#2794 Specific urinary peptidomic profiling in patients with post-COVID-19: a case-control study (UriCoV)

Abstract Background and Aims Post-COVID-19 (PC) is challenging to diagnose and treat, and its molecular pathophysiology is unclear. Urinary peptidomics is an unexplored area in PC research and could address this knowledge gap by providing information on urine peptides that may enable diagnosis at a molecular level. Method In this study with a 1:1 ratio matched case-control design, the urinary peptidome of 50 patients with PC using a standardized capillary-electrophoresis-mass-spectrometry (CE-MS) approach was investigated. The cohort was randomized into training (30 PC, 30 controls) and test sets (20 PC, 20 controls). Based on differences in peptide abundances between cases and controls, we aimed to develop a diagnostic model using a support vector machine. Results The median symptom duration for cases was 293 days [(IQR: 165–396), 19.2% missing]. In comparison with healthy controls and ME/CFS of non-COVID-19 origin, the abundance of 195 urine peptides was significantly different, with a predominant abundance in collagen alpha chains. This molecular signature (PC195), effectively distinguished PC from non-PC in the derivation [AUC of 0.949 (95% CI 0.900–0.998, P < .0001)] and in the validation cohort with an AUC of 0.962 [95% CI (0.897–1.00), P < .0001]. Sub-analyses showed associations between classification scores and word retrieval (P = 0.028), concentration impairment and short-term memory consolidation (P = 0.048), as well as exertional dyspnea (P = 0.047). Conclusion Here, we provided a novel and non-invasive diagnostic model for PC. Collagen dysregulation, ongoing inflammatory responses, hemostatic imbalances, and endothelial damage were identified through alterations in the excretion of endogenous peptides, which are partly generated along the nephron. Reflecting PC pathophysiology and renal processes, PC195 has the potential to advance diagnostics and inform therapeutic interventions.