Postoperative hypoxaemia: comparison of extradural, i.m. and patient-controlled opioid analgesia.

Arterial oxygen saturation (SaO2) was analysed continuously before and for 24 h after lower abdominal surgery in 30 patients breathing air using one of three postoperative analgesic regimens: i.v. diamorphine using a patient-controlled analgesia system (PCAS), extradural diamorphine or i.m. morphine. Hypoxaemia was defined as SaO2 less than 94% for more than 6 min h-1. Before operation there was no difference between the three analgesia groups assessed by the duration when SaO2 was less than 94%. After operation the pattern of SaO2 vs time distribution was either stable, with little variation from hour to hour with no hypoxaemia, or unstable with large variation with 30% of patients hypoxaemic. Thus three patterns of SaO2 distribution were seen in the postoperative period: stable without hypoxaemia (4/10 PCAS, 0/10 extradural, and 1/10 i.m. patients), unstable without hypoxaemia (4/10 PCAS, 5/10 extradural and 7/10 i.m. patients) and unstable with prolonged nocturnal periods with SaO2 less than 94% for a mean of 17.7 min h-1, 95% confidence limits (CL) 10-25 min h-1, (2/10 PCAS, 2/10 i.m. and 5/10 extradural patients). Before operation, the unstable group with hypoxaemia spent longer at less than 94% SaO2 (mean 4.8 min h-1, 95% CL 1.0-8.6 min h-1) than the stable group (mean 0.4 min h-1, 95% CL 0.17-0.61 min h-1) and this was a predictor of postoperative hypoxaemia. Hypoxaemia occurred in all analgesia groups, but extradural diamorphine tended to cause longer periods. Some patients at risk of postoperative hypoxaemia may be predicted by preoperative monitoring of SaO2 although extradural diamorphine boluses were associated with hypoxaemia in patients with normal preoperative values.