Background Peak oxygen consumption (VO₂) has an important prognostic role in chronic heart failure (CHF), but its discriminatory power is limited in patients with intermediate exercise capacity (peak VO₂ between 10-18 mL/kg/min). Thus, supplementary exertional indexes are greatly needed. Methods Six hundred patients with CHF with left ventricular ejection fraction (LVEF) ≤40% who performed a symptom-limited cardiopulmonary exercise testing were screened and followed up for 780 ± 450 days. Results Eighty-seven patients had major cardiac events (77 cardiac deaths and 10 urgent heart transplantations). Multivariate analysis revealed the rate of increase of minute ventilation per unit of increase of carbon dioxide production (VE/VCO₂ slope) (χ², 79.3, P <.0001), LVEF (χ², 24.6, P <.0001), and peak VO₂ (χ², 9.4, P <.0001) as independent and additional predictors of major cardiac events. VE/VCO₂ slope was the strongest independent predictor of outcome (χ², 20.9, P =.0001) in patients with intermediate peak VO₂ (n = 403), and the best cutoff value was 35 (χ², 25.8; relative risk = 3.2, 95% CI 2.0-5.1, P <.0001). Total mortality rate was 30% in patients with VE/VCO₂ slope ≥35 (n = 103, 26%) and 10% in those with VE/VCO₂ slope <35 (n = 300, 74%) (P <.0001). Patients with VE/VCO₂ slope ≥35 had a similar total mortality rate to those with peak VO₂ ≤10 mL/kg/min (30% vs 37%, P not significant). Conclusions A rational and pragmatic risk stratification process with symptom-limited cardiopulmonary exercise testing in CHF should include both peak VO₂ and VE/VCO₂ slope, the latter index effectively predicting outcome in almost one fourth of patients with intermediate exercise capacity. (Am Heart J 2002;143:418-26.)