Abstract 19496: A New Exercise Testing Model Performs Better than the Duke Treadmill Score to Identify Patients at Increased Risk for All-cause Mortality
Background: Cardiovascular mortality has declined with advances in primary and secondary prevention, yet risk assessment for patients undergoing exercise treadmill testing is generally still based upon the Duke Treadmill Score (DTS). We hypothesized that a new risk score would perform better than the DTS.
Methods: We studied 59,877 consecutive patients who had exercise testing between Januray 1, 2000 and October 27, 2011 with the sample divided equally into derivation and validation cohorts. The primary outcome was all-cause mortality. Risk scores for men and women were developed from Cox proportional hazards models. The DTS was compared to new risk scores in the validation cohort with C-statistics, category-free net reclassification improvement (NRI), and integrated discrimination improvement (IDI).
Results: Over a median follow-up of 8 years, there were 2,521 deaths. In the multivariable model, age, exercise capacity, heart rate recovery, weight, and renal disease were associated with mortality in men and women. Diabetes mellitus and history of smoking were associated with mortality in women whereas heart failure, current smoking, and hypertension were associated with mortality in men. C-statistics, NRI, and IDI were all improved with the new risk scores (Table). In particular, the new risk scores correctly reclassified patients with events, especially in women. Patients in the highest tertile of risk score also had substantial mortality during follow-up (Figure).
Conclusions: We have developed new risk scores for men and women that perform better than the DTS, particularly in patients at the highest risk for all-cause mortality.
Author Disclosures: P.C. Cremer: None. Y. Wu: None. L. Pierson: None. D. Brennan: None. L. Cho: None.
- © 2015 by American Heart Association, Inc.