One hundred seventeen patients with severe chronic obstructive pulmonary disease whose PaO2 in a stable clinical state ranged from 41 to 59 mmHg, were included in a multicentric controlled study on the effects of long-term O2 therapy (LTO), and the results of the 3-month probationary period are given here. Arterial blood gases were checked every month in order to assess the stability of hypoxemia. Time 0 represented the beginning and T3 the end of the probationary period. Patients whose PaO2 at T3 was still between 41 and 59 mmHg could be accepted for the long-term study (Group A), whereas patients whose PaO2 had increased over 59 mmHg were excluded (Group B). Only 77 patients could be followed up regularly from T0 to T3 (40 patients were lost to follow-up, missed some visits, or suffered from acute exacerbations of the disease) and 23 of them had a PaO2 at T3 that exceeded 59 mmHg (Group B), allowing the inclusion of 54 patients (Group A) in the long-term study. Patients in Groups A and B did not differ with regard to pulmonary volumes, PaCO2 at T0 or a past history of right heart failure. Arterial O2 tension at T0 was slightly and significantly (p less than 0.02) lower in Group A (51.3 +/- 5.5 mmHg) than in Group B (54.5 +/- 3.8 mmHg), but there was an important overlapping of individual results from one group to another.(ABSTRACT TRUNCATED AT 250 WORDS)