Evaluation of large esophageal varices in cirrhotic patients by transient elastography: a meta-analysis.

Background And Purpose: Transient elastography (TE) has been shown to be a valuable tool for the prediction of large esophageal varices. However, the conclusions have not been always consistent throughout the different studies. Therefore, we performed a further meta-analysis in order to evaluate the diagnostic accuracy of transient elastography for the prediction of large esophageal varices.

Methods: We performed a systematic literature search in PubMed, EMBASE, Web of Science, and CENTRAL in The Cochrane Library without time restriction. The strategy we used was "(fibroscan OR transient elastography OR stiffness) AND esophageal varices". Accuracy measures such as pooled sensitivity, specificity, among others, were calculated using Meta-DiSc statistical software.

Results: Twenty studies (2,994 patients) were included in our meta-analysis. The values of pooled sensitivity, specificity, positive and negative likelihood ratios and diagnostic odds ratio were as follows: 0.81 (95% CI, 0.79-0.84), 0.71 (95% CI, 0.69-0.73), 2.63 (95% CI, 2.15-3.23), 0.27 (95% CI, 0.22-0.34) and 10.30 (95% CI, 7.33-14.47). The area under the receiver operating characteristics curve was 0.83. The Spearman correlation coefficient was 0.246 with a p-value of 0.296, indicating the absence of any significant threshold effects. In our subgroup analysis, the heterogeneity could be partially explained by the geographical origin of the study or etiology; or it could be partially explained blindingly, through the appropriate interval and cut-off value of the liver stiffness (LS).

Conclusions: Transient elastography could be used as a valuable non-invasive screening tool for the prediction of large esophageal varices. However, since LS cut-off values vary throughout the different studies and significant heterogeneity also exists among them, we need more reasonable approaches or flow diagram in order to improve the operability of this technology.