Cardiorespiratory Fitness Reclassifies Mortality Risk Beyond Traditional Risk Factors: The ETHOS Study

Abstract

Background: Cardiorespiratory fitness (CRF) is a powerful predictor of mortality, but its incremental value for improving clinical risk classification beyond traditional cardiovascular risk factors remains incompletely defined. We evaluated whether the addition of CRF improves 10-year mortality risk prediction using contemporary reclassification metrics. Methods: We analyzed 562,234 adults from a clinical exercise testing cohort with complete covariate data. CRF was assessed by peak metabolic equivalents (METs) achieved during treadmill testing and dichotomized as fit versus unfit, with unfit defined as the least fit quintile (<5.0 METs). Cox proportional hazards models were constructed to estimate 10-year all-cause mortality risk using traditional risk factors (age, sex, hypertension, diabetes, dyslipidemia, and smoking), with and without CRF. Incremental prognostic value of CRF was assessed using categorical and continuous net reclassification improvement (NRI), integrated discrimination improvement (IDI), and Harrell’s C-statistic. Results: During a mean follow-up of 10.2±5.1 years, CRF was independently associated with mortality, with each 1-MET higher CRF associated with a 16% lower hazard of death (hazard ratio 0.84; 95% CI 0.838–0.842; p<0.001). The addition of CRF significantly improved risk prediction, yielding a categorical NRI of 0.06 and a continuous NRI of 0.37 (both p<0.001), with significant improvement in IDI. Model discrimination increased from 0.662 to 0.682 after addition of CRF (ΔC=0.020; p<0.001). Adjusted 10-year mortality risk was substantially higher among unfit versus fit individuals (32.9% vs 13.4%). Conclusions: CRF significantly improves 10-year mortality risk prediction beyond traditional cardiovascular risk factors and meaningfully enhances individual-level risk stratification. Incorporation of CRF into routine clinical risk assessment frameworks should be considered.