【摘要】 目的 探究老年膝关节置换术后患者发生贫血的影响因素，并构建预测模型。方法 选择2021年10月—2023年8月在山东大学齐鲁医院行膝关节置换术的229例老年患者，按照术后是否贫血将患者分为无贫血组与贫血组。采用Lasso回归、logistic回归分析筛选术后发生贫血的独立危险因素，由此构建预测模型并评估模型的准确性。结果 Lasso回归及多因素logistic回归结果显示，中度衰弱（OR=5.195,95%CI:1.462～18.456,P=0.011）、术前贫血（OR=21.754,95%CI:2.680～176.608,P=0.027）、尿素氮异常（OR=3.021,95%CI:1.214～7.518,P=0.017）是术后发生贫血的危险因素。将这些因素纳入预测模型并绘制列线图。受试者工作特征曲线显示该预测模型具有良好的区分度，其曲线下面积为0.714(95%CI:0.644～0.785)。校准曲线证实模型与实际观测结果具有较好的一致。临床决策曲线表明，当术后贫血发生概率为0.18～0.74时，使用本研究得出的列线图具有临床有效性。结论 中度衰弱、术前贫血及尿素氮异常是老年患者膝关节置换术后发生贫血的影响因素，基于上述因素构建的临床预测模型对患者术后贫血的发生具有一定的预测价值。更多还原

【Abstract】 Objective To explore the risk factors of postoperative anemia in elderly patients undergoing knee replacement and construct the predictive model. Methods A total of 229 older patients who underwent knee replacement from October 2021 to August 2023 in Qilu Hospital of Shandong University were selected and divided into anemia group and no anemia group according to whether they were anaemic after surgery. Lasso regression analyses and logistic regression analyses were used to identify independent risk factors for postoperative anemia, and the prediction model was constructed and its accuracy was evaluated. Results Lasso regression and multivariate logistic regression suggested that moderate frailty(OR=5. 195, 95% CI: 1. 462-18. 456, P=0. 011), preoperative anemia(OR=21. 754, 95% CI: 2. 680-176. 608, P=0. 027), and abnormal BUN(OR=3. 021, 95% CI: 1. 214-7. 518, P=0. 017) were risk factors for postoperative anemia. A nomogram was established based on these factors. ROC curve indicated that the prediction model had good discrimination, and AUC was 0. 714(95% CI: 0. 644-0. 785). The calibration curve confirmed a good consistency between the model prediction results and the actual observations. DCA showed that when the risk threshold of postoperative anemia was 0. 18-0. 74, the use of this nomograph in the current study was more clinically valuable. Conclusions Moderate frailty, preoperative anemia, and abnormal BUN are the influence factors for postoperative anemia of knee replacement. The clinical prediction model based on these factors has a certain predictive value for postoperative anemia.更多还原