The optimal cut-off for predicting large oesophageal varices using transient elastography is disease specific.

The diagnosis of cirrhosis requires screening for oesophageal varices by upper gastrointestinal endoscopy. In many countries, serological tests and elastography are replacing liver biopsy for diagnosing cirrhosis. The aims of this study were to see whether there was an optimal cut-off of liver stiffness that could predict the presence of large (>F2) oesophageal varices and whether this was disease specific. A total of two hundred and twenty-two patients with all cause cirrhosis (Child class A) were screened, and 211 had successful elastography and are included in the analysis. Of the patients studied, one hundred and thirty-two patients had no or small F1 varices and 79 had large varices. Liver stiffness of 19.8 kPa had a negative predictive value of 91% and a positive predictive value of 55% with an area under the curve (AUC) on receiver operating characteristics (ROC) of 0.73 in differentiating between small and large varices. Seven patients with large varices would have been incorrectly classified. In the 157 patients with hepatitis C as the aetiology of cirrhosis, the negative predictive value was 98% and only one patient was misclassified. Liver stiffness was superior in diagnostic accuracy to platelet count in all patients. A liver stiffness of >19.8 kPa could be utilized as a cut-off for endoscopy and beta blocker utilization, particularly in patients with hepatitis C.