OBJECTIVE: To examine the association between cardiac performance during recovery and the severity of heart failure, as determined by clinical and cardiopulmonary exercise test responses. METHODS: As part of a retrospective cohort study, 46 heart failure patients and 13 normal subjects underwent cardiopulmonary exercise testing while cardiac output was measured using a noninvasive device. Cardiac output in recovery was expressed as the slope of a single exponential relationship between cardiac output and time; the recovery-time constant was assessed in relation to indices of cardiac function, along with clinical, functional, and cardiopulmonary exercise responses. RESULTS: The recovery time constant was delayed in patients with heart failure compared with normal subjects (296.7 + 238 vs. 110.1 +27 seconds, p <0.01), and the slope of the decline of cardiac output in recovery was steeper in normal subjects compared with heart failure patients (p,0.001). The slope of the decline in cardiac output recovery was inversely related to peak VO2 (r = -0.72, p<0.001) and directly related to the VE/VCO2 slope (r = 0.57, p,0.001). Heart failure patients with abnormal recovery time constants had lower peak VO2, lower VO2 at the ventilatory threshold, lower peak cardiac output, and a heightened VE/VCO2 slope during exercise. CONCLUSIONS: Impaired cardiac output recovery kinetics can identify heart failure patients with more severe disease, lower exercise capacity, and inefficient ventilation. Estimating cardiac output in recovery from exercise may provide added insight into the cardiovascular status of patients with heart failure.