Alcohol consumption and the risk of type 2 diabetes mellitus: effect modification by hypercholesterolemia: the Third Korea National Health and Nutrition Examination Survey (2005).

OBJECTIVE While the protective nature of moderate alcohol consumption against diabetes mellitus is well known, inconsistent findings continue to be reported. The possibility of different mixes of effect modifiers has been raised as a reason for those inconsistent findings. Our study aim was to examine potential effect modifiers that can change the effect of alcohol consumption on type 2 diabetes. METHODS From data in the third Korea National Health and Nutrition Examination Survey, 3,982 individuals over the age of 30 years who had not been diagnosed with diabetes were selected for inclusion in the study population. Breslow and Day's test and the Wald test between hypercholesterolemia and alcohol consumption in a multiple logistic regression model were used to assess effect modification. RESULTS Odds ratios for diabetes stratified by alcohol consumption strata and assessed using Breslow and Day's tests for homogeneity indicated that hypercholesterolemia was not a significant confounding factor (p=0.01). However, the Wald test for interaction terms, which is a conservative method of effect modification, was significant (p=0.03). CONCLUSIONS The results indicate that moderate alcohol consumption is not necessarily protective for type 2 diabetes mellitus, if a person has hypercholesterolemia. People who have hypercholesterolemia should be aware of the risk associated with alcohol consumption, a risk that contrasts with the reported protective effect of moderate alcohol consumption on diabetes.