Endobronchial valve placement improves pulmonary function in some patients with chronic obstructive pulmonary disease, but its effects on exercise physiology have not been investigated. In 19 patients with a mean (SD) FEV(1) of 28.4 (11.9)% predicted, studied before and 4 weeks after unilateral valve insertion, functional residual capacity decreased from 7.1 (1.5) to 6.6 (1.7) L (p = 0.03) and diffusing capacity rose from 3.3 (1.1) to 3.7 (1.2) mmol . minute(-1) . kPa(-1) (p = 0.03). Cycle endurance time at 80% of peak workload increased from 227 (129) to 315 (195) seconds (p = 0.03). This was associated with a reduction in end-expiratory lung volume at peak exercise from 7.6 (1.6) to 7.2 (1.7) L (p = 0.03). Using stepwise logistic regression analysis, a model containing changes in transfer factor and resting inspiratory capacity explained 81% of the variation in change in exercise time (p < 0.0001). The same variables were retained if the five patients with radiologic atelectasis were excluded from analysis. In a subgroup of patients in whom invasive measurements were performed, improvement in exercise capacity was associated with a reduction in lung compliance (r(2) = 0.43; p = 0.03) and isotime esophageal pressure-time product (r(2) = 0.47; p = 0.03). Endobronchial valve placement can improve lung volumes and gas transfer in patients with chronic obstructive pulmonary disease and prolong exercise time by reducing dynamic hyperinflation.