Background Anti-citrullinated protein antibodies (ACPA) and rheumatoid factor (RF) are important diagnostic markers in rheumatoid arthritis (RA). These antibodies are predominantly of the IgM (RF) or IgG (ACPA) isotype. The added diagnostic and prognostic value of IgA autoantibodies is being debated.

Objectives To determine the prevalence of IgA-RF and IgA-ACPA in patients with RA and to investigate their potential predictive value regarding response to treatment with methotrexate (MTX) and TNF inhibitors.

Methods A total of 255 patients were tested for the presence of IgA-RF, IgA-ACPA and IgG-ACPA by EliA® (Thermo Fisher Scientific); IgM-RF was measured by nephelometry. Therapeutic responses to MTX and TNF blocking biologicals were calculated in an inception cohort (n=104) who had started their DMARD therapy at our clinic. To define therapeutic responses simplified disease activity index (SDAI) 50 and American College of Rheumatology (ACR) 20 responses were calculated.

Results Among the 255 patients tested 125 (49%) had at least one type of IgA autoantibody: 114 (44.7%) were found to be IgA-RF positive and of these 10.5% were negative for IgM-RF and 5.2% were double negative for both IgM-RF and IgG-ACPA; thus, in these patients IgA-RF was the only detectable antibody. IgA-ACPA were detected in 79 (31%) patients and apart from one exception all of them had also IgG-ACPA. Remarkably, the percentage of patients showing a SDAI50 response to TNF inhibitors was significantly lower in patients positive for IgA-RF and/or IgA-ACPA (p<0.0001) compared to IgA negative patients. Thus, 58% of IgA negative (but IgM-RF and/or IgG ACPA positive) patients showed a SDAI50 response whereas only 25% of the IgA-RF and/or IgA-ACPA positive ones were responders. Interestingly, while the presence of both IgA specificities did not further change the percentage of responders, patients positive for IgA-ACPA but negative for IgA-RF showed the lowest response rate to anti-TNF treatment. Completely seronegative patients also showed a significantly lower SDAI50 response (p<0.0001) to TNF inhibitors compared with the IgA negative (but IgM-RF and/or IgG-ACPA positive) patients. Similar results were obtained when ACR20 was used as response criteria. No differences between the various serological groups were seen with respect to treatment with MTX.

Conclusions While the added diagnostic value of IgA antibody measurement was moderate, IgA-RF and particularly IgA-ACPA appear to be associated with poorer therapeutic responses to TNF inhibitory biological drugs and therefore may help in further stratification of RA patients and therapeutic decision making.

Disclosure of Interest None declared