To determine the effects of exercise with expiratory flow-limitation (EFL) on systemic O(2) delivery, seven normal subjects performed incremental exercise with and without EFL at approximately 0.8 l s(-1) (imposed by a Starling resistor in the expiratory line) to determine maximal power output under control (W'(max,c)) and EFL (W'(max,e)) conditions. W'(max,e) was 62.5% of W'(max,c), and EFL exercise caused a significant fall in the ventilatory threshold. In a third test, after exercising at W'(max,e) without EFL for 4 min, EFL was imposed; exercise continued for 4 more minutes or until exhaustion. O(2) consumption (V'(O)(2)) was measured breath-by-breath for the last 90 s of control, and for the first 90 s of EFL exercise. Assuming that the arterio-mixed venous O(2) content remained constant immediately after EFL imposition, we used V'(O)(2) as a measure of cardiac output (Q'(c)). Q'(c) was also calculated by the pulse contour method with blood pressure measured continuously by a photo-plethysmographic device. Both sets of data showed a decrease of Q'(c) due to a decrease in stroke volume by 10% (p < 0.001 for V'(O)(2)) with EFL and remained decreased for the full 90 s. Concurrently, arterial O(2) saturation decreased by 5%, abdominal, pleural and alveolar pressures increased, and duty cycle decreased by 43%. We conclude that this combination of events led to a decrease in venous return secondary to high expiratory pressures, and a decreased duty cycle which decreased O(2) delivery to working muscles by approximately 15%.