Autoantibodies against citrullinated and native proteins and prediction of rheumatoid arthritis-associated interstitial lung disease: a nested case–control study

Background Rheumatoid arthritis-associated interstitial lung disease (ILD) is one of the leading causes of premature death among patients with rheumatoid arthritis. Improving prediction of rheumatoid arthritis-associated ILD is crucial to allow for earlier diagnosis and treatment. We aimed to identify fine-specificity anti-citrullinated protein antibodies (ACPAs) associated with incident rheumatoid arthritis-associated ILD. Methods In this nested case–control study within the prospective Brigham Rheumatoid Arthritis Sequential Study (BRASS), we matched cases of incident rheumatoid arthritis-associated ILD diagnosed between March 1, 2003, and April 14, 2016, to control patients with rheumatoid arthritis without ILD on the following characteristics: time of blood collection, age, sex, rheumatoid arthritis duration, and rheumatoid factor status. We measured ACPA and anti-native protein antibodies using a multiplex assay on stored serum collected before onset of rheumatoid arthritis-associated ILD. We used logistic regression models to calculate odds ratios (ORs) with 95% CIs for rheumatoid arthritis-associated ILD, adjusting for prospectively collected covariates. We estimated the optimism-corrected area under the curves (AUCs) using internal validation. We used model coefficients to generate a risk score for rheumatoid arthritis-associated ILD. Findings We identified 84 incident rheumatoid arthritis-associated ILD cases (mean age 67 [SD 10] years, 65 [77%] female and 19 [23%] male, 76 [90%] White) and 233 rheumatoid arthritis controls without ILD (mean age 66 [11] years, 186 [80%] female and 47 [20%] male, 219 [94%] White). We identified six fine-specificity antibodies that were associated with rheumatoid arthritis-associated ILD. The antibody isotypes and targeted proteins were IgA2 to citrullinated histone 4 (adjusted OR 0·08 [95% CI 0·03–0·22] per log-transformed unit), IgA2 to citrullinated histone 2A (4·03 [2·03–8·00]), IgG to cyclic citrullinated filaggrin (3·47 [1·71–7·01]), IgA2 to native cyclic histone 2A (5·52 [2·38–12·78]), IgA2 to native histone 2A (4·60 [2·18–9·74]), and IgG to native cyclic filaggrin (2·53 [1·47–4·34]). These six antibodies predicted the risk of rheumatoid arthritis-associated ILD better than did all clinical factors combined (optimism-corrected AUC 0·84 versus 0·73). We developed a risk score for rheumatoid arthritis-associated ILD by combining these antibodies with clinical factors (smoking, disease activity, glucocorticoid use, and obesity). At 50% predicted probability of developing rheumatoid arthritis-associated ILD, the risk scores both without (2·6) and with (5·9) antibody biomarkers achieved a specificity of 93% or higher for rheumatoid arthritis-associated ILD. Interpretation Specific ACPAs and anti-native protein antibodies improve prediction of rheumatoid arthritis-associated ILD. These findings implicate synovial protein antibodies in the pathogenesis of rheumatoid arthritis-associated ILD and, once validated in external studies, suggest that these antibodies might have clinical utility in predicting the development of ILD in patients with rheumatoid arthritis. Funding US National Institutes of Health/National Institute of Arthritis and Musculoskeletal and Skin Diseases.