An analysis of the determinants of hepatic allograft rejection using stepwise logistic regression.

Hepatic allograft rejection is currently diagnosed using both histologic and clinical criteria. The purpose of this study is to extract data from the allograft biopsy, which can establish the diagnosis of rejection with more precision. A total of 566 allograft biopsies were examined in 56 patients following orthotopic liver transplantation, and 35 variables were examined. Using stepwise logistic regression, only six of these variables affected the diagnosis of rejection. These were portal tract spillover, portal tract eosinophilia, portal vein endothelialitis, portal tract neutrophilia, central vein endothelialitis, and cholestasis. Coefficients of these variables were determined, and could be used to calculate the probability of rejection for a given biopsy. Using this model the probability of having rejection can be calculated using histologic data alone.

Duke Scholars

Author John Wiley Williams Jr. Medicine, General Internal Medicine