Aims: This study sought to examine the predictive values of changes over time in exercise capacity and echocardiographic measurements of ventricular dimensions or function in predicting mortality in patients with chronic heart failure. Methods and Results: Sixty-two patients with chronic heart failure (58 men, mean [± SD] age 60 ± 10 years, mean peak oxygen consumption (VO₂) 18.2 ± 5.9 ml.kg^-1.min^-1, mean left ventricular ejection fraction 38.9 ± 15.8) who underwent both treadmill exercise testing and echocardiographic examination on two occasions, separated by 19 ± 15 months were followed-up for a mean of 17 months (interquartile range 9-30 months). During the follow-up period, 19 patients (30) died and three (4.8) underwent heart transplantation. Of measurements taken at a single time-point (visit 2) exercise duration, peak VO₂ ventilatory response to exercise (VE/VCO₂), left atrial diameter and left ventricular ejection fraction were found, by Cox proportional-hazard analysis, to predict the outcome in these patients tall P < 0.05). Of the changes in parameters between visit 1 to visit 2, only changes in peak VO₂ per year (P = 0.026) predicted non-transplanted survival (independent of changes in left ventricular ejection fraction and VE/VCO₂). In Kaplan-Meier survival analysis patients with increased peak VO₂ over time (n = 28) showed a better prognosis at 2 years (cumulative survival 75 [95 confidence interval: 56-95] than those with a decrease in peak VO₂ (n = 34, cumulative survival 50 [95 confidence interval: 31-68]). Conclusions: Although single estimates of peak VO₂,VE/VCO₂ and left ventricular ejection fraction have significant prognostic importance in patients with chronic heart failure, when monitoring changes over time only peak VO₂ remains a significant predictor of outcome.