Background The slope of the relationship between ventilation and carbon dioxide production ((V) over dotE/(V) over dot CO2 slope), obtained during symptom-limited ramp exercise testing, reflects exercise ventilatory efficiency. Importantly, the (V) over dotE/(V) over dot CO2 slope is related to prognosis in patients with congestive heart failure (CHF). The aim of the present study was to determine the relationship between the institution of beta-blockers, carvedilol or metoprolol, and the (V) over dotE/(V) over dot CO2 slope during exercise in patients with CHF. Methods and Results Fifty-seven patients with New York Heart Association functional class II or III with a radionuclide left ventricular ejection fraction (LVEF) of less than 40% received carvedilol or metoprolol in a randomized fashion. The (V) over dotE/(V) over dot CO2 slope, LVEF and plasma brain natriuretic peptide (BNP) concentration were determined before and after 16 weeks of treatment. LVEF improved (p<0.01), but the (V) over dotE/(V) over dot CO2 slope and BNP did not. A significant improvement in the (V) over dotE/(V) over dot CO2 slope was observed in patients with LVEF <29% or BNP >63 pg/ml (respective baseline median values) (p<0.05, p<0.05). In patients with BNP >63 pg/ml, the improvement effect on the (V) over dotE/(V) over dot CO2 slope with carvedilol was significantly greater than that with metoprolol (p<0.05) and a significant improvement in the (V) over dotE/(V) over dot CO2 Slope was observed only in those who took carvedilol (p<0.01). Conclusions The (V) over dotE/(V) over dot CO2 slope was not improved after beta-blocker therapy in any of the patients. However, it did improve in patients with a lower LVEF or higher BNP level at baseline, and carvedilol was more effective than metoprolol in improving the (V) over dotE/(V) over dot CO2 Slope inpatients with higher BNP levels at baseline.