IHD双段枸橼酸抗凝时非计划下机预测模型的构建

Construction of unplanned offline prediction model for IHDtwo-stage citrate anticoagulation

目的研究间歇性血液透析(IHD)双段枸橼酸抗凝的危险因素并建立非计划下机预测模型。方法采用回顾性分析,将2019年1月至2023年2月该院行IHD使用双段枸橼酸抗凝34例患者的118例次治疗纳入研究。根据该例次治疗是否由于体外循环管路凝血而未达到治疗时间,将118例次治疗分为计划下机组(n=111)和非计划下机组(n=7)。采用单因素及多因素logistic回归分析非计划下机的危险因素,并建立风险预测模型,采用受试者工作特征(ROC)曲线分析回归模型的预测价值。结果单因素分析结果显示,计划下机组与非计划下机组血细胞比容(HCT)、PLT、活化部分凝血活酶时间(APTT),以及治疗模式等方面比较,差异有统计学意义(P<0.05)。多因素logistic回归分析结果显示,HCT、APTT是非计划下机的独立影响因素(P<0.05)。以A表示HCT水平,B表示APTT水平,预测模型Logit(P)=1.304+0.206×A-0.378×B。预测模型ROC曲线下面积(AUC)为0.912(95%CI:0.825~0.995,P<0.001),约登指数最大值为0.782,cut off值为0.113,灵敏度为85.7%,特异度为92.5%。结论通过多因素logistic回归分析建立的预测模型可对双段法IHD治疗是否发生凝血作出初步判断。

Objective To study the risk factors of two-stage citrate anticoagulation in intermittent hemodialysis(IHD)and to establish an unplanned offline prediction model.Methods A retrospective analysis was conducted to include 34 patients and 118 times of treatment with two-stage citrate anticoagulation for IHD in the hospital from January 2019 to February 2023.According to whether the treatment did not reach the treatment time due to the coagulation of the extracorporeal circulation pipeline,118 treatments were divided into the planned units(n=111)and the unplanned units(n=7).Univariate and multivariate logistic regression analysis were used to analyze the risk factors of unplanned weaning,and a risk prediction model was established.The receiver operating characteristic(ROC)curve was used to analyze the predictive value of the regression model.Results Univariate analysis showed that there were statistically significant differences in hematocrit(HCT),platelet count(PLT),activated partial thromboplastin time(APTT),and treatment mode between the planned and unplanned units(P<0.05).Multivariate logistic regression analysis showed that HCT and APTT were independent influencing factors for unplanned weaning(P<0.05).The HCT level was represented by A,the APTT level was represented by B,and the prediction model was:Logit(P)=1.304+0.206×A-0.378×B.The area under the ROC curve(AUC)of the prediction model was 0.912(95%CI:0.825-0.995,P<0.001),the maximum Youden index was 0.782,the cut off value was 0.113,the sensitivity was 85.7%,and the specificity was 92.5%.Conclusion The prediction model established by multivariate logistic regression analysis can make a preliminary judgment on whether coagulation occurs in two-stage IHD treatment.