Introduction: Prediction of mortality after liver transplantation is important in determining the utility of transplantation and is therefore necessary in estimating the survival benefit of liver transplantation.  Current models of post-transplant survival are imprecise.  Inflammatory markers have been shown to predict outcome of various diseases, including cirrhosis and hepatocellular carcinoma.  We hypothesised that c-reactive protein (CRP) measured before liver transplantation is predictive of post-transplant mortality. 
Methods: The prospectively maintained database of the Liver Transplant Unit Victoria was retrospectively analysed.  The study included adult liver transplants performed between 1 June 2001 and 7 November 2011.  Variables assessed included those known to be associated with post-transplant outcome, including recipient age, disease, MELD score, donor risk index, year of transplant, as well as the most recent CRP level prior to transplantation.  The outcome measure was post-transplant mortality.  ROC curves were generated for continuous variables, cut-off levels were determined using Youden’s J statistic, Kaplan-Meier curves were generated, dividing the population by the cut-off levels and outcomes were compared using log-rank. Cox regression was used to create a model of independently significant variables that predict post-transplant mortality.
Results: A total of 596 patients were included in the study.  Survival was significantly lower in patients whose pre-transplant CRP was ≥ 13.25 mg/L (1-, 5-, 10- and 15-year survival 90%, 82%, 75% and 69%, respectively) than those whose pre-transplant CRP was < 13.25 mg/L (1-, 5-, 10- and 15-year survival 94%, 90%, 84% and 82%, respectively, P = 0.008).  Independent predictors of post-transplant mortality included fulminant hepatic failure (OR 1.037, 95% CI 1.012-1.062, P = 0.034), recipient age (OR 1.037, 95% CI 1.012-1.062, P = 0.003), donor risk index (OR 1.770, 95% CI 1.240-2.526, P = 0.002), earlier year of transplant (OR 0.922, 95% CI 0.876-0.970, P = 0.002) and CRP (OR 1.007, 95% CI 1.004-1.011, P < 0.001).
Discussion: This study has confirmed known predictors of mortality after liver transplantation and has demonstrated that CRP is also predictive of mortality.  This is a single centre retrospective study, with the inherent limitations associated with such studies, and requires confirmation with large, prospective studies. 
Conclusion: CRP before liver transplantation is an independent predictor of post-transplant mortality and can be included in models that assess the utility of liver transplantation.