The relationship between self-reported physical activity and cardiovascular risk factors was evaluated in a population-based sample. The sample included 541 premenopausal women recruited for a study of the natural history of risk factor change associated with change in sex hormone status. Physical activity was assessed using the Paffenbarger Activity Questionnaire. Women were classified according to quartile of weekly energy expenditure into groups of 0-500, 501-999, 1,000-1,999, and 2,000 kcal or greater. Results showed that the more active the women, the lower their blood pressure and heart rate. More active women had lower cholesterol and triglycerides, and higher high-density lipoprotein (HDL) cholesterol. Tricep and suprailiac skinfolds, fasting insulin, and insulin/glucose levels were also lower among the more active women. When the analysis was repeated controlling for the effect of education and body mass index, the statistical test for linear trend remained significant except for the trend for heart rate, total cholesterol, and triglycerides. Women reporting activity of 1,000 kcal/week had higher high-density lipoprotein cholesterol and lower diastolic blood pressure, body mass index, skinfolds, fasting insulin, and fasting insulin/glucose ratios compared with women reporting lower levels of activity. Only those women who reported 2,000 kcal/week had significantly lower total cholesterol, triglycerides, and low-density lipoprotein cholesterol, and higher HDL2 cholesterol; women reporting less activity did not differ with regard to these lipids and lipoproteins. The study suggests that physical activity is associated with improved cardiovascular risk profiles among middle-aged women and that the beneficial effects of activity are seen at different levels for specific risk factors.