No interaction of body mass index and smoking on diabetes mellitus risk in elderly women

Objective: We sought to assess the interaction of smoking and body mass index (BMI) on diabetes risk. Methods: We analyzed data from a community-based prospective cohort of 41,836 women from Iowa who completed a baseline survey in 1986 and five subsequent surveys through 2004. The final analysis included 36,839 participants. Results: At baseline (1986), there were 66% never smokers, 20% former smokers, and 14% current smokers. Subjects represented 40% normal weight, 38% overweight, and 22% obese individuals. Compared to normal weight women, the hazard ratio (HR) for diabetes was increased in overweight (HR 1.96; 95% CI 1.75-2.19) and obese subjects (HR 3.58: 95% CI 3.19-4.02). The hazard ratio for diabetes increased in a dose-dependent manner with smoking intensity. Compared to never smokers, former smokers had a higher risk for diabetes (HR 1.22; 95% CI 1.11-1.34). Among current smokers, the hazard ratio for diabetes was 1.21 (95% CI 0.95-1.53) for 1-19 pack-year smokers, 1.33 (95% CI 1.12-1.57) for 20-39 pack-year smokers, and 1.45 (95% CI 1.23-1.71) for >= 40 pack-year smokers. Similar trends were observed when the results were stratified by BMI. A test of interaction between BMI and smoking on diabetes risk was not statistically significant. Conclusions: Our findings suggest that smoking increases diabetes risk through a BMI-independent mechanism. (C) 2008 Elsevier Inc. All rights reserved.