多发性骨髓瘤化疗医院感染风险预测模型的建立及应用价值

Establishment of prediction model of risk for nosocomial infection in multiple myeloma patients undergoing chemotherapy and its application value

目的建立多发性骨髓瘤化疗医院感染风险预测模型,并对其应用价值进行分析。方法回顾性分析上海交通大学医学院附属瑞金医院血液科2017年3月-2020年1月收治的150例多发性骨髓瘤患者的临床资料,根据其化疗期间是否发生医院感染分为感染组和未感染组。对患者感染部位分布特点进行调查;通过病历调查收集患者资料,采用单因素分析、多元Logistic回归分析多发性骨髓瘤患者化疗医院感染危险因素;根据分析结果建立医院感染预测模型,采用受试者工作特征曲线(Receiver operating characteristic,ROC)对其诊断价值进行评价。结果150例多发性骨髓瘤患者共93例患者发生医院感染,医院感染率为62.00%(93/150),上呼吸道感染、肺部感染、皮肤感染、胃肠道感染发生率较高;多元Logistic回归分析结果显示年龄≥70岁、DS分期较高、ISS分期较高、ECOG评分为3~4分、CONUT评分≥2分、PNI评分<45分、CRP≥10 mg/L、PCT≥0.5 ng/ml是MM患者化疗医院感染的独立危险因素(P<0.05);ROC分析结果显示,以该模型得出的感染模型评分进行医院感染诊断,其曲线下面积(Area under the curve,AUC)为0.829,高于单指标诊断水平,此时感染模型评分截断值为5.86;所建立感染模型经检验总准确率为(79+30)/150=72.67%,总体判别能力较好。结论年龄、疾病状态、营养水平和全身炎症因子水平等是多发性骨髓瘤患者化疗期间发生医院感染的危险因素,据此构建的预测模型对患者医院感染进行判断具有较好的准确度。

OBJECTIVE To establish the prediction model of risk for nosocomial infection in the multiple myeloma patients undergoing chemotherapy and analyze its application value.METHODS The clinical data were collected from 150 patients with multiple myeloma who were treated in Ruijin Hospital Affiliated to Medical College of Shanghai Jiaotong University from Mar.2017 to Jan.2020 and were retrospectively analyzed.The patients were divided into the infection group and the non-infection group according to the status of nosocomial infection during the chemotherapy.The characteristics of infection sites were investigated.The data were collected through review of medical records,univariate analysis and multivariate logistic regression analysis were performed for risk factors for the nosocomial infection in the multiple myeloma patients undergoing chemotherapy.The prediction model for nosocomial infection was established based on the result of analysis,and the diagnostic values were evaluated by means of receiver operating characteristic(ROC)curve.RESULTS Of the 150 patients with multiple myeloma,totally 93 had nosocomial infection,with the infection rate 62.00%(93/150);the incidence rates of upper respiratory tract infection,pulmonary infection,skin infection and gastrointestinal tract infection were relatively high.The result of multivariate logistic regression analysis showed that the no less than 70 years of age,high DS stage,high ISS stage,ECOG score ranging between 3 and 4 points,CONUT score no less than 2 points,PNI score less than 45 points,CRP no less than 10 mg/L and PCT no less than 0.5 ng/mL were the independent risk factors for the nosocomial infection in the MM patients undergoing chemotherapy(P<0.05).ROC analysis indicated that the area under curve(AUC)of the infection model score that was obtained based on the model was 0.829 in diagnosis of the nosocomial infection,higher than that of the single index,the cutoff clue of the infection model score was 5.86;the total accuracy of the established infection model after testing was(79+30)/150=72.67%,and the overall discriminant ability was better.CONCLUSION The age,state of disease,nutrition level and levels of systemic inflammatory factors are the risk factors for nosocomial infection in the multiple myeloma patients during the chemotherapy,the established prediction model has high accuracy in diagnosis of nosocomial infection.