A retrospective review of 184 pancreatic biopsies in 178 patients was performed to assess the prevalence of severe postbiopsy pancreatitis. The size, contour, and pathology of the lesions biopsied; the course of the needle (i.e., through bowel or other viscera); the size of the needle; the number of needle passes made; and the guidance technique used were analyzed. Severe pancreatitis developed in five cases (five patients) 5/184 or 3% of the biopsies), usually within 24-48 hr. Three of the five patients who developed pancreatitis had true-negative biopsies (normal pancreas) proved either at surgery (two) or at clinical follow-up (one). The diagnoses for the two remaining patients were adenocarcinomas. In four of the five patients, the diagnosis of severe pancreatitis was made by inspection at surgery. The fifth case was diagnosed by CT. Three patients who underwent surgery and one patient who had percutaneous drainage recovered from the pancreatitis. The fifth patient died despite surgical intervention. All five patients with pancreatitis had masses 3 cm or smaller as compared with the overall group, in which 71 (39%) of the 184 biopsies performed had masses smaller than 3 cm. Overall, 18% of the biopsies were true negative, compared with the 60% true-negative rate in the pancreatitis group. The bowel was transgressed in 21% of all 178 patients, including three of the five pancreatitis patients. We conclude that although the risk of pancreatitis is exceedingly small in percutaneous needle biopsies, it may occur, and at a higher rate than previously reported.