老年髋部骨折患者人工关节置换术后发生谵妄的网页版列线图构建研究

Construction of an online nomogram of risk factors for postoperative delirium in elderly patients with hip fracture

目的探讨老年髋部骨折患者人工关节置换术后发生谵妄的危险因素及网页版列线图的构建。方法回顾性分析南京医科大学金陵临床医院骨科-关节外科2020年5月至2021年8月期间采用人工关节置换术治疗的483例老年髋部骨折患者资料。男166例, 女317例;年龄为61~99岁, 平均82.1岁;骨折类型:股骨颈骨折333例, 股骨转子间骨折150例。按照患者术后是否发生谵妄分为谵妄组(n=149)和非谵妄组(n=334), 比较两组患者的年龄、性别、体重指数、骨折类型、合并内科疾病等一般资料, 以及术前白蛋白、血红蛋白及术后C-反应蛋白(CRP)等资料的差异。将P < 0.05的因素纳入多因素logistic回归分析, 筛选出术后发生谵妄的危险因素。利用R软件的"rms"程序包绘制列线图;采用Bootstrap法重复抽样1 000次进行验证, 计算一致性指数, 绘制受试者工作特征(ROC)曲线及校正曲线;利用"rmda"软件包绘制决策曲线。结果谵妄组与非谵妄组患者的年龄、术前焦虑情况、阿尔茨海默病、脑血管疾病史、术前白蛋白、术中低血压及术后CRP等比较差异均有统计学意义(P < 0.05)。多因素logistic回归分析结果显示:高龄、术前焦虑、阿尔茨海默病、术前白蛋白< 35 g/L及术后CRP≥90 mg/L是老年髋部骨折患者人工关节置换术后发生谵妄的危险因素(P < 0.05)。纳入术后谵妄危险因素构建的列线图的ROC曲线下面积为0.894(95%CI:0.865~0.923), 一致性指数为0.889;该模型的校正曲线趋近于理想曲线。决策曲线分析结果显示:阈值为0.01~1.00, 该列线图模型预测老年髋部骨折患者人工关节置换术后发生谵妄的净获益率> 0。结论患者高龄、术前焦虑、阿尔茨海默病、术前白蛋白< 35 g/L及术后CRP ≥90 mg/L是老年髋部骨折患者人工关节置换术后发生谵妄的危险因素, 基于此构建的网页版列线图对术后谵妄的发生具有良好的预测价值。

Objective:To investigate the risk factors for postoperative delirium in the elderly patients with hip fracture and to construct an online nomogram of the risk factors.Methods:Retrospectively analyzed were the data of 483 elderly patients with hip fracture who had been treated with artificial joint replacement from May 2020 to August 2021 at Department of Orthopaedics (Department of Joint Surgery), Jinling Hospital Affiliated to Medical College of Nanjing University. There were 166 males and 317 females, aged from 61 to 99 years (average, 82.1 years). Fracture types: 333 femoral neck fractures and 150 intertrochanteric fractures. The patients were divided into a delirium group ( n=149) and a delirium-free group ( n=334) according to whether postoperative delirium occurred after surgery. The 2 groups were compared in terms of general data like age, gender, body mass index, and concomitant diseases, as well as in terms of indexes like pre-operative albumin, preoperative hemoglobin, and postoperative C-reactive protein (CRP). Factors with P < 0.05 were included in the multi-factor logistic regression analysis to screen out the risk factors for postoperative delirium. The "rms" package of R software was used to draw the nomogram;the Bootstrap method was used to repeat the sampling 1,000 times for evaluation, calculation of the consistency index ( CI), and drawing of the ROC curve and correction curve;the decision curve was plotted using the "rmda" package. Results:There were significant differences between the delirium group and the delirium-free group in age, preoperative anxiety, Alzheimer's disease, history of cerebrovascular disease, preoperative albumin, intraoperative hypotension and postoperative CRP ( P < 0.05). The multifactorial logistic regression analysis showed that high age, preoperative anxiety, Alzheimer's disease, preoperative albumin < 35 g/L, and postoperative CRP ≥90 mg/L were the risk factors for postoperative delirium in the elderly patients with hip fracture after artificial joint replacement ( P < 0.05). The area under the ROC curve of the nomogram constructed by incorporating the risk factors for postoperative delirium was 0.894 (95% CI: 0.865 to 0.923) with a CI of 0.889;the calibration curve showed that the calibration curve of this nomogram model tended to be close to the ideal curve. The decision curve analysis showed that the threshold value was 0.01 to 1.00, showing the net benefit rate of this nomogram model > 0 when used to predict the postoperative delirium in the elderly patients with hip fracture. Conclusions:High age, preoperative anxiety, Alzheimer's disease, preoperative albumin < 35 g/L, and postoperative CRP ≥90 mg/L may be the risk factors for postoperative delirium in the elderly patients with hip fracture after artificial joint replacement. The online nomogram based on these factors demonstrates a good value in prediction of postoperative delirium.