Objectives To evaluate the predictive value of autoantibodies (Ab) directed against citrullinated recombinant rat filaggrin (ACRFA) detected by a new ELISA for prognosis in a cohort of community cases of RA.

Methods 127 patients (mean age: 50.5 yr, F/M: 2.9) with early RA (median disease duration: 2 yr) were enrolled prospectively in 1996 (T1), primarily from a population based-recruitment (80%), and were followed until 1999 (T2). At T1 were tested rheumatoid factors determined by the latex fixation test (LFT) and anti-filaggrin Abs identified by different assays, i.e. anti-keratin Ab (AKA), and anti-perinuclear factor (APF) by indirect immunofluorescence, ACRFA by an ELISA test whose results are expressed as the difference of OD values given by sera tested against the citrullinated and non-citrullinated forms of filaggrin. At T1 and T2: radiographs of the hands and wrists. Prognosis was defined by (1) radiological damage (van der Heijde modified Sharp method) at T2; (2) progression of radiological damage between T1 and T2.

Results At T1: RA had mild activity (Ritchie articular index: 11/78, mean CRP: 15 mg/l), mild functional disability (HAQ: 0.8/3), mild X-ray destruction (Sharp: 9.2/280); 96% of patients were treated at T1 (DMARDs: 95%; prednisone: 72%); frequency of LFT, AKA, APF and ACRFA were respectively: 50%, 33%, 45% and 45%. At T2: total Sharp score: 22.8/280. The only parameter at T1 associated with a higher Sharp score at T2 (p = 0.03) and particularly with a more important radiological progression (p = 0.03) was the LFT.

Conclusion Whatever the method used, indirect immunofluorescence or ELISA, anti-filaggrin do not allow to predict radiological damage in community cases of RA of limited duration.