Hyperinsulinemia is associated with insulin resistance and with the development of diabetes, hypertension, and coronary heart disease. Physical activity appears to be negatively associated with insulin resistance, although the mechanism is unclear. The relationship between physical activity and insulin resistance could be mediated, in part, by direct effects on skeletal muscle, a significant site for insulin-mediated glucose disposal. This report examines the relationship between skeletal muscle strength (as measured by handgrip dynamometry) and fasting insulin levels in a cohort of men in the ongoing Normative Aging Study (NAS). Handgrip strength was negatively associated (P = .013) with logarithmic (log) fasting insulin in cross-sectional data from 655 subjects after adjustment for potential confounders including age, body mass index (BMI), ratio of abdominal girth to hip breadth (AG/HB), usual physical activity level, and alcohol intake in a multiple regression model. In data collected prospectively among 1,195 subjects, handgrip strength measured at study entry was negatively predictive of log fasting insulin (P = .017) measured 22.9 +/- 2.6 years later, after adjustment for age, BMI, and AG/HB at study entry in a multiple linear regression model. A cross-sectional association was confirmed in an analysis of prospective data on the relationship between handgrip strength and fasting insulin levels. The findings suggest that skeletal muscle weakness may precede and predict the development of insulin resistance, and raise the intriguing possibility of some common cause in skeletal muscle pathophysiology.