脑卒中后吞咽障碍患者吸入性肺炎的影响因素分析及预测模型构建

Predicting aspiration pneumonia among stroke survivors with dysphagia

目的分析脑卒中后吞咽障碍患者发生吸入性肺炎(AP)的影响因素并构建预测模型。方法采用回顾性调查方法,共纳入符合标准的脑卒中后吞咽障碍患者148例,根据住院期间是否发生AP将其分为AP组和非AP组,收集入选患者的一般人口学信息、临床资料等进行单因素分析和多因素Logistic回归分析,并构建列线图模型,同时采用受试者工作特征(ROC)曲线分析各相关因素及列线图模型对AP的预测价值。结果AP组共有患者46例,非AP组共有患者102例,AP发生率为31.08%;通过多因素Logistic回归分析发现,年龄、Barthel指数评分、洼田饮水试验评级、营养支持方式、中性粒细胞/淋巴细胞比值(NLR)是脑卒中后吞咽障碍患者发生AP的重要影响因素(P<0.05);通过分析ROC曲线发现,年龄、洼田饮水试验评级、营养支持方式、NLR、Barthel指数评分及列线图模型预测AP的曲线下面积(AUC)分别为0.598、0.651、0.676、0.723、0.775、0.902,上述指标预测AP的灵敏度及特异度分别为91.30%和31.40%、97.80%和32.40%、78.30%和56.90%、67.40%和71.60%、82.60%和61.80%、84.80%和80.40%。结论年龄、洼田饮水试验评级、营养支持方式、NLR、Barthel指数评分均是影响脑卒中后吞咽障碍患者发生AP的重要因素,采用列线图模型预测AP的能力显著优于上述各因素单独预测。

Objective To analyze the risk factors for aspiration pneumonia(AP)among stroke survivors with dysphagia and construct a prediction model.Methods The records of one hundred and forty-eight stroke survivors with dysphagia admitted since 2021 were retrospectively analyzed.They were divided into an AP group(n=46)and a non-AP group(n=102).Univariate and multivariate logistic regressions were evaluated seeking independent risk factors for AP.the receiver-operating characteristic(ROC)curves were used to analyze predictive power.Results The incidence of AP was 31%.Multivariate logistic regression analysis showed that age,Barthel index,Kubota′s water drinking test grading,nutritional support methods and the neutrophil/lymphocyte ratio(NLR)could be useful independent predictors of AP.The areas under the ROC curves suggested that Kubota grading,nutritional support methods,the NLR and the Barthel index could be used to prepare a prediction nomogram with good sensitivity and specificity.Conclusions Age,Barthel index,Kubota water drinking test grades,nutritional support methods and NLR are independent predictors of AP for stroke survivors with dysphagia.Combining them in a nomogram gives better predictive power than using the factors alone.