摘要目的 筛选食管胃底静脉曲张破裂出血的危险因素，为预测曲张静脉破裂出血提供更为经济且创伤更小的手段。方法 选取诊断明确的肝硬化腹水患者108例，肝硬化合并原发性肝癌患者60例。追踪随访1年,观察其发生食管胃底静脉曲张破裂出血的情况，并运用统计学方法进行分析。结果 单因素非条件Logistic回归分析显示腹水白蛋白、血清腹水白蛋白梯度(SAAG)、血小板、部分凝血活酶时间(APTT)、门静脉宽度、脾脏长度和厚度是危险因素，年龄和血白蛋白是保护因素。多因素分析显示SAAG、APTT和门静脉宽度是独立危险因素,OR值分别为3.559、2.468和2.608。构建受试者工作特征曲线后，SAAG取值18.50 g/L时为最佳临界值，敏感度和特异度分别为96.3％和56.3％。结论 SAAG对预测食管胃底静脉曲张破裂出血具有较好的价值。

abstractsObjective To screen the risk factors of esophageal-gastric fundus variceal bleeding,in order to provide a more economical and less invasive method for predicting esophageal-gastric fundus variceat bleeding. Methods A total of 168 diagnosed liver cirrhosis patients accompanied with ascites and 60 cases of liver cirrhosis patients with primary hepatic carcinoma were enrolled. Followed up for one year, the esophageal-gastric fundus variceal bleeding was observed and analyzed with statistic methods. Results Unconditional single factor logistic regression model analysis indicated that albumin level of ascites, serum-ascites albumin gradients (SAAG), platelets, activated partial thromboplastin time (APTT), portal vein width, length and thickness of the spleen were independent risk factors,age and serum albumin were protective factors. Multifactor analysis indicated SAAG, APTT, and portal vein width were the independent risk factors, OR was 3.559 , 2.468 and 2. 608 respectively.After building receiver operating characteristic curve, the best SAAG cut-off value was 18.50 g/L, of which the sensitivity was 96.3％ and specificity was 56.3％. Conclusion SAAG has good value in predicting esophageal-gastric fundus variceal bleeding.