肝硬化食管静脉曲张无创性评估模型的研究  

Non-invasive predictive models of esophageal varices in liver cirrhosis

作　　者：王锦俊 刘欣艳[1] 杨根妹[1] 王丽晶[2] 熊伍军[2] 

机构地区：[1]上海中医药大学附属普陀医院老年科,上海200062 [2]同济大学附属东方医院消化内科

出　　处：《上海医学》2014年第11期949-952,共4页Shanghai Medical Journal

摘　　要：目的 探讨肝硬化患者食管静脉曲张的无创性预测因素并建立预测模型。方法 收集104例肝硬化失代偿期患者资料,分入非重度曲张组（50例）和重度曲张组（54例）,对患者的各项血生化指标和影像学检查指标进行单因素分析和多因素Logistic回归分析,建立回归模型,利用受试者工作特征曲线（ROC曲线）评估模型的预测效率。结果 重度曲张组的血小板计数、血红蛋白、血小板计数/脾脏长度、血小板计数/门静脉宽度均显著低于非重度曲张组（P值均〈0.01）,γ-谷氨酰转肽酶、门静脉宽度、脾脏长度、肝右叶长度/白蛋白均显著高于非重度曲张组（P值均〈0.01）。进一步作多因素Logistic回归分析,血红蛋白[r=-0.031,P=0.003,95%可信区间（CI）为0.95~0.99]、门静脉宽度（r=0.32,P=0.047,95%CI为1.005~1.887）和血小板计数/脾脏长度（r=-5.633,P=0.001,95%CI为0~0.089）为重度食管静脉曲张的独立危险因素,据此得出回归模型预测指数=1.674-0.031×血红蛋白水平（g/L）-5.633×血小板计数/脾脏长度（×109·L-1·mm-1）＋0.32×门静脉宽度（mm）。所构建模型ROC曲线的曲线下面积为0.814,95%CI为0.732~0.896,敏感度和特异度分别为77.8%和72.0%,P〈0.001。结论 血红蛋白、门静脉宽度、血小板计数/脾脏长度是预测肝硬化重度食管静脉曲张的独立危险因素,构建的回归模型预测指数对食管静脉曲张具有较好的预测价值。Objective To explore non-invasive predictors of esophageal varices in liver cirrhosis and to establish a predictive model. Methods Totally 104 patients with liver cirrhosis were enrolled in the study and divided into two groups： group A （50 cases without severe esophageal varices） and group B （54 cases with severe esophageal varices）. All parameters of blood tests and imaging tests were analyzed with univariate analysis and multivariate logistic regression analysis. A logistic regression model was established. Receiver operating curve （ROC） was used to assess the efficacy of the predictive model. Results Platelet count, hemoglobin level, ratio of platelet count to spleen diameter and ratio of platelet count to portal vein width were significantly decreased in group B as compared with group A （P〈0.01）. Gamma glutamyl transpeptidase, portal vein width, diameter of spleen and ratio of right liver lobe diameter to albumin were increased in group B as compared with group A （P〈 0.01）. Multivariate logistic regression analysis showed that hemoglobin （r = -0. 031, P = 0. 003, 95% CI= 0. 95--0. 99） , portal vein width （ r = 0.32, P = 0. 047, 95 % CI = 1. 005--1. 887） and ratio of platelet count to spleen diameter （r = - 5. 633, P = 0. 001, 95% CI= 0--0. 089） were the independent risk factors of severe esophageal varices. Therefore, predictive index= 1. 674 - 0. 031 x hemoglobin （g/L） - 5. 633 × ratio of platelet count to spleen diameter （ × 10^9 ·L^-1·mm^-1 ） ＋ 0.32 × portal vein width （mm） was concluded. A ROC curve from the established predictive model had an area under the curve （AUC） of 0. 814 （95% CI= 0. 732-0. 896, sensitivity= 77.8%, specificity= 72.0%, P〈0. 001）. Conclusion Hemoglobin, width of portal vein and ratio of platelet count to spleen diameter are the independent predictors of severe esophageal varices. The established logistic regression model can accurately predict esophageal varices.

关 键 词：食管静脉曲张 肝硬化 无创性预测因素 

分 类 号：R575.2[医药卫生—消化系统]