Abstract: SA-PO578
Metabolites Associated with Mortality in Hemodialysis Patients
Session Information
- Hemodialysis: Biomarkers, Translational Research
November 04, 2023 | Location: Exhibit Hall, Pennsylvania Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 801 Dialysis: Hemodialysis and Frequent Dialysis
Authors
- Al Awadhi, Solaf, Houston Methodist, Houston, Texas, United States
- Rhee, Eugene P., Massachusetts General Hospital, Boston, Massachusetts, United States
- Wulczyn, Kendra E., Massachusetts General Hospital, Boston, Massachusetts, United States
- Kalim, Sahir, Massachusetts General Hospital, Boston, Massachusetts, United States
- Segev, Dorry L., NYU Langone Health, New York, New York, United States
- McAdams-DeMarco, Mara, NYU Langone Health, New York, New York, United States
- Moe, Sharon M., Indiana University School of Medicine, Indianapolis, Indiana, United States
- Moorthi, Ranjani N., Indiana University Bloomington, Bloomington, Indiana, United States
- Hostetter, Thomas H., University of North Carolina Wilmington, Wilmington, North Carolina, United States
- Himmelfarb, Jonathan, University of Washington, Seattle, Washington, United States
- Meyer, Timothy W., Stanford Medicine, Stanford, California, United States
- Powe, Neil R., University of San Francisco, San Francisco, California, United States
- Tonelli, Marcello, University of Calgary, Calgary, Alberta, Canada
- Shafi, Tariq, Houston Methodist, Houston, Texas, United States
Background
Uremic toxins contributing to the increased risk of death in hemodialysis patients remain largely unknown. We used untargeted metabolomics profiling to identify plasma metabolite associated with mortality in hemodialysis patients.
Methods
We created a cohort of 465 participants from the Longitudinal US/Canada Incident Dialysis (LUCID) study in which we profiled 498 known plasma metabolites measured on an untargeted platform. We assessed the association between metabolites and 1-year mortality adjusting for age, sex, race, cardiovascular disease, diabetes, BMI, albumin, KT/V, dialysis duration, and country. We used limma, a metabolite-wise linear model with empirical Bayesian inference, and two machine learning models, LASSO and random forest (RF), for analysis. We corrected for batch effects in metabolite abundances using the removal of unwanted variation (RUV) method, and we accouted for multiple testing by false discovery rate (q) below 10%. We defined mortality-metabolite associations as robust if significant in the limma model (q<0.1) and at least of medium importance in both LASSO and RF models (metabolite is above the 70th percentile for the variable importance measure).
Results
The mean age was 61 years, 88% were male, 54% had diabetes and 48% had cardiovascular disease. There were 44 deaths (9.5%). The mean duration from dialysis initiation to metabolomic profiling was 62 days. We identified two metabolites significantly associated with 1-year mortality; mesaconate (HMDB0000749) and quinolate (HMDB0000232) (q<0.1 and high importance by both LASSO and RF). We identified 29 additional metabolites associated with 1-year mortality (q≥0.1) with high and/or medium importance by both LASSO and RF.
Conclusion
We identified two metabolites significantly associated with an increased 1-year mortality risk in incident hemodialysis patients. Mesaconate is an intermediate in the glutamate degradation pathway and has not been previously designated as a uremic toxin. Quinolate is a product in the kynurenine pathway and has been previously considered as a uremic toxin. Our study identifies additional metabolites that could be further investigated for mechanisms of uremic toxicity and potetial targeted interventions to prevent poor outcomes.