老年髋部骨折患者住院期间发生术后心力衰竭的列线图预测模型的构建及验证

Construction and validation of a nomogram for predicting in-hospital postoperative heart failure in elderly patients with hip fracture

目的:构建预测老年髋部骨折患者住院期间发生术后心力衰竭(postoperative heart failure, PHF)的列线图预测模型并验证其效能。方法:采用回顾性队列研究设计,将2015年7月至2023年12月在北京大学第三医院住院接受手术治疗的老年髋部骨折患者(≥65岁)纳入研究,以2015年7月至2021年12月入院的患者为开发队列,以2022年1月至2023年12月入院的患者为验证队列,通过电子病历系统收集患者的临床资料。采用单因素及多因素Logistic回归分析筛选老年髋部骨折患者PHF的预测因素,并通过R语言构建列线图预测模型,采用Bootstrap方法进行内外部验证。分别采用受试者工作特征(receiver operating characteristic, ROC)曲线下面积(area under the curve, AUC)、校准曲线、Hosmer-Lemeshow拟合优度检验、决策曲线分析(decision curve analysis, DCA)评价模型的区分度、校准度和临床适用性。结果:开发队列共944例患者,其中54例(5.7%)住院期间发生PHF;验证队列共469例患者,其中18例(3.8%)住院期间发生PHF。多因素Logistic回归分析显示,年龄(OR=1.071, 95%CI:1.019~1.127,P=0.008)、合并心脏病史(OR=5.360, 95%CI:2.808~10.234,P<0.001)、术前血红蛋白(OR=0.979, 95%CI:0.960~0.999,P=0.041)、术前血肌酐(OR=1.007, 95%CI:1.001~1.013,P=0.015)、髋关节置换术(OR=2.513, 95%CI:1.259~5.019,P=0.009)和全身麻醉(OR=2.024, 95%CI:1.053~3.890,P=0.034)是老年髋部骨折患者发生PHF的独立预测因素。基于4个术前预测因素构建列线图预测模型,模型内外部验证的AUC值分别为0.818(95%CI:0.768~0.868)、0.873(95%CI:0.805~0.929),显示其预测效能良好。校准曲线和Hosmer-Lemeshow拟合优度检验(内部验证χ^(2)=9.958,P=0.354;外部验证χ^(2)=5.477,P=0.791)显示模型的预测准确性较高。DCA表明该模型具有一定的临床适用性。结论:本研究建立了一个简单易用的术前列线图预测模型,可有效预测老年髋部骨折患者发生PHF的风险,有助于早期识别PHF高危患者,优化围手术期管理。

Objective:To construct and validate a nomogram for prediction of in-hospital postoperative heart failure(PHF)in elderly patients with hip fracture.Methods:This was a retrospective cohort study.The patients aged≥65 years undergoing hip fracture surgery in Peking University Third Hospital from July 2015 to December 2023 were enrolled.The patients admitted from July 2015 to December 2021 were divided into a development cohort,and the others admitted from January 2022 to December 2023 in to a validation cohort.The patients’clinical data were collected from the electronic medical record system.Univariate and multivariate Logistic regression were employed to screen the predictors for PHF in the patients.The R software was used to construct a nomogram.Internal and external validation were performed by the Bootstrap method.The discriminatory ability of the model was determined by the area under the receiver operating characteristic curve(AUC).The calibration was evaluated by the calibration plot and Hosmer-Lemeshow goodness-of-fit test.Decision curve analysis(DCA)was performed to assess the clinical utility.Results:In the study,944 patients were eventually enrolled in the development cohort,and 469 were in the validation cohort.A total of 54(5.7%)patients developed PHF in the deve-lopment cohort,and 18(3.8%)patients had PHF in the validation cohort.Compared with those from non-PHF group,the patients from PHF group were older,had higher prevalence of heart disease,hypertension and pulmonary disease,had poorer American Society of Anesthesiologists(ASA)classification(Ⅲ-Ⅳ),presented with lower preoperative hemoglobin level,lower left ventricular ejection fraction,higher preoperative serum creatinine,received hip arthroplasty and general anesthesia more frequently.Multivariate Logistic regression analysis showed that age(OR=1.071,95%CI:1.019-1.127,P=0.008),history of heart disease(OR=5.360,95%CI:2.808-10.234,P<0.001),preoperative hemoglobin level(OR=0.979,95%CI:0.960-0.999,P=0.041),preoperative serum creatinine(OR=1.007,95%CI:1.001-1.013,P=0.015),hip arthroplasty(OR=2.513,95%CI:1.259-5.019,P=0.009),and general anesthesia(OR=2.024,95%CI:1.053-3.890,P=0.034)were the independent predictors for PHF in elderly patients with hip fracture.Four preoperative predictors were incorporated to construct a preoperative nomogram for PHF in the patients.The AUC values of the nomogram in internal and external validation were 0.818(95%CI:0.768-0.868)and 0.873(95%CI:0.805-0.929),indicating its good accuracy.The calibration plots and Hosmer-Lemeshow goodness-of-fit test(internal validation:χ2=9.958,P=0.354;external validation:χ2=5.477,P=0.791)showed its satisfactory calibration.Clinical usefulness of the nomogram was confirmed by decision curve analysis.Conclusion:An easy-to-use nomogram for prediction of in-hospital PHF in elderly patients with hip fracture is well developed.This preoperative risk assessment tool can effectively identify patients at high risk of PHF and may be useful for perioperative management optimization.