Diagnostic accuracy of ultrasonography for the detection of hepatic steatosis: an updated meta-analysis of observational studies

Abstract

Aim: we examined the diagnostic accuracy of ultrasonography to detect any HS (defined as steatotic hepatocytes ≥ 5% on histology) and moderate-severe HS (defined as steatotic hepatocytes ≥ 30% on histology) by performing a systematic review and meta-analysis.

Methods: we systematically searched PubMed, Web of Science, and Scopus databases, from January 2011 to February 2021, to identify studies conducted in adults investigating the diagnostic accuracy of ultrasonography vs. histology for detecting either ≥ 5% histologically defined HS or moderate-severe HS (≥ 30%). Meta-analysis was performed using random-effects modeling.

Results: twelve studies were included involving a total of 2921 individuals, 1710 (58.5%) of whom had HS ≥ 5% by histology. The overall sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of ultrasonography for the detection of ≥ 5% histologically defined HS, compared to histology, were 82% (95% confidence interval 76%-86%), 80% (72%-86%), 4.0 (2.90-5.70), and 0.23 (0.18-0.30), respectively. Based on the pooled analysis of seven studies, the overall sensitivity, specificity, positive likelihood ratio, and negative likelihood ratio of ultrasonography for the detection of ≥ 30% histologically defined HS were 85% (72%-92%), 85% (73%-93%), 5.72 (3.06-10.7), and 0.18 (0.10-0.33) , respectively. Funnel plots did not reveal any significant publication bias.

Conclusion: conventional ultrasonography allows for reliable and accurate detection of ≥ 5% histologically defined HS compared to histology. These findings call for an extensive use of conventional ultrasonography in the clinical arena.

More information

Identifiers

ORCID for Christopher Byrne: orcid.org/0000-0001-6322-7753

Catalogue record

Export record

Altmetrics