Background Autoantibodies (AAb) (RF/ACPA) are hallmarks of RA. ACPA have been regarded more specific than RF diagnostic and prognostic utility. While it is known for long that AAb+ patients (pts) have more active and severe disease than AAb-, it is not clear if this is due to RF, ACPA or both given the high (>80%) overlap of RF and ACPA.
Objectives To understand the impact of RF and/or ACPA on disease activity in RA.
Methods We kindly received two large datasets of pt data from multicenter RA clinical trials. In one of them which we used for initial analysis and hypothesis generation, MTX was compared with rituximab(RTX)+ MTX1. The other, used to validate the hypothesis generated, comprised patients from golimumab(GOL) trials2-4. Pts had high disease activity (DA) at inclusion into trials1;2-4. Clinical and serological core set variables as well as lRF and ACPA were comprised in the databases. For this study we focused only at baseline data and pooled the pts for all RTX and separately all GOL trial arms. The following groups were formed for each dataset: RF-/ACPA- (n=64 and 268, resp.); RF+/ACPA+ (n=611 and 938); RF+/ACPA- (n=40 and 92); RF-/ACPA+ (n=29 and 68). We compared individual disease activity variables and composite scores (disease activity score, DAS28; simplified and clinical disease activity index, SDAI and CDAI) across these groups by Kruskal Wallis test for overall assessment and Wilcoxon test for subsequent pairwise comparisons if appropriate. A main focus was the comparison of RF+/ACPA- and RF-/ACPA+ pts.
Results At baseline, DA in the RTX trial was 6.9, 7.1, 7.1 and 6.6, resp, by DAS28-ESR, and 47.8, 49.8, 53.1, 42.3 by SDAI, with significant differences among the four groups. By both scores, the lowest values of DA were seen for the RF-/ACPA+ population, and significantly lower than in the RF+/ACPA- group (p=0.014 for the DAS28, and p=0.004 for the SDAI). Similar findings were made for CDAI, swollen and tender joint counts, while for ESR and CRP there was a numerical trend in this direction. In the golimumab-dataset, these findings could be fully confirmed: DAS28-CRP was 5.6, 5.5, 5.6 and 5.1; and SDAI was 40.7, 39.4, 39.6 and 33.4 (p=0.0022), again with the lowest values seen in RF-/ACPA+ (mean ACPA levels: 112 U/ml) and the highest in RF+/ACPA- (mean RF levels: 90 IU/ml) pts; moreover, significant differences in the same direction were also seen for CRP in the GOL dataset, mean CRP: 1.9, 2.2; 2.4 and 1.8mg/dl, resp (p=0.0004. Interestingly, in both datasets RF+/ACPA+ pts tended to have slightly lower DA than RF+/ACPA-, and much higher than RF-/ACPA+ ones.
Conclusions The data presented suggest that RF contributes more strongly to DA than ACPA and thus may have higher prognostic utility. The presented data rest on two large, separate clinical datasets that independently revealed confirmatory results. These data further suggest that new light may have to be shed on the role and value of RF.
Tak PP et al, Ann Rheum Dis 2011; 70(1):39-46.
Emery P et al, Arthritis Rheum 2009; 60(8):2272-2283.
Keystone EC et al, Ann Rheum Dis 2009; 68(6):789-796.
Smolen JS et al, Lancet 2009; 374(9685):210-221.
Acknowledgements We thank Roche and Janssen for kindly providing the data of this study.
Disclosure of Interest None Declared