Modified spleen stiffness measurement by transient elastography is associated with presence of large oesophageal varices in patients with compensated hepatitis C virus cirrhosis

To evaluate the accuracy of liver transient elastography (TE), spleen TE and other noninvasive tests (AAR, APRI score, platelet count, platelet/spleen ratio) in predicting the presence and the size of oesophageal varices in compensated hepatitis C virus (HCV) cirrhosis, we studied 112 consecutive patients with compensated HCV cirrhosis who underwent biochemical tests, gastrointestinal endoscopy, liver TE and spleen TE by Fibroscan((R)) (Echosens, Paris, France) using a modified software version with a range between 1.5 and 150kPa. Spleen TE was not reliable in 16 patients (14.3%). Among the 96 patients with a valid measurement (69.8% men, mean age: 63.2 +/- 9.5years), 43.7% had no oesophageal varices, 29.2% had grade 1% and 27.1% had grade 2 or grade 3 oesophageal varices. Patients with values of 75kPa by standard spleen TE had mean values of modified spleen TE of 117kPa (range: 81.7-149.5). Linear regression revealed a significant correlation between modified spleen TE and oesophageal varix size (r=0.501; beta: 0.763, SE: 0.144; P<0.001). On univariate analysis, the variables associated with grade 2/grade 3 oesophageal varices were AAR score, APRI score, platelet/spleen ratio, liver TE and modified spleen TE. On multivariate analysis, only modified spleen TE (OR: 1.026; 95% CI: 1.007-1.046; P=0.006) and AAR (OR: 14.725; 95% CI: 1.928-112.459; P=0.010) remained independently associated with grade 2/grade 3 oesophageal varices. Platelet/spleen ratio was the best predictor of oesophageal varices area under the ROC curve (AUROC: 0.763, cut-off: 800, sensitivity: 74%, specificity: 70%), while modified spleen TE was more accurate in predicting grade 2/grade 3 oesophageal varices (AUROC: 0.82, cut-off: 54.0kPa, sensitivity: 80%, specificity: 70%). Portal hypertension increases spleen stiffness, and the measurement of modified spleen TE is an accurate, noninvasive tool for predicting the presence of large oesophageal varices in patients with compensated HCV cirrhosis.