To verify if the improvement in gas exchange observed at rest with almitrine in patients with chronic air-flow obstruction is maintained at exercise, we studied 10 patients on a randomized, crossover, double-blind exercise protocol. After assessment of steady state of their disease, patients with FEV 1 less than 1.5 L.s-1 and PaCO2 less than 70 mmHg cycled 8 to 11 min at 80% of their previously determined maximal O2 uptake. Blood gases, arterial lactate (La), expired ventilation, (VE), O2 uptake (Vo2), Co2 production (VCO2), heart rate (HR), and arterial blood pressure (BP) were measured at rest, at mid-exercise, at end-exercise, and at 2, 14, 15, and 35 min of recovery. We found that at all stages of the study PaO2 was significantly higher in those who had received almitrine than in those who had received placebo; the mean increment was 11.9 torr at rest, 8.9 at end-exercise, and 13.7 at 15 min of recovery. Changes in PaCO2 and arterial (H+) were the opposite of those in PaO2; at end-exercise PaCO2 was 5.9 torr lower in those who had received almitrine, and pH was 7.36 versus 7.32 in those who had received placebo. No significant difference was found at all stages of the study between drug and placebo with respect to VE, VO2, VCO2, HR, and BP. But the ventilatory equivalent for oxygen, VE/VO2, was slightly higher in those who had received almitrine than in those who had received placebo during exercise.(ABSTRACT TRUNCATED AT 250 WORDS)