Type 2 diabetes (T2D) develops through the long-term interplay of genetic predisposition, lifestyle, and environmental exposures. Residential greenness is increasingly studied as a measurable environmental exposure that may be related to diabetes risk. However, prospective evidence from Chinese populations remains limited, and whether genetic susceptibility modifies the greenness–T2D association, particularly in terms of absolute risk reduction, has remained unclear.
A prospective cohort study published in Life Metabolism by investigators from the 4C Study Group addresses this question using satellite-derived Normalized Difference Vegetation Index (NDVI). The study included 7,861 adults aged 40 years or older from Jiashan County, Zhejiang Province (Figure 1), all free of diabetes at baseline, and followed them for a median of 3.8 years. During follow-up, 499 incident T2D cases were documented.
Residential greenness was quantified using MODIS satellite-derived NDVI and spatially linked to the coordinates of each participant’s home address. The primary exposure was the mean NDVI within a 500-meter residential buffer, updated annually from 2011 to 2016 to reflect temporal changes in greenness exposure. Incident T2D was defined using standardized criteria, including fasting plasma glucose ≥ 7.0 mmol/L, 2-hour oral glucose tolerance test glucose ≥ 11.1 mmol/L, glycated hemoglobin (HbA1c) ≥ 6.5%, or self-reported physician-diagnosed diabetes, rather than relying solely on self-report.
After comprehensive adjustment for age, sex, education, smoking, alcohol drinking, physical activity, diet quality, body mass index, blood pressure, blood lipids, and time-varying PM2.5 exposure, each 0.1-unit increase in NDVI was associated with a 44% lower risk of incident T2D (hazard ratio [HR] = 0.56, 95% confidence interval [CI]: 0.48–0.66). Compared with participants in the lowest greenness tertile, those in the highest tertile had a 51% lower risk of incident T2D (HR = 0.49, 95% CI 0.38–0.63). The inverse association was broadly consistent in sensitivity analyses using 250-meter and 1250-meter buffers.
Higher residential greenness was also associated with more favorable glycemic traits. Restricted cubic spline analyses showed that the HR for incident T2D decreased monotonically with increasing NDVI. Higher NDVI was associated with lower homeostasis model assessment of insulin resistance (HOMA-IR), suggesting lower insulin resistance. The association with homeostasis model assessment of β-cell function (HOMA-B) was non-linear: HOMA-B increased as NDVI rose to approximately 0.35 and then plateaued, suggesting a threshold-type association with β-cell function markers.
The researchers then evaluated T2D-specific genetic susceptibility among 5,389 participants with DNA data. A weighted genetic risk score (GRS) was constructed from 89 genome-wide significant single-nucleotide polymorphisms identified in East Asian populations and weighted by published effect estimates. In the medium and high genetic risk groups, higher greenness was associated with lower incident T2D risk. Participants in the highest greenness tertile had 57% and 61% lower T2D risk, respectively, compared with those in the lowest greenness tertile within the same genetic risk stratum. By contrast, no significant trend was observed among participants with low genetic risk.
Interaction analyses further supported this pattern. The multiplicative interaction between residential greenness and GRS was statistically significant (P for interaction = 0.019), indicating that genetic susceptibility modified the relative association between greenness and T2D risk. Additive interaction was also significant (relative excess risk due to interaction [RERI] = 0.19, 95%: CI 0.05–0.32; P = 0.006), indicating that higher greenness corresponded to a larger absolute risk reduction among adults with medium or high genetic risk. This additive-scale finding is particularly relevant to public health because absolute risk differences are directly informative for risk stratification and prevention planning.
The study was observational and therefore cannot prove that increasing green space would, by itself, prevent T2D. Nevertheless, the results are consistent with several plausible pathways. Adjustment for physical activity and body mass index had little influence on the inverse association, whereas additional adjustment for time-varying PM2.5 strengthened it, suggesting that air pollution exposure may shape the observed greenness–T2D relationship. Associations with HOMA-IR and HOMA-B further suggest that insulin sensitivity and β-cell function may represent important metabolic phenotypes linking residential environment to diabetes risk. Other pathways, including stress reduction, sleep quality, and noise exposure, were not directly measured.
By integrating satellite-based environmental exposure assessment, standardized glycemic outcome ascertainment, and T2D-specific genetic risk scoring, this study provides prospective evidence from China that greener residential environments are associated with lower T2D risk, especially among individuals with higher genetic susceptibility. The findings support considering residential greenness as part of population-level diabetes prevention strategies, while underscoring the need for longer follow-up, more diverse populations, refined measures of green-space quality and accessibility, and mechanistic studies to further validate these observations.
DOI
10.1093/lifemeta/loag019