We have studied the histopathologic correlates of a significantly decreased urinary amylase excretion rate (UAER) to determine its reliability in predicting the presence of cellular rejection within pancreas allografts drained via a duodenocystostomy. Significant hypoamylasuria in pancreas allograft recipients was defined as a diminution of greater than 50% in UAER sustained for greater than 36 hr and not associated with a decrease in serum amylase activity. We observed 18 such episodes of hypoamylasuria in 13 of 18 patients receiving pancreas allografts. Pancreaticoduodenal material was obtained during 11 of these episodes, one attempt failed, and for the remaining 6 episodes we obtained 3 renal allograft biopsy specimens. Histopathologic examination of the 14 specimens revealed cellular rejection in 9 (64%), fibrosis in 2 (14%), enzymatic necrosis in 1 (7%), cytomegaloviral pancreatitis in 1 (7%), and no abnormal features in 1 (7%). During these 14 episodes, a genetically identical renal allograft was present for 11 and showed signs of dysfunction in 9; however, the pancreatic histologic features suggested rejection in only 7 of the 9. Thus even the combination of hypoamylasuria and renal dysfunction in recipients of genetically identical organs was not fully reliable in predicting pancreas allograft rejection. In addition, the interval between organ implantation and onset of hypoamylasuria did not predict the histologic diagnosis. As with other solid-organ allografts, biopsy is a useful adjuvant for determining patient management in the presence of organ dysfunction.