个体化预测脓毒血症患儿并发急性肾损伤列线图模型的建立及验证
Establishment and validation of a personalized line graph model for predicting acute kidney injury in children with sepsis
王敏 1梁珍花 1刘桂良 1黄翰武 1张健1
作者信息
- 1. 广西壮族自治区人民医院儿科,广西南宁 530021
- 折叠
摘要
目的 构建及验证脓毒血症患儿并发急性肾损伤(AKI)的列线图预测模型,并进行验证.方法 本研究设计为回顾性研究,收集2019 年8 月至2021 年8 月在我院治疗的108 例脓毒症患儿的临床资料.通过单因素和多因素Logistic回归分析脓毒血症患儿并发AKI的独立危险因素,利用R软件(R 4.0.3)构建列线图模型,采用Bootstrap法进行内部验证,采用校正曲线、Hosmer-Lemeshow检验和ROC曲线评价模型的预测能力.结果 108 例脓毒症患儿中,30 例(27.78%)并发AKI(AKI组),78 例(72.22%)无发生AKI(无AKI组).AKI组与无AKI组的机械通气时间、急性生理与慢性健康评分(APACHEⅡ评分)、降钙素原(PCT)、C-反应蛋白(CRP)、白蛋白(ALB)、肌酸激酶(CK)等比较,差异均有统计学意义(均P<0.05).多因素Logistic回归分析显示,机械通气时间(OR=1.245,95%CI:1.004～1.545)、APACHEⅡ评分(OR=1.529,95%CI:1.068～2.187)、PCT(OR=1.597,95%CI:1.237～2.061)、CRP(OR=1.046,95%CI:1.012～1.080)、ALB(OR=0.815,95%CI:0.721～0.979)、CK(OR=0.814,95%CI:0.730～0.942)是脓毒血症患儿并发AKI的独立影响因素(P<0.05).基于上述影响因素建立列线图预测模型,经内部验证,校准曲线的平均绝对误差(MAE)为 0.031,预测值与实际值基本一致;Hosmer-Lemeshow拟合优度检验显示模型拟合度较好(χ2=2.134,P=0.976).曲线下面积(AUC)为0.965(95%CI:0.932～0.996),灵敏度为0.867,特异度为0.949.结论 机械通气时间、APACHEⅡ评分、PCT、CRP、ALB、CK是脓毒血症患儿并发AKI的独立影响因素,基于上述因素建立的脓毒血症患儿并发AKI列线图预测模型具有良好的校准度和区分度,可为临床早期筛查脓毒血症并发AKI的高风险患儿及个体化干预方案的制定提供参考依据.
Abstract
Objective To construct and validate a column chart prediction model for acute kidney injury(AKI)in children with sepsis.Methods Clinical data from 108 children with sepsis admitted to our hospital from August 2019 to August 2021 were select-ed.Independent risk factors for AKI in children with sepsis were analyzed through univariate and multivariate Logistic regression.A col-umn chart model was constructed using R software(R 4.0.3),and internal validation was conducted using Bootstrap.The predictive a-bility of the model was evaluated using calibration curves,Hosmer Lemeshow tests,and ROC curves.Results Among the 108 chil-dren with sepsis,30 cases(27.78%)developed AKI(AKI group),78 cases(72.22%)had no AKI(non AKI group).There were significant differences in mechanical ventilation time,acute physiological and chronic health scores(APACHEⅡ score),procalcitonin(PCT),C reactive protein(CRP),albumin(ALB)and creatine kinase(CK)between AKI group and non AKI group(P<0.05).Multivariate Logistic regression analysis showed that mechanical ventilation time(OR=1.245,95%CI:1.004-1.545),APACHE II score(OR=1.529,95%CI:1.068-2.187),PCT(OR=1.597,95%CI:1.237-2.061),CRP(OR=1.046,95%CI:1.012-1.080),ALB(OR=0.815,95%CI:0.721-0.979)and CK(OR=0.814,95%CI:0.730-0.942)were independent influencing factors for AKI in children with sepsis(P<0.05).Based on the above influencing factors,a column chart prediction model was established.After in-ternal verification,the average absolute error(MAE)of the calibration curve was 0.031,and the predicted value was basically consist-ent with the actual value.Hosmer-Lemeshow goodness of fit test showed that the model had a good fit(χ2=2.134,P=0.976).The area under the curve(AUC)was0.965(95%CI:0.932-0.996),with a sensitivity of 0.867 and a specificity of 0.949.Conclusion Mechanical ventilation time,APACHE II score,PCT,CRP,ALB,and CK are independent influencing factors for sepsis complicated AKI in children.The column chart prediction model for sepsis complicated AKI established based on these factors has good calibration and discrimination,and can be used to screen high risk children with sepsis complicated AKI in the early clinical stage and develop person-alized intervention plans.
关键词
脓毒血症/急性肾损伤/危险因素/列线图模型/个体化干预Key words
Sepsis/Acute kidney injury/Risk factors/Column chart model/Individualization引用本文复制引用
基金项目
广西壮族自治区卫生健康委员会自筹经费科研课题(Z-A20220033)
出版年
2024
NETL
NSTL
万方数据