Share this article on:

Association of Changes in Fitness and Body Composition with Cancer Mortality in Men


Medicine & Science in Sports & Exercise: July 2014 - Volume 46 - Issue 7 - p 1366–1374
doi: 10.1249/MSS.0000000000000225

Introduction Both baseline cardiorespiratory fitness and adiposity predict the risk of cancer mortality. However, the effects of changes in these two factors over time have not been evaluated thoroughly. The aim of this study was to examine the independent and joint associations of changes in cardiorespiratory fitness and body composition on cancer mortality.

Methods The cohort consisted of 13,930 men (initially cancer-free) with two or more medical examinations from 1974 to 2002. Cardiorespiratory fitness was assessed by a maximal treadmill exercise test, and body composition was expressed by body mass index (BMI) and percent body fat. Changes in cardiorespiratory fitness and body composition between the baseline and the last examination were classified into loss, stable, and gain groups.

Results There were 386 deaths from cancer during an average of 12.5 yr of follow-up. After adjusting for possible confounders and BMI, change hazard ratios (95% confidence intervals) of cancer mortality were 0.74 (0.57–0.96) for stable fitness and 0.74 (0.56–0.98) for fitness gain. Inverse dose–response relationships were observed between changes in maximal METs and cancer mortality (P for linear trend = 0.05). Neither BMI change nor percent body fat change was associated with cancer mortality after adjusting for possible confounders and maximal METs change. In the joint analyses, men who became less fit had a higher risk of cancer mortality (P for linear trend = 0.03) compared with those who became more fit, regardless of BMI change levels.

Conclusions Being unfit or losing cardiorespiratory fitness over time was found to predict cancer mortality in men. Improving or maintaining adequate levels of cardiorespiratory fitness appears to be important for decreasing cancer mortality in men.

1Department of Sports Medicine, Beijing Sport University, Beijing, CHINA; 2Department of Exercise Science, University of South Carolina, Columbia, SC; and 3Department of Epidemiology and Biostatistics, University of South Carolina, Columbia, SC

Address for correspondence: Xuemei Sui, PhD, 921 Assembly Street, Columbia,SC 29208; E-mail:

Submitted for publication August 2013.

Accepted for publication November 2013.

© 2014 American College of Sports Medicine