671 FO15 – Identification of biomarkers of illness incidence in national team cross-country skiers using metabolomics

Background Metabolomics has potential to highlight novel pathways and biomarkers of interest to provide early warning signs of health problems, such as opportunistic infection susceptibility among athletes. Objective To investigate differences in metabolites at rest and in response to exercise among athletes susceptible and not susceptible to opportunistic infections. Design Prospective cohort study. Setting Laboratory tests and illness diaries from national-team cross-country skiers. Participants Thirteen male and ten female, aged 21 ± 1 years. Interventions Capillary whole blood samples collected at rest and following a maximal incremental exercise test. Metabolite profiles at rest and post-exercise obtained using targeted and untargeted liquid and gas-chromatography analysis. Main Outcome Measurements Illness incidence, prospectively monitored for 15 weeks using the OSTRC-H questionnaire. Discriminating metabolites between the two groups at rest and in response to exercise, analysed using OPLS-DA followed by enrichment analysis of metabolic pathways in the KEGG database. Results Participants reporting ≥1 illnesses during the observation period were classified as illness susceptible (n = 14) and those that remained healthy as not susceptible (n = 8). The initial OPLS-DA model including all 346 identified metabolites (at rest) did not reach statistical significance but significance was achieved (R2Y = 0.916, Q2 = 0.617, p = 0.030) when refined to include only metabolites with a variable influence on projection of >1 (n = 129). Glucose-6-phosphate, squalene and deoxycholic acid contributed most to the separation. Enrichment analysis indicated no significant differences in metabolic pathway activity between groups. No significant OPLS-DA model was achieved for metabolite responses to exercise. Conclusions While our models achieved separation between susceptible and non-susceptible athletes using resting metabolic profiles, the primary identified metabolites do not overlap with existing findings in the literature. Further exploration of biomarkers of infection risk using omics approaches in larger and more heterogeneous cohorts may be considered.