Recertification for long-term oxygen therapy (LTOT) has been recommended for patients who are clinically unstable when home oxygen therapy is begun. Periods of observation for clinical stability have ranged from three weeks to three months in large multicenter clinical trials. There is concern, however, that an increase in arterial oxygen tension occurring after three months may be related to the beneficial effects of oxygen rather than to continued changes in clinical stability. In a review of 20 patients receiving transtracheal oxygen (TTO2) therapy, it was found that four (20 percent) did not qualify for oxygen therapy at the end of six months because the PaO2 breathing ambient air had increased to levels above 55 mm Hg. All patients were clinically stable at the time of insertion of the transtracheal catheter and all had been receiving nasal oxygen for at least seven months (mean, 25.8 months) before entering the study. A retrospective analysis of data published by Weitzenblum et al disclosed that four (25 percent) of 16 patients had a similar increase in PaO2 when reexamined after one year of oxygen therapy. All of the patients had been studied at least one year before oxygen therapy was initiated and each had three consecutive arterial blood gas measurements done monthly to ensure clinical stability. The increase in PaO2 to levels above 55 mm Hg observed in patients receiving TTO2 therapy was associated with a reduction in alveolar-arterial oxygen gradient; however, arterial oxygen desaturation with walking persisted. The specific mechanisms for improvement in PaO2 during oxygen therapy require further study. Any recommendation for recertification of LTOT must recognize that an increase in PaO2 after three months may be due to the beneficial effects of the oxygen therapy and does not provide prima facie justification for termination of therapy.