Abstract
Background: The aspartate aminotransferase-to-platelet ratio index (APRi) has been proposed as a non-invasive biomarker of liver injury in biliary atresia (BA). However, conflicting evidence and varying cutoff values have left its association with histologically confirmed liver damage uncertain. This study assessed the relationship between APRi and liver injury severity (Meta-Analysis of Histological Data in Viral Hepatitis [METAVIR] F0–F4) in BA and evaluated its predictive value.
Methods: A retrospective analytical review of electronic records was conducted for all BA patients treated at a South African academic hospital between 01 January 2009 and 31 December 2019.
Results: Sixty-seven patients were included, 74.6% of whom were female. The most common subtypes were cytomegalovirus (CMV) immunoglobulin M-positive (IgM+) BA (34.3%) and isolated BA (31.3%); about one-third could not be classified because CMV serology was unavailable. Liver biopsy METAVIR scores were F1 (8.3%), F2 (50.0%), F3 (18.3%) and F4 (23.3%). In a sub-analysis (n = 39), APRi modestly differentiated ≤ F2 from ≥ F3 fibrosis at a cutoff of 2.69 (area under the receiver operating characteristic curve 0.61). Decision curve analysis suggested modest clinical utility despite a non-significant logistic regression model (odds ratio 1.25; 95% confidence interval 0.94–1.92; p = 0.2). Of the 55 patients with known outcomes, 94.5% (n = 52) had confirmed mortality or were referred for palliative care.
Conclusion: Aminotransferase-to-platelet ratio index should not replace established clinical decision-making but may provide useful adjunctive information alongside clinical, biochemical and histological assessment. These findings provide contextual evidence from a South African resource-limited setting and support prospective multicentre validation before routine clinical implementation.
Contribution: The study contributes evidence supporting continued evaluation of APRi as an adjunctive non-invasive biomarker in BA in the South African setting.