Fitness, weight, blood-pressure, and lipids were measured in 290 women joining a fitness programme. Five years later, 110 (38%) had remained active, and the others had become sedentary again. The net effect of being active was to increase fitness by 18% and to reduce weight by 1.9 kg, diastolic blood pressure by 3 mmHg, and triglycerides by 0.15 mmol/L-1. After controlling for age, weight and fitness, physical activity accounted for 1% and 3% of the variance of changes in systolic and diastolic blood pressure respectively, and was not associated with changes in lipids. Change in fitness accounted for only 1% of change in diastolic blood pressure. More 'active' women gave up smoking than 'inactive' (70% versus 10%). Hypertensive women were more sensitive to changes in weight and fitness than normotensive women. It was concluded that physical activity and change in fitness were only weakly related to changes in blood pressure and were unrelated to changes in lipids.