缺血性脑卒中后吞咽障碍患者肺部感染的影响因素及预测模型分析OA

Objective To explore the influencing factors of pulmonary infection in patients with dysphagia after ischemic stroke,and to develop a predictive model.Methods A retrospective cohort study was conducted.We retrospectively analyzed 125 patients with dysphagia after ischemic stroke at the Nanjing Drum Tower Hospital,the Affiliated Hospital of Nanjing University Medi-cal School,from August 2021 to December 2024,and the patients were divided into a pulmonary infection group(57 cases)and a non-pulmonary infection group(68 cases)based on the presence or absence of pulmonary infection.General clinical data of all sub-jects were colleted from the electronic medical record system of the Nanjing Drum Tower Hospital.All the included variables were analyzed using univariate logistic regression to screen for potential influencing factors.Variables with statistically significant differ-ences(P<0.05)were included in the multivariate logistic regression analysis to identify risk factors for pulmonary infection in pa-tiens with dysphagia after ischemic stroke.Subsequently a Nomogram predictive model was constructed.The receiver operator char-acteristic(ROC)curve was used to assess the discriminative ability of the model,and the area under the curve(AUC),sensitivity,specificity,and Youden index were calculated to compare the discriminative ability of different models.The Akaike information cri-terion(AIC)and Bayesian information criterion(BIC)values were calculated to further compare the predictive performance of dif-ferent models.The Delong test was utilized to compare the AUC differences among different predictive models.Finally,leave-one-out cross validation was performed for internal validation,and discriminat analysis(reporting Wilks'λ and classfication accuracy)was used to evaluate the stability of the model.Results Univariate logistic regression analysis indicated that hypoproteinemia,stroke location,neutrophil-to-lymphocyte ratio(NLR),water swallowing test(WST)grade,and impaired safety on the volumevis-cosity swallowing test(V-VST)were potential influencing factors for pulmonary infection in patients with dysphagia after ischemic stroke,with statistically significant differences(P<0.05).Multivariate logistic regression analysis demonstrated that elevated NLR(OR=1.424,P=0.016),WST grade Ⅳ(OR=5.657,P=0.002)and impaired safety on V-VST(OR=5.488,P=0.003)were independent risk factors for pulmonary infection.A Nomogram predictive model was constructed by utilizing NLR,WST grade,and V-VST.The AUC values for the three single-indicator prediction models(NLR,WST grade,and V-VST)were 0.733,0.774,and 0.764,respec-tively.The combined prediction model(NLR+WST+V-VST)attained an AUC of 0.878.Compared with the three single-indicator models,the combined model exhibited lower AIC and BIC values,suggesting a superior goodness-of-fit(P<0.05).The Delong test demonstrated that the combined model had a significantly higher AUC value compared with each single-indicator model(P<0.05).Leave-one-out cross-validation indicated that the combined model had a Wilks'λ value of 0.624(P<0.05),with a classification accu-racy of 84.21%for the pulmonary infection group,80.88%for the non-pulmonary infection group,and an overall classification accu-racy of 82.40%.Conclusion The NLR,WST grade,and impaired safety on V-VST can predict the incidence of pulmonary infec-tions in patients with dysphagia after ischemic stroke,and the combined model demonstrates superior predictive value and clinical utility.