非透析慢性肾脏病患者发生肾性贫血风险预测模型的开发与验证

Development and validation of a risk prediction model for renal anemia in non-dialysis chronic kidney disease patients

目的探讨非透析慢性肾脏病(CKD)患者发生肾性贫血的影响因素,建立非透析CKD患者发生肾性贫血的风险模型,并验证预测模型的有效性。方法回顾性分析2017年1月—2021年12月收治的非透析CKD患者的临床资料,按7∶3比例随机分为训练集(n=388)与验证集(n=165)。利用单因素和多因素Logistic回归分析筛选影响肾性贫血的影响因素,基于赤池信息准则(AIC)最小值选取最终预测因素构建列线图,并验证模型效能。结果单因素Logistic回归分析显示,17个变量[白细胞计数、中性粒细胞百分比、红细胞分布宽度(RDW)、血小板压积、碱性磷酸酶、血清白蛋白(ALB)、血胱抑素C、尿素、CKD分期、碳酸氢钠、血钾、血钙、血磷、C反应蛋白(CRP)、甘油三酯、尿白蛋白、尿隐血]与贫血的发生相关(P<0.05)。多因素Logistic回归分析结果提示,白细胞计数、CRP、RDW、ALB、血胱抑素C、尿素、CKD分期、碳酸氢盐、血钙是贫血独立影响因素(P<0.05)。利用危险因素构建列线图,经验证模型具有较好的区分度[训练集ROC曲线下面积(AUC)为0.915(95%CI:0.870~0.959),验证集为0.949(95%CI:0.927~0.971)]。校准曲线及H-L检验均提示模型预测值与实际值差异无统计学意义(P>0.05)。临床决策曲线分析(DCA)表明该模型具有良好的临床应用价值。结论本研究开发的预测模型能较好地预测非透析CKD患者肾性贫血的发生风险,为临床决策提供参考。

Objective To investigate the influencing factors of renal anemia in non-dialysis chronic kidney disease(CKD)patients,and to establish a risk model of renal anemia in non-dialysis CKD patients and verify its validity.Methods The clinical data of non-dialysis CKD patients admitted from January 2017 to December 2021 were retrospectively analyzed.The patients were randomly divided into training set(n=388)and validation set(n=165)at a ratio of 7∶3.Univariate and multivariate Logistic regression analysis was used to screen out the influencing factors for renal anemia.Based on the minimum Akaike Information Criterion(AIC)criterion,the final predictors were selected to construct the nomogram,and the efficiency of the model was verified.Results Multivariate Logistic regression analysis showed that 17 variables including white blood cell,percentage of neutrophil granulocyte,erythrocyte distribution width(RDW),thrombocytocrit,alkaline phosphatase,serum albumin(ALB),cystatin C,urea,CKD stage,bicarbonate,potassium,calcium,phosphorus,C-reactive protein(CRP),triglyceride,urinary albumin,occult urine were associated with anemia(P<0.05).Multiple Logistic regression analysis showed that white blood cell count,CRP,RDW,ALB,cystatin C,urea,CKD stage,bicarbonate and blood calcium were independent influencing factors for anemia(P<0.05).Using risk factors to construct a nomogram,the validation model had good discrimination[the area under ROC curve(AUC)of the training set was 0.915(95%CI,0.870 to 0.959),and the validation set was 0.949(95%CI,0.927 to 0.971)].The calibration curve and H-L test showed that there was no significant difference between the predicted value and the actual value of the model(P>0.05).The clinical decision curve analysis(DCA)showed that the model had good clinical application value.Conclusion The model developed in this study can better predict the risk of renal anemia in non-dialysis CKD patients,and provide reference for clinical decision-making.