Associations of green space and air pollution with insulin resistance: Analysis of repeated-measurements health examination data.

Environmental green space and air pollution frequently coexist in urban settings and may have opposing associations with metabolic health. However, longitudinal evidence regarding their associations with insulin resistance (IR) and potential inflammatory pathways remains limited. We conducted a longitudinal study of 17,731 adults undergoing routine health examinations in Shijiazhuang, China, from 2021 to 2023, yielding 41,525 person-year observations. Green space exposure was assessed using the normalized difference vegetation index (NDVI) within a 500-m buffer around workplaces, and annual mean concentrations of PM2.5, PM10, NO2, SO2, and CO were obtained from the China High Air Pollutants dataset. IR was primarily assessed using the metabolic score for insulin resistance (METS-IR), with triglyceride-glucose index (TyG), the triglyceride-glucose index combined with body mass index (TyG-BMI), and the ratio of triglycerides to high-density lipoprotein cholesterol (TG/HDL-C) included as supplementary indicators. Linear mixed-effects models were used to estimate associations and interaction effects, and exploratory mediation analyses evaluated the potential role of platelet-to-lymphocyte ratio (PLR). Higher NDVI was associated with lower METS-IR, whereas associations with TyG, TyG-BMI, and TG/HDL-C were not statistically significant. Each standard deviation increase in NDVI was associated with a decrease in METS-IR (β = -0.030; 95% CI: -0.039 to -0.021). Higher air pollutant concentrations generally showed positive associations with IR indices, with stronger and more frequently statistically significant associations observed for METS-IR. Interaction analyses suggested heterogeneous associations across levels of green space exposure. Exploratory mediation analyses indicated that PLR statistically accounted for a small proportion (approximately 1-10%) of associations involving several pollutants and IR indices. These findings suggest that workplace-based environmental exposures may be associated with insulin resistance, although patterns varied across surrogate indices. Further studies using refined exposure assessment and longitudinal mechanistic measurements are needed to clarify these associations.