Low-flow continuous oxygen can lead to significant improvement in exercise capacity in selected patients with stable hypoxemic pulmonary disease. Although the mechanisms of improvement are incompletely understood, two tenable hypotheses are (1) the relief of hypoxic pulmonary vasoconstriction and (2) improved peripheral oxygen delivery. This prospective study was performed to examine these two hypotheses. Stable patients with hypoxemic lung disease performed symptom-limited ergometry with hemodynamic monitoring before and after greater than one month of long-term home oxygen therapy. There were 22 patients who increased exercise capacity by greater than 100 KPM/min and greater than 4 min duration and 11 patients who did not increase their exercise capacity greater than 2 min or in KPM/min. Neither the responders nor the nonresponders had significant changes in pulmonary pressures or resistances or effective arterial volume elastances at exercise after long-term O2 therapy. The responders had significant increases in oxygen delivery during exercise (148 +/- 33 ml/min vs 184 +/- 51 ml/min, p = 0.004), cardiac output (7.7 +/- 1.6 L/min vs 8.7 +/- 2.1 L/min, p = 0.003), and arterial oxygen content (18.6 +/- 2.6 ml/dl vs 20.2 +/- 2.8 ml/dl, p = 0.02). The nonresponders had no significant changes in these parameters during exercise. Increased exercise capacity in response to long-term O2 therapy is associated with increased O2 delivery not relief of hypoxic vasoconstriction (in terms of pressure or resistance or arterial elastance). Increased O2 delivery can accrue from both increased cardiac output and increased arterial O2 content. Increase in arterial O2 content is unique to O2 relative to all vasodilator drugs.