全髋关节置换术后谵妄发生的危险因素及风险预测模型的构建

Risk factors of delirium after total hip arthroplasty and construction of a risk prediction model

目的探讨全髋关节置换术后谵妄的影响因素,并构建预测列线图模型。方法回顾性分析2018年4月—2023年4月行全髋关节置换术治疗的320例患者的临床资料,根据术后谵妄发生情况将其分为发生谵妄组72例和未发生谵妄组248例。采用单因素方差分析比较2组一般资料;将有统计学意义的指标纳入二元Logistic回归模型,分析影响全髋关节置换术后谵妄的相关因素,并构建预测模型。结果二元Logistic回归分析显示,年龄、合并慢性阻塞性肺疾病、股骨颈骨折、手术时间、术中出血量、术后4 h的C反应蛋白(CRP)水平、术后低氧血症、复苏时间(≥1 h)、术后疼痛评分(≥4分)、伴有营养障碍、伴有睡眠障碍、伴有认知功能障碍是影响全髋关节置换术后谵妄的危险因素(P<0.05)。纳入上述因素构建的预测模型的整体预测准确率为99.7%,Hosmer-Lemeshow拟合度检验分析显示模型具备较好的拟合优度。Bootstrap法内验证预测模型的曲线下面积(AUC)为0.992,敏感度为99.6%,特异度为93.1%。结论年龄、合并慢性阻塞性肺疾病、股骨颈骨折、手术时间、术中出血量、术后4 h的CRP水平、术后低氧血症、复苏时间(≥1 h)、术后疼痛评分(≥4分)、伴有营养障碍、伴有睡眠障碍、伴有认知功能障碍是影响全髋关节置换术后谵妄的危险因素,基于上述相关因素构建的风险预测模型具有较好的稳定性与预测性。

Objective To explore the influencing factors of delirium after total hip arthroplasty and construct a predictive Nomogram model.Methods Clinical materials of 320 patients with total hip arthroplasty from April 2018 to April 2023 were analyzed retrospectively,and they were divided into delirium group(n=72)and no delirium group(n=248)according to incidence of postoperative delirium.One-way ANOVA was used to analyze the general materials between two groups;the significant indicators were enrolled into a binary Logistic regression model,the relevant factors affecting delirium after total hip arthroplasty were analyzed,and a predictive model was established.Results Binary Logistic regression analysis showed that age,concomitant chronic obstructive pulmonary disease,femoral neck fracture,operation time,intraoperative blood loss,level of C-reaction protein(CRP)at 4 hours after operation,postoperative hypoxemia,resuscitation time(≥1 h),postoperative pain score(≥4 points),concomitant dystrophia,concomitant sleep disorders and concomitant cognitive impairment were the risk factors affecting delirium after total hip arthroplasty(P<0.05).The overall prediction accuracy of the prediction model constructed by incorporating the above factors was 99.7%,and the Hosmer-Lemeshow fit test analysis indicated that the model had good goodness of fit.The area under the curve(AUC)of the Bootstrap method for internal validation of prediction model was 0.992,with a sensitivity of 99.6%and a specificity of 93.1%.Conclusion Age,concomitant chronic obstructive pulmonary disease,femoral neck fracture,operation time,intraoperative blood loss,level of CRP at 4 hours after operation,postoperative hypoxemia,resuscitation time(≥1 h),postoperative pain score(≥4 points),concomitant dystrophia,concomitant sleep disorders and concomitant cognitive impairment are the risk factors affecting delirium after total hip arthroplasty.The risk prediction model constructed based on the above relevant factors has good stability and predictability.