基于血栓弹力图和凝血功能指标构建输血原位肝移植患者临床转归的预测模型

A predictive model for the clinical outcome of patients undergoing orthotopic liver transplantation with blood transfusion was constructed based on thromboelastography and coagulation function

摘要:

目的 探讨血栓弹力图（TEG）和凝血功能指标对输血原位肝移植（OLT）患者临床转归（以术后严重并发症为主要观察终点）的影响。方法 回顾性分析2022年2月—2024年9月西安大兴医院和西安交通大学第一附属医院106例行OLT术患者的临床资料。依据9∶1定量将患者随机分为训练集（95例）与验证集（11例）。根据患者术后3个月内是否发生Clavien-Dindo分级≥3级的并发症，将其分为严重并发症组与无严重并发症组。比较两组患者基础资料。采用多因素一般Logistic回归模型分析OLT术患者术后发生严重并发症的影响因素。基于影响因素构建OLT术患者术后发生严重并发症的列线图预测模型并加以验证。结果 训练集中，95例行OLT术患者中，发生严重并发症22例，占23.2%。验证集中，11例行OLT术患者中，发生严重并发症2例，占18.2%。训练集中严重并发症组和无严重并发症组性别构成、肝移植病因、终末期肝病模型评分、年龄、体质量指数、血小板计数、白细胞计数、活化部分凝血活酶时间、凝血酶时间、凝血因子反应时间、凝血综合指数及术中出血量比较，差异均无统计学意义（P >0.05）。严重并发症组纤维原蛋白（FIB）、最大幅度（MA）、夹角均高于无严重并发症组（P <0.05），凝血酶原时间（PT）低于无严重并发症组（P <0.05）。FIB高、MA高、夹角高和PT低均为OLT术患者术后发生严重并发症的影响因素（P <0.05）。训练集模型预测OLT术患者术后发生严重并发症的最佳截断值为0.412，此阈值下对应的敏感性为88.5%（95% CI：0.781，0.989），特异性为85.0%（95% CI：0.746，0.954），曲线下面积（AUC）为0.885（95% CI：0.811，0.958）。验证集风险总分的最佳截断值为0.358，此阈值下对应的敏感性为85.3%（95% CI：0.768，0.938），特异性为86.7%（95% CI：0.782，0.952），AUC为0.879（95% CI：0.814，0.944）。结论 基于TEG和凝血功能指标建立的列线图预测模型对OLT术患者术后发生严重并发症具有良好的预测效能。

Abstract:

Objective To analyze the influencing factors of thromboelastography (TEG) and coagulation function on the clinical outcome of patients undergoing orthotopic liver transplantation (OLT) after blood transfusion, using postoperative major complications as the primary observational endpoint, and to construct a predictive model.Methods A retrospective method was adopted. A total of 106 patients who underwent OLT in Xi'an Daxing Hospital and the First Affiliated Hospital of Xi'an Jiaotong University between February 2022 and September 2024 were selected. According to 9:1, they were quantitatively and randomly divided into the training set (n = 95) and the validation set (n = 11). Based on the occurrence of Clavien-Dindo grade ≥ 3 complications within 3 months after surgery, patients were categorized into a major complications group and a no major complications group. The basic data of the two groups were compared, the factors influencing the occurrence of major complications after OLT were analyzed, and a predictive model for major complications after OLT was constructed and verified.Results In the training set, among 95 patients who underwent OLT, 22 cases had major complications, accounting for 23.2%. In the validation set, among the 11 patients who underwent OLT, 2 cases had major complications, accounting for 18.2%. In the training set, there were no statistically significant differences (P > 0.05). the major complications group had higher levels of FIB, MA, and Angle, and a lower PT compared to the no major complications group (P < 0.05). Higher FIB, higher MA, higher Angle, and lower PT were all identified as influencing factors for the occurrence of major complications after OLT (P < 0.05). In the training set, the optimal cut-off value of the model for predicting major complications after OLT was 0.412, with a corresponding sensitivity of 88.5% (95% CI: 0.781, 0.989) and specificity of 85.0% (95% CI: 0.746, 0.954). The area under the curve (AUC) was 0.885 (95% CI: 0.811, 0.958). In the validation set, the optimal cut-off value for the risk score was 0.358, with a sensitivity of 85.3% (95% CI: 0.768, 0.938) and a specificity of 86.7% (95% CI: 0.782, 0.952). The AUC was 0.879 (95% CI: 0.814, 0.944).Conclusion TEG and coagulation function are associated with the occurrence of major complications after OLT. The establishment of a nomogram prediction model has a good predictive effect on the major complications after OLT.