Background: Peak exercise oxygen consumption provides valuable short-term prognostic information in patients with heart failure. However, peak exercise oxygen consumption is determined not only by the cardiac output response to exercise but also by age, gender, and muscle mass. We investigated whether percentage of predicted maximal exercise oxygen consumption rather than an absolute value may be a better predictor of survival.
Methods: Peak exercise oxygen consumption was measured and percentage of predicted maximal exercise oxygen consumption was derived from two standard formulas (Wasserman and Astrand) in 272 ambulatory patients referred for transplant evaluation. The predictive ability of these variables was determined by comparison of Kaplan-Meier curves, univariable proportional-hazards models, and receiver operating characteristic curves.
Results: Neither method of determining percentage of predicted maximal exercise oxygen consumption significantly improved the prediction of survival over peak exercise oxygen consumption alone. Overall model discrimination, as assessed by area under the receiver operating characteristic curve, was not significantly improved with percentage of predicted maximal exercise oxygen consumption (Wasserman) rather than weight-normalized peak exercise oxygen consumption (0.71 +/- 0.04 versus 0.66 +/- 0.04; Z = 1.60, p = 0.11). All of the difference between percentage of predicted maximal exercise oxygen consumption-Wasserman and peak exercise oxygen consumption resulted from differences in women (areas under receiver operating characteristic curve = 0.68 +/- 0.09 and 0.74 +/- 0.09; p = 0.14); results for men were the same (both areas = 0.68 +/- 0.04).
Conclusions: Normalization of peak exercise oxygen consumption for predicted values adds only minimal prognostic information. A peak exercise oxygen consumption < 14 ml/kg/min remains a reasonable guideline by which to time heart transplantation.