Background Anti-carbamylated protein antibodies (anti-CarP) are a new type of autoantibodies specific of patients with RA. Reactivity has generally been tested with in vitro carbamylated proteins from fetal calf serum (FCS) as anti-CarPF. It is likely that other sources of antigens could be more suitable than FCS unless the reactivity of the autoantibodies is directed primarily against the carbamylated residues independently of the protein. Objectives To evaluate the anti-CarP antibodies against in vitro carbamylated synovial tissue proteins (anti-CarPS) and to compare them with anti-CarPF antibodies. Methods A pool of sinovial tissue proteins from 3 osteoarthritis patients was in vitro carbamylated and used as antigen for ELISA. Anti-CarPS antibodies were determined in 520 sera from patients with RA and 278 healthy controls, which have been previously tested for anti-CarPF antibodies (1). Logistic regression analysis was used to evaluate the association of the antibodies with SE alelles, PTPN22, smoking and erosions. Relationship between antibodies titers was analyzed with Spearman rank correlation (rs) and concordance between positivity of the different antibodies was analyzed with the Goodman and Kruskal γ. Results Similar percentages of patients were positive for anti-CarPS (31.7%) and for anti-CarPF antibodies (29.4%). However, many patients were discordant (28.9%). Notably, the concordance between anti-CarPS and anti-CCP antibodies was higher than the concordance between anti-CarPS and anti-CarPF antibodies (γ =0.73, 95% CI: 0.60–0.86 vs. γ =0.60, 95% CI: 0.47–0.73, respectively). Moreover, titers of anti-CarPS antibodies were more correlated with anti-CCP than with anti-CarPF antibodies (rs =0.41, 95% CI: 0.33–0.48 vs. rs=0.29, 95% CI: 0.21–0.37, respectively). Accordingly, presence of anti-CarPS antibodies showed a clearer trend to be associated with the genetic and environmental RA risk factors than anti-CarPF antibodies, although only the association with smoking was significant (smoking OR=2.0, p=0.02 vs. OR=1.1, p=0.8, respectively; OR=1.3 vs. OR=1.1 for HLA-SE and OR=1.3 vs. OR=0.9 for PTPN22). There were no differences in the association with erosions, which was significant for both antibodies analyzed separately (OR=2.3, p=6.8 x 10-4 for anti-CarPS and OR=2.4 p=2.0 x 10-4 for anti-CarPF) and together (OR=1.8, p=0.02 and OR=2.0, p=0.005, respectively). Conclusions Our results indicate that the proteins used as antigens in the determination of anti-CarP antibodies are relevant for the obtained positivity and titer. The differences could affect the pattern of characteristics associated with these antibodies, making it necessary to identify the specific carbamylated autoantigens targeted by them in vivo. References Montes A, et al. PLoS One 2016;11:e0161141. Acknowledgements Supported by PI14/01651 and RD12/0009/0008 of the ISCIII (Spain) with participation of FEDER. Disclosure of Interest None declared