Correlation of Spleen Stiffness with Esophageo-gastric Variceal Bleeding in Patients with Non-cirrhotic Portal Hypertension
摘要
Bleeding from esophageo-gastric varices is a serious complication related to portal hypertension. In patients with cirrhosis, liver and splenic stiffness measurements (LSM and SSM) are useful to predict high-risk varices, bleeding and decompensation. Data regarding the utility of SSM for predicting bleeding in patients with non-cirrhotic portal hypertension (NCPH) is scarce.
MethodsConsecutive patients with NCPH (n = 150, 129 extrahepatic portal vein obstruction; 21 non-cirrhotic portal fibrosis) with baseline available LSM, SSM and HVPG values were included. Half of the patients had recent variceal bleeding (RVB, within 3 months). We aimed to study the association of baseline SSM, LSM, HVPG and the presence of hypersplenism with respect to the history of RVB and the long-term follow-up changes in LSM and SSM values.
ResultsBaseline demographic and lab characteristics were comparable in patients with NCPH with/without RVB. Mean SSM and SSM/LSM ratios were higher in patients with RVB as compared to patients without RVB [90 ± 7.5 vs. 70 ± 7.8 kPa(p < 0.0001) and 16.51 ± 2.97 vs. 13.49 ± 2.66(p < 0.001) respectively]. The SSM (OR,1.085(95%CI:1.029–1.144) and the presence of hypersplenism [OR,0.379(0.157–0.91) were associated with RVB in multivariate analysis. The SSM (cut-off 75 kPa) has a sensitivity of 92% and specificity of 90% for being associated with RVB(AUROC-0.975). There were no significant changes in SSM on follow-up both in bleeders and those without bleeding; however, SSM at 12 months (88.3 ± 7.3 vs. 68.4 ± 7.8 kPa) and 24 months (87.4 ± 7.2 vs. 68.7 ± 7.2 kPa) remained significantly higher in bleeders than those without bleeding (p < 0.001).
ConclusionRecent esophageo-gastric variceal bleeding is associated with significantly higher SSM in patients with non-cirrhotic portal hypertension, with a best cut-off of 75kPa. The SSM values remain high in bleeders up to 24 months of follow-up.