Long-term mortality with multiple treadmill exercise test abnormalities: comparison between patients with and without cardiovascular disease.

BACKGROUND: Poor exercise capacity, abnormal heart rate responses, and electrocardiographic abnormalities during treadmill exercise testing independently predict mortality. The combined relationship of these 3 variables to determine the incremental increase in mortality was compared in groups with and without known cardiovascular disease (CVD).

METHODS: Patients referred for treadmill exercise testing during 1986 to 1991 were included. Exercise capacity <74% (of age- and gender-predicted value), heart rate reserve of <68 beat/min, and horizontal or down-sloping ST depression of > or =1 mm were considered abnormal. Cox proportional hazards regression was used to determine all-cause mortality (average follow-up of 16 years) based on the number of exercise test abnormalities (0, 1, 2, or all 3).

RESULTS: Among 10,897 patients, 20.9% (n = 2,277) had CVD. Poor exercise capacity and limited heart rate reserve were associated with increased risk of mortality (P < .0001) in both groups; however, abnormal exercise electrocardiogram was associated with an increased risk of mortality in the no-CVD group only (P < .0001). A graded increase in mortality was observed with increase in number of abnormal exercise test results in both groups. Patients without CVD having 2 or 3 abnormal exercise test results had a similar age-adjusted risk of long-term mortality as those with CVD but normal exercise test results, with a hazard ratio comparing these groups = 1.01 (95% CI 0.79-1.28).

CONCLUSIONS: The combinatorial approach validates the prognostic significance of multiple exercise test variables. The presence of > or =2 exercise test abnormalities may constitute a "CVD risk equivalent" in patients without CVD.