In 159 chronic obstructive pulmonary disease (COPD) patients (139 males, mean age 62 +/- 8 yrs, arterial oxygen tension (PaO2) 7.2 +/- 0.9 kPa), on long-term oxygen therapy (LTOT), we evaluated the effects of portable oxygen therapy both on the daily duration of oxygen therapy and on daily activities. They were given two types of LTOT at random: group A (n = 75), oxygen concentrators only (OC); group B (n = 84), either small oxygen cylinders plus OC (B1 = 51) or liquid oxygen (B2 = 33). The patients were followed-up for one year by means of: a) medical examination every three months; b) monthly home interviews concerning the daily duration of oxygen therapy, the utilization of the devices and the daily activities of the patients; c) a measurement of the daily oxygen usage. The results show that: 1) there are no significant clinical and functional differences between groups A and B at the onset of and throughout the study; 2) in group B the daily use of oxygen therapy is significantly longer than in group A (17 +/- 3.5 h.day-1 vs 14 +/- 3 h.day-1, p less than 0.01) without any difference between groups B1 and B2; 3) outdoor walking activities are different between groups A and B, at least in those patients using oxygen more than 18 h.day-1. Only 60% of patients in group B (55% of B1; 67% of B2) use their portable devices outdoors and for walking. No strict predictive criterion of this use is found in our study.(ABSTRACT TRUNCATED AT 250 WORDS)