Prediction of Cardiovascular Mortality by Estimated Cardiorespiratory Fitness Independent of Traditional Risk Factors: The HUNT Study

Abstract Objective To assess the predictive value of estimated cardiorespiratory fitness (eCRF) and evaluate the additional contribution of traditional risk factors in cardiovascular disease (CVD) mortality prediction. Participants and Methods The study included healthy men (n=18,721) and women (n=19,759) aged 30 to 74 years. A nonexercise algorithm estimated cardiorespiratory fitness. Cox proportional hazards models evaluated the primary (CVD mortality) and secondary (all-cause, ischemic heart disease, and stroke mortality) end points. The added predictive value of traditional CVD risk factors was evaluated using the Harrell C statistic and net reclassification improvement. Results After a median follow-up of 16.3 years (range, 0.04-17.4 years), there were 3863 deaths, including 1133 deaths from CVD (734 men and 399 women). Low eCRF was a strong predictor of CVD and all-cause mortality after adjusting for established risk factors. The C statistics for eCRF and CVD mortality were 0.848 (95% CI, 0.836-0.861) and 0.878 (95% CI, 0.862-0.894) for men and women, respectively, increasing to 0.851 (95% CI, 0.839-0.863) and 0.881 (95% CI, 0.865-0.897), respectively, when adding clinical variables. By adding clinical variables to eCRF, the net reclassification improvement of CVD mortality was 0.014 (95% CI, −0.023 to 0.051) and 0.052 (95% CI, −0.023 to 0.127) in men and women, respectively. Conclusion Low eCRF is independently associated with CVD and all-cause mortality. The inclusion of traditional clinical CVD risk factors added little to risk discrimination and did not improve the classification of risk beyond this simple eCRF measurement, which may be proposed as a practical and cost-effective first-line approach in primary prevention settings.