The benefit of hepatic surgery for benign or malignant conditions is a balance between peri-operative morbidity/mortality and long-term potential for cure or palliation. The aim of this retrospective study was to illustrate that the safety of liver resection is a function of the frequently of performance of the procedure. Between 1973 and 1992, 327 elective liver resections were performed. The indication for surgery was malignant tumour in 275 cases of which 170 (62%) and 105 (38%) were for metastatic and primary disease, respectively, and non-malignant conditions in 52 cases. The series included the complete spectrum of hepatectomies. There were nine deaths (2.7%). Mortality was 8% (3/38) before 1985, 3.4% (3/89) between 1985-88 and 1.5% (3/200) between 1989-92. In non-jaundiced/non-cirrhotic patients, mortality was 1.4% (4/270). Morbidity, defined as the incidence of at least one major complication, occurred in 87 patients (26.6%) with a re-operation rate of 6.4%. During the same time periods, the morbidity rate was 42, 35 and 20%, respectively, and the median blood transfusion requirement and postoperative stay progressively decreased to 2 units and 9 days, respectively. In conclusion, as experience was gained, the need for blood transfusion diminished, morbidity and mortality improved and the hospital stay shortened.