Background: Liver graft viability assessment has long been considered a limit of hypothermic oxygenated machine perfusion (HOPE). Aim of this study was assessing correlations of easily available perfusate parameters (PP) (aspartate aminotransferase, alanine aminotransferase, lactate dehydrogenase, glucose, lactate, and pH) with graft features and outcome.
Methods: In the period October 2018-February 2020, perfusate samples were obtained every 30 minutes during 50 dual-HOPE (D-HOPE) procedures. Correlations of PP with graft factors, 90-day graft loss, early allograft dysfunction (EAD), L-GrAFT score, acute kidney injury, and comprehensive complication index were analyzed using Pearson coefficient, receiver-operating characteristics analysis and by univariable and multivariable regression.
Results: Median D-HOPE time was 122 minutes. All parameters were normalized to liver weight. Only macrovesicular steatosis (MaS) significantly impacted PP levels and slope. Grafts with ≥30% MaS exhibited significantly different PP values and slope. Graft loss and EAD rate were 2% (n = 1) and 26% (n = 13). All PP except lactate correlated with EAD, 90-minute alanine aminotransferase showing the highest area under the receiver-operating characteristics curve (0.84). However, at multivariable analysis, the only factor independently associated with EAD was MaS (odds ratio, 5.44; confidence interval, 1.05-28.21; P = 0.04). Ninety minutes lactate dehydrogenase had the strongest correlation with L-GrAFT (R = 0.70; P < 0.001). PP correlated poorly with comprehensive complication index and grades 2-3 acute kidney injury rate.
Conclusions: PP were predictive of graft function after transplant, but their association with graft survival and clinical outcomes requires further evaluation. MaS influenced levels of PP and was the only independent predictor of EAD.