分析神经重症康复患者多重耐药菌（MDRO）感染呼吸机相关性肺炎（VAP）的危险因素及1年预后。

收集2020年12月至2022年9月首都医科大学附属北京康复医院神经重症康复病房住院的神经重症康复患者共248例，入院时患者无明显全身感染，根据住院期间是否被诊断为MDRO感染VAP分为病例组（80例）和对照组（168例）。各观察指标中计量资料呈正态分布者采取两个独立样本t检验，非正态分布者采取非参数检验。计数资料采用卡方检验。经单因素分析后有统计学差异的指标为自变量，住院期间是否诊断MDRO感染VAP为因变量，行多因素Logistic回归分析神经重症康复患者MDRO感染VAP的危险因素。绘制受试者工作特征（ROC）曲线，评估各危险因素预测住院期间获得MDRO感染VAP的价值。绘制生存曲线比较病例组与对照组患者1年生存率。

符合入组条件的神经重症患者共248例，住院期间明确为MDRO感染的VAP患者（病例组）80例（32.26%），168例（67.74%）患者未诊断MDRO感染VAP（对照组）。两组患者年龄（Z = 0.221、P = 0.026）、GCS评分（Z = 3.153、P = 0.002）、发病至入院前时间（Z = 2.238、P = 0.025）、入院时血红蛋白（Z = 2.502、P = 0.012）、使用机械通气（χ² = 18.750、P < 0.001）、慢性肾功能不全病史（χ² = 7.394、P = 0.007）、静脉使用抑酸药（χ² = 8.556、P = 0.003）、使用中心静脉导管（CVC）（χ² = 15.262、P < 0.001）和单次气管切开套管囊上引流量（χ² = 17.908、P < 0.001），差异有统计学意义。多因素Logistic回归分析显示：使用机械通气（OR = 2.514、95%CI：1.326～4.767、P = 0.005），使用CVC（OR = 2.400、95%CI：1.210～4.759、P = 0.012），GCS评分（OR = 0.845、95%CI：0.766～0.932、P = 0.001）和单次气管切开套管囊上引流量（OR = 2.323、95%CI：1.494～3.613、P = 0.032）均为神经重症康复患者MDRO感染VAP的独立危险因素。ROC曲线显示预测MDRO感染VAP的危险因素：机械通气时间、中心静脉导管使用时间、GCS评分和单次气管切开套管囊上引流量的灵敏度分别为43.7%、37.5%、68.7%和86.2%，特异性分别为94.0%、91.1%、54.8%和53.0%；曲线下面积分别为0.687、0.632、0.623和0.728，最佳截断值分别为96 h、16 d、8和2.8 ml。生存曲线Log-rank检验显示，病例组较对照组1年生存率较低，差异有统计学意义（χ² = 39.004、P < 0.001）。

缩短机械通气时间、减少非必要导管日、促醒、减少气管切开套管囊上引流量可有效预防神经重症患者MDRO感染VAP。

To investigate the risk factors and one-year prognosis of ventilator-associated pneumonia (VAP) infected by multidrug-resistant organism (MDRO) in patients treated in severe neurological rehabilitation ward.

Total of 248 patients with severe neurological rehabilitation who were hospitalized in the neurological intensive care rehabilitation unit of Beijing Rehabilitation Hospital, Capital Medical University from December 2020 to September 2022 were collected. There were no obvious systemic infection at admission. They were divided into case group (80 cases) and control group (168 cases) according to whether diagnosed with VAP infected by MDRO during hospitalization. Two independent sample t tests were taken for the measurement data with normal distribution and Mann-Whitney U tests were taken for the measurement data with non-normal distribution. The classified data were analyzed by Chi-square analysis. The indicators with significant differences in univariate analysis were independent variables, and the diagnosis of VAP infected by MDRO during hospitalization was the dependent variable. Multivariate Logistic regression analysis was performed to find the risk factors of VAP infected by MDRO in patients with severe neurological rehabilitation. Receiver operating characteristic (ROC) curves were plotted to assess the value of risk factors to predict the acquisition of VAP caused by MDRO during hospitalization. Survival curves were drawn to compare one-year survival rates between VAP group and control group.

Among the 248 patients with severe neurological conditions eligible for enrollment, 80 patients (32.26%) were confirmed of VAP with MDRO during hospitalization (case group), and 168 patients (67.74%) were undiagnosed (control group). The age (Z = 0.221, P = 0.026), GCS score (Z = 3.153, P = 0.002), time from onset to admission (Z = 2.238, P = 0.025), hemochrome at admission (Z = 2.502, P = 0.012), use of mechanical ventilation (χ² = 18.750, P < 0.001), history of chronic renal insufficiency (χ² = 7.394, P = 0.007), intravenous use of acid inhibitors (χ² = 8.556, P = 0.003), use of central vein catheter (CVC) (χ² = 15.262, P < 0.001) and single drainage volume of subglottic secretion (DVSS) (χ² = 17.908, P < 0.001) between the two groups were all significantly different. Multivariate Logistic regression analysis indicated: mechanical ventilation (MV) (OR = 2.514, 95%CI: 1.326-4.767, P = 0.005), use of CVC (OR = 2.400, 95%CI: 1.210-4.759, P = 0.012), GCS score (OR = 0.845, 95%CI: 0.766-0.932, P = 0.001), and DVSS (OR = 2.323, 95%CI: 1.494-3.613, P = 0.032) were independent risk factors for VAP infected by MDRO in patients with severe neurological conditions. ROC curves showed the predicting risk factors for MDRO infection with VAP: the sensitivity of time of MV, time of CVC, GCS and DVSS were 43.7%, 37.5%, 68.7% and 86.2%; and the specificity of them were 94.0%, 91.1%, 54.8% and 53.0%; the areas under the curve of them were 0.687, 0.632, 0.623 and 0.728; and the optimal cut-off values were 96 h, 16 d, 8 min and 2.8 ml, respectively. Comparative survival analysis by Log-rank test displayed that the VAP group infected by MDRO had a lower one-year survival rate than the control group, with significant differences (χ² = 39.004, P < 0.001).

Shortening the time of mechanical ventilation, reducing unnecessary catheter days, promoting wakefulness and reducing DVSS could effectively prevent VAP infected by MDRO in patients with severe neurological conditions.