Abstract
Liver transplantation (LT) is a therapeutic option for patients suffering from end-stage liver disease. Recent research has probed the prognostic significance of biomarkers to predict graft function and mortality post-transplant, yet few candidates are recommended in clinical practice. We employed a pipeline that integrates meta-analysis (PRISMA 2020), followed by Kaplan-Meier (KM)-based individual patient data (IPD) analysis, aiming to identify potential novel prognostic biomarker panels for LT recipients. Ovid Medline, Embase, and Cochrane databases were searched. Twenty-one prognostic and 8 diagnostic studies were eligible, pooling 34,922 patients. Single biomarkers sampled at an early stage (≤15 d after LT) were significantly associated with graft-related outcomes (HR/OR 0.95 [0.94-0.97]) but did not predict mortality (HR/OR 1.00 [0.97-1.04]) or composite outcomes (HR/OR 1.02 [0.98-1.07]). Biomarkers in combination (GGT/bilirubin ratio, ALT+AST or ALT+AST+bilirubin+INR) predicted composite outcomes (graft failure or mortality, aHR/aOR 4.37 [2.65-7.21]). Biomarkers assessed at late stage (>15) did not show association with mortality (HR/OR 1.02 [1.00-1.04]) or composite outcomes (HR/OR 1.00 [0.99-1.01]). KM-based IPD analysis showed that coagulation factor V combined with ALT predicted graft survival (HR 2.12 [1.44-3.12]), and coagulation factor V+insulin-like growth factor 1 stratified the risk of patient survival (HR 2.97 [1.79-4.91]). Therefore, we were able to compare various scoring systems in predicting graft-related outcomes and mortality following LT. Additionally, we identified novel combinations of biomarkers that exhibited prognostic value for LT patients. Finally, we demonstrate that combined analytical tools for assessing large clinical datasets effectively evaluate multi-modal markers for risk stratification of early and late outcomes for LT.