Background/Purpose: Determining modifiable risk factors for interstitial lung disease (ILD) is crucial given its substantial morbidity/mortality. Treatment to target of remission/low disease activity improves articular RA outcomes, but the association with ILD is unclear. Prior studies correlated active RA with ILD, but were limited by cross-sectional designs with prevalent ILD. Therefore, we aimed to investigate RA disease activity and incident ILD risk.

Methods: We studied RA disease activity and incident ILD in a prospective cohort study at a single center (2003-2016). All subjects had RA according to ACR criteria. Disease activity score with 28 joints and C-reactive protein (DAS28-CRP3) and covariates were measured annually. Two pulmonologists and one radiologist adjudicated every clinically-indicated chest computed tomography (CT) scan. Cases were defined as consensus agreement with ILD diagnosis. We analyzed subjects with no clinically apparent ILD at baseline. We used Cox regression to estimate HRs and 95%CIs for ILD by DAS28-CRP3, adjusting for known ILD risk factors (age, sex, smoking, RA duration, and serostatus). We investigated DAS28-CRP3 categories at baseline in the primary analysis. As a secondary analysis, we used cumulative average updated DAS28-CRP3, which took into account all previous disease activity measures to predict incident ILD in the subsequent year.

Results: Among 1,281 subjects at baseline, mean age was 56.0 years (SD 14.1), 82.1% were female, 69.6% were seropositive, median RA duration was 9 years, and 58.3% had high/moderate disease activity. During 13,141 patient-years (median follow-up 8 years), we identified 86 cases of incident clinically-apparent ILD. ILD risk significantly increased across baseline DAS28-CRP3 categories. The multivariable HR (95%CI) for ILD by DAS28-CRP3 categories were: 1.00 (reference) for remission, 0.87 (0.30-2.54) for low, 2.31 (1.15-4.46) for moderate, and 2.27 (1.09-4.73) for high; p for trend=0.001). Compared to low/remission, moderate/high disease activity had HR for ILD of 2.41 (95%CI 1.37-4.25). When analyzing cumulative average updated DAS28-CRP3 with fewer ILD outcomes and shorter follow-up, there was a trend towards significance of active RA increasing ILD risk (p=0.09). Seropositivity was strongly associated with ILD risk (HR 2.69, 95%CI 1.38-5.27).

Conclusion: In this large prospective cohort using adjudicated ILD outcomes, active RA was associated with increased risk for clinically-apparent ILD. Replication studies in other prospective cohorts and clinical trials are needed to firmly establish the role of RA treat-to-target approaches and ILD risk, particularly for patients with seropositive RA.