基于主成分分析-Logistic回归方法的人体农兽药化学污染物残留与高胆固醇血症关系的研究
作者:
作者单位:

1.复旦大学公共卫生学院,上海 200032;2.中国疾病预防控制中心营养与健康所/国家卫生健康委 微量元素与营养重点实验室,北京 100050

作者简介:

李梦梦 女 硕士研究生 研究方向为卫生统计学 E-mail:20211020117@fudan.edu.cn

通讯作者:

张兵 男 教授 研究方向为营养与食品卫生 E-mail:zhangbing@chinacdc.cn

中图分类号:

R155

基金项目:

国家重点研发计划(2019YFC1605100);国家自然科学基金(81573155,82173613);上海市卫生健康委科研项目(202140018)


Principal component analysis based Logistic regression analysis of the association between pesticide and veterinary chemical pollutant residues and hypercholesterolemia in humans
Author:
Affiliation:

1.School of Public Health, Fudan University, Shanghai 200032, China;2.National Institute for Nutrition and Health, Chinese Center for Disease Control and Prevention/Key Laboratory of Trace Element Nutrition of National Health Commission, Beijing 100050, China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    目的 基于主成分分析-Logistic(PCA-Logistic)回归组合分析策略探索人体农兽药化学污染物残留与高胆固醇血症之间的关系。方法 研究数据来源于“降低成年超重者营养相关慢性病风险的适宜身体活动量研究”2018—2019年调查数据。选择有人口学信息、体格测量和血清农兽药化学污染物及血脂检测等数据的496名成年居民作为研究对象。采用常规Logistic与PCA-Logistic分别建立回归模型分析高胆固醇血症的危险因素,利用受试者工作特征曲线下面积(AUC)评价模型判别效果。结果 PCA-Logistic回归分析结果显示对高胆固醇血症具有显著影响的农兽药化学污染物包括呋喃它酮、2-乙基己基二苯基磷酸酯(EHDPP)、全氟己烷磺酸(PFHxS)和4-氯苯氧乙酸(4-CPA)。PCA-Logistic回归模型对高胆固醇血症预测性能(AUC=0.736)优于常规Logistic回归模型的预测性能(AUC=0.695)。结论 呋喃它酮、EHDPP、PFHxS、4-CPA等农兽药化学污染物残留是影响高胆固醇血症的重要危险因素,PCA-Logistic回归分析方法可有效应用于探索农兽药化学污染物残留与高胆固醇血症的关联性分析。

    Abstract:

    Objective To explore the association of human pesticide and veterinary chemical pollutant residues on hypercholesterolemia based on the principal component analysis (PCA)-Logistic regression model.Methods The data were collected from a survey done in a 2018 to 2019 study on the appropriate physical activity to reduce the risk of nutrition-related chronic disease in overweight adults. A total of 496 adult residents with demographic information, physical measurements, serum pesticides and veterinary chemical pollutants, and blood lipid testing data were selected as the research subjects. Logistic and PCA-Logistic regression analyses were used to establish the models separately to explore the risk factors of hypercholesterolemia. The discrimination ability of the models was evaluated by the area of under curve (AUC).Results PCA-Logistic regression analysis revealed that pesticide and veterinary chemical pollutant residues, including furaltadone, 2-ethylhexyl diphenyl phosphate(EHDPP), perflurohexane sulfonate(PFHxS), and 4-chlorophenoxyacetate(4-CPA), had significant impacts on hypercholesterolemia. The PCA-Logistic regression model showed better distinguishing performance than the ordinary Logistic model (AUC 0.736 vs 0.695).Conclusion Pesticide and veterinary chemical pollutant residues that include furaltadone, EHDPP, PFHxS, 4-CPA are important risk factors for hypercholesterolemia. PCA-Logistic regression analysis can be effective to explore the association between human pesticide and veterinary chemical pollutant residues and hypercholesterolemia.

    参考文献
    相似文献
    引证文献
引用本文

李梦梦,牛然,宋琪哲,苏畅,王惠君,张兵,武振宇.基于主成分分析-Logistic回归方法的人体农兽药化学污染物残留与高胆固醇血症关系的研究[J].中国食品卫生杂志,2023,35(5):639-644.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2022-05-13
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2023-08-14
  • 出版日期:
《中国食品卫生杂志》邮寄地址与联系方式变更通知
关闭