成人社区获得性肺炎进展为重症肺炎的临床预测模型构建及验证

Construction and validation of a clinical prediction model for the progression of community⁃acquired pneumonia to severe pneumonia in adults

目的探究影响成人社区获得性肺炎(CAP)进展为重症肺炎(SP)的危险因素,构建成人CAP进展为SP的临床预测模型并验证,为临床早期识别并预测重症肺炎及指导临床治疗策略提供依据。方法收集2022年6月-2023年9月在锦州市中心医院住院诊断为成人CAP患者376例的临床资料,并根据入院时间顺序将入组病例分为训练集301例和验证集75例。将训练集CAP患者以出院诊断分为重症组和非重症组。采用单因素分析、多因素logistic回归分析筛选预测因子,构建CAP进展为重症肺炎的预测模型,并绘制列线图。模型的验证分别采用受试者工作特征(ROC)曲线、校正曲线及临床决策曲线评估模型的区分度、校准度及临床适用度。结果训练集的重症肺炎患者107例,非重症肺炎患者194例。对训练集各临床指标进行单因素分析显示,意识障碍、肺部浸润>50%、呼吸频率、中性粒细胞(Neu)、C-反应蛋白(CRP)、乳酸脱氢酶(LDH)、血沉(ESR)、乳酸(Lac)这8个临床指标差异有统计学意义(P均<0.05)。二元多因素logistic回归分析显示,呼吸频率(OR=1.528,95%CI:1.220~1.913),Neu(OR=1.179,95%CI:1.087~1.278),CRP(OR=1.012,95%CI:1.005~1.019),LDH(OR=1.008,95%CI:1.004~1.011),ESR(OR=1.020,95%CI:1.004~1.037),最终模型纳入了这5个变量作为预测因子构建模型,并绘制CAP进展为重症肺炎的风险列线图。训练集的ROC曲线下面积(AUC)为0.847,95%CI:0.804~0.729(P<0.05),验证集的AUC为0.861,95%CI:0.958~0.704。校正曲线显示,进展为重症肺炎的预测概率与观察概率一致性较好,临床决策曲线表明列线图具有临床价值。结论呼吸频率、Neu、CRP、LDH、ESR与成人CAP患者进展为重症肺炎相关,基于上述因素建立的模型可以较好地预测成人CAP进展为重症肺炎的风险,可早期识别并预测重症肺炎,为指导临床治疗策略提供依据。

Objective To explore the risk factors affecting the progression of adult community⁃acquired pneumonia(CAP)to severe pneumonia(SP),construct a clinical prediction model of SP in adults,and provide a basis for early clinical identification and prediction of severe pneumonia in clinical treatment strategies.Methods The clinical data of 376 patients diagnosed with adult CAP who were hospitalized in Jinzhou Central Hospital from June 2022 to September 2023 were collected,and the enrolled cases were divided into a training set of 301 cases and a validation set of 75 cases according to the order of admission time.The CAP patients in the training set were divided into severe group and non⁃severe group based on discharge diagnosis.Single⁃factor analysis and multi⁃factor logistic regression analysis were used to screen predictive factors,build a prediction model for the progression of CAP to severe pneumonia,and draw a nomogram.The receiver operating characteristic(ROC)curve,calibration curve and clinical decision curve were used to evaluate the discrimination,calibration and clinical applicability of the model.Results The training set included 107 patients with severe pneumonia and 194 patients with non⁃severe pneumonia.Univariate analysis of each clinical index in the training set showed that consciousness disorder,pulmonary infiltration>50%,respiratory rate,neutrophils(Neu),C⁃reactive protein(CRP),lactate dehydrogenase(LDH),blood sedimentation(ESR),and lactic acid(Lac)were statistically different(P<0.05).The binary multivariate logistic regression analysis showed that,respiratory rate(OR=1.528,95%CI:1.220-1.913),Neu(OR=1.179,95%CI:1.087-1.278),CRP(OR=1.012,95%CI:1.005-1.019),LDH(OR=1.008,95%CI:1.004-1.011),ESR(OR=1.020,95%CI:1.004-1.037),the final model included these five variables as predictors to construct the model,and to plot the risk nomograms of CAP progression to severe pneumonia.The area under the ROC curve(AUC)in the training set was 0.847,95%CI:0.804-0.729(P<0.05),and the AUC in the validation set was 0.861,95%CI:0.958-0.704.The correction curve showed that the predicted probability of progression to severe pneumonia agreed well with the observed probability,and the clinical decision curve indicated that the nomogram has clinical value.Conclusions Respiratory rate,Neu,CRP,LDH,and ESR are associated with progression to SP in adult patients with CAP.The model established based on the above factors can better predict the risk of CAP progression to SP in adults,identify and predict SP early,and provide a basis for guiding clinical treatment strategies.