Investigating the interaction of diet quality on the association between air pollution exposure and diabetes incidence in the Hispanic Community Health Study/Study Of Latinos

Abstract

BackgroundType 2 Diabetes (T2D) is a chronic disease linked to increased morbidity and mortality. Hispanic and Latino groups have a disproportionately high prevalence of diagnosed T2D cases among ethnic and racial groups in the United States. Dietary patterns that feature the consumption of processed meats and sugar-sweetened beverages have been linked to development of chronic disease, including T2D. Increased exposure to ambient air pollutants have also been linked to the development of T2D. It remains uncertain whether diet may modify the risk of diabetes associated with air pollution. Methods We investigated the association between air pollution, diet, and incident T2D between visit 1 and visit 2 of the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). We utilized a spatio-temporal air pollution model to estimate outdoor residential exposure to particulate matter with an aerodynamic diameter of 2.5 microns (PM2.5) and nitrogen dioxide (NO2) over the follow-up period at each participant’s residential address. Diet was assessed at baseline using the Alternative Healthy Eating Index-2010 (AHEI-2010). Using covariate adjusted Poisson regression models, we investigated the association with air pollution and T2D, diet and T2D, and the potential effect modification of diet on the association between air pollution and T2D, adjusting for age, sex, income, education, family history of diabetes, heritage-study center, and waist-to-hip ratio. Results In our analytic sample of 7,534 study participants without diabetes at baseline and with complete data available, 876 incident cases of diabetes were observed. Primary models of PM2.5 (IRR: 1.21; 95% CI: 1.01, 1.44) and NO2 (IRR: 1.11; 95% CI: 1.02, 1.22) demonstrated an association with incidence of diabetes, which was attenuated with adjustment for adiposity and study site. Primary models between AHEI-2010 and incidence of diabetes demonstrate a suggestive but not statistically significant association between healthy eating and diabetes (IRR: 0.93; 95% CI: 0.83, 1.04). We did not find evidence that diet modified the association between ambient air pollution and incident diabetes. Conclusion Though initial models demonstrate the association we expected, further adjustment attenuated the estimates we observed. Additional research, including different approaches to dietary quality ascertainment, may be necessary to better understand these relationships.