肝硬化门静脉血栓形成的相关危险因素及预测模型

发布时间：2018-10-08 19:37

【摘要】：目的:探讨肝硬化门静脉血栓(PVT)形成的相关危险因素,建立预测模型,评估模型的预测价值。方法:选取我院2012年1月至2016年2月确诊肝硬化门静脉血栓形成的病人39例为病例组,选择相同时间肝硬化无门静脉血栓的病人384例为对照组。收集病人的一般资料、实验室及影像学检查结果,比较两组患者相关并发症发生情况,筛选出相关的影响因素,计量资料比较采用独立样本的t检验;计数资料比较采用卡方检验;采用非条件的logistic回归进行独立危险因素分析。然后根据多元回归分析结果建立回归预测模型,绘制ROC曲线,检验预测价值。结果:病例组与对照组在年龄、体重质量指数、使用非选择性β受体阻滞剂、脾脏手术、食道静脉曲张套扎术、Child-Pugh评分、白细胞、血小板分布宽度、血红蛋白、γ-谷氨酰转肽酶、白蛋白、门静脉主干内径、脾脏长径、脾静脉内径、脾脏厚度、腹水量上差异有统计学意义(P值均0.05)。并且PVT组更容易发生自发性腹膜炎、门静脉高压性胃病、食管胃底静脉重度曲张、上消化道出血等并发症(P值均0.001)。进一步采用非条件logistic回归分析结果提示γ-谷氨酰转肽酶降低(P0.05,OR=0.976)为PVT形成的保护因素,门静脉内径增宽(P0.05,OR=1.419)、门静脉高压性胃(P0.05,OR=6.766)、上消化道出血病史(P0.05,OR=5.546)是肝硬化PVT形成的独立危险因素。根据多元回归结果建立预测模型:LogitP=2.031-0.025γ-GT(U/L)+0.35MPV(mm)+1.912PHG(是=1,否=0)+1.713UGIB(是=1,否=0),方程准确度为82.742%,ROC曲线分析结果提示LogitP的AUC最大,为0.817。结论:γ-谷氨酰转肽酶降低可使肝硬化PVT形成的风险减少,门静脉主干内径增宽、有门静脉高压性胃病和上消化道出血病史可增加PVT形成的风险。建立的Logit P回归预测模型预测价值较高,临床工作中可根据预测模型选择合理的治疗方法预防PVT的形成。
[Abstract]:Objective: to investigate the risk factors of portal vein thrombosis (PVT) in cirrhotic patients, to establish a predictive model and to evaluate the predictive value of the model. Methods: from January 2012 to February 2016, 39 patients with cirrhotic portal vein thrombosis were selected as the case group, and 384 patients with cirrhosis without portal vein thrombosis at the same time as the control group. Collect the general data of patients, laboratory and imaging results, compare the two groups of patients with related complications, screen out the relevant factors, the measurement of data comparison using independent samples t-test; The counting data were compared by chi-square test, and independent risk factors were analyzed by non-conditional logistic regression. Then according to the results of multiple regression analysis, the regression prediction model is established, the ROC curve is drawn, and the predictive value is tested. Results: age, BMI, non-selective 尾 receptor blocker, splenectomy, esophageal variceal ligation, Child-Pugh score, WBC, platelet distribution width, hemoglobin, 纬 -glutamyl transpeptidase were used in the case group and control group. There were significant differences in albumin, portal vein diameter, spleen length diameter, splenic vein diameter, spleen thickness and ascites volume (P < 0.05). In PVT group, spontaneous peritonitis, portal hypertensive gastropathy, severe esophageal and gastric varices and upper gastrointestinal bleeding were more likely to occur (P < 0.001). The results of non-conditional logistic regression analysis showed that the decrease of 纬 -glutamyl transpeptidase (P0.05OR0.976) was the protective factor for the formation of PVT, the enlargement of portal vein diameter (P0.05OR-1.419), the portal hypertensive stomach (P0.05OR6.766) and the history of upper gastrointestinal bleeding (P0.05OR5.546) were independent risk factors for the formation of PVT in liver cirrhosis. According to the result of multivariate regression, a prediction model was established, that is, the 0.35MPV (mm) 1.912PHG of LogitP is 2.031-0.025 纬 -GT (UP / L) 1.713UGIB. The accuracy of the equation is 82.742ROC curve. The result shows that the AUC of LogitP is the largest (0.817.7). Conclusion: the decrease of 纬 -glutamyl transpeptidase can reduce the risk of PVT formation in cirrhosis, widen the inner diameter of portal vein trunk, and increase the risk of PVT formation with history of portal hypertension gastropathy and upper gastrointestinal bleeding. The predictive value of Logit P regression prediction model is high, and reasonable treatment methods can be selected to prevent the formation of PVT in clinical work.
【学位授予单位】：新疆医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R575.2

【相似文献】