Background/Purpose: Recent studies have shown that  anti–citrullinated protein antibodies (ACPA) are detectable years before rheumatoid arthritis (RA) diagnosis, representing a potential early marker of RA pathogenesis. The goal of this study was to investigate the presence of several ACPAs targeted to epitopes found in the rheumatoid synovium years prior to RA diagnosis, within a nested case-control study in the Nurses’ Health Study (NHS) and Nurses’ Health Study II (NHSII) cohorts.

Methods: We confirmed 196 incident RA cases (NHS: 139; 1976-2008, NHSII: 57; 1989-2009) with blood collected prior to RA symptoms by medical record review. Three controls were matched to each case on year of birth, race, menopausal status and post-menopausal hormone use, time of day, fasting status at blood draw and timing in the menstrual cycle (pre-menopausal women in NHSII only). Sixteen ACPA antigens previously identified by protein mass spectrometry were coupled to spectrally distinct beads for analysis using the BioPlex platform using a Luminex 200 instrument. These peptides included epitopes derived from clusterin, enolase, fibrinogen, histone 2A and vimentin. Cutpoints for positivity were defined by ROC analyses. Conditional logistic regression models were used to estimate risk ratios and 95% confidence intervals (RR; [95% CI]) in each cohort separately and combined using meta-analysis random effects models for 13 of the ACPAs. Multivariable-adjusted models included alcohol intake and pack-years smoking, collected by questionnaire before blood draw. We adjusted for multiple comparisons using Bonferroni adjustment. Spearman correlation was used to determine whether the number of positive ACPA reactivities was associated with the time between blood draw and diagnosis (4 mo- 14 yr) among cases only.

Results: Mean time (±SD) from blood draw to diagnosis among cases was 8.4 (±4.8) years in NHS and 5.0 years (±2.7) in NHS2. Mean age (±SD) at RA diagnosis was 64.1(±8.0) and 49.2 (±5.1) years in NHS and NHSII. In NHS, 57.9% and, in NHSII, 53.0% of cases were seropositive at diagnosis . In multivariable-adjusted conditional logistic regression analyses, antibody reactivity against clusterin, enolase, 3 of the fibrinogen, both of the histone 2A epitopes and citrullinated vimentin were strongly associated with RA risk (Table). Cases with blood drawn closer in time to diagnosis had a higher number of positive ACPAs (NHS: ρ corr =0.35, p<0.0001, NHSII: ρ corr=0.25, p=0.07).

Conclusion: We have demonstrated that several of a panel of ACPAs targeted to the rheumatoid synovium are strongly associated with risk of future RA.  The number of ACPAs recognized is associated with time to diagnosis, which suggests that epitope spreading occurs within 4 years of diagnosis and thus temporally proximal to the time of diagnosis.