Tissue diagnosis of pancreas graft dysfunction is desirable. Bladder-drained pancreaticoduodenal transplants allow tissue diagnosis by cytoscopic biopsy procedures of the pancreas and duodenum. To assess the diagnostic utility of duodenal biopsies, we reviewed all cystoscopically obtained pancreas and duodenal biopsy tissues at our institution (July 1, 1989 through September 30, 1993). Adequate tissue for histologic examination was obtained from 75 biopsies in 58 recipients. Indications for cytoscopic biopsies were relative hypoamylasuria in 85%, hematuria in 6%, hyperamylasemia in 3%, and other causes in 6%. Duodenal specimens were available from 52 biopsies (25 with, and 27 without, concurrent pancreas biopsies). Of the 27 duodenal biopsies alone, 3 were diagnostic of rejection, 15 had features consistent with rejection, 6 were normal, 1 showed fibrosis, 1 showed necrosis, and 1 was ulcerated. Thus, two-thirds of the duodenal biopsies alone yielded clinically relevant information resulting in antirejection treatment. In 25 of the duodenal biopsies, pancreas tissue was also available (11 simultaneous pancreas-kidney, 9 pancreas transplant alone, and 5 pancreas after kidney recipients). Findings in both organs completely agreed in 9 (36%) of the biopsies. In 7 (28%), rejection was suggested or diagnosed in both organs, although the organs were discrepant with regard to the presence of vascular rejection (6 pancreas, 1 duodenum). In 2 (11%), minor nonrejection discrepant findings were present. Therefore, in 18 of 25 (72%) pancreas-duodenal biopsies, treatment would not have been different if only one graft had been biopsied. But in the other 7 (28%), treatment would have been different if only the organ with negative findings had been biopsied. In 6 cases (4 duodenal, 2 pancreas), rejection was seen in one organ but not the other. In 1 case, cytomegalovirus (CMV) inclusions were present in the duodenum, but the pancreas was normal. We conclude that (1) the duodenum and pancreas can reject independently of each other, and a negative biopsy does not preclude rejection of the other organ; (2) duodenal biopsies determined therapeutic decisions one-fifth of the time when both tissues were available for examination, and two-thirds of the time when only duodenal tissue was available; and (3) since cystoscopy allows easy access to the duodenum, both the pancreas and duodenum should be biopsied whenever possible; tissue samples of one organ alone are sufficient only with positive findings.