天津医药

天津医药 ›› 2021, Vol. 49 ›› Issue (6): 641-645.doi: 10.11958/20201943

• 应用研究 • 上一篇    下一篇

老年髋部骨折患者术后谵妄的预测模型构建及初步应用 #br#

张明媚，朱星波 ，黄立新   

  1. 1苏州大学附属第一医院骨外科（邮编215006），2普外科
  • 收稿日期:2020-07-12 修回日期:2021-03-11 出版日期:2021-06-15 发布日期:2021-06-15
  • 通讯作者: 张明媚 E-mail:zhangmingmei1987@163.com
  • 基金资助:
    苏州市科技计划项目

Prediction model construction and preliminary application of postoperative delirium in elderly hip fracture patients #br#

ZHANG Ming-mei, ZHU Xing-bo, HUANG Li-xin #br#   

  1. 1 Department of Orthopedic Surgery, 2 Department of General Surgery, the First Affiliated Hospital of Soochow University,
    Suzhou 215006, China
  • Received:2020-07-12 Revised:2021-03-11 Published:2021-06-15 Online:2021-06-15

PDF (PC)

3614

摘要：

目的 构建预测模型评估老年髋部骨折患者术后谵妄（POD）发生风险的预测模型，并验证其可行性。方法 便利抽取确定行择期髋部骨折手术治疗的老年患者205例，收集患者年龄、性别、合并疾病、营养障碍、衰弱程度、术前白蛋白、术中输血、手术方式、手术时长、美国麻醉师协会（ASA）评分、麻醉方式、术后C反应蛋白（CRP）含量。采用多因素Logistic回归筛选POD的影响因素并建立风险预测模型，绘制列线图，受试者工作特征（ROC）曲线分析模型的预测效果。另外抽取80例行髋部手术的老年患者初步对预测模型进行外部验证。 结果205例老年髋部骨折手术患者中73例（35.61%）发生POD，多因素Logistic回归分析显示年龄、手术时长、合并心脑血管疾病、麻醉方式、衰弱程度、ASA分级是髋部骨折患者术后POD的影响因素。纳入上述因素构建的列线图模型预测POD的ROC曲线下面积为0.915（95%CI：0.869~0.950），敏感度为78.49%，特异度为85.97%。外部验证队列中，80例患者实际共21例发生POD，模型预测24例发生POD，预测敏感度为90.48%，特异度为91.53%。 结论 本研究建立的预测模型适用于老年髋部骨折患者术后POD的风险评估，预测效果较好，可为临床医护人员及时识别高风险人群提供参考。

关键词: 髋骨折, 老年人, Logistic模型, 列线图, 术后谵妄, 风险评估

Abstract:

Objective To establish an effective predictive model for assessing the risk of postoperative delirium (POD)
in elderly patients with hip orthopedics and verify its feasibility. Methods A total of 205 elderly patients undergoing hip
surgery in the department of orthopedics were selected. The data of patient age, gender, comorbidities, dystrophia,
debilitation, preoperative albumin, intraoperative blood transfusion, surgical method, operative duration, the American
Society of Anesthesiologists (ASA) grade, anesthesia method and postoperative C-reactive protein (CRP) were collected.
Multivariate Logistic regression was used to screen the influencing factors of POD, and a risk prediction model was
established. The histogram was drawn. Receiver operating characteristic (ROC) curve was used to analyze the prediction
effect of the model. In addition, 80 elderly patients who underwent hip surgery were selected to preliminarily verify the
effectiveness of the prediction model. Results A total of 73 cases occurred POD in 205 elderly patients with hip surgery,
and the incidence rate was 35.61%. Univariate and multivariate Logistic regression analysis showed that age, duration of
operation, complicated cardiovascular and cerebrovascular diseases, anesthesia method, frailty and ASA grade were
influencing factors of POD. The AUC under ROC curve was 0.915 (95%CI: 0.869-0.950), sensitivity was 78.49% and
specificity was 85.97%. In the preliminary validation cohort, 21 cases (80 patients) developed POD, and the model
predicted 24 cases. The sensitivity and specificity were 90.48% and 91.53%, respectively. ConclusionThe prediction
model established in this study is suitable for the risk assessment of perioperative evaluation of elderly hip fracture patients,
which provides reference for clinical medical staff to identify high-risk population and provides targeted preventive
intervention.

Key words: hip fractures, aged, Logistic models, nomograms, postoperative delirium, risk assessment