- (1) University of Southern Denmark, grid.10825.3e, SDU
- (2) Herlev Hospital, grid.411900.d, Capital Region
- (3) Rigshospitalet, grid.475435.4, Capital Region
- (4) University of Copenhagen, grid.5254.6, KU
- (5) Kong Christian X's Gigthospital, grid.490243.8, Southern Denmark Region
- (6) ReumaKlinik Roskilde, Støden 3, Roskilde, Denmark
BACKGROUND: Undifferentiated arthritis (UA) is a label applied to patients with joint complaints which cannot be classified according to current criteria, which implies a need for precision diagnostic technologies. We studied serum galectin-3, a proinflammatory mediator, and seromarkers of structural joint elements in patients with early, UA and their associations with disease profile and biochemical and imaging findings. METHODS: One hundred and eleven UA patients were followed-up for at least 12 months and reclassified according to appropriate criteria (TUDAR). At baseline, demographics and laboratory and clinical disease measures, as well as wrist magnetic resonance imaging (MRI) synovitis, erosion, and bone marrow edema scorings, were recorded. Galectin-3, the type IIA collagen N-terminal propeptide (PIIANP), which is a marker of regenerative cartilage formation, and hyaluronan (HYA), which is prevalent in synovial tissue swellings, were measured by enzyme-linked immunosorbent assay (ELISA). Receiver operating characteristic (ROC) curve analysis was carried out to assess the discriminant capacity of galectin-3 against arthritis subsets. RESULTS: Galectin-3 was increased in pre-rheumatoid arthritis (RA) (4.6 μg/l, interquartile range (IQR) 3.8-5.5) versus non-RA (4.0 μg/l, IQR 3.1-4.9; p = 0.03) and controls (3.8 μg/l, IQR 3.0-4.8; p = 0.009). PIIANP was equally depressed in either subset (p < 0.01). Galectin-3 in non-RA and HYA in UA did not differ from healthy controls. In the entire UA cohort, galectin-3 correlated with the MRI bone marrow edema score, while PIIANP correlated with the MRI erosion score, and HYA with the synovitis and erosion scores. ROC curve analysis showed that baseline galectin-3 discriminated well between pre-RA and non-RA with univariate area under the curve (AUC) of 0.64 (95% confidence interval (CI) 0.53-0.76) while AUC for galectin-3 + anti-CCP increased to 0.71 (95% CI 0.59-0.83). CONCLUSIONS: Galectin-3 in serum was increased in patients with early UA of pre-RA origin. Cartilage remodeling assessed by PIIANP was diminished in UA irrespective of subsequent clinical differentiation, while HYA did not differ from controls. ROC analysis showed a potential for galectin-3 to discriminate between pre-RA and non-RA. TRIAL REGISTRATION: KF 11 315829. Registered 25 July 2006.