Alanine aminotransferase influencing performances of routine available tests detecting hepatitis B-related cirrhosis.

ABSTRACT

The performances of routine tests such as FIB-4 and APRI in detecting cirrhosis and significant fibrosis in chronic hepatitis B (CHB) have been shown to be discrepant between studies. Novel tests such as red cell distribution width-platelet ratio (RPR), γ-glutamyl transpeptidase to platelet ratio (GPR) and easy liver fibrosis test (eLIFT) are introduced recently. To evaluate the aminotransferase influence on the performance of these routine tests, a total of 1005 CHB patients who underwent liver biopsies and routine tests were retrospectively analysed. The diagnostic cut-offs referring to likelihood ratio were determined for excluding or including cirrhosis diagnosis and also for ruling in significant fibrosis diagnosis. The performances of RPR, FIB-4, eLIFT and APRI in detecting cirrhosis seemed improved at higher ALT levels, while GPR was conversely impaired. The likelihood ratio was âˆ for APRI cut-off 2 diagnosing cirrhosis in ALT < 2 upper limit of normal (ULN), 14.6 for APRI cut-off 1.5 determining significant fibrosis in ALT ≤ 5ULN and 20.6 for FIB-4 cut-off 3.2 diagnosing ≥ F3 in the total cohort, respectively. The optimal cut-offs for cirrhosis diagnosis were increased with higher ALTs by tests which included aminotransferase, but not for RPR. The proportions of patients classified as having cirrhosis or no cirrhosis stratified by ALT level cut-offs were superior. Stepwise applying RPR, GPR and eLIFT would determine 60% of patients as having cirrhosis or no cirrhosis with an accuracy of 93.0%. In conclusion, the performance of aminotransferase comprising tests in detecting cirrhosis in CHB were influenced by ALT levels. Thus, ALT stratified cut-offs may be a preferred alternative. In resource-limited settings, stepwise applying routine tests could be recommended as a preferred measurement for cirrhosis detection.