Meta-analysis of the accuracy of transient elastography in measuring liver stiffness to diagnose esophageal varices in cirrhosis

BACKGROUD: To assess the diagnostic performance of transient elastography (TE) in detecting the presence and size of esophageal varices (EV) in cirrhotic patients. METHODS: We searched PubMed, Web of Science, Wiley Online Library, Science Direct, China National Knowledge Infrastructure, WeiPu, WanFang database, and Baidu Scholar to identify studies that evaluated the diagnostic accuracy of TE in liver stiffness measurement, compared with esophagogastroduodenoscopy (EGD), for the detection of the presence and degree of EV in cirrhosis. RESULTS: We included 32 studies in the presence of any EV (grade 1–3; n = 4082), 27 studies on substantial EV (grade 2–3; n = 5221) and 5 studies on large EV (grade 3). The pooled sensitivity, specificity, and diagnostic odds ratio (DOR) were 0.8 (95% CI, 0.78–0.86), 0.68 (95% CI, 0.62–0.74), and 10 (95% CI, 7–14) for any EV; 0.81 (95% CI, 0.77–0.85), 0.72 (95% CI, 0.66–0.77), and 11 (95% CI, 8–15) for substantial EV; and 0.92 (95% CI, 0.83–0.96), 0.78 (95% CI, 0.70–0.85), and 40 (95% CI, 15–107) for large EV. Subgroup analysis revealed that the heterogeneity among studies on any EV could potentially be explained by study location, proportion of Child A, and time interval between TE and EGD; for substantial EV, the proportion of Child A, etiology of cirrhosis, and the time interval between TE and EGD were important heterogeneity factors. Publication bias was found among studies evaluating diagnostic performance of TE for any EV. CONCLUSION: TE is a good tool for detecting the presence and degree of EV; however, in determination of the liver stiffness cutoff values means that TE is only cautiously used in clinical practice.