To find out which prognostic factors were important in predicting postoperative mortality and length of hospital stay in patients with perforated peptic ulcers. Retrospective study. Teaching hospital, Lund, Sweden. 246 patients with perforated peptic ulcer who presented between January 1974 and December 1992. Cox proportional hazards analysis. Influence of age, sex, coexisting disease, duration of symptoms, site of perforation and operative technique on mortality and length of hospital stay. Age over 75 years (p = 0.002), coexisting cardiac or pulmonary disease (p = 0.02), perforation of the cardia or body of the stomach (p = 0.02), lapse of more than 12 hours between start of symptoms and operation (p = 0.006) and type of operation (p < 0.0001) had a significant influence on hospital mortality. Age over 75 years (p < 0.0001) and lapse of more than 12 hours between start of symptoms and operation (p = 0.03) significant influenced the likelihood of a prolonged stay in hospital. Patients with perforated peptic ulcers should be operated on as soon as possible. Simple closure is simple and safe with relatively low mortality and short stay in hospital.