Abstract 2598: Do Exercise Test Abnormalities Fully Account for the Difference in Risk Between Patients with and without a History of Cardiovascular Disease?
Poor functional capacity and heart rate and electrocardiographic abnormalities during exercise testing occur more often in patients with cardiovascular disease (CVD), and CVD patients have poorer survival than those free of CVD history. We sought to determine whether the exercise test can accurately establish prognosis independent of history of CVD. Subjects consisted of patients referred for exercise testing during 1986–1991. They were stratified by history of CVD established prior to the exercise test. Mortality was determined by examination of the National Death Index in August, 2005. Cox proportional hazards regression was used to determine the significance of CVD history for all-cause mortality after adjustment for age, sex, beta blocker use, functional capacity, heart rate reserve, and normal exercise ECG. A second age, sex, and beta blocker-adjusted analysis used only subjects with a “good” exercise test result defined as functional capacity ≥ 100% of age- and gender- predicted with heart rate reserve of at least 50 bpm and normal exercise ECG. A total of 10,940 patients were included in the analysis. CVD history was established in 1865 of 8215 men (22.7%) and 416 of 2725 women (15.3%). Follow-up averaged 16 ± 3.2 years during which time 649/1865 (34.8%) men and 114/416 (27.4%) women with baseline CVD died compared to 537/6350 (8.5%) men and 159/2309 (6.9%) women without CVD history.
Results of the Cox regression for all subjects showed a hazard ratio =2.26 (95% CI 2.01–2.54, p<0.0001) for CVD history after controlling for age, gender, beta blocker, functional capacity, heart rate reserve, and normal exercise EGC (each of which were significant, independent predictors of all-cause mortality). We identified 537/2281 CVD patients (23.5%) and 4550/8659 healthy subjects (52.5%) with “good” exercise test results. Overall mortality rates were 25.5% for CVD patients and 4.9% for non-CVD subjects within this “good” test subgroup with an adjusted hazard ratio for CVD history of 3.19 (95% CI 2.53–4.05, p<0.001). Results suggest that exercise test results, though strongly predictive of all-cause mortality, cannot fully account for CVD history. Clinicians must use separate risk equations for establishing prognosis in patients with and without CVD history.