Background and aimsNon-invasive fibrosis assessment has been highly recommended in many liver diseases. However, comparative diagnostic accuracy of laboratory markers, ultrasound and magnetic resonance elastography (MRE) for fibrosis in autoimmune hepatitis (AIH) patients has not been established.MethodsMedline, Embase and Cochrane Library were searched. Primary outcome was significant fibrosis (SF), advanced fibrosis (AF) and cirrhosis, defined as Metavir stage F2, F3 and F=4 according to liver biopsy. Hierarchical summary receiver operating characteristic curve (ROC) model was used to evaluate diagnostic accuracy of non-invasive methods. Summary area under ROC (AUROC) and diagnostic odds ratio (DOR) with 95% confidence interval (CI) were calculated. The Grading of Recommendations Assessment, Development and Evaluation system was used to assess quality of evidence.ResultsOverall, 16 studies with 861 patients were included, comparing aspartate aminotransferase to platelet ratio index (APRI), fibrosis-4 index (FIB-4), aspartate aminotransferase/alanine aminotransferase ratio, transient elastography (TE), acoustic radiation force impulse, shear wave elastography and MRE versus liver biopsy. Among all non-invasive markers, TE had good performance for fibrosis staging. Summary AUROCs and DORs of TE were 0.90 (95% CI 0.87, 0.92) and 23.7, 0.91 (95% CI 0.89, 0.93) and 31.6, 0.89 (95% CI 0.86, 0.92) and 80.5 for staging SF, AF and cirrhosis, whereas APRI and FIB-4 showed poor performance for detecting AF (DOR, 4.6 and 4.7) and cirrhosis (DOR, 5.5 and 12.9).ConclusionsTE performs well to stage liver fibrosis in patients with AIH, compared with other laboratory non-invasive indexes. Nevertheless, diagnostic accuracy of APRI and FIB-4 is poor.