Long-term exposure to transportation noise and air pollution in relation to incident diabetes in the SAPALDIA study

Background Epidemiological studies have inconsistently linked transportation noise and air pollution (AP) with diabetes risk. Most studies have considered single noise sources and/or AP, but none has investigated their mutually independent contributions to diabetes risk. Methods We investigated 2631 participants of the Swiss Cohort Study on Air Pollution and Lung and Heart Diseases in Adults (SAPALDIA), without diabetes in 2002 and without change of residence between 2002 and 2011. Using questionnaire and biomarker data, incident diabetes cases were identified in 2011. Noise and AP exposures in 2001 were assigned to participants' residences (annual average road, railway or aircraft noise level during day-evening-night (Lden), total night number of noise events, intermittency ratio (temporal variation as proportion of event-based noise level over total noise level) and nitrogen dioxide (NO2) levels. We applied mixed Poisson regression to estimate the relative risk (RR) of diabetes and their 95% confidence intervals (CI) in mutually-adjusted models. Results Diabetes incidence was 4.2%. Median [interquartile range (IQR)] road, railway, aircraft noise and NO2 were 54 (10) dB, 32 (11) dB, 30 (12) dB and 21 (15) μg/m3, respectively. Lden road and aircraft were associated with incident diabetes (respective RR: 1.35; 95% CI: 1.02-1.78 and 1.86; 95% CI: 0.96-3.59 per IQR) independently of Lden railway and NO2 (which were not associated with diabetes risk) in mutually adjusted models. We observed stronger effects of Lden road among participants reporting poor sleep quality or sleeping with open windows. Conclusions Transportation noise may be more relevant than AP in the development of diabetes, potentially acting through noise-induced sleep disturbances.