Systematic review and meta-analysis: non-invasive detection of non-alcoholic fatty liver disease related fibrosis in the obese

Research output: Contribution to journalReview ArticleResearchpeer-review

25 Citations (Scopus)

Abstract

Background: Non-alcoholic fatty liver disease (NAFLD) is a significant disease burden in obesity. Liver fibrosis is an important prognostic factor in NAFLD, and detection is vital. The pathophysiological changes of obesity can alter the accuracy of non-invasive NAFLD tests. We aimed to review current evidence for common non-invasive tests for NAFLD-related fibrosis in obesity. Methods: We systematically searched for studies assessing the diagnostic accuracy of 11 biomarker panels and elastography techniques for NAFLD-related fibrosis in obesity. Meta-analyses were performed where possible. Results: Thirty-eight studies were identified assessing the selected tests in obese populations. Simple biomarker panels (e.g. NAFLD fibrosis score) were the most validated. Evidence showed better accuracy of complex biomarker panels (NAFLD fibrosis score: summary receiver operator characteristic [SROC] 0.795–0.813 vs. enhanced liver fibrosis: SROC 0.962); however, these were poorly validated in obesity. Elastography techniques were better studied and had high diagnostic accuracy (transient elastography: SROC 0.859; magnetic resonance elastography: SROC 0.965) but were limited by BMI-dependent failure. Limited evidence was found to validate the accuracy of any test in exclusively obese populations. Conclusion: In obese subjects, complex biomarker panels and elastography have been reasonable to good accuracy for NAFLD-related fibrosis; however, these methods have not been well validated. Further study in this high-risk population is needed.

Original languageEnglish
Pages (from-to)281-294
Number of pages14
JournalObesity Reviews
Volume19
Issue number2
DOIs
Publication statusPublished - 1 Feb 2018

Keywords

  • Elastography
  • liver fibrosis
  • non-alcoholic fatty liver disease
  • obesity

Cite this