Evaluation of serum cytokeratin-18 as prognostic parameter for acute liver failure

Background and aims: Acute liver failure (ALF) is a devastating clinical syndrome associated with a high mortality rate. In many cases, the etiology remains unknown and despite clinical criteria, it is challenging to pinpoint individuals with ALF that are able to survive without liver transplantation (LTx). The aim of the study was to determine whether cell death markers have the potential to predict the clinical course of ALF and/or may be utilized to differentiate etiologies.

Methods: In a prospective monocenter study (11/2006–09/2008), we recruited 52 consecutive ALF patients (56% females/44% males), who met the King's College criteria. Clinical and laboratory data were collected for a period of 4 weeks after study admission or until discharge/death of the individual patient. Serum cytokeratin-18 (CK-18) was quantified by employing the M65 (total) and M30 (cleaved caspase-3) antibody ELISA Kits. Peak values were used for statistical analysis.

Results: Total CK-18 (M65) was significantly higher in deceased (27,686±7,232 U/L) compared to patients who recovered spontaneously (15,534±2,187 U/L;p<0.001). A weaker correlation was found for the AST/ALT ratio (1.08±0.65 (restitutio) vs. 2.22±0.83 (deceased);p<0.05). Bilirubin, transaminases, GGT, serum bile acid levels and M30 were of no prognostic value. Utilizing aGST and M30/M65 ratio, we were able to discriminate acute HBV infection from acetaminophen intoxication (AI), other toxic causes and secondary liver failure following congestive heart failure with a high specificity. Especially an increased M30/M65 ratio, indicating a high apoptosis rate, was attributable to acute HBV infection, but not to acetaminophen intoxication or secondary liver failure.

Conclusions: This study on 52 patients with ALF of different causes revealed that M65 is a suitable marker for survival, and thus might alleviate the allocation of donor livers and patient management. Assessing the apoptosis rate, we are currently working on a study protocol to establish a clinical algorithm to determine the etiology in cases where anamnestic information remained ineffective (19% of 52 patients). Given the low incidence of ALF, we suggest multicenter studies to validate these results and to confirm these parameters on a larger cohort of patients.