Air pollution exposure and incidence of type 2 diabetes in women: A prospective analysis from the Mexican Teachers' Cohort.

Abstract Background Air pollution is a risk factor for type 2 diabetes (T2D). However, scarse longitudinal studies have evaluated this association in low- and middle-income countries, where 80% of the worldwide cases of T2D occur. Objective Our aim was to estimate the association between PM2.5 and NO2 exposure and incident T2D, in the Mexican Teachers' Cohort (MTC). Methods We selected a subsample of female teachers from the MTC from Mexico City metropolitan area (MCMA), recruited in 2008 and with active follow-up every three years. We assigned the monthly time-weighted exposures (PM2.5 and NO2) using home and work addresses, until failure, censoring or death. We developed two high resolution (1 × 1-km) spatiotemporal predictive generalized additive models of PM2.5 and NO2. Incident diabetes was identified through self-report and two administrative databases of registered diabetes patients. We fitted time-varying Cox models to estimate hazard ratios of the relation between PM2.5 and NO2 and incident T2D, adjusting for confounding variables that were identified using a causal model. Results A total of 13,669 teachers were followed-up for a maximum of 11.5 years, over which 996 incident T2D cases (88 cases per 100,000 person-months) occurred. Incident T2D increased by 72% (HR = 1.72 [1.47–2.01]) for each 10 μg/m3 increase of PM2.5, and 52% for each 10 ppb of NO2 (HR = 1.52 [1.37–1.68]). Discussion Mid-term exposure to PM2.5 and NO2 was associated with a higher risk of T2D after adjusting for indoor wood smoke, socioeconomic status, and physical activity. These associations were attenuated in two-pollutant models but remained positive when evaluated long-term exposure. This is the first prospective study to evaluate T2D risk by exposure to both pollutants, PM2.5 and NO2 in a population from an upper middle-income country in the Americas.