ASN's Mission

ASN leads the fight to prevent, treat, and cure kidney diseases throughout the world by educating health professionals and scientists, advancing research and innovation, communicating new knowledge, and advocating for the highest quality care for patients.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on Twitter

Kidney Week

Abstract: PO2305

Novel Approach to the Relation of Environmental Exposure and Kidney Dysfunction: Data Analysis from Korean National Environmental Health Survey (KoNEHS), 2015-2017

Session Information

Category: CKD (Non-Dialysis)

  • 2101 CKD (Non-Dialysis): Epidemiology, Risk Factors, and Prevention

Authors

  • Yoo, Kyung Don, Ulsan University Hospital, Ulsan, Korea (the Republic of)
  • Noh, Junhyug, Lawrence Livermore National Laboratory, Livermore, California, United States
  • Lee, Inae, Seoul National University, Gwanak-gu, Seoul, Korea (the Republic of)
  • An, Jung Nam, Hallym University Sacred Heart Hospital, Anyang, Gyeonggi-do, Korea (the Republic of)
  • Park, Jae Yoon, Dongguk University Ilsan Hospital, Goyang, Gyeonggi-do, Korea (the Republic of)
  • Lee, Jeonghwan, Seoul National University Seoul Metropolitan Government Boramae Medical Center, Dongjak-gu, Seoul, Korea (the Republic of)
  • Choi, Kyungho, Seoul National University, Gwanak-gu, Seoul, Korea (the Republic of)
  • Lee, Jung Pyo, Seoul National University Seoul Metropolitan Government Boramae Medical Center, Dongjak-gu, Korea (the Republic of)
Background

Forecasting patient outcomes with kidney disease using standard statistical techniques is complex to estimate effects of the environmental chemicals. Herein, we aim to assess risk prediction for kidney disease in the general population using novel methods.

Methods

Serum POPs, serum creatinine or urinary albumin were measured in subpopulation (n=1,266) among the general adult participants from the 3rd Korean National Environmental Health Survey (KoNEHS) (n=3,787). Classification algorithms were used for the prediction of chronic kidney disease (CKD), defined by estimated glomerular filtration rate (eGFR) and urinary albumin-to-creatinine ratio (UACR). In addition, weighted quantile sum (WQS), which provides weights to the components of the mixture, was used to assess multi-pollutant effects.

Results

Of 1,266 adult subjects and 44 variables, including baseline characteristics and laboratory findings, were analyzed for the modeling process. The risk prediction of CKD was presented by machine-learning algorithms using bagging, ridge, lasso, random forest and was compared to conventional algorithms using logistic regression. A decision-tree algorithm was presented that outperformed a conventional method such as logistic regression (AUC 0.653 vs. 0.621). Among various decision-tree models, the lipid-corrected polychlorinated biphenyl congener 153 (PCB 153) was selected as the best predictor of CKD. Because persistent organic pollutants (POPs) accumulate with age, stratification analysis was conducted based on age. In the WQS model, PCB 153 showed the highest weight in its contribution to lower eGFR after adjusting covariates in the middle-aged group (under 50 yrs) (p=0.0135). If subjects with young age (under 50 yrs) were hemoglobin level > 13.25 g/dl, the CKD was predicted as 71.4% in the high serum PCB153 group.

Conclusion

We propose a machine learning-based prediction model. POPs and age were interrelated as notable risk factors for CKD in healthy Korean volunteers.