Published in:
01-01-2019 | Original Article
Systematic review: diagnostic accuracy of non-invasive tests for staging liver fibrosis in autoimmune hepatitis
Authors:
Shanshan Wu, Zhirong Yang, Jialing Zhou, Na Zeng, Zhiying He, Siyan Zhan, Jidong Jia, Hong You
Published in:
Hepatology International
|
Issue 1/2019
Login to get access
Abstract
Background and aims
Non-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.
Methods
Medline, Embase and Cochrane Library were searched. Primary outcome was significant fibrosis (SF), advanced fibrosis (AF) and cirrhosis, defined as Metavir stage F ≥ 2, F ≥ 3 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.
Results
Overall, 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).
Conclusions
TE 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.