Podium Presentation Osteoarthritis

10.3.10 - Metabolomic profile and metabolic syndrome in patients with knee osteoarthritis: a sub-group analysis.

Presentation Number
10.3.10
Presentation Topic
Osteoarthritis
Lecture Time
14:51 - 15:00
Session Type
Free Papers
Corresponding Author
  • E. Sousa (Rio de Janeiro, BR)
Authors
  • E. Sousa (Rio de Janeiro, BR)
  • M. Pacheco Junior (Rio de Janeiro, BR)
  • D. Aguiar (Rio de Janeiro, BR)
  • R. Aguiar (Rio de Janeiro, BR)
  • G. Dos Santos Junior (Rio de Janeiro, BR)

Abstract

Purpose

To correlate the metabolomic profile of synovial fluid from patients with and without knee OA and clinical factors predisposing to OA development.

Methods and Materials

Synovial fluid (SF) had been collected from patients submitted to knee arthroscopy (n=11; K-L 0/1) or knee replacement (n=37, 9 K-L 3 and 26 K-L 3/4), centrifuged to isolate cells from the fluid, and then submitted to metabolomics analysis through NMR spectra acquired on Bruker 400MHZ spectrometer. 1H one-dimensional spectra were acquired with ZGESGP, 1K scans, TD of 64K, and SW of 20ppm. Two-dimensional 1H-1HTOCSY was acquired for assignments. Inversion-recovery-T1 was acquired to each sample to observe viscosity. All spectra were processed, phase and baseline corrected, on TopSpin 3.2 software (Bruker). CCPNMR V2 software, with metabolomics package installed1, HMDB 3.02, and BMRB3, were used to assignments. For multivariate analysis, 1D spectra were aligned, normalized by sum of intensities, 0.02ppm binned and pareto scaled, on AMIX software (Bruker). PLS-DA, VIP-score, and cross-validation, were done on MetaboAnalyst 3.04. For univariate analysis, we did multiple t-test, with 95% CI, and same SD between control/disease samples, on GraphPad Prism 6 software. Using records from a database of a previous study and medical records form the patients included, we performed a multivariate analysis of the data obtained. Clinical data as glucose levels, body mass index and OA severity were obtained from medical and radiological records.

Results

We found that among the variables studied, only OA severity levels were able to distinguish between different classes of patients according to the metabolomic profile. We found no differences regarding glucose levels and body mass index.

Conclusion

Our preliminary data suggests that although metabolic syndrome and OA are close related, glucose levels and BMI were not able to distinguish metabolomics profiles of OA patients.

Collapse