Exercise, Fitness, and Mortality-Reply

In Reply.—  Dr Goldstein is correct in his suggestion that the inverse relation in our study between physical fitness and cancer mortality might be confounded by undiagnosed cancer present at baseline examination. Preclinical cancer could result in poor treadmill test performance and increase the likelihood of death due to cancer during follow-up. To address this issue, we excluded all cancer deaths that occurred in the cohort during the first 3 years of follow-up and reanalyzed the relation between fitness and cancer mortality. This exclusion left 55 cancer deaths in men and 13 in women. Age-adjusted cancer death rates per 10 000 person-years of observation in low-, moderate-, and high-fitness groups were 16.4, 6.1, and 4.5 in men and 13.5, 5.9, and 1.0 in women, respectively. These findings reduce the likelihood that preclinical cancer is the sole cause of the relation between fitness and cancer. However, residual confounding from prior cigarette