Noninvasive scores are poorly predictive of histological fibrosis in paediatric fatty liver disease

Objectives: Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease in children. Roughly a quarter of paediatric patients with NAFLD develop nonalcoholic steatohepatitis and fibrosis. Here, we evaluated the diagnostic accuracy of previously published noninvasive fibrosis scores to predict liver fibrosis in a large European cohort of paediatric patients with NAFLD. Methods: The 457 patients with biopsy-proven NAFLD from 10 specialized centers were included. We assessed diagnostic accuracy for the prediction of any (F >= 1), moderate (F >= 2) or advanced (F >= 3) fibrosis for the AST/platelet ratio (APRI), Fibrosis 4 score (FIB-4), paediatric NAFLD fibrosis score (PNFS) and paediatric NAFLD fibrosis index (PNFI). Results: Patients covered the full spectrum of fibrosis (F0: n = 103; F1: n = 230; F2: n = 78; F3: n = 44; F4: n = 2). None of the scores were able to accurately distinguish the presence of any fibrosis from no fibrosis. For the detection of moderate fibrosis, area under the receiver operating characteristic curve (AUROC) were: APRI: 0.697, FIB-4: 0.663, PNFI: 0.515, PNFS: 0.665, while for detection of advanced fibrosis AUROCs were: APRI: 0.759, FIB-4: 0.611, PNFI: 0.521, PNFS: 0.712. Fibrosis scores showed no diagnostic benefit over using ALT <= 50/ > 50 IU/L as a cut-off. Conclusions: Established fibrosis scores lack diagnostic accuracy to replace liver biopsy for staging of fibrosis, giving similar results as compared to using ALT alone. New diagnostic tools are needed for Noninvasive risk-stratification in paediatric NAFLD.