Forty otherwise healthy patients (29 women and 11 men), undergoing elective cholecystectomy, were randomly allocated to be ventilated during the operation either with a positive end-expiratory pressure of 1 kPa (10 cmH2O) (PEEP group) or with intermittent positive pressure ventilation without PEEP (control group). During the operation the mean arterial oxygen tension (PaO2) in the PEEP group increased from 14.6 to 16.5 kPa, while no changes occurred in the control group (13.5 and 13.6 kPa). On the first postoperative day, PaO2 decreased by 12% of the preoperative values in the PEEP group; the decrease was 20% in the control group. On the third postoperative day, the PaO2 in the control group was still 9% below the preoperative values, but on the fifth day, both groups had reached their preoperative PaO2 values. In the postoperative period, no statistically significant difference in PaO2 could be demonstrated between the groups. Determinations of the forced vital capacity and forced expiratory volume in the first second showed no difference between the groups pre- or postoperatively. The present study demonstrated no clinically relevant beneficial effect of peroperative PEEP ventilation on the postoperative arterial hypoxaemia after an upper abdominal laparotomy.