Personality and diabetes mellitus incidence in a national sample

To test whether personality traits were prospectively associated with type 2 diabetes incidence. The sample (n = 6798) was derived from the National Health and Nutrition Examination Survey Epidemiological Follow-up Study cohort. We fit four logistic regression models to test whether neuroticism, extraversion, openness to experience, or the Type A behavior pattern predicted type 2 diabetes incidence. Model 1 included sex, age, and race/ethnicity. Model 2 added personality traits, Model 3 added depressive symptoms, and Model 4 added body mass index (BMI), hypertension, and cigarette smoking status as predictors. In Model 1 age was associated with increased risk of diabetes (2% per year); being black as opposed to white was associated with a three-fold increase in risk. In Model 2 age and being black were still significant and extraversion was associated with decreased risk (17% per standard deviation [SD]). In Model 3 age, being black, and extraversion were still significant. In addition, neuroticism was associated with decreased risk (26% per SD) and depressive symptoms were associated with increased risk (28% per SD). In Model 4 age, being black, neuroticism, and depressive symptoms were still significant. BMI was associated with increased risk (14% per SD) and extraversion was no longer significant. Higher neuroticism was associated with reduced type 2 diabetes risk even after controlling for race/ethnicity, age, depressive symptoms, and BMI. Extraversion and Type A behavior were not significant after including covariates.