There is a continuing debate on the role of pulmonary hypertension in the natural history of COPD. The aim of the study was to compare 2 groups of COPD pts, one investigated before domiciliary oxygen treatment era (Group I) with the other benefiting from LTOT (Group II). Both were followed-up for 5 years. METHODS pure COPD, PaO2 < 60 mmHg, PAP > 20 mmHg. Group I consisted of 16 pts (13M, 3F), mean age 55 +/- 9y. On initial evaluation their FEV1 averaged 0.93 +/- 0.3L and PaO2 54 +/- 6 mmHg. Their mean pulmonary arterial pressure (PAP) was 37.9 +/- 9 mmHg. During five years 12 patients (75%) died. Mean survival time of pts who died was 18.9 +/- 14.7 months. Group II consisted of 34 pts (27M, 7F) mean age 56 +/- 8y (NS vs group I), FEVI 0.83 +/- 3L (NS). PaO2 53 +/- 6 mmHg (NS), PAP 30.9 +/- 8.5 mmHg (p < 0.02). During five years 26 pts (76.5%) died. Mean survival time was 25 +/- 16 months (NS). Mean oxygen breathing time was 14h/ day. The Cox's survival analysis did not show difference between patients group I and II. We conclude that high initial PAP in group I patients did not affect survival compared to patients with similar severity of respiratory failure and lower PAP. The finding that LTOT did not improve survival may depend on many other variables not included into analysis.