维持性血液透析患者死亡危险因素分析及预测模型构建

Analysis of mortality risk factors and establishment of predictive model for patients with maintenance hemodialysis

目的探讨维持性血液透析(maintenance hemodialysis,MHD)患者死亡风险的影响因素,并构建预测模型。方法在安徽医科大学附属巢湖医院血液净化中心,选取了2019年1月至2022年1月接受MHD治疗的198例患者作为研究对象。按照是否存活,将其分为生存组(152例)和死亡组(46例)。收集患者的临床资料,并应用Logistic回归分析影响MHD患者死亡风险的独立危险因素。利用R软件构建列线图模型,并应用受试者工作特征曲线(receiver operating characteristic,ROC)曲线下面积(area under the curve,AUC)和校准曲线(calibration curve,CA)来验证模型的区分能力和校准程度,并绘制决策曲线分析(decision curve analysis,DCA)以评估其在实际临床中的应用价值。结果本研究纳入了198例MHD患者,发现死亡率为23.23%(46例)。通过Logistic回归分析,我们确定了血镁(OR=0.585,95%CI:0.375~0.912)、hs-CRP(OR=1.173,95%CI:1.063~1.294)、mCI(OR=0.691,95%CI:0.566~0.844)和SII(OR=1.030,95%CI:1.016~1.045)均为MHD患者死亡风险的独立危险因素。我们以这4个因素为基础,建立了预测MHD患者死亡风险的列线图模型,其AUC为0.884(95%CI:0.819~0.949),具有80.26%的特异度和89.13%的敏感度,表明该模型具有良好的区分度。通过1000次bootstrap方法进行内部验证后,AUC为0.876(95%CI:0.812~0.933),校准曲线与实际曲线吻合良好,提示该模型的校准度较高。绘制的DCA曲线显示,当阈值概率在0.03~0.99范围内,净收益均大于0,进一步证实了该模型的有效性。结论基于血镁、hs-CRP、mCI和SII这4项危险因素构建的列线图模型对MHD患者死亡风险的预测展现出了出色的区分度和校准度,具备一定的临床应用价值。

Objective To explore the factors affecting mortality risk among maintenance hemodialysis(MHD)patients and to develop a predictive model.Methods A total of 198 patients who underwent MHD treatment in the Hemopurification Center of Chaohu Hospital of Anhui Medical University from January 2019 to January 2022 were selected.Based on survival status,they were categorized into two groups:the survival group(152 cases)and the death group(46 cases).The clinical data were collected and logistic regression analysis was conducted to identify independent risk factors influencing mortality among MHD patients.A column diagram model was constructed using R software.Receiver operating characteristic(ROC)curve analysis,including area under the curve(AUC)and calibration curve were employed to assess the model’s discriminative ability and calibration accuracy.Additionally,decision curve analysis(DCA)was performed to evaluate the model’s practical clinical utility.Results In this study,we enrolled 198 MHD patients and observed a mortality rate of 23.23%(46 cases).Through logistic regression analysis,we identified blood magnesium(OR=0.585,95%CI:0.375~0.912),high-sensitivity C-reactive protein(hs-CRP,OR=1.173,95%CI:1.063~1.294),modified Creatinine Index(mCI,OR=0.691,95%CI:0.566~0.844)and systemic immune-inflammation index(SII,OR=1.030,95%CI:1.016~1.045)as independent risk factors for mortality in MHD patients.A column diagram model was developed using these four factors,which showed an AUC of 0.884(95%CI:0.819~0.949),specificity of 80.26%and sensitivity of 89.13%,indicating excellent discriminative ability.Following internal validation using 1000 bootstrap methods,the AUC remained high at 0.876(95%CI:0.812~0.933),with the calibration curve showing strong agreement with the actual curve,indicative of robust model calibration.The DCA curve demonstrated positive net benefit across threshold probabilities ranging from 0.03 to 0.99,further confirming the model’s effectiveness.Conclusion The column diagram model constructed using the four risk factors of blood magnesium,hs-CRP,mCI and SII demonstrates excellent discriminative ability and calibration in predicting mortality risk among MHD patients,indicating its potential clinical utility.