老年髋部骨折患者术后30 d发生主要并发症的风险预测模型构建及验证研究

Construction and validation of a risk prediction model for major complications 30 days after surgery in elderly patients with hip fracture

目的构建老年髋部骨折患者术后30 d内发生主要并发症的风险预测模型并进行验证。方法回顾性分析2019年6月至2021年12月期间承德医学院附属医院创伤骨科收治的276例老年髋部骨折患者资料。男96例,女180例;年龄(74.5±9.3)岁;骨折类型:股骨颈骨折139例,股骨转子间骨折137例。以术后30 d内是否发生主要并发症为研究结局。应用多因素logistic回归分析老年髋部骨折患者术后30 d内发生主要并发症的危险因素。采用向前逐步法及似然比检验进行最佳预测模型的筛选,构建并使用列线图展示模型。采取受试者工作特征曲线(ROC)、Hosmer-Lemeshow拟合优度检验、临床决策曲线和临床影响曲线分析对模型的稳定性和有效性进行评价。结果多因素logistic回归分析结果显示:术前血红蛋白减少(P<0.05)、入院至手术时间>72 h(OR=3.001,95%CI:1.564~5.758,P<0.001)、控制营养状态评分>4分(OR=3.394,95%CI:1.724~6.680,P<0.001)、年龄调整的改良衰弱指数评分>2分(OR=2.875,95%CI:1.548~5.339,P=0.001)、手术时间增加(OR=1.016,95%CI:1.006~1.025,P=0.001)、手术出血量>60 mL(OR=2.373,95%CI:1.016~5.540,P=0.046)是老年髋部骨折患者术后30 d内发生主要并发症的危险因素。本次构建的logistic风险预测模型ROC曲线下面积为0.846(95%CI:0.799~0.889),Hosmer-Lemeshow拟合优度检验结果显示:χ^(2)=8.080,P=0.426,临床决策曲线和临床影响曲线提示该模型准确性和有效性较好。结论本研究基于患者术前血红蛋白、入院至手术时间、控制营养状态评分、年龄调整的改良衰弱指数评分、手术时间及手术出血量构建了老年髋部骨折患者术后30 d发生主要并发症的风险预测模型,有助于预测患者术后主要并发症的发生。

Objective To construct and validate a risk prediction model for major complications 30 days after surgery in the elderly patients with hip fracture.Methods A retrospective study was conducted to analyze the clinical data of 276 elderly patients with hip fracture who had been admitted to Department of Trauma and Orthopaedics,The Hospital Affiliated to Chengde Medical University from June 2019 to December 2021.There were 96 males and 180 females with an age of(74.5±9.3)years,and 139 femoral neck fractures and 137 intertrochanteric fractures.The outcome of this study was whether major complications occurred within 30 days after surgery.Multiple logistic regression analysis identified the risk factors for major complications in the elderly patients with hip fracture within 30 days after surgery.The forward step-by-step method and likelihood ratio test were used to screen the best prediction model.A nomogram was constructed to display the model.The stability and effectiveness of the model were evaluated by the receiver operating characteristic curve(ROC),Hosmer-Lemeshow goodness-of-fit test,clinical decision curve and clinical impact curve analysis.Results Logistic regression analysis showed that decreased preoperative hemoglobin(P<0.05),time from admission to surgery>72 hours(OR=3.001,95%CI:1.564 to 5.758,P<0.001),control of nutritional status(CONUT)score>4 points(OR=3.394,95%CI:1.724 to 6.680,P<0.001),and age-adjusted modified frailty index(aamFI)>2 points(OR=2.875,95%CI:1.548 to 5.339,P=0.001),increased operation time(OR=1.016,95%CI:1.006 to 1.025,P=0.001),and surgical bleeding>60 mL(OR=2.373,95%CI:1.016 to 5.540,P=0.046)were independent risk factors for major complications within 30 days after surgery in the elderly patients with hip fracture.The area under the ROC curve in the logistic risk prediction model was 0.846(95%CI:0.799 to 0.889),and the results of Hosmer-Lemeshow goodness-of-fit test showed(χ^(2)=8.080,P=0.426).The clinical decision curve and clinical impact curve showed that the prediction model was accurate and effective.Conclusion Based on the patients'preoperative hemoglobin,time from admission to surgery,control of nutritional status score,age-adjusted modified frailty index,operation time and surgical blood loss,this study has constructed successfully a risk prediction model for complications 30 days after surgery in the elderly patients with hip fracture which enables medical staff to predict the occurrence of major postoperative complications.