To study the effects of long-term, self-monitored exercise on the serum lipid profile and body composition of middle-aged non-smoking males, a controlled study was conducted in 61 sedentary, middle-class Swiss men. Thirty-nine men were randomly allocated to jog 2 h/wk for 4 months on an individually prescribed, heart rate-controlled basis, whereas 22 men served as controls. Despite varying adherence to the exercise regimen, the following 4-month net changes (effect in exercise group minus effect in control group) in lipids were seen: HDL cholesterol (C) +0.12 mmol/l (95% CI 0.02, 0.22; P = 0.028), LDL-C +0.08 mmol/l (ns), VLDL-C -0.26 mmol/l (-0.45, -0.07; P = 0.009), total triglycerides (TT) -0.21 mmol/l (ns), HDL-C/total C +0.02 (0.001, 0.05; P = 0.047). The net changes in endurance capacity and resting heart rate in favour of exercisers were significant as well, whereas no significant changes in apolipoprotein levels were seen. Exploratory analyses revealed, for example, associations of the increase in total physical activity with an increase in the HDL-C/total C ratio (r = 0.46; P less than 0.001), and of the change in estimated body fat content with an opposed change in the HDL-C/total C ratio (r = -0.40; P less than 0.001), or an inverse relationship of the change in subcutaneous fat with a change in the HDL2-C level (r = -0.39; P less than 0.001). Multivariable regression analysis suggested that much of the effect of jogging on HDL-C was apparently mediated through a decrease in body fat content. A change in the waist/hip ratio was unrelated to lipoprotein changes but was related to the change of TT level (r = 0.22; P less than 0.05). This study confirms that individually prescribed, unsupervised jogging can increase HDL-C levels and improve the serum lipoprotein profile in self-selected nonsmoking males. Although the effect is modest, it may be relevant to preventive cardiology, given the evidence for a reduction in cardiovascular risk even after apparently small decreases in risk factor levels.