RESUMO
BACKGROUND & AIMS: The current definition of the pattern of liver injury in hepatotoxicity (DILI) is given by the R (ratio) value, dividing alanine aminotransferase (ALT) and alkaline phosphatase (ALP) in upper limits of normal at DILI onset. We aimed to explore the validity of using aspartate aminotransferase (AST) and gamma-glutamyl transpeptidase (GGT) as biomarkers of hepatocelullar and cholestatic damage, respectively, when calculating the R value. METHODS: Clinical, laboratory and histological data from 588 DILI episodes included in the Spanish DILI Registry were analyzed. Linear regression analysis was performed to establish the most appropriate cut-off points for hepatocellular and cholestatic patterns when calculating R with AST and GGT. RESULTS: The overall agreement between ALT/ALP and AST/ALP was 76%, with 96%, 61% and 41% agreement in the hepatocellular (R ≥ 5), cholestatic (R ≤ 2) and mixed groups respectively (P < 0.001). Classified by the causative drug, the agreement was higher (87-95%) among drug classes that mainly present with hepatocellular damage and lower (48-58%) for those in which cholestatic-mixed injury predominate (P < 0.001)). The overall agreement between ALT/ALP and ALT/GGT was weak (59%), except for in hepatocellular cases that showed a good agreement (94%) (P = 0.001). Pattern of injury according to liver histology demonstrated 65%, 68% and 47% agreement for ALT/ALP, AST/ALP and ALT/GGT ratios respectively. CONCLUSIONS: AST can reliably replace ALT when calculating pattern of liver injury in DILI, while GGT can only substitute ALP when the R value scores as hepatocellular. The biochemical signature of causative drugs does influence the validity of the ratios with AST or GGT.