预测急诊脓毒性休克患者预后的列线图构建与验证

Construction and validation of a nomogram for predicting the prognosis of patients with septic shock in department of emergency medicine

目的构建预测急诊脓毒性休克患者28 d病死率的列线图模型,并进行预测效能验证。方法基于天津医科大学朱宪彝纪念医院、天津医科大学总医院和天津医科大学第二医院的急诊医学科病案机构数据库,收集2017年1月至2020年10月急诊收治的913例脓毒性休克患者的信息,包括人口基线资料和临床特征、实验室指标及主要结局(28 d病死率)。根据简单随机抽样法将患者分为训练集和验证集。将训练集单因素二元Logistic回归分析中所有重要的变量纳入多因素Logistic回归分析,分析急诊脓毒性休克患者28 d死亡的危险因素,并构建列线图模型。采用校准曲线和受试者工作特征曲线(ROC曲线)评估列线图模型的预测效能。结果最终纳入860例符合标准的脓毒性休克患者,其中训练集472例,验证集388例。训练集和验证集的28 d病死率分别为52.5%(248/472)、54.1%(210/388)。在训练集中,死亡组的年龄、呼吸频率(RR)、C-反应蛋白(CRP)、D-二聚体、白细胞计数(WBC)、中性粒细胞计数(NEU)、中性粒细胞/淋巴细胞比值(NLR)、单核细胞/淋巴细胞比值(MLR)、平均血小板体积(MPV)、血小板计数(PLT)水平均明显高于存活组,碱剩余(BE)、淋巴细胞计数(LYM)、血红蛋白(Hb)水平和慢性阻塞性肺疾病(COPD)比例均明显低于存活组(均P<0.05)。多因素Logistic回归分析显示,NLR〔优势比(OR)=0.0230,95%可信区间(95%CI)为-0.2044~0.1130〕、MPV(OR=0.1798,95%CI为-0.8776~0.1727)、Hb(OR=0.0078,95%CI为0.0103~0.0408)、降钙素原(PCT;OR=1.9570,95%CI为1.2430~3.0810)和D-二聚体(OR=0.0001,95%CI为-0.0004~0.0001)是急诊脓毒性休克患者28 d病死率的独立预测因子(均P<0.05)。基于上述变量建立列线图模型,ROC曲线显示,训练集和验证集中列线图模型预测脓毒性休克患者28 d病死率的ROC曲线下面积(AUC)分别为0.907(95%CI为0.864~0.940)、0.822(95%CI为0.781~0.863)。校准曲线显示,训练集及验证集的预测结果与观察结果之间均具有良好的一致性。结论基于NLR、MPV、Hb、PCT和D-二聚体构建的列线图模型对预测急诊脓毒性休克患者的28 d病死率具有重要的临床价值。

Objective To construct a nomogram model for predicting the 28-day mortality of patients with septic shock in the emergency medicine department and to validate the predictive efficacy.Methods Based on the database of the emergency medicine department of Chu Hsien-I Memorial Hospital of Tianjin Medical University,Tianjin Medical University General Hospital and the Second Hospital of Tianjin Medical University,the data of 913 patients with septic shock admitted to the emergency medicine department from January 2017 to October 2020 were collected,including baseline demographic information and clinical characteristics,laboratory indices,and the main endpoints(28-day mortality).The patients were divided into a training set and a validation set based on simple random sampling.All significant variables from the one-way binary Logistic regression analysis of the training set were included in the multivariate Logistic regression analysis to analyze the risk factors for 28-day mortality in patients with septic shock and to construct a column-line graphical model.The predictive efficacy of the nomogram model was assessed using calibration curves and receiver operator characteristic curve(ROC curve).Results A total of 860 patients with septic shock meeting the criteria were finally enrolled,including 472 in the training set and 388 in the validation set.The 28-day mortalities were 52.5%(248/472)and 54.1%(210/388)for the training and validation sets,respectively.In the training set,age,respiratory rate(RR),the levels of C-reactive protein(CRP),D-dimer,white blood cell count(WBC),neutrophil count(NEU),neutrophil/lymphocyte ratio(NLR),monocyte/lymphocyte ratio(MLR),mean platelet volume(MPV),and platelet count(PLT)in the death group were significantly higher than those in the survival group,and the levels of base remaining(BE),lymphocyte count(LYM),hemoglobin(Hb)and the proportion of chronic obstructive pulmonary diseases(COPD)were significantly lower than those in the survival group(all P<0.05).Multifactorial Logistic regression analysis showed that NLR[odds ratio(OR)=0.0230,95%confidence interval(95%CI)was-0.2044 to 0.1130],MPV(OR=0.1798,95%CI was-0.8776 to 0.1727),Hb(OR=0.0078,95%CI was 0.0103 to 0.0408),procalcitonin(PCT;OR=1.9570,95%CI was 1.2430 to 3.0810),and D-dimer(OR=0.0001,95%CI was-0.0004 to 0.0001)were independent predictors of 28-day mortality in patients with septic shock in the emergency department(all P<0.05).A column-line graph model was established based on the above variables,and the ROC curves showed that the area under the ROC curve(AUC)of the nomogram model in the training set and validation set for predicting the 28-day mortality of patients with septic shock was 0.907(95%CI was 0.864 to 0.940)and 0.822(95%CI was 0.781 to 0.863),respectively.The calibration curves showed good agreement between the predicted and observed results for both the training and validation sets.Conclusion The nomogram model constructed based on NLR,MPV,Hb,PCT and D-dimer has significant clinical value in predicting the 28-day mortality of patients with septic shock in the emergency medicine department.