Enhanced liver fibrosis test for the non-invasive diagnosis of fibrosis in patients with NAFLD: A systematic review and meta-analysis

Yasaman Vali, Jenny Lee, Jérôme Boursier, René Spijker, Jürgen Löffler, Joanne Verheij, M. Julia Brosnan, Zsolt Böcskei, Quentin M. Anstee, Patrick M. Bossuyt, Mohammad Hadi Zafarmand, Michael Pavlides (+16 others)
2020 Journal of Hepatology  
The Enhanced Liver Fibrosis (ELF) test is a non-invasive biomarker, suggested as an appropriate test for advanced liver fibrosis in non-alcoholic fatty liver disease (NAFLD). This systematic review aimed to provide summary estimates of the accuracy of this test against biopsy. In this systematic review, we searched MEDLINE, Embase, Web of Science and the Cochrane Library, for studies included NAFLD patients and undertook both liver biopsy as the reference standard and the ELF test. Two authors
more » ... ndependently screened the references, extracted the data and assessed the quality of included studies. Due to the variation in reported thresholds, we used a multiple thresholds random effects model for meta-analysis (diagmeta R-package). The meta-analysis of 11 studies reporting advanced fibrosis and five studies reporting significant fibrosis showed sensitivity of >0.90 of the ELF test for excluding fibrosis at threshold of 7.7. However, as a diagnostic test at high thresholds, the test showed specificity and positive predictive value >0.80, only in very high-prevalence settings (>50%). Desiring specificity of 0.90 for advanced and significant fibrosis resulted in thresholds of 10.18 (sensitivity: 0.57) and 9.86 (sensitivity: 0.55), respectively. The ELF test showed high sensitivity but limited specificity to exclude advanced and significant fibrosis at low cutoffs. The diagnostic performance of the test at higher thresholds was found to be more limited in low prevalence settings. We conclude that clinicians should carefully consider the likely disease prevalence in their practice setting and adopt suitable test thresholds to achieve the desired test performance.
doi:10.1016/j.jhep.2020.03.036 pmid:32275982 fatcat:gvma4z5bt5ghdkfediavph3ove