In this study we analysed the all-cause mortality over a period of maximal 6 years in 60 male patients (age: 63.4+/-8.3 years, mean+/-S.D.), suffering from chronic heart failure with resting left ventricular ejection fraction and E/O2 slope as independent factors. We assessed functional NYHA class (II: n=36, III: n=24), radionuclide left ventricular ejection fraction (29.2+/-10.4%) and peak values of heart rate, O2, CO2, E, anaerobic threshold and exercise duration with an incremental work load test on the treadmill. O2 relative to E was based on the individual slopes of the regression of O2 on E during the first 6 min of exercise. These slopes with other exercise-related variables and factors such as etiology, medication, and NYHA class were analysed with a Cox's Regression Method. A survival time analysis (Kaplan-Meier survival curve) was done to establish the influence of E/O2 slope and left ventricular ejection fraction (both split into above and below median values), as well as their interaction, on survival. From all investigated exercise-related variables. E/O2 slope is the most powerful variable regarding prediction of all-cause mortality in our group of chronic heart failure patients. Concerning risk stratification, the subgroup (n=18) with a relatively high left ventricular ejection fraction (>28%) and flat E/O2 slope (<27.6) had most survivors (77.8%) after about 3 years, while the subgroup (n=12) with a relatively high left ventricular ejection fraction (>28%), but a steep E/O2 slope (>27.6) had least survivors (33.3%). This difference in percentage is highly significant (P=0.0025). The fact that E/O2 slope and left ventricular ejection fraction show comparable main and interaction effects between measures of exercise tolerance (e.g., anaerobic threshold, peak O2, exercise duration) on the one hand, and all-cause mortality on the other, suggests the existence of common sources of variance. Based on our analysis, it is unlikely that effects on all-cause mortality are mediated through phenomena related to exercise tolerance. Therefore, we hypothesize that the effects on exercise tolerance and all-cause mortality both depend on common factors, which cause both cardiac and peripheral organ (c.q. muscular) dysfunctions. Moreover, this study clearly shows that E/O2 slope during incremental exercise is an important prognostic marker for risk stratification in chronic heart failure patients, NYHA class II and III.