RESUMEN
BACKGROUND: Peritoneal dialysis (PD) helps prevent lethal complications of end-stage renal disease (ESRD). However, the clinical outcomes are affected by PD-related complications. We investigated metabolic biomarkers to estimate the clinical outcomes of PD and identify patients at high risk of downstream complications and recurrent/relapsing infections. METHODS: Metabolites of normal control and ESRD patient were compared via an untargeted metabolomic analysis. Potential metabolic biomarkers were selected and quantified using a multiple reaction monitoring-based target metabolite detection method. A nomogram was built to predict the clinical outcomes of PD patients using clinical features and potential metabolic biomarkers with the least absolute shrinkage and selection operator Cox regression model. RESULTS: Twenty-five endogenous metabolites were identified and analyzed. ESRD-poor clinical outcome-related metabolic modules were constructed. Adenine, isoleucine, tyramine, xanthosine, phenylacetyl-L-glutamine, and cholic acid were investigated using the weighted gene correlation network analysis blue module. Potential metabolic biomarkers were differentially expressed between the NC and ESRD groups and the poor and good clinical outcomes of PD groups. A 3-metabolite fingerprint classifier of isoleucine, cholic acid, and adenine was included in a nomogram predicting the clinical outcomes of PD. CONCLUSION: Metabolic variations can predict the clinical outcomes of PD in ESRD patients.
