Metabolomics biomarkers and the risk of overall mortality and ESRD in CKD: Results from the Progredir Cohort

Introduction Studies on metabolomics and CKD have primarily addressed CKD incidence defined as a decline on eGFR or appearance of albuminuria in the general population, with very few evaluating hard outcomes. In the present study, we investigated the association between metabolites and mortality and ESRD in a CKD cohort. Setting and methods Data on 454 participants of the Progredir Cohort Study, Sao Paulo, Brazil were used. Metabolomics was performed by GC-MS (Agilent MassHunter) and metabolites were identified using Agilent Fiehn GC/MS and NIST libraries. After excluding metabolites present in <50% of participants, 293 metabolites were analyzed. An FDR q value <0.05 criteria was applied in Cox models on the composite outcome (mortality or incident renal replacement therapy) adjusted for batch effect, resulting in 34 metabolites associated with the outcome. Multivariable- adjusted Cox models were then built for the composite outcome, death, and ESRD incident events. Competing risk analysis was also performed for ESRD. Results Mean age was 68 +/- 12y, mean eGFR-CKDEPI was 38.4 +/- 14.6 ml/min/1.73m(2) and 57% were diabetic. After adjustments (GC-MS batch, sex, age, DM and eGFR), 18 metabolites remained significantly associated with the composite outcome. Nine metabolites were independently associated with death: D-malic acid (HR 1.84, 95% CI 1.32-2.56, p = 0.0003), acetohydroxamic acid (HR 1.90, 95% CI 1.30-2.78, p = 0.0008), butanoic acid (HR 1.59, 95% CI 1.17-2.15, p = 0.003), and docosahexaenoic acid (HR 0.58, 95% CI 0.39-0.88, p = 0.009), among the top associations. Lactose (SHR 1.49, 95% CI 1.04-2.12, p = 0.03), 2-O-glycerol-alpha-D-galactopyranoside (SHR 1.76, 95% CI 1.06-2.92, p = 0.03), and tyrosine (SHR 0.52, 95% CI 0.31-0.88, p = 0.02) were associated to ESRD risk, while D-threitol, mannitol and myo-inositol presented strong borderline associations. Conclusion Our results identify specific metabolites related to hard outcomes in a CKD population. These findings point to the need of further exploration of these metabolites as biomarkers in CKD and the understanding of the underlying biological mechanisms related to the observed associations.