Reproducibility of peak oxygen uptake and other cardiopulmonary exercise parameters: implications for clinical trials and clinical practice.

Abstract

Clinical trialists involved in projects aimed at assessing the impact of a device or therapy on exercise capacity must select an end point that can detect a therapy-derived change and do so in a manner that is both sensitive to the change and valid. Using baseline data from the Prospective Evaluation of Elastic Restraint to Lessen the Effects of Heart Failure (PEERLESS-HF) trial, we describe the coefficient of variation (CV) for peak oxygen uptake (Vo2) and other parameters measured during a cardiopulmonary exercise (CPX) test. Prior to conducting CPX testing, a CPX core laboratory conducted training with testing staff at each site, and a standardized quality assurance protocol was completed. During screening, subjects performed two symptom-limited CPX tests (CPX-0 and CPX-1) within 14 days. The CVs for peak Vo2, ventilatory efficiency, and ventilatory-derived anaerobic threshold were 5.9%, 4.8%, and 6.8%, respectively. During CPX-0, the percentage of subjects with a respiratory exchange ratio (RER) < 1.05 who increased their peak Vo2 by ≥ 6% during CPX-1 was 47%, compared with 12% of subjects also with an RER < 1.05 who experienced a decrease in peak Vo2 ≥ 6%. In a multisite clinical trial setting, we showed that the achieved test-retest reproducibility for peak Vo2 can be acceptable, as evidenced by a CV at 5.9%. Until more data are available to identify patients who may demonstrate greater test-retest variance and to avoid repeat testing in all subjects at baseline, clinical trialists might consider repeating tests when RER < 1.05. clinicaltrials.gov; Identifier: NCT00382863.