Associations between residential traffic noise exposure and smoking habits and alcohol consumption-A population-based study.

Abstract

Traffic noise stresses and disturbs sleep. It has been associated with various diseases, and has recently also been associated with lifestyle. Hence, the association between traffic noise and disease could partly operate via a pathway of lifestyle habits, including smoking and alcohol intake. We investigated associations between modelled residential traffic noise and smoking habits and alcohol consumption. In a cohort of 57,053 participants, we performed cross-sectional analyses using data from a baseline questionnaire (1993-97), and longitudinal analyses of change between baseline and follow-up (2000-02). Smoking status (never, former, current) and intensity (tobacco, g/day) and alcohol consumption (g/day) was self-reported at baseline and follow-up. Address history from 1987-2002 for all participants were found in national registries, and road traffic and railway noise was modelled 1 and 5 years before enrolment, and from baseline to follow-up. Analyses were performed using logistic and linear regression, and adjusted for demographics, socioeconomic variables, leisure-time sports, and noise from the opposite source (road/railway). Road traffic noise exposure 5 years before baseline was positively associated with alcohol consumption (adjusted difference per 10dB: 1.38g/day, 95% confidence interval (CI): 1.10-1.65), smoking intensity (adjusted difference per 10dB: 0.40g/day, 95% CI: 0.19-0.61), and odds for being a current vs. never/former smoker at baseline (odds ratio (OR): 1.14; 95% CI: 1.10-1.17). In longitudinal analyses, we found no association between road traffic noise and change in smoking and alcohol habits. Railway noise was not associated with smoking habits and alcohol consumption, neither in cross-sectional nor in longitudinal analyses. The study suggests that long-term exposure to residential road traffic is associated with smoking habits and alcohol consumption, albeit only in cross-sectional, but not in longitudinal analyses.