F Fu un nc ct ti io on na al l o ou ut tc co om me e o of f p pa at ti ie en nt ts s w wi it th h c ch hr ro on ni ic c o ob bs st tr ru uc ct ti iv ve e p pu ul lm mo on na ar rySixteen stable COPD patients (13 males and 3 females, aged 60±5 yrs, mean±SD) who had previously undergone pulmonary function tests and progressive exercise testing with arterial blood sampling at rest and maximal capacity, entered the study. At first evaluation (E1), subjects were normocapnic at rest (arterial carbon dioxide tension (Pa,CO 2 ): 4.9-5.7 kPa, (37-43 mmHg)) and all presented exercise-induced hypercapnia (end-exercise Pa,CO 2 >5.7 kPa (43 mmHg) with a minimal 0.5 kPa (4 mmHg) increase from resting value). The subjects were re-evaluated 24-54 months later (34±8 months) (second evaluation (E2)).At E2, forced expiratory volume in one second (FEV1) had decreased from 42±13 to 38±15% of predicted values, and mean resting Pa,CO 2 had increased from 5.2±0.3 to 5.7 + 0.4 kPa. Maximal exercise capacity (Wmax) decreased between E1 and E2 from 76±30 to 56±22 W. Even if Wmax was lower at E2, end-exercise Pa,CO 2 was higher than at E1 (6.6±0.8 vs 6.4±0.5 kPa). At E2, eight subjects presented resting hypercapnia (group H), whilst the others remained normocapnic (Group N). Group H subjects had higher Pa,CO 2 , at Wmax than Group N and lower Wmax than Group N at E2. Group H and N were not significantly different for physiological dead space/tidal volume ratio (VD/VT), FEV1, lung volumes and transfer factor of the lungs for carbon monoxide (TL,CO), both at E1 and E2.In half of the patients studied, exercise hypercapnia was a step in the progression of COPD towards resting hypercapnia, and was associated with severe exercise limitation. During exercise, patients who responded to a deterioration in their lung function by increasing minute ventilation remained normocapnic at rest, whilst those who did not increase their ventilation developed chronic hypercapnia at rest during the 2-4 year follow-up period.