Long-term exposure to air pollutants (PM2.5and NO2) and incidence of type 2 diabetes in the Mexican Teachers’ Cohort.

Abstract

Background/Aim: Air pollution is a risk factor for type 2 diabetes (T2D). However, no longitudinal studies have evaluated this association in a low- and middle-income countries, where 80% of the worldwide cases of T2D occur. Our aim was to evaluate the association between incident T2D and long-term exposure to PM2.5 and NO2 in the Mexican Teachers’ Cohort (MTC). Methods: We selected a subsample of female teachers from the MTC who lived and worked in the Mexico City Metropolitan Area, recruited in 2008 and with active follow-up every 3 years (n = 14,194). We assigned the time-weighted exposures (PM2.5 and NO2) in home and work addresses, from 12 months before baseline to failure or censoring. We developed two high resolution (1km x 1km) spatio-temporal predictive Generalized Additive Models of PM2.5 and NO2. We defined incident T2D when teachers self-reported at least two of the following questions: have a medical diagnosis, use of medical treatment, or date of diagnosis. 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: age, passive and active smoking status, physical activity, SES, wood smoking and time spent outdoors. Results: Teachers were followed up for 11.5 years. There were 795 incident T2D cases (84 cases per 100,000 person-months). Incident T2D increased by 80% (HR = 1.8[1.6-2.1]) for each 10 μg/m3 increase of PM2.5 and 60% for every 10 ppb of NO2 (HR = 1.6[1.4-1.9]). In two-pollutant models, effects were attenuated. Conclusions: PM2.5 and NO2 increased the risk of T2D in Mexico, in a larger magnitude than previous published studies. Despite improvements in air quality, there are still health risks related to current pollutant levels, arguing for even stricter regulation.