Objective
To evaluate rheumatoid arthritis (RA) disease activity and risk of RA‐associated interstitial lung disease (RA‐ILD).
Methods
We investigated disease activity and risk of RA‐ILD using the Brigham RA Sequential Study (BRASS, 2003–2016). All patients were diagnosed as having RA according to accepted criteria. Disease Activity Scores in 28 joints (DAS28) and covariate data were measured prospectively at annual study visits. Diagnosis of RA‐ILD was determined by review of images from clinically indicated chest computed tomography scans. We analyzed patients without RA‐ILD at baseline. We used Cox regression to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for RA‐ILD, using annually updated DAS28 data, with adjustment for known RA‐ILD risk factors (age, sex, smoking status, RA duration, and serologic status). We performed alternative analyses that did not censor at the time of missing DAS28 data and included adjustment for use of methotrexate, use of glucocorticoids, presence of bone erosions, and presence of rheumatoid nodules.
Results
Among 1,419 participants, the mean ± SD age was 55.8 ± 14.2 years, and 68.6% were seropositive for either cyclic citrullinated peptide or rheumatoid factor. We identified 85 incident cases of RA‐ILD during a mean ± SD follow‐up duration of 8.9 ± 4.2 years per patient. The moderate/high disease activity group had a multivariable HR of 2.22 (95% CI 1.28–3.82) for RA‐ILD compared to the remission/low disease activity group. Risk of RA‐ILD increased across disease activity categories: multivariable HR 1.00 (reference) for remission, 1.41 (95% CI 0.61–3.28) for low disease activity, 2.08 (95% CI 1.06–4.05) for moderate disease activity, and 3.48 (95% CI 1.64–7.38) for high disease activity (P for trend = 0.001). For each unit increase in the DAS28, the risk of RA‐ILD increased by 35% (95% CI 14–60%). Results were similar in analyses that included follow‐up for missing DAS28 data and with adjustment for use of methotrexate, use of glucocorticoids, presence of bone erosions, or presence of rheumatoid nodules.
Conclusion
Active articular RA was associated with an increased risk of developing RA‐ILD. These results suggest that decreasing systemic inflammation may alter the natural history of RA‐ILD development.