Predictive model of portal venous system thrombosis in cirrhotic portal hypertensive patients after splenectomy.

OBJECTIVE This study is to investigate the risk factors of portal venous system thrombosis (PVT) in patients with cirrhotic portal hypertension after splenectomy and to establish a Logistic regression prediction model. METHODS A total of 119 patients with cirrhotic portal hypertension were enrolled. Their clinical data was retrospectively analyzed. They were divided into PVT group (n = 18) and non-PVT group (n = 101). One-way analysis and multivariate Logistic regression analysis were performed to analyze the independent risk factors of PVT. Logistic regression prediction model was established. The receiver operating characteristic curve was generated and correlation analysis was conducted. RESULTS Platelet count (PLT), mean platelet volume (MPV) and D-Dimer were independent risk factors affecting PVT. Anticoagulation therapy (UAT) and usage of reducing portal pressure therapy (URPT) were independent protective factors of PVT. Logistic regression prediction model was expressed as Logit P = -9.165 + 0.664 × PLT (× 10(11)/L) + 0.413 × MPV (fL) + 0.662 × D-Dimer (mg/L) -1.674 × UAT (Yes = 1, No = 0) -1.518 × URPT (Yes = 1, No = 0). And, the cut-off value of Logit P was -1.14. The area under the receiver operating characteristic curve and the accuracy were 0.865 and 84.03%. The cut-off value of PLT, MPV and D-Dimer were 4.42 × 10(11)/L, 13.30 fL and 2.55 mg/L, respectively. MPV and D-Dimer were positively correlated. CONCLUSION PLT, MPV and D-Dimer are independent risk factors while UAT and URPT are independent protective factors of PVT. Logistic regression prediction model can predict PVT with a high sensitivity, specificity and accuracy. It provides theoretical foundation and cut-off value for predicting PVT after splenectomy.