Background.Although the value of duodenal histology as a means to diagnose acute rejection in pancreaticoduodenal allografts has been validated, it is not known how the duodenum responds to antirejection treatment in comparison with the pancreas.Methods.Diabetic Lewis rats received a pancreaticoduodenal allograft. Cyclosporine was given for 5 days and then discontinued for 2 days (group 1), for 4 days (group 2), for 6 days (group 3), for 8 days (group 4), for 9 days (group 5), and for 10 days (group 6). Two animals of each group were killed for histology at the end of immunosuppressive-free intervals. In the remaining rats, rejection was treated with methylprednisolone on 3 consecutive days. Duodenal histology was compared with pancreatic morphology before and after treatment of rejection.Results.Duodenal histology had a positive and negative predictive value of 100% for detection of acute rejection in the pancreatic portion of the graft. After antirejection treatment, duodenal morphology was however less accurate (positive predictive value, 96%; negative predictive value, 67%). The Spearman correlation coefficient (p) of duodenal and pancreatic rejection grades was higher before antirejection treatment (p=1.0) than thereafter (p=0.724). Considering interstitial and vascular changes separately, vascular rejection correlated to a higher extent than interstitial rejection between the two portions of the graft (p=0.725 vs.p=0.677).Conclusions.Duodenal histology accurately predicts the initial diagnosis of rejection of the pancreas. However, after treatment of acute rejection, duodenal morphology is more likely to recover from rejection than the pancreas. Awareness of this phenomenon might be important for the interpretation of duodenal follow-up biopsies.