Predictors of hepatic steatosis in living liver donors

Fatty liver disease (steatosis) is considered a risk factor in donor liver transplantation (LT). Macrosteatosis (>50%) is associated with primary graft dysfunction and may reduce long-term recipient survival.Objective: to identify predictors of macrovesicular steatosis (>50%) by analyzing donor characteristics.Materials and methods. The retrospective study included 525 potential liver donors between January 1, 2019 and December 31, 2020. Clinical and morphological characteristics of donors were studied using logistic regression and receiver operating characteristic (ROC) analysis. Threshold values of parameters demonstrating statistical significance in multivariate analysis as predictors of >50% hepatic steatosis were obtained by ROC analysis based on calculation of the optimal cutoff point.Results. Diabetes mellitus (DM), cause of donor’s death (traumatic brain injury), alanine transaminase (ALT) >90 units/L and aspartate transaminase (AST) >110 units/L were predictors of >50% steatosis, revealed by time-zero biopsy in the donor. Almost identical sensitivity and specificity indicators were determined in ROC analysis for liver enzymes – ALT and AST – which were 69.1 and 80.6; 72.2 and 81.1, respectively. Given the obtained values, we can say that with elevated levels of liver enzymes in the donor’s blood, there is a high degree of probability of liver parenchymal damage, but low sensitivity indicates possible multifactoriality of liver damage, and fatty liver disease may be one of the factors, but there may also be no damage to the liver parenchyma. At the same time, the rather high specificity revealed in ROC analysis for liver enzymes is a reliable sign of the absence of fatty liver disease at enzyme values less than the threshold.Conclusion. The thresholds established for ALT and AST and their corresponding levels of sensitivity and specificity indicate that these parameters have a relatively low predictive level in the context of the presence of severe fatty liver disease in a donor. This allows, nevertheless, to use models built on their basis as screening models in the primary evaluation of liver donors.