中国煤炭工业医学杂志2024,Vol.27Issue(1) :12-17.DOI:10.11723/mtgyyx 1007-9564 202401003

连续性血液净化患者并发血管通路不畅危险因素及列线图预测模型的构建

Analysis of risk factors of vascular obstruction in patients with continuous blood purification and construction of a column chart prediction model

林建美 张瑶琳 兰秀君 汪燕 马川入 孙晓敏 周罗平 卿永洪
中国煤炭工业医学杂志2024,Vol.27Issue(1) :12-17.DOI:10.11723/mtgyyx 1007-9564 202401003

连续性血液净化患者并发血管通路不畅危险因素及列线图预测模型的构建

Analysis of risk factors of vascular obstruction in patients with continuous blood purification and construction of a column chart prediction model

林建美 1张瑶琳 1兰秀君 1汪燕 1马川入 1孙晓敏 2周罗平 3卿永洪4
扫码查看

作者信息

  • 1. 641300 四川省资阳市第一人民医院
  • 2. 资中县人民医院
  • 3. 青神县人民医院
  • 4. 富顺县中医院
  • 折叠

摘要

目的 分析连续性血液净化(CBP)患者并发血管通路不畅危险因素,并构建列线图预测模型.方法 选择2022年1月-2023年6月该院收治的168例CBP患者,依据CBP患者是否并发血管通路不畅分为病例组(n=41)与对照组(n=127),采用Logistic回归分析筛选影响CBP患者并发血管通路不畅的危险因素;采用R软件构建预测CBP患者并发血管通路不畅的列线图模型,且使用ROC曲线、校准曲线、Hosmer-Lemeshow拟合优度检验评估列线图模型.结果 168例CBP患者41例并发血管通路不畅,占24.40%.病例组与对照组患者的性别、疾病类型、高血脂病史比较,差异无统计学意义(P>0.05),而二组间的年龄、CBP时间、封管液使用、置管位置、血栓形成、穿刺经验以及高血压、糖尿病病史比较,差异有统计学意义(P<0.05).Logistic回归分析结果显示,年龄≥60岁(95%CI:2.222~14.565)、封管液使用不适宜(95%CI:2.319~20.136)、置管位置不合适(95%CI:2.010~13.600)、有血栓形成(95%CI:1.800~11.263)是 CBP 患者并发血管通路不畅的独立危险因素.ROC曲线下面积为0.835(95%CI:0.774~0.896).校准曲线斜率接近1,且Hosmer-Lemeshow拟合优度检验x2=7.063,P=0.422.根据最大约登指数0.542筛选列线图模型预测并发血管通路不畅的临界值为P(预测概率)=13.94%[或分值(Total Points)=174分].以P(预测概率)>13.94%(临界值)为预测标准,样本的预测结果显示,列线图模型预测的灵敏度和特异度为95.12%和59.06%.结论 CBP患者并发血管通路不畅影响因素众多,本研究基于年龄≥60岁、封管液使用不适宜、置管位置不合适、有血栓形成构建的列线图对CBP患者并发血管通路不畅有较好的预测价值.

Abstract

Objective To analyze the risk factors of vascular obstruction in patients with continuous blood purification(CBP)and construct a column chart prediction model.Methods From January 2022 to June 2023,one hundred and sixty-eight patients with CBP admitted to Ziyang First People's Hospital were separa-ted into two groups based on whether they had concurrent vascular obstruction,case group(n=41)and con-trol group(n=127).Logistic regression analysis was applied to screen for risk factors that affected vascular access obstruction in CBP patients.R software was applied to construct a column chart model for predicting vascular obstruction in patients with CBP,and ROC curves,calibration curves,and Hosmer-Lemeshow good-ness of fit tests were applied to evaluate column chart models.Results Among 168 CBP patients,41 cases(24.40%)were complicated with vascular obstruction.There was no statistically significant difference between case group and control group in terms of gender,disease type,and hyperlipidemia history(P>0.05),but there was a statistically significant difference between the two groups in terms of age,CBP time,use of sealing fluid,placement position,thrombosis,puncture experience,and hypertension and diabetes history(P<0.05).The results of Logistic regression analysis showed that age≥60 years old(95%CI:2.222~14.565),inappro-priate use of sealing fluid(95%CI:2.319~20.136),inappropriate placement of catheter(95%CI:2.010~13.600),and thrombosis(95%CI:1.800~11.263)were independent risk factors for vascular obstruction in CBP patients.The area under ROC curve was 0.835(95%CI:0.774~0.896).The slope of cali-bration curve was close to 1,and Hosmer-Lemeshow goodness of fit test x2=7.063,P=0.422.The critical value for predicting concurrent vascular access obstruction using the column chart model based on the maxi-mum Jordan index of 0.542 is P(prediction probability)=13.94%(or total points=174 points).Using P(prediction probability)>13.94%(critical value)as the prediction criterion,the prediction results of the sam-ple show that sensitivity and specificity of the column chart model prediction are 95.12%and 59.06%,respec-tively.Conclusion There are many influencing factors for occurrence of vascular obstruction in CBP patients.This study constructs a column chart based on age≥60 years old,inappropriate use of sealing fluid,inappro-priate placement of catheter,and thrombosis,which has good predictive value for occurrence of vascular ob-struction in CBP patients.

关键词

连续性血液净化/血管通路/危险因素/列线图预测模型

Key words

Continuous blood purification/Vascular pathways/Risk factors/Column chart prediction model

引用本文复制引用

基金项目

四川省基层卫生事业发展研究中心科研项目(SWFZ22-C-100)

出版年

2024
中国煤炭工业医学杂志
河北联合大学

中国煤炭工业医学杂志

CSTPCD
影响因子:0.692
ISSN:1007-9564
参考文献量21
段落导航相关论文