Chronic obstructive pulmonary disease (COPD) is thought to limit exercise capacity through a decreased ventilatory reserve, with cardiovascular factors playing a minimal role. We assessed respiratory muscle (RM) and cardiopulmonary function during exercise in very severe COPD (FEV1 0.79 +/- 0.17 L). We determined minute ventilation (VE), oxygen consumption (VO2), carbon dioxide production (VCO2), heart rate (HR), respiratory rate (RR), and O2 pulse with a metabolic cart. RM function was assessed from esophageal and gastric pressures. Dyspnea was assessed with a visual analog scale (VAS). Exercise capacity (peak VO2 = 36 +/- 31%), ventilatory reserve (VE/maximum voluntary ventilation [MW] = 89 +/- 31%), HR = 76 +/- 15%, and O2 pulse (O2Pmax = 45 +/- 15%) were abnormal. Peak VO2 correlated with O2Pmax(r = 0.82), the change in end-inspiratory pleural pressure (deltaPpli) (r = -0.74), maximal transdiaphragmatic pressure (Pdimax) (r = 0.68), and VEmax (r = 0.58). There were similar correlations with exercise endurance time. Multiple regression analysis revealed O2Pmax to be the best predictor of peak VO2. Thereafter, only VEmax and deltaPpli remained significant (r2 = 0.87). O2Pmax correlated with inspiratory muscle function (Pplimax, r = -0.58; Pdimax, r = 0.53; deltaPpli, r = -0.47; and PImax, r = -0.47). By multiple regression analysis, the predictors of O2Pmax were Pplimax and deltaPpli (r2 = 0.47). In very severe COPD, the impressive swings in intrathoracic pressure resulting from deranged ventilatory mechanics are the most likely cause of exercise limitation and reduced O2 pulse. The contributions of factors such as deconditioning, hypoxemia, and concurrent heart disease remain unknown.