OBJECTIVE To determine the incidence of hypoxaemia and the efficacy of oxygen therapy in the early postoperative period. METHODS Prospective, randomized study. METHODS Regional hospital. METHODS 89 patients undergoing general anaesthesia divided into two groups, I (n = 45) and II (n = 44). METHODS Patients of group I breathed room air while patients of group II received supplementary oxygen. RESULTS Oxygenation was monitored continuously for 60 minutes with a pulse oximeter after the end of surgery; hypoxaemia was graded into four values of SpO2: mild (86-90%), moderate (81-85%), severe (76-80%), extreme (< or = 76%). One or more episodes of mild hypoxaemia were recorded in 68% of patients in group I and in 36% of patients in group II (p < 0.005); moderate hypoxaemia were recorded in 51% of patients in group I compared with 13% of patients in group II (p < 0.001). Episodes of severe or extreme hypoxaemia were recorded in 22% and in 6% of patients in group I; no patients in group II exhibited such small values of saturation. The smallest recorded SpO2 values were 81 +/- 5.94% in group I and 89 +/- 8.63% in group II (p < 0.001). CONCLUSIONS There is a high incidence of clinically unsuspected hypoxaemia in the early postoperative period, the use of supplemental oxygen reduces but does not prevent hypoxaemic episodes, every patient should be monitored continuously with a pulse oximeter in the recovery room.