Predictors of graft and patient survival in hepatitis C virus (HCV) recipients: model to predict HCV cirrhosis after liver transplantation

Abstract

Background: Hepatitis C virus (HCV) recurrence after liver transplantation (LT) is almost universal, but the natural history of recurrent HCV in the allograft is highly variable. Our study had two aims: 1) to assess the impact of different pre- and postLT factors on graft and patient survival in HCV transplant recipients and 2) to create a model which may predict the patients at risk for HCV-related graft cirrhosis at 5 years postLT.

Methods: A total of 168 LTs were considered for this study. Univariate and multivariate Cox proportional hazards regression model was used, as well as logistic regression analysis to create a model of prediction of HCV cirrhosis within 5 years after LT.

Results: Predictive factors for both decreased graft and patient survival included patients recently transplanted (2000-2004), induction without azathioprine, short-term therapy with mycophenolate mofetil and prednisone (< or =6 months), presence of early cholestasis, histologically proven early recurrence of hepatitis C. Recipient human leukocyte antigen DR3 positivity, presence of early cholestasis, and donor age >50 years were identified as independent predictors of graft cirrhosis within 5 years. A predictive model was established in order to calculate at 6 months a risk score for graft HCV cirrhosis within 5 years postLT using a formula that included the identified independent predictors. The area under receiver operating characteristic curve was 0.83, indicating a good ability to predict medium-term HCV allograft cirrhosis.

Conclusion: This model may be a useful tool for better identifying high-risk HCV patients who should be selected for early initiation of antiviral therapy.