探讨肝移植受者术后下肢深静脉血栓形成(LEDVT)的危险因素，建立并验证肝移植受者术后LEDVT风险预测模型。

回顾性分析2020年1月至2023年10月在吉林大学第一医院接受同种异体原位肝移植的336例受者临床资料。采用sample.split函数将受者以7 ∶3的比例随机分为建模组(235例)和验证组(101例)。通过文献回顾、课题小组讨论及临床知识确定肝移植受者术后LEDVT相关预测因素。计量资料组间比较采用成组t检验或Mann-Whitney U检验；计数资料组间比较采用卡方检验或Fisher确切概率法。将单因素分析结果中P<0.05的变量纳入多因素Logistic回归分析。使用R软件(4.3.2)绘制列线图，并在shinyapps.io上开发肝移植受者术后LEDVT风险预测模型网页版计算器，采用受试者操作特征曲线(ROC曲线)下面积、Hosmer-Lemeshow拟合优度检验及临床决策曲线评价肝移植受者术后LEDVT风险预测模型的区分度、准确度及临床效益。P<0.05为差异有统计学意义。

建模组235例受者中LEDVT组49例，非LEDVT组186例，LEDVT发生率为20.8%(49/235)。LEDVT组和非LEDVT组受者年龄、术前肝性脑病、日常生活能力水平、凝血因子反应时间及术后AST、ALT、Na^＋水平、Ca^2＋水平、凝血酶原时间(PT)、国际标准化比值差异均有统计学意义(Z＝－3.552、－2.808、－2.567、－2.161、－2.297、－1.986、－3.815和－2.395，χ²＝13.822和36.213，P均<0.05)。多因素Logistic回归分析结果示：年龄(OR＝1.048，95%CI：1.002~1.096)、术前患有肝性脑病(OR＝2.484，95%CI：1.041~5.930)、术前日常生活能力(中度依赖)(OR＝5.266，95%CI：1.685~16.458)、术前日常生活能力(重度依赖)(OR＝8.342，95%CI：1.748~39.802)、术后Na^＋水平(OR＝1.105，95%CI：1.001~1.220)及术后PT(OR＝0.827，95%CI：0.737~0.928)是肝移植受者术后LEDVT的独立危险因素(P均<0.05)。建模组和验证组肝移植受者术后LEDVT风险预测模型ROC曲线下面积分别为0.811(95%CI: 0.745~0.876)和0.736(95%CI：0.615~0.856)，Hosmer-Lemeshow检验结果示χ²＝5.166和10.378，P均>0.05；两组受者临床决策曲线均表现出较好的临床效益。

本研究建立的风险预测模型预测效果良好，可为临床医护人员评估肝移植受者术后LEDVT发生风险提供参考。

To investigate the risk factors of lower extremity deep vein thrombosis(LEDVT) in liver transplant recipients, and to establish and validate a risk prediction model for LEDVT in recipients after liver transplantation.

A total of 336 recipients who underwent allogeneic orthotopic liver transplantation at the First Hospital of Jilin University from January 2020 to October 2023 were selected as the research subjects. The sample.split function was used to randomly divide the recipients into the modeling group (n=235) and the validation group (n=101) at a ratio of 7∶3. Through literature review, group discussion and clinical knowledge, the predictors of LEDVT in recipients after liver transplantation were determined. The group t test or Mann-Whitney U test was used for comparison of measurement data between groups. Comparison of counting data between groups was performed using the chi-square test or the Fisher exact probability method. Those predictors with P<0.05 in the univariate analysis were included in multivariate Logistic regression analysis to clarify the independent risk factors of LEDVT in recipients after liver transplantation.The nomogram was drawn using R (version 4.3.2) software, and a web-based calculator of the postoperative LEDVT risk prediction model for liver transplant recipients was developed on the shinyapps.io, and the area under the receiver operating characteristic (ROC) curve, Hosmer-Lemeshow goodness-of-fit test and clinical decision curve were used to evaluate the discrimination, accuracy and clinical benefit of the LEDVT risk prediction model in liver transplant recipients. A P<0.05 was considered statistically significant.

Among the 235 recipients in the modeling group, there were 49 cases in the LEDVT group and 186 cases in the non LEDVT group, with a LEDVT incidence rate of 20.8%(49/235). There were statistically significant differences in age, preoperative hepatic encephalopathy, daily living ability level, coagulation factor response time, postoperative AST, ALT, Na⁺ level, Ca²⁺ level, prothrombin time (PT), and international standardized ratio between the LEDVT group and non LEDVT group (Z=-3.552, -2.808, -2.567, -2.161, -2.297, -1.986, -3.815 and -2.395, χ²=13.822 and 36.213, all P<0.05).The results of multivariate Logistic regression analysis showed that the age (OR=1.048, 95%CI: 1.002-1.096), preoperative presence of hepatic encephalopathy (OR=2.484, 95%CI: 1.041-5.930), preoperative daily living ability (moderate dependence) (OR=5.266, 95%CI: 1.685-16.458), preoperative daily living ability (severe dependence) (OR=8.342, 95%CI: 1.748-39.802), postoperative Na⁺ level (OR=1.105, 95%CI: 1.001-1.220), and postoperative PT (OR=0.827, 95%CI: 0.737-0.928) were independent risk factors for postoperative LEDVT in liver transplant recipients (all P<0.05).The area under the ROC curve of the LEDVT risk prediction model of the modeling group and validation group were 0.811 (95%CI: 0.745-0.876) and 0.736 (95%CI: 0.615-0.856), respectively, the Hosmer Lemeshow test result showed χ²=5.166 and 10.378, all P<0.05. Good clinical benefits were shown both in the modeling group and validation group.

The risk prediction model established in this study has a good prediction effect and can provide a reference basis for clinical medical staff to evaluate the risk of postoperative LEDVT in liver transplant recipients.