Background Previous studies indicate that anti-citrullinated peptide antibodies (ACPAs) may be detected in individuals who later develop rheumatoid arthritis (RA) years before onset, and that smoking is a predictor of RA.

Objectives To investigate circulating autoantibodies and their relation to smoking before the onset of RA.

Methods Incident cases of RA were identified among participants (n=30447) in a community based health survey, which was linked to local and national registers, followed by a structured review of the medical records. One control, matched for age, sex and year of inclusion in the health survey, was selected for each validated case. Using a multiplex assay system, developed in collaboration with Phadia AB (Uppsala, Sweden), sera were investigated for 10 different ACPA fine specificities and antibodies to 3 other autoantigens. For all investigated antibodies, previous studies had shown good correlations between results obtained using this system and standard ELISAs. Corrected net values for ACPA specificity were calculated by subtracting values for arginine control peptides from the corresponding citrulline peptides. Cut offs were set at the 98th percentile, based on investigation of sera from the matched controls. Analyses were stratified by time from inclusion in the health survey to RA diagnosis (1-3, 4-5, 6-8 and 9-13 years, respectively), and by history of current or ever smoking.

Results Serum was available from 167 cases (79% women), who were diagnosed with RA a median of 5 years (IQR 3-8; range 1-13) after inclusion in the health survey. Eighty-four (50%) of the pre-RA cases were positive for ≥1 ACPA. The most frequently detected ACPAs were antibodies to citrullinated vimentin 60-75 (cit Vim 60-75) (27%) and antibodies to citrullinated alpha-enolase peptide 1 (CEP-1) (22%). Significantly greater numbers of positive ACPA specificities were detected among pre-RA cases sampled closer to RA diagnosis (p for trend 0.003). Most individual ACPA specificities, including those directed against cit Vim 60-75 and CEP-1, were more frequent among those with a shorter time to diagnosis. Current (p=0.005) and ever (p=0.02) smoking were associated with RA. There were no major differences in detected ACPAs before RA diagnosis between current smokers and current non-smokers, whereas ever smokers were more likely to be positive for most ACPAs and had higher numbers of positive ACPA specificities compared to never smokers (p=0.03).

Conclusions The increase in ACPA specificities with a shorter time to diagnosis indicates that epitope spreading may precede disease onset. A history of smoking may contribute to the development of ACPAs before the clinical onset of RA.

Disclosure of Interest None Declared