老年骨科大手术后谵妄的危险因素分析及预测模型的构建

Analysis of risk factors and construction of prediction model of postoperative delirium after major orthope⁃dic surgery in elderly patients

目的目前国内关于骨科大手术后谵妄(POD)风险预测模型的相关研究鲜有报道。文中主要通过评估老年骨科大手术患者POD的危险因素,构建风险评分预测模型并进行验证。方法收集2022年1月至2022年8月在宁波市第六医院接受骨科大手术的802例患者临床资料。其中分析队列561例,预测队列241例。根据分析队列术后根据意识模糊评估法分为谵妄组(n=102)和非谵妄组(n=459)。在分析队列中先进行单因素分析筛选相关的风险因素,然后采用逐步Logistic回归分析筛选出独立风险因素并根据优势比(OR)的大小进行赋值,建立风险评分预测模型。根据预测模型对分析队列和预测队列分别进行评分,对该模型的有效性进行验证。结果年龄≥73岁、心脏疾病、糖尿病、睡眠紊乱、皮质醇≥411 nmol/L、CRP≥14 mg/L为老年骨科大手术患者POD的危险因素(P<0.05)。上述6个危险因素组成老年骨科大手术患者POD实用预测模型。在分析队列中,该模型受试者工作特征曲线(ROC)下面积AUC为0.863,灵敏度为78.43%,特异度为78.87%,准确率为78.79%;在预测队列中,AUC为0.941,灵敏度为95.24%,特异度为89.45%,准确率为90.46%。结论由年龄≥73岁、心脏疾病、糖尿病、睡眠紊乱、CRP≥14 mg/L、皮质醇≥411 nmol/L构建的风险评分预测模型对于老年骨科大手术患者具有一定的的预测效能。

Objective At present,there are few reports about the risk prediction model of postoperative delirium(POD)af⁃ter major orthopaedic surgery in China.In this paper,the risk factors of POD in elderly patients undergoing major orthopedic surgery were evaluated,and a risk score prediction model was constructed and verified.Methods Clinical data of 802 patients who under⁃went major orthopaedic surgery in Ningbo Sixth Hospital from January 2022 to August 2022 were collected.There were 561 cases in the analysis cohort and 241 cases in the prediction cohort.The analysis cohort was divided into delirium group(n=102)and non⁃delirium group(n=459)according to the Confusion Assessment Method.In the analysis cohort,the relevant risk factors were first screened by single factor analysis,and then independent risk factors were screened by stepwise Logistic regression analysis and assigned according to the odds ratio(OR)to establish a risk score prediction model.According to the prediction model,the analysis cohort and the predic⁃tion cohort are scored respectively to verify the validity of the model.Results Age≥73 years old,heart disease,diabetes,sleep disor⁃der,cortisol≥411 nmol/L,CRP≥14 mg/L were the risk factors for POD in elderly patients undergoing major orthopaedic surgery(P<0.05).The above six risk factors constituted the POD practical pre⁃diction model for elderly patients undergoing the surgery.In the analy⁃sis cohort,the area AUC under receiver operating characteristic curve(ROC)of the model was 0.863,sensitivity 78.43%,specificity 78.87%,and accuracy 78.79%.In the prediction cohort,the AUC,sensitivity,specificity and accuracy was 0.941,95.24%,89.45%,and 90.46%,respectively.Conclusion The risk score predic⁃tion model established by age≥73 years,heart disease,diabetes,sleep disorder,CRP≥14 mg/L,cortisol≥411 nmol/L has certain predictive efficacy for elderly patients undergoing major orthopaedic surgery.