Background: Peak oxygen uptake ((V) over dotO(2)peak) is a well- established prognostic marker in chronic heart failure (CHF). Cardiopulmonary exercise testing (CPET) provides physiological parameters other than (V) over dotO(2)peak that might have prognostic value. We aimed at determining whether exercise recovery data kinetics have prognostic implications over (V) over dotO(2)peak and Heart Failure Survival Score. Methods: Exercise data from 200 consecutive CHF patients evaluated for possible heart transplantation and received CPET at our institution between 2004 and 2011 were analyzed. The rate of recovery of oxygen uptake ((V) over dotO(2)) at 2 minutes after exercise ( (V) over dotO(2)- REC2) was calculated using the difference between (V) over dotO(2)peak and (V) over dotO(2) at minute 2 of recovery and expressed as a percentage of (V) over dotO(2)peak. The composite primary end point was the time from CPET to the first event including death, heart transplant, or mechanical heart implantation. Results: Mean follow- up period was 1271 +/- 61 days during which there were 108 first events including 35 deaths, 66 heart transplants, and 7 mechanical heart implantations. The strongest prognostic factors in the univariate analysis were (V) over dotO(2)-REC2, (V) over dotO(2)peak, (V) over dotO(2) efficiency slope, and ventilation to carbon dioxide excretion ratio slope (all P < 0.0001). Multivariate analysis showed that (V) over dotO(2)- REC2 (P < 0.0001), ventilation to carbon dioxide excretion ratio slope (P = 0.0022), use of angiotensin- converting enzyme inhibitors or angiotensin receptor blockers (P = 0.0042), presence of a defibrillator (P = 0.0127), and mean arterial pressure (P = 0.0151) were independent predictors of event-free survival time. Conclusions: (V) over dotO(2)-REC2 was the strongest prognostic marker of death, heart transplantation, and mechanical heart implantation in severe CHF. This finding should be confirmed prospectively.