【摘要】 目的 分析胫骨平台骨折患者围手术期异体输血(ABT)的影响因素,构建风险预测模型,并验证其预测效能。方法 回顾性分析2020年1月至2024年12月我院收治的508例胫骨平台骨折患者的临床资料。按7∶3比例分层抽样分为训练集(356例)和验证集(152例)。训练集数据用于模型构建,以是否接受ABT治疗为因变量,通过LASSO回归筛选出自变量,纳入多因素Logistic回归分析确定独立危险因素,根据独立危险因素构建预测模型。验证集数据用于模型验证,绘制受试者工作特征(ROC)曲线,校准曲线和临床决策曲线(DCA)评价该模型的预测效能和临床应用价值。结果术前Hb水平、出血量和氨甲环酸使用量是胫骨平台骨折患者围手术期ABT的独立影响因素。ROC曲线显示,训练集和验证集的曲线下面积(AUC)分别为0.973(95%CI 0.942～1.000)和0.960(95%CI 0.901～1.000)。校准曲线显示,训练集和验证集均具有较好的一致性。DCA曲线显示训练集和验证集在阈值范围内均具有临床净收益。结论 本研究构建的预测模型具有较好的区分度,一致性和临床净收益,值得临床推广。更多还原

【Abstract】 Objective To analyze the influencing factors of allogeneic blood transfusion(ABT)in patients with tibial plateau fracture during the perioperative period,construct a risk prediction model,and verify its predictive efficacy.Methods From January 2020 to December 2024,a retrospective analysis was conducted on clinical data of 508 patients with tibial plateau fractures admitted to our hospital. They were randomly divided into training set(n = 356)and validation set(n = 152)at a ratio of 7∶3 by stratified sampling. The training set data was used for model construction. Whether ABT treatment was received was regarded as the dependent variable. Independent variables were screened out via LASSO regression and incorporated into the multivariate Logistic regression analysis to identify the independent risk factors. The prediction model was constructed based on these independent risk factors. The validation set data was used for model validation. The receiver operating characteristic(ROC)curve,the calibration curve,and the decision curve analysis(DCA)were employed to assess the predictive efficacy and clinical application value of the model.Results Preoperative Hb,blood loss,and tranexamic acid dosage were independent influencing factors for perioperative ABT in patients with tibial plateau fractures. The ROC curve indicated that the Area under the curve(AUC)of the training set and the validation set were 0. 973(95% CI 0. 942～1. 000)and 0. 960(95% CI 0. 901～1. 000),respectively. The calibration curve demonstrated that both the training set and the validation set had good consistency. The DCA curve revealed that both the training set and the validation set had net clinical benefits within the threshold range.Conclusion The prediction model constructed has good discrimination,consistency and clinical net benefit,which is worthy of clinical promotion.更多还原