高血压并发急性脑梗死患者介入术后呼吸机相关性肺炎预测模型构建与验证
投稿时间:2024-10-15  修订日期:2024-11-02  点此下载全文
引用本文:何小花,梁文菲,朱晶玲,何秋杏,莫换好,宁为民,赵湛,陈敬毅.高血压并发急性脑梗死患者介入术后呼吸机相关性肺炎预测模型构建与验证[J].医学研究杂志,2025,54(4):96-101, 129
DOI: 10.11969/j.issn.1673-548X.2025.04.018
摘要点击次数: 170
全文下载次数: 664
作者单位
何小花 广州中医药大学东莞医院脑病科 523000 
梁文菲 广州中医药大学东莞医院脑病科 523000 
朱晶玲 广州中医药大学东莞医院脑病科 523000 
何秋杏 广州中医药大学东莞医院脑病科 523000 
莫换好 广州中医药大学东莞医院脑病科 523000 
宁为民 广州中医药大学东莞医院脑病科 523000 
赵湛 广州中医药大学东莞医院脑病科 523000 
陈敬毅 广州中医药大学东莞医院脑病科 523000 
基金项目:广东省自然科学基金资助项目(2022A1515011665);东莞市社会发展科技项目(20231800915372)
中文摘要:目的 通过分析高血压并急性大血管闭塞性脑梗死(acute ischemic stroke with large vessel occlusion,AIS-LVO)患者介入术后并发呼吸机相关性肺炎(ventilator associated pneumonia, VAP)的临床特征,构建并验证风险预测模型。方法 回顾性分析广州中医药大学东莞医院2020年7月~2023年8月因高血压并AIS-LVO住院治疗患者107例的临床资料,根据患者住院期间是否发生VAP,将其分为VAP(n=64)组和非VAP组(n=43)。采用最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)回归筛选预测因子,再采用多因素Logistic回归分析构建Nomogram模型,绘制受试者工作特征(receiver operating characteristic,ROC)曲线下面积(area under curve,AUC)、校准曲线、和决策曲线(decision curve analysis,DCA)对预测模型的区分度、校准度和临床效用进行评价。结果 107例高血压合并AIS-LVO患者行介入术后,VAP发生率为59.8%,两组患者中年龄、心房颤动、入院美国国立卫生研究院脑卒中量表评分(National Institute of Health Stroke Scale,NIHSS)、入院格拉斯哥昏迷量表(Glasgow Coma Scale,GCS)评分、意识障碍、吞咽功能障碍、手术时间、术后24h全身炎性反应指数(systemic inflammation response index, SIRI)、机械通气时间、ICU住院天数、留置胃管比较,差异有统计学意义(P<0.05),且VAP组不良功能结局发生率为57.8%,非VAP组为14.0%,提示VAP组预后更差。将LASSO回归筛选变量结果纳入多因素Logistic分析,结果显示,术后24hSIRI、机械通气时间(P<0.05)是高血压并AIS-LVO患者介入术后发生VAP的独立危险因素,入院GCS评分为保护因素(P<0.05)。用上述3项指标构建预后预测列线图模型,该模型采用Bootstrap法经过1000次重抽样后对模型进行内部验证,校准曲线与拟合线接近,平均绝对误差为0.029,曲线准确度良好,采用Hosmer and Lemeshow 法评估Logistic 回归模型拟合度(χ2=5.38,P=0.716)良好;ROC曲线下面积为0.812(95% CI:0.730~0.894);DCA显示该模型的临床应用价值较好。结论 根据入院GCS评分、术后24hSIRI、机械通气时间构建的预测模型对高血压并AIS-LVO患者介入术后发生VAP具有较好的预测价值。
中文关键词:高血压 急性大血管闭塞性脑梗死 血管内治疗 呼吸机相关性肺炎 列线图
 
Construction and Validation of a Predictive Model for Ventilator-associated Pneumonia after Endovascular Treatment in Patients with Hypertension Complicating Acute Cerebral Infarction.
Abstract:Objective Through the analysis of the clinical characteristics of post-interventional complications of ventilator-associated pneumonia (VAP) in patients with hypertension and acute large vessel occlusive cerebral infarction (AIS-LVO), we aim to develop and validate a risk prediction model.Methods We conducted a retrospective analysis of clinical data from 107 patients hospitalized for hypertension with AIS-LVO at Dongguan Hospital of Guangzhou University of Traditional Chinese Medicine between July 2020 and August 2023. Patients were categorized into two groups based on their hospitalization:the VAP group, consisting of 64 patients who developed VAP, and the non-VAP group, comprising 43 patients who did not experience VAP. Least absolute shrinkage and selection operator regression(LASSO) analysis was employed to identify potential predictors. Subsequently, a nomogram model was constructed using multifactorial Logistic regression analysis. The model′s discriminative ability, calibration, and clinical utility were assessed using the area(AUC)under the receiver operating characteristic(ROC)curve, calibration curves, and decision curve analysis(DCA). Results In 107 patients with hypertension and AIS-LVO who underwent interventional surgery, the incidence of VAP was 59.8%.The results revealed statistically significant differences between the two groups in terms of age, presence of atrial fibrillation, admission national institutes of health stroke scale (NIHSS) score, admission glasgow coma scale (GCS) score, impaired consciousness, swallowing dysfunction, operative time, postoperative 24-hour systemic inflammatory response index (SIRI), duration of mechanical ventilation, length of ICU stay, and the presence of an indwelling gastrostomy tube (P < 0.05). Additionally, the incidence of poor functional outcomes was 57.8% in the VAP group compared to 14.0% in the non-VAP group, indicating a worse prognosis for patients in the VAP group.The variables identified through LASSO regression screening were included in a multifactorial Logistic regression analysis. This analysis revealed that the SIRI and the duration of mechanical ventilation were independent risk factors for the development of VAP following intervention in patients with hypertension and AIS-LVO (P < 0.05). Conversely, the admission GCS score was identified as a protective factor (P < 0.05). These three indicators were utilized to construct a prognostic nomogram model. The model′s internal validity was assessed using the Bootstrap method with 1000 resamples, and the calibration curves were found to be in close agreement with the ideal fitted line, exhibiting a mean absolute error of 0.029, indicating good curve accuracy. The Logistic regression model′s goodness of fit was evaluated using the Hosmer and Lemeshow test, which yielded a χ2 value of 5.38 and a P-value of 0.716, suggesting a good fit. The AUC of the model was 0.812 (95% confidence interval 0.730-0.894), and the DCA indicated that the model has favorable clinical applicability.ConclusionA prediction model based on admission GCS score, postoperative 24-hour SIRI, and duration of mechanical ventilation demonstrates a robust predictive value for the incidence of VAP following intervention in patients with hypertension and acute ischemic stroke with AIS-LVO.
keywords:Hypertension  Acute large vessel occlusive cerebral infarction  Endovascular therapy  Ventilator-associated pneumonia  Nomogram
查看全文  查看/发表评论  下载PDF阅读器

京公网安备 11010502037822号