体外循环心脏瓣膜置换术后谵妄的危险因素及其交互效应

Risk factors and interaction effect of delirium after cardiopulmonary bypass heart valve replacement

目的:探讨体外循环心脏瓣膜置换术患者术后发生谵妄的危险因素及其交互效应。方法:选取2019年1月至2022年6月在广西壮族自治区人民医院心胸血管外科行体外循环心脏瓣膜置换术的358例患者,分为谵妄组(n=92)和非谵妄组(n=266)。采用Logistic回归分析法分析患者术后发生谵妄的危险因素。采用多因子降维法(MDR)分析危险因素之间的相互作用。绘制受试者工作特征(ROC)曲线,评估模型的灵敏度和特异度。结果:两组在年龄、糖尿病、术后使用镇静药物、机械通气时间、体外循环时间、阻断时间等方面比较,差异均有统计学意义(P<0.05)。术后使用镇静药物、糖尿病、体外循环时间三者之间存在显著交互效应(P<0.05)。加入交互效应的回归模型ROC曲线下面积(AUC)为0.859(95%CI:0.815~0.903),最佳截断值为0.267,特异度为0.744,灵敏度为0.739。未调整交互效应的模型AUC为0.851(95%CI:0.806~0.895),最佳截断值为0.266,特异度为0.759,灵敏度为0.761。结论:年龄、术后使用镇静药物、机械通气时间、体外循环时间、阻断时间均为谵妄发生的危险因素,术后使用镇静药物、糖尿病、体外循环时间3个因素间的交互效应可能是术后谵妄发生的重要预测因素。

Objective:To explore the risk factors and interaction effect of postoperative delirium in patients with cardiopulmonary bypass heart valve replacement.Methods:A total of 358 patients with cardiopulmonary bypass heart valve replacement admitted to the Department of Cardiovascular Surgery of Guangxi Zhuang Autonomous Region People’s Hospital from January 2019 to June 2022 were selected and divided into delirium group(n=92)and non-delirium group(n=266).Logistic regression was used to analyze the risk factors of postoperative delirium in patients with cardiopulmonary bypass heart valve replacement.Multifactor dimension reduction(MDR)was used to analyze the interactions among risk factors.Receiver operating characteristic curve(ROC)was drawn to evaluate the sensitivity and specificity of the model.Results:There were statistically significant differences in age,diabetes,postoperative use of sedatives,duration of mechanical ventilation,duration of cardiopulmonary bypass and block duration between the two groups(P<0.05).The use of postoperative sedatives,diabetes and the duration of cardiopulmonary bypass had an significant interaction effect(P<0.05).The area under ROC curve(AUC)of the regression model with interaction effect was 0.859(95%CI:0.815-0.903),the optimal cut-off value was 0.267,the specificity was 0.744,and the sensitivity was 0.739.The AUC of the model without adjustment for interaction effect was 0.851(95%CI:0.806-0.895),the optimal cut-off value was 0.266,the specificity was 0.759,and the sensitivity was 0.761.Conclusion:Age,postoperative use of sedatives,duration of mechanical ventilation,duration of cardiopulmonary bypass and block duration are all risk factors for delirium.The interaction effect of postoperative use of sedatives,diabetes and duration of cardiopulmonary bypass may be important predictors for postoperative delirium.