Timing of Elevations of Autoantibody Isotypes Prior to Diagnosis of Rheumatoid Arthritis

Abstract

Objective: To evaluate patterns of elevations of isotypes of rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPAs) pre-rheumatoid arthritis (RA) diagnosis and post-RA diagnosis.

Methods: Using the Department of Defense Serum Repository we identified 214 RA cases and 210 matched controls. Up to 3 pre-RA diagnosis and 1 post-RA diagnosis serum samples per subject were tested for RF and for IgA, IgG, and IgM ACPAs. The timing and trajectories of elevations of autoantibodies were evaluated.

Results: Autoantibody levels were elevated in cases versus controls a mean of 17.9 years before RA diagnosis for IgG ACPA, 14.2 years for IgA-RF, 7.2 years for IgM-RF, 6.2 years for IgA ACPA, and 5.0 years for both IgM ACPA and IgG-RF (P < 0.01 for all comparisons). There were similar relationships for positive or negative autoantibody status, with cases first showing positivity for IgG ACPA 1.9 years pre-RA and for IgA-RF 1.7 years pre-RA, followed by the other isotypes. Only IgA ACPA positivity was significantly increased in post-RA diagnosis samples (19% 0-2 years pre-RA versus 39% >2 years post-RA diagnosis; P = 0.04). All autoantibody levels demonstrated an early initial elevation, a period of stability, then an increase immediately before RA diagnosis. A pre-RA endotype of early elevation of autoantibodies was associated with increased use of biologic therapy, and a higher prevalence of sicca symptoms and lung disease post-RA diagnosis.

Conclusion: Differences in patterns of elevations of autoantibody isotypes have implications for understanding the pathophysiology of RA development. These include understanding what factors drive initial autoantibody elevations compared to what factors (including mucosal) drive later increases in autoantibody levels and a transition to clinically apparent RA, and how pre-RA endotypes may influence post-RA diagnosis phenotypes.

Figure 1.. RF and ACPA IgA and IgG levels in RA cases and controls

Using mixed model estimates, levels of RF-IgA diverged significantly between cases and controls at 14.2 years prior to RA diagnosis (mean (SE) log2 levels 1.02 (0.12) in cases vs. 0.69 (0.11) in controls, p<0.05), while for RF-IgG, a significant divergence between cases and controls occurred at 5.0 years pre-RA diagnosis (mean (SE) log2 levels 2.62 (0.08) in cases vs. 2.36 (0.07) in controls, p<0.05) (Figure 1A, 1B) While not shown, RF-IgM diverged from controls at 7.2 years. Levels of ACPA-IgG diverged between cases and controls at 17.9 years pre-RA diagnosis compared to a divergence of ACPA- IgA at 6.2 years (Figure 1C, 1D) While not shown, ACPA-IgM diverged from controls at 5.0 years. The horizontal black line in each figure represents the cut-off for positivity for that autoantibody isotype set in 156 controls.

Figure 2.. Two endotypes identified among RA cases based on trajectories of autoantibody isotypes

In the ‘early’ endotype (116 RA cases, green lines), RF and ACPA IgG and A autoantibodies began to increase from controls (blue lines) early in the pre-RA period (Figure 2A-D). RF-IgM and ACPA-IgM trajectories are not pictured but are within the range of IgG and A autoantibodies for each endotype. In contrast, in the ‘late’ endotype (98 cases, red lines) RF and ACPA IgG and A autoantibodies showed an increase from controls (blue lines) just proximal to the diagnosis of RA (Figure 2A-D, red lines).