Patients with chronic heart failure (CHF) have an increased ventilatory cost of CO2 excretion on exercise (the linear VENCO2 slope — m). This represents mismatch of perfusion to ventilation Q/V) during exercise and may reflect dysfunction of pulmonary endothelium and microvessels, as is present in systemic vessels in CHF. We have studied the effect of L-arginine, substrate for endogenous endothelial nitric oxide (NO). on slope m in CHF. Male patients (n = 11) with stable NYHA class lib or III CHF were enrolled into a cross-over study of iv I-arginine (ARC) (0,5 mg/kg over 30 min) vs. sublingual nifedipine (endothelium independent vasodilator). Patients had no other known risk factors for endothelial dysfunction, no inducible myocardial ischaemia and all were receiving ACE inhibitors and diuretics. Patient groups were characterised according to peak VO2 < 15 ml/kg/min. Exercise tests were performed 15 min pre and immediately post infusion. Outcome measures were m before and after treatment, exercise duration (ExD, secs) and peak oxygen consumption (VO2. ml/kg/min). m pre Post arginine m pre Post nifedipine m Ex D V02 m ExD VO2 V02 < 15 41 ± 7 34 ± 7 * 340 ± 90 14 ± 1 37 ± 5 36 ± 6 360 ± 50 13 ± 1 V02> 15 31 ± 2 31 ± 2 360 ± 50 18 ± 1 31 ± 3 30 ± 1 690 ± 30 18 ± 1 * p < 0.05, pre vs post infusion Mean blood pressure during exercise was similar following both treatments. Patients with peak V02 < 15 had an elevated VE/VCO2 slope which was reduced by ARG. These preliminary data suggest that exercise related ON mismatch in severe CHF is acutely reversible and may involve endothelial dysfunction.