban cohort. Methods We included adults aged ≥30 years from the Constances cohort enrolled between 2012 and 2019 and followed through 2021, excluding rural residents. Outcome-specific study populations were defined after removing participants with the condition of interest at baseline. Residential greenness exposure was derived from the Normalized Difference Vegetation Index (NDVI) within circular buffers (300-m, 500-m, and 1,000-m), based on the residential address history throughout follow-up. Incident T2DM and hypertension were ascertained using clinical measurements and medication data; CVDs were identified from hospital discharge records. Cox proportional hazards models estimated associations between 0.1-unit increments in annual mean residential greenness exposure and each outcome. Models were adjusted for sociodemographic factors and smoking status. Results are expressed as hazard ratios (HRs) with 95% confidence intervals (95%CIs). Results Between 2013 and 2021, 744 participants developed T2DM (0.7%; N=109,518) and 5,519 developed hypertension (7.1%; N=78,265). CVD incidence was below 0.5%, including 612 CHD, 549 stroke, and 149 HF events. Higher residential greenness was inversely but not statistically significantly associated with T2DM, with similar estimates across buffers (e.g., 300-m: 0.95, 0.89–1.02). Associations for hypertension and other CVDs were generally close to the null and non-significant (e.g., for hypertension and 300-m: 1.03, 1.00–1.05). Results were similar in different sensitivity analyses: restricting analyses to non-movers, excluding Paris, accounting for competing risks or using multiple imputation to treat missing values. Conclusions Residential surrounding greenness was not consistently associated with incident cardiometabolic diseases in this large urban population, with small effect estimates close to the null. Further research is needed to clarify context-specific effects and the role of sociodemographic factors in these associations

Objective We explored associations between long-term residential surrounding greenness and incident type 2 diabetes (T2DM), hypertension, and cardiovascular diseases (CVDs)—coronary heart disease (CHD), stroke, and heart failure (HF)—in a large French urban cohort. Methods We included adults aged ≥30 years from the Constances cohort enrolled between 2012 and 2019 and followed through 2021, excluding rural residents. Outcome-specific study populations were defined after removing participants with the condition of interest at baseline. Residential greenness exposure was derived from the Normalized Difference Vegetation Index (NDVI) within circular buffers (300-m, 500-m, and 1,000-m), based on the residential address history throughout follow-up. Incident T2DM and hypertension were ascertained using clinical measurements and medication data; CVDs were identified from hospital discharge records. Cox proportional hazards models estimated associations between 0.1-unit increments in annual mean residential greenness exposure and each outcome. Models were adjusted for sociodemographic factors and smoking status. Results are expressed as hazard ratios (HRs) with 95% confidence intervals (95%CIs). Results Between 2013 and 2021, 744 participants developed T2DM (0.7%; N=109,518) and 5,519 developed hypertension (7.1%; N=78,265). CVD incidence was below 0.5%, including 612 CHD, 549 stroke, and 149 HF events. Higher residential greenness was inversely but not statistically significantly associated with T2DM, with similar estimates across buffers (e.g., 300-m: 0.95, 0.89–1.02). Associations for hypertension and other CVDs were generally close to the null and non-significant (e.g., for hypertension and 300-m: 1.03, 1.00–1.05). Results were similar in different sensitivity analyses: restricting analyses to non-movers, excluding Paris, accounting for competing risks or using multiple imputation to treat missing values. Conclusions Residential surrounding greenness was not consistently associated with incident cardiometabolic diseases in this large urban population, with small effect estimates close to the null. Further research is needed to clarify context-specific effects and the role of sociodemographic factors in these associations