We set out to examine the long-term effect of work-related and leisure time physical activity on risk of death from coronary heart disease (CHD) and other causes. Data were obtained from a large prospective population study of 7142 participant men aged 47 to 55 years at baseline in 1970-1973 and without symptomatic CHD. Data on physical activity at work and at leisure were assessed by way of a questionnaire. After 20 years follow-up men with physically demanding work had a slightly higher mortality from all causes, but not from CHD. The association with all-cause mortality disappeared after controlling for smoking, occupational class, and alcohol abuse. Men who were physically active during leisure time had a lower risk of death from CHD, cancer, and all causes. After controlling for smoking, diastolic blood pressure, serum cholesterol, body mass index, diabetes alcohol abuse, and low occupational class, the most active men had a relative risk (RR) of dying from CHD of 0.72 (0.56-0.92 95% CI) and of all causes of 0.70 (0.61-0.80 95% CI). The adjusted relative risk for cancer death was of borderline significance. Noncardiovascular deaths other than cancer showed a strong relationship with low physical activity in the present study that remained after controlling for smoking, alcohol abuse, and other factors (adjusted RR 0.55 (0.42-0.73 95% CI)). The protective effect of CHD and on mortality was present after excluding the first 8 years of follow-up, and at all levels of other risk factors, such as smoking, diastolic blood pressure, serum cholesterol, and body mass index. In conclusion, this long-term study demonstrates the protective effect of leisure time physical activity on CHD death, cancer death, and on death from all causes. The effect on coronary death and on death from all causes was independent of other risk factors.