血液透析患者发生高磷血症风险预测模型的构建与评估

Construction and Evaluation of Risk Prediction Model for Hyperphosphatemia in Hemodialysis Patients

目的分析尿毒症患者血液透析期间发生高磷血症的危险因素,构建列线图模型,并验证模型的预测效果。方法选取2019年1月~2022年5月在三峡大学第一临床医学院(宜昌市中心人民医院)规律血液透析的患者为研究对象,收集其血液透析的临床资料,经最小绝对收缩和选择算子(Lasso)回归、十折交叉验证法获得高磷血症最佳危险预测因子子集,并采用多因素Logistic回归分析确定高磷血症的危险预测因子,建立预测模型。采用受试者工作特征(receiver operating characteristic,ROC)曲线、C指数、校准曲线图和决策曲线分析来评估预测模型的预测能力、区分度、校准和临床实用性。结果共纳入200例血液透析患者,发生磷高磷血症166例,发生率为83%。多因素Logistic回归分析结果显示,甲状旁腺素、血肌酐、转铁蛋白饱和度为血液透析患者发生高磷血症的独立危险因素。基于以上影响因素建立列线图模型,构建的列线图预测模型预测尿毒症患者血液透析期间发生高磷血症的曲线下面积为0.824(95%CI:0.750~0.897),经内部验证C指数可达到0.784,具有良好的区分度与一致性。结论基于尿毒症患者血液透析期间发生高磷血症的危险因素建立列线图预测模型,可为临床医生评估血液透析患者高磷血症发生率提供理论依据,具有临床指导价值。

Objective To analyze the risk factors of hyperphosphatemia in patients with uremia during hemodialysis,construct a nomogram model,and verify the prediction effect of the model.Methods Patients receiving maintenance hemodialysis in the First College of Clinical Medical Science,China Three Gorges University(Yichang Central People′s Hospital)from January 2019 to May 2022 were enrolled,and their clinical data of hemodialysis were collected.The optimal risk predictor subset of hyperphosphatemia were obtained by minimum absolute contraction and selection operator(Lasso)regression,and 10-fold cross validation method.Multivariate Logistic regression analysis was used to determine the risk predictors of hyperphosphatemia,and the prediction model was established.Receiver operating characteristic(ROC)curves,consistency index(C-index),calibration curve,and decision curve analysis were used to evaluate the predictive power,differentiation,calibration,and clinical utility of the prediction model.Results Among 200 hemodialysis patients,166 cases with hyperphosphatemia occurred,with an incidence of 83%.The results of multivariate Logistic regression analysis showed that parathyroid hormone,serum creatinine and transferrin saturation were independent risk factors for hyperphosphatemia in hemodialysis patients.A nomogram model was established based on the above influencing factors.and it demonstrated that the area under the curve of hyperphosphatemia in patients with uremia during hemodialysis was 0.824(95%CI:0.750-0.897),and the C-index was up to 0.784 after internal verification,with good differentiation and consistency.Conclusion Based on the risk factors of hyperphosphatemia in patients with uremia during hemodialysis,the establishment of a nomogram prediction model can provide a theoretical basis for clinicians to evaluate the incidence of hyperphosphatemia in hemodialysis patients,which has clinical guiding value.