消化道出血行内镜治疗后再出血影响因素分析及列线图预测模型构建

Analysis of influencing factors of rebleeding after endoscopic treatment of gastrointestinal bleeding and construction of nomogram prediction model

目的:分析消化道出血行内镜治疗后再出血的危险因素并建立列线图预测模型。方法:收集接受内镜治疗的357例消化道出血患者的资料并进行回顾性分析,将纳入的患者按照7∶3分为建模组(250例)及验证组(107例)。建模组根据患者是否再出血分为发生组(74例)和未发生组(176例),对建模组患者资料行单因素及多因素Logistic回归分析,并根据消化道出血行内镜治疗后发生再出血的影响因素分析结果建立列线图预测模型,绘制校准曲线、受试者工作特征(ROC)曲线对模型的有效性及区分度进行评估。结果:单因素分析结果显示,有无休克、出血病变直径、Forrest分级、出血量、血尿素氮、血红蛋白、血小板、凝血酶原时间、Rockall危险评分、Glasgow-Blatchford评分为影响消化道出血行内镜治疗后再出血的因素(P<0.05);多因素Logistic回归分析显示,休克、Forrest分级、出血量、血尿素氮、血红蛋白、血小板、凝血酶原时间为消化道出血行内镜治疗后再出血的危险因素(P<0.05);校准曲线结果显示,建模组χ^(2)=7.242,P=0.511;验证组χ^(2)=7.008,P=0.536;ROC曲线结果显示,建模组ROC曲线下面积为0.880,灵敏度为89.19%,特异度为78.41%;验证组ROC曲线下面积为0.837,灵敏度为81.25%,特异度为77.52%。结论:消化道出血行内镜治疗后再出血的危险因素为休克、Forrest分级、出血量、血尿素氮、血红蛋白、血小板、凝血酶原时间,建立的列线图预测模型区分度及有效性较好,能够作为早期预测消化道出血行内镜治疗后发生再出血的工具。

Objective:To analyze the risk factors of rebleeding after endoscopic treatment of gastrointestinal bleeding,and to establish a nomogram prediction model.Methods:The data of 357 patients with gastrointestinal bleeding who received endoscopic treatment were collected for retrospective analysis,and the included patients were divided into modeling group(250 cases)and validation group(107 cases)according to 7∶3.The modeling group was divided into occurrence group(74 cases)and non-occurrence group(176 cases)according to whether the patient had rebleeding.Univariate and multivariate Logistic regression analyses were performed on the data of the patients in the modeling group,and a nomogram prediction model was established based on the analysis results of the influencing factors of rebleeding after endoscopic treatment of gastrointestinal bleeding.Calibration curve and ROC curve were drawn to evaluate the validity and discrimination of the model.Results:Univariate analysis showed that shock,bleeding lesion diameter,Forrest grading,blood loss,blood urea nitrogen,hemoglobin,platelet,prothrombin time,Rockall risk score and Glasgow-Blatchford score were the factors affecting gastrointestinal bleeding after endoscopic treatment(P<0.05);multivariate Logistic regression analysis showed that shock,Forrest grade,bleeding volume,blood urea nitrogen,hemoglobin,platelets,and prothrombin time were the risk factors of rebleeding after endoscopic treatment(P<0.05).The calibration curve results showed that the modeling group χ^(2)=7.242,P=0.511;the validation group χ^(2)=7.008,P=0.536;the ROC curve results showed that the area under the ROC curve of the modeling group was 0.880,the sensitivity was 89.19%,and the specificity was 78.41%;the area under the ROC curve of the validation group was 0.837,the sensitivity was 81.25%,and the specificity was 77.52%.Conclusion:The risk factors affecting rebleeding after endoscopic treatment for gastrointestinal bleeding are shock,Forrest grade,bleeding volume,blood urea nitrogen,hemoglobin,platelets,and prothrombin time.The established nomogram prediction model has good discrimination and validity,and can be used as a tool for early prediction of rebleeding after endoscopic treatment for gastrointestinal bleeding.