Association of dietary composition with fasting serum insulin level: The ARIC Study

Insulin resistance and subsequent hyperinsulinemia are risk markers for diabetes and have been implicated in the development of cardiovascular disease; therefore, modifiable determinants of the serum insulin level are of interest. Several studies have suggested that certain dietary components, particularly fat and alcohol, may be associated with both fasting and postprandial insulin concentrations independent of obesity. The relation of dietary intake, assessed by a modified version of the Willett 61-item food frequency questionnaire, to fasting serum insulin was examined in 13,167 nondiabetic men and women, aged 45-64. Univariately, fasting insulin was associated positively (p<0.01) with energy intake. After adjustment for gender, race, age, energy intake, body mass index, waist-to-hip ratio, smoking, and physical activity in multivariate models, alcohol was negatively associated with insulin, with the predicted difference in insulin for the interquartile range being 5.0% (95% CI = 6.3%-3.6%). A significant inverse association was also observed between dietary fiber and insulin among women, but not men. Total fat, saturated fat, monounsaturated fat, and carbohydrates were associated positively (p<0.05) with fasting insulin, but the magnitude of these associations was small. The predicted percent difference in insulin for persons at the 25th versus the 75th percentile of total fat, saturated fat, monounsaturated fat, and carbohydrates was 2.3%, 2.3%, 2.2%, and 1.4%, respectively. These data suggest that dietary composition may have modest effects on fasting insulin level, beyond the contributions of energy imbalance and obesity.