Background: Active smoking has been linked to type 2 diabetes mellitus (T2DM) but only few recent studies have shown environmental tobacco smoke (ETS) to be associated with DM in never-smokers. We assessed the association between long term ETS exposure and DM, and explored effect modifications of this association in our sample.
Methods: We analysed 6392 participants of the Swiss study on air pollution and lung and heart diseases in adults (SAPALDIA). We used mixed logistic regression models to assess the cross-sectional association between ETS and DM. Selected variables were tested for effect modification and several sensitivity analyses were performed, mostly treating participants' study area as a random effect.
Results: The prevalence of DM and ETS in the sample was 5.5% and 47% respectively. There were 2779 never-smokers with 4% diabetes prevalence. Exposure to ETS increased risk of DM in never-smokers by 50% [95% confidence interval (CI): 1.00, 2.26], and we observed a positive dose-response relationship between ETS exposure level and DM in never-smokers. Associations were strengthened (more than three-folds) by older age and chronic obstructive pulmonary disease, and were stronger in post-menopausal, obese, hypertriglyceridaemic and physically inactive participants. Estimates of association were robust across all sensitivity analyses (including inverse probability weighting for participation bias and fixed-effect analysis for study area). ETS had no substantial associations in current and ex-smokers in our study.
Conclusions: We found a positive association between ETS exposure and DM in never smokers. Additional longitudinal studies involving biomarkers are needed to further explore underlying mechanisms and susceptibilities.