简易无创评分系统对乙型肝炎肝硬化患者食管静脉曲张出血的评估

Assessment of the presence of esophageal varices bleeding in patients with hepatitis B virus-related cirrhosis by a noninvasive score system

目的回顾性分析以瞬时弹性成像技术（TE）、血清学检测和影像学检查指标为基础的简易无创评分系统对乙型肝炎肝硬化患者食管静脉曲张出血的诊断价值。方法2011年4月至2012年12月天津市第三中心医院肝内科门诊或住院的乙型肝炎肝硬化患者172例，其中男120例，女52例，平均年龄（52．9±10．6）岁。用胃镜观察患者有无食管静脉曲张，若有食管静脉曲张，按食管静脉曲张形态及出血危险程度分轻、中、重3级。采用TE中的Fibroscan进行肝脏硬度及脾脏硬度检测。彩色多普勒超声检查患者门静脉宽度、脾静脉宽度和脾脏厚度。同时检测患者白细胞和血小板计数。运用受试者工作特征（ROC）曲线及曲线下面积（AUC）判断以肝脏硬度、脾脏硬度、脾脏厚度等为基础的简易无创评分系统对食管静脉曲张破裂出血的评估价值。两样本均数比较采用￡检验，相关性分析采用Pearson相关分析。结果172例患者均行胃镜检查，其中首次食管静脉曲张出血患者41例，食管静脉曲张未出血患者131例。172例患者中无食管静脉曲张39例，轻度食管静脉曲张30例，中度食管静脉曲张47例，重度食管静脉曲张56例。食管静脉曲张出血与肝脏硬度值、脾脏硬度值、门静脉宽度、脾脏厚度、脾静脉宽度呈正相关（r值分别为0．224、0．771、0．214、0．425和0．364，均P〈0．05），与血小板计数呈负相关（r=-0．408，P=0．000），与白细胞计数无关（r=0．126，P=0．215）。食管静脉曲张出血患者肝脏硬度值、脾脏硬度值、门静脉宽度、脾脏厚度和脾静脉宽度均明显高于未出血患者，而血小板计数低于未出血患者，差异均有统计学意义（均P〈0．05）。简易无创评分系统预测食管静脉曲张及出血的AUC分别为0．953和0．882，最佳界值分别为7分和10分（均P=0．000），其诊断食管静脉曲张的敏感度为96％，特异度为85％，诊断食管静脉曲张破裂出血的敏感度为78％，特异度为89％。结论应用以肝硬度、脾硬度、脾脏厚度、脾静脉宽度、门静脉宽度、血小板计数为基础的简易无创评分系统对乙型肝炎肝硬化食管静脉曲张出血的诊断具有重要的指导意义，可作为临床筛选方法。

Objective To retrospectively analyze the diagnostic value of a noninvasive score system based on transient elastography （TE）, serological test and imaging examination on esophageal variceal bleeding （EVB） in patients with hepatitis B virus （HBV）-related cirrhosis. Methods Between April 2011 and December 2012, 172 patients with HBV-related cirrhosis including 120 males and 52 females who visited clinic or hospitalized at the Department of Hepatology, Tianjin Third Central Hospital, were retrospectively enrolled. The mean age was （52.9 ± 10. 6） years. Patients underwent upper gastrointestinal endoscopy to evaluate esophageal varices （EV） and were further categorized into three stages of mild, moderate and severe according to the morphology of EV and the risk of bleeding. Liver stiffness and spleen stiffness measurement were performed using Fibroscan. Portal vein width, splenic width and spleen thickness were measured using color Doppler ultrasound. All the patients were tested for white blood cell counts and platelet counts. With endoscopy as the gold standard, receiver operating characteristic （ROC） curves and the areas under curves （AUC） were used to assess the performance o{ the noninvasive score system in predicting EV by liver stiffness, spleen stiffness, portal vein width, spleen thickness and platelet counts. Student＇s t-test was performed to determine differences between continuous variables. Pearson＇s correlation was used to evaluate the association between EVB and these parameters. Results All these 172 patients underwent endoscopy. Among them, 41 were EVB patients and 131 with no bleeding of EV. Among 172 EV patients, 39 without EV, 30 were mild EV, 47 were moderate EV and 56 were severe EV. EVB was all positively correlated with liver and spleen stiffness, portal vein width, spleen thickness, splenic vein width （r=0. 224, 0. 771, 0. 214, 0. 425 and 0. 364, respectively; all P〈0.05）. EVB was negatively correlated with platelet counts （r=- 0. 408, P〈 0. 000）. Liver stiffness, spleen stiffness, portal vein width, spleen thickness and splenic vein width in EVB patients were significantly higher than those in EV patients （P〈0. 05）. In contrast, platelet counts level was lower in EVB patients with difference of statistical significance （P〈0.05）. AUC of non-invasive score system for EV and EVB were 0. 953 and 0. 882, respectively （P〈0. 05）. The optimal cut-off level of noninvasive score system for prediction of EV and EBV were 7 （sensitivity = 96 %, specificity = 85%） in EV patients and 10 （sensitivity = 78%, specificity： 89%） in EVB patients. Conclusion Non-invasive score system based on liver stiffness, spleen stiffness, spleen thickness, width of splenic and portal vein and platelet counts is of clinical importance in assessing the presence of EV in patients with HBV-related cirrhosis, which is higher clinically valuable in the diagnosis for EV.