All-cause death rates in normotensive and hypertensive men were examined across physical fitness levels. Subjects were 10,224 healthy normotensive men and 1,832 men who reported a history of hypertension, but were otherwise healthy. Physical fitness was determined by maximal treadmill exercise testing. Baseline preventive medical examinations were given during 1970-1981, and mortality surveillance was conducted on the cohort through 1985. There were 240 deaths in the normotensive men and 78 deaths in hypertensive men. Age-adjusted all-cause mortality rates per 10,000 man-years of follow-up in normotensive men ranged from 64.0 in the least fit quintile to 18.6 in the most fit quintile. Corresponding rates for hypertensive men were 110.5 to 24.8. Subjects were further classified into lower and higher blood pressure groups by baseline resting systolic blood pressure (less than 140 mmHg and greater than or equal to 140 mmHg). Normotensive and hypertensive men who were more fit had lower death rates compared to less fit men within both of the measured blood pressure strata. The relation between fitness and all-cause mortality held in multiple logistic regression analyses after adjustment for the influence of age, serum cholesterol, resting systolic blood pressure, body mass index, current smoking habit, and length of follow-up. We conclude that low levels of physical fitness result in an increased risk for all-cause mortality in normotensive and hypertensive men.