消化性溃疡并发上消化道出血的危险因素及预测模型构建

Risk factors of peptic ulcer complicated with upper gastrointestinal bleeding and construction of a prediction model

目的:探讨消化性溃疡患者非静脉曲张性上消化道出血的危险因素,构建并验证预测模型。方法:回顾分析于我科内镜下确诊为消化性溃疡的患者,收集并记录患者的临床资料,根据住院期间是否发生上消化道出血,分为出血组(107例)和对照组(371例)。采用多因素Logistic回归进行危险因素分析,并初步构建预测模型,采用受试者工作特征(ROC)曲线进行模型验证。结果:多因素分析显示高龄、幽门螺杆菌感染、抗血小板和(或)非甾体类抗炎药物的使用、不良生活习惯、Rockall评分高危和精神紧张是上消化道出血的独立危险因素;基于危险因素建立的回归方程模型为Logit(p)=-8.320+0.586×年龄+2.974×使用抗血小板/非甾体抗炎药(否=0,是=1)+1.678×不良生活习惯(否=0,是=1)+2.339×Rockall评分(低中危=0,高危=1)+0.059×HAMA评分;ROC曲线分析表明该模型具有良好的分化能力和一致性。结论:不同于临床医生使用内镜参数进行患者风险评价和分层,本研究风险预测模型有助于临床护理人员对高危患者做出直观评价,并有针对性地采取相应的护理对策,对预防患者发生上消化道出血具有一定的价值。

Objective:To explore the risk factors of non-variceal upper gastrointestinal bleeding in patients with peptic ulcer,and to construct and validate a prediction model.Methods:Patients with peptic ulcer diagnosed by endoscopy in our department were retrospectively analyzed.The clinical data of the patients were collected and recorded.According to whether upper gastrointestinal bleeding occurred during hospitalization,the patients were divided into bleeding group(107 cases)and control group(371 cases).Logistic regression was used for risk factor analysis,and the prediction model was preliminarily constructed.The receiver operating characteristic(ROC)curve was used to verify the model.Results:Multivariate analysis showed that advanced age,Helicobacter pylori infection,use of antiplatelet and/or non-steroidal anti-inflammatory drugs,bad living habits,high-risk Rockall score and mental stress were independent risk factors for upper gastrointestinal bleeding.The regression model based on the risk factors was Logit(p)=-8.320+0.586×age+2.974×use of antiplatelet/non-steroid anti-inflammatory drugs(no=0,yes=1)+1.678×bad living habits(no=0,yes=1)+2.339×Rockall score(low and moderate risk=0,high risk=1)+0.059×HAMA score.ROC curve analysis showed that the model had good differentiation ability and consistency.Conclusion:Different from clinicians using endoscopic parameters for risk assessment and patients' stratification,the risk prediction model helps clinical nurses to make intuitive evaluation of high-risk patients and take corresponding nursing countermeasures,which is of certain value in preventing patients from upper gastrointestinal bleeding.