Background Metabolomics is an emergent research field within the omics sciences aimed at characterizing the metabolome in complex biological samples. The advances in metabolomics are leading to a better knowledge of the regulatory processes on biological systems and, ultimately, of the metabolic disturbances related to pathological processes.

The present study represents the first metabolomic analysis on a large cohort including multiple immune-mediated inflammatory diseases (IMIDs).

Objectives This study is aimed at identifying disease diagnostic and disease activity biomarkers through the analysis of the whole urine metabolome on 1,200 IMID patients and 100 control individuals.

Methods Spectral proton nuclear magnetic resonance (¹H-NMR) data of urine samples was acquired among 6 disease cohorts (n=200 Spanish patients per cohort) including: rheumatoid arthritis (RA), psoriasis (PS), psoriatic arthritis (PA), Crohn's disease (CD), ulcerative colitis (UC) and systemic lupus erythematosus (SLE). Patients on each cohort were selected in order to be equally distributed among groups of low and high disease activity. Furthermore, ¹H-NMR data of 100 control individuals were also acquired in this study.

The quantification of the spectral peaks containing the metabolic information was performed by a powerful methodology¹ to process the spectral data. Mann-Whitney U test was used to evaluate the association of metabolic peaks with each IMID cohort. Subsequently, association with disease activity was also evaluated.

Results After processing the spectral data n=473 peaks were identified, from which n=145 passed the quality control filters.

The statistical analysis identified 45 metabolic peaks significantly associated (P-Value<1E-4) in at least one of the performed disease diagnostic or disease activity tests. The aggregated IMID analysis identified n=16 peaks associated with disease diagnostic and n=9 peaks associated with disease activity. When analyzing each IMID cohort separately, RA, SLE, CD and UC obtained the largest number of diagnostic biomarker candidates: n=27, n=14, n=18 and n=21 respectively. PS and PA showed lower differences against controls. On the other side, when comparing low to high disease activity patients a lower number of significant associations (n=5; P-Value<5E-3) were observed, mainly related to CD and RA.

Conclusions Significant urine metabolic differences have been identified on this first large cohort of jointly analyzed IMIDs. The results have shown large differences on some metabolic peaks when comparing the IMID cohorts against controls. These results evidence the high potential of the diagnostic properties of the urine metabolome. Although the differences are lower when evaluating disease activity, the identified candidates are of great interest.

References

1. Alonso, A., et al. Analytical Chemistry 2013, 86, 1160.

Disclosure of Interest None declared

DOI 10.1136/annrheumdis-2014-eular.4927