The liver section of the British Society of Gastroenterology and the research unit of the Royal College of Physicians collaborated to set up a nationwide audit to investigate the practice of percutaneous liver biopsy in England and Wales. Each of 189 health districts in England and Wales was approached to provide a list of 10 consecutive percutaneous biopsies performed during 1991, and details of demographic data, indications, suspected diagnosis, investigations, biopsy technique, outcome, and influence on patient management were collected. Data were retrieved on 1500 (79%). The age distribution showed 6% of biopsies were done in those over 80 years of age and as many over 90 as under 10 years of age. Suspected malignancy and chronic liver disease each contributed one third of the indications. In 34% the procedure was carried out by radiologists under ultrasound image control. The remainder were done by general physicians and gastroenterologists, with the operator in the second group being more senior and experienced. The Trucut biopsy needle accounted for two thirds of biopsies, the remainder being the Menghini type. For both needles the samples were recorded as excellent or satisfactory in 83% and inadequate in only 5%. Bleeding complicated 26 procedures (1.7%), requiring transfusion in 11, and was commoner when clotting was impaired or serum bilirubin raised. There were two definite and three possible procedure related, given an overall mortality of 0.13-0.33%. The diagnosis made before biopsy was confirmed in 63% of patients, and the clinician found the biopsy helpful in treatment in 75%. Day case biopsy and techniques to reduce the risk of bleeding were surprisingly rare in this series, which has given a unique opportunity to examine everyday practice across a wide range of hospitals.