Dose-response relationship between tobacco or alcohol consumption and the development of diabetes mellitus in Japanese male workers.

Abstract Background The aim of this study was to examine the dose–response relationships between tobacco or alcohol consumption and the development of diabetes mellitus. Methods An 8-year prospective cohort study was conducted in 8423 male workers who received annual health check-ups between 2002 and 2010 at a Japanese steel company. The endpoints were defined as an HbA lc  ≥ 6.1% or taking any anti-diabetic medication. The dose–response relationships of tobacco or alcohol consumption were investigated using a proportional hazards regression with time-dependent covariates selected from baseline age, body mass index, mean arterial pressure, total serum cholesterol, aspartate aminotransferase, creatinine and uric acid, shift work or day work, and habitual exercise by stepwise selection method. Results A positive dose–response relationship between tobacco consumption and the development of diabetes mellitus was observed, with a significantly higher hazard ratio (HR) seen with higher tobacco consumption (11–20 cigarettes/day, HR 1.26 [95% confidence interval (CI), 1.00–1.59], ≥21 cigarettes/day, HR 1.54 [95%CI, 1.20–1.97]). In contrast, we observed a negative dose–response relationship between alcohol consumption and the development of diabetes mellitus, with a significantly lower HR with higher weekly alcohol consumption (7.0–13.9 gou/week [154–307 g/week], HR 0.73 [95% CI, 0.55–0.97], ≥14.0 gou/week [308 g/week], HR 0.75 [95% CI, 0.57–0.98]). Conclusions The results indicated that decreasing tobacco consumption will achieve significant prevention of diabetes mellitus. On the other hand, we observed a significant, negative dose–response relationship between alcohol consumption and the development of diabetes mellitus, in contrast to previous studies that reported a positive relationship in the Japanese population.