Physical Activity, Cardiorespiratory Fitness, and Population-Attributable Risk

Objective: To determine population-attributable risk (PAR) and exposure impact number (EIN) for mortality associated with impaired cardiorespiratory fitness (CRF), physical inactivity, and other risk markers among veteran subjects. Methods: The sample included 5890 male subjects (mean age 58±15) who underwent a maximal exercise test for clinical reasons between January 1, 1992, and December 31, 2014. All-cause mortality was the end point. Cox multivariable hazard models were performed to determine clinical, demographic, and exercise-test determinants of mortality. Population-attributable risks and EIN for the lowest quartile of CRF and for inactive behavior were analyzed, accounting for competing events. Results: There were 2728 deaths during a mean ± standard deviation follow-up period of 9.9±5.8 years. Having low CRF (<5.0 metabolic equivalents [METs]) was associated with an approximate 3-fold higher risk of mortality and a PAR of 12.9%. Each higher MET achieved on the treadmill was associated with a 15% reduction in mortality (hazard ratio [HR]=0.85; 95% confidence interval [CI], 0.83 to 0.88; P<.001). Nearly half the sample was inactive, and these subjects had a 23% higher mortality risk and a PAR of 8.8%. The least fit quartile (<5.0 METs) had relative risks of ≈6.0 compared with the most-fit group (HR=5.99; 95% CI, 4.9 to 7.3). The least-active tertile had ≈2-fold higher risks of mortality vs the most active subjects (HR=1.9; 95% CI, 0.91 to 4.1). The lowest EIN was observed for low fitness (3.8; 95% CI, 3.4 to 4.3, P<.001), followed by diabetes, smoking, hypertension, and physical inactivity (all P<.001 except for diabetes, P=.008). Conclusion: Both higher CRF and physical activity provide protection against all-cause mortality in subjects referred for exercise testing for clinical reasons. Encouraging physical activity with the aim of increasing CRF would have a significant impact on reducing mortality.