The effects of the methylxanthine drugs doxofylline, theophylline and bamifylline were investigated on the basis of clinical and gasometric parameters in hypoxic patients with chronic heart failure. A parallel, double-blind, randomized study was conducted in 48 in-patients with NYHA II-IV chronic heart failure with normo- or hypercapnic hypoxaemia. They were divided into three groups and then submitted to a 4-day run-in with placebo. Either doxofylline 800 mg b.i.d., theophylline slow-release 400 mg b.i.d. or bamifylline 1200 mg b.i.d. were administered orally in each group of 16 patients for 10 days. Exercise capacity was estimated through NYHA class modification. Gasometric determinations, including arterial oxygen tension (PaO2), carbon dioxide tension (PaCO2) and oxygen saturation (SaO2), were measured from arterial blood samples at the time of enrollment (T-4), at the onset of xanthine therapy (T0) and at the end of the trial (T9). After 10 days' treatment, the NYHA class was found to be diminished in 50% of the doxofylline group, 50% of the bamifylline group and 44% of the theophylline group. PaO2 showed a > 15% increase in 75% of the doxofylline group, 56% of the theophylline group and 43% of the bamifylline group (responders). In all three groups the responders presented a highly significant enhancement in PaO2 and SaO2 (p < 0.01 T0 vs T9). Doxofylline exhibited the highest percent increase in PaO2 and SaO2 with respect to T0. The effects on cardiac rhythm showed a progressive heart-rate reduction in the doxofylline group, whereas patients receiving theophylline presented an increase rate of beating. In conclusion, the use of methylxanthines in patients with chronic heart failure seems to be particularly effective especially when a significant ventilatory dysfunction is present. Doxofylline appears to be specially useful because of its ability not to interfere with cardiac rhythm.