Objectives: we aimed to investigate the accuracy of liver (LSM) spleen stiffness measurement (SSM) with transient elastography (TE) to predict varices in the presence of non-selective beta-blockers (NSBB).
Methods: In this cross-sectional study of consecutive patients with chronic hepatitis B (CHB) and cirrhosis, all patients underwent TE and upper endoscopic examinations. LSM and SSM in predicting varices in patients receiving and not receiving NSBB were evaluated.
Results: Altogether 144 CHB patients (29 receiving NSBB; 35 with any varices, 31 and 11 with esophageal and gastric varices, respectively) were recruited. Their mean LSM and SSM were 13.3 ± 9.0 kPa and 32.8 ± 19.2 kPa, respectively. The correlation between LSM and SSM was better in the NSBB subgroup (r = 0.525, P = 0.003) than its counterpart (r = 0.329, P < 0.001). The area under receiver operating characteristic curve (AUROC) of LSM and SSM for any varices was 0.675 and 0.685 (P = 0.002 and 0.001), respectively. SSM of 18.9 kPa had a negative predictive value of 92.1% and negative likelihood ratio of 0.27 for ruling out any varices; and SSM of 54.9 kPa had a positive predictive value of 56.5% and a positive likelihood ratio of 4.05 to rule in varices. The AUROC of LSM for varices was 0.742 and 0.549 in patients with or without NSBB, respectively; the corresponding AUROC of SSM was 0.572 and 0.603, respectively.
Conclusions: SSM only has modest accuracy to predict varices independent of NSBB use. An SSM cutoff value of 18.9 kPa may be adopted to achieve a high negative predictive value to rule out varices.
Keywords: cirrhosis; splenomegaly; transient elastography; varices.
© 2016 Chinese Medical Association Shanghai Branch, Chinese Society of Gastroenterology, Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine and John Wiley & Sons Australia, Ltd.