Development and Internal Validation of a Novel Pre-Transplant Biomarker Panel to Predict Post-Transplant Mortality in Liver Transplant Recipients

Abstract Background and Aims Current scoring systems accurately predict risk of pre-liver transplant (LT) mortality but fall short in estimation of post-LT survival. This study seeks to identify biomarkers of pre-LT immune dysfunction that predict mortality following LT. Methods From 10/1/13 − 3/23/21, 279 cirrhotic patients underwent assessment of plasma biomarker (Luminex) and clinical variables immediately prior to LT (T 0 ). Results Cox-proportional hazards modeling identified HCV IgG, Fractalkine, and MMP3 as multivariate predictors of 1-year mortality, with covariate selection by clinical importance and LASSO methodology. These were utilized to comprise the novel Liver Immune Frailty Index (LIFI), which stratifies recipients into -low, -moderate, and –high risk tertiles. One-year mortality was 1.4%, 12.7%, and 58.3% for LIFI-low, -moderate, and -high, respectively. Internal validation through bootstrap resampling with 2000 replicates demonstrates LIFI predicts early post-LT mortality with C-statistic = 0.84 and Brier score of 0.04. LIFI-high is not dependent upon HCV status. The relative contribution of the continuous variables in the LIFI calculation exceeded the discrete HCV IgG binary variable contribution. Stratification into either LIFI-high or moderate requires cumulative contribution of elevated MMP3 and Fractalkine levels. Excluding HCV IgG + as a covariate similarly stratifies patients at high-, moderate-, and low-risk of early futility after LT based only on MMP3 and Fractalkine levels (C-statistic 0.83). Conclusions LIFI may identify patients at risk for persistent severe immune dysfunction and early mortality following LT.