基础医学与临床 ›› 2024, Vol. 44 ›› Issue (1): 98-102.doi: 10.16352/j.issn.1001-6325.2024.01.0098

• 临床研究 • 上一篇    下一篇

老年单侧全髋关节置换术围术期输血风险预测模型的构建

臧晗1, 胡嫒1, 许轩奇2,3, 许力1*   

  1. 1.中国医学科学院 北京协和医学院 北京协和医院 麻醉科,北京 100730;
    2.高可信软件技术教育部重点实验室(北京大学), 北京 100084;
    3.北京大学 计算机学院,北京 100084
  • 收稿日期:2023-10-18 修回日期:2023-11-21 出版日期:2024-01-05 发布日期:2023-12-25
  • 通讯作者: *:pumchxuli@163.com
  • 基金资助:
    中央高水平医院临床科研业务费资助(2022-PUMCH-B-119)

Development of a predictive model for perioperative blood transfusion in elderly patients undergoing unilateral total hip arthroplasty

ZANG Han1, HU Ai1, XU Xuanqi2,3, XU Li1*   

  1. 1. Department of Anesthesiology, Peking Union Medical College Hospital, CAMS & PUMC, Beijing 100730;
    2. Key Laboratory of High Confidence Software Technologies(Peking University), Ministry of Education, Beijing 100084;
    3. School of Computer Science, Peking University, Beijing 100084, China
  • Received:2023-10-18 Revised:2023-11-21 Online:2024-01-05 Published:2023-12-25
  • Contact: *:pumchxuli@163.com

摘要: 目的 分析老年单侧全髋关节置换术围术期输血危险因素并建立风险预测模型。方法 回顾性收集于2013年1月至2021年10月在北京协和医院接受单侧初次全髋关节置换术的老年患者467例。将全部数据的70%划分为训练集,30%划分为测试集。根据是否接受围术期输血将训练集中的患者划分为输血组和非输血组。通过单因素与多因素Logistic回归分析患者的人口学特征、手术信息和术前实验室指标,识别围术期输血的危险因素,结合临床经验构建预测模型并绘制列线图。在测试集中使用受试者工作特征(ROC)曲线和校准曲线评估模型性能。结果 在纳入的467例患者中,91例(19.5%)患者接受围术期输血。多因素Logistic回归分析显示合并冠心病、手术时间增加和术前低血红蛋白是围术期输血的危险因素(P<0.05)。根据统计分析结果与临床经验,纳入是否合并冠心病、手术时间、术前血红蛋白、年龄和是否美国麻醉医师协会(ASA)分级>Ⅱ级等因素构建预测模型,模型的受试者工作特征曲线下面积(AUC)为0.809。结论 老年单侧全髋关节置换术围术期输血风险预测模型的表现良好,可以为临床工作提供帮助。

关键词: 围术期输血, 全髋关节置换术, 老年患者, 输血风险预测

Abstract: Objective To analyze risk factors for perioperative blood transfusion in elderly patients undergoing unilateral primary total hip arthroplasty and develop a prediction model. Methods The study retrospectively collected 467 elderly patients receiving unilateral primary total hip arthroplasty between January 2013 and October 2021 at Peking Union Medical College Hospital. The 70% of the data were used as the training set and the 30% of the data were used as the testing set. Patients were divided into the transfusion and no-transfusion groups based on the presence or absence of perioperative blood transfusion. Univariate analysis and multivariable logistic regression were conducted to analyze patient demographic characteristics, surgical information, and preoperative laboratory tests for identifying risk factors. Clinical experience was combined to establish a prediction model and draw the nomogram. The receiver operating characteristic(ROC) curve and calibration curve were used to evaluate the model in the testing set. Results A total of 91 patients(19.5%) received perioperative blood transfusion. Multivariable logistic regression suggested the history of coronary artery disease, prolonged operation time, and lower preoperative hemoglobin were risk factors for perioperative blood transfusion(P<0.05). The prediction model was constructed based on the results of statistical analysis and clinical experience, including the history of coronary artery disease, operation time, preoperative hemoglobin, age, and American Society of Anesthesiologists(ASA) physical status>Ⅱ. The area under the receiver operating characteristic curve(AUC) of the model was 0.809. Conclusions The prediction model for perioperative blood transfusion in elderly patients undergoing unilateral total hip arthroplasty had a good performance and could assist in clinical practice.

Key words: perioperative blood transfusion, total hip arthroplasty, elderly patients, blood transfusion risk prediction

中图分类号: