Real‐World Treatment Patterns in Rheumatoid Arthritis–Associated Interstitial Lung Disease

Objective We evaluated real‐world treatment patterns in rheumatoid arthritis–associated interstitial lung disease (RA‐ILD) within the Veterans Health Administration (VA). Methods We performed a cohort study of RA‐ILD treatment patterns in the VA between 2006 and 2021. Patients with RA‐ILD were identified using validated administrative algorithms. Pharmacy databases were queried for medication dispensing before and after ILD diagnosis. The primary comparison was the proportion of patients prescribed each medication before and after ILD diagnosis. Multivariable Cox regression was used to evaluate predictors of initiating an American College of Rheumatology (ACR)/American College of Chest Physicians (CHEST) guideline recommended medication after ILD diagnosis among patients not using these therapies at baseline. Results Among 4,108 patients with RA‐ILD, mean age was 68.6 years and 93.4% were male. Prednisone was commonly prescribed before (64.6%) and after (76.7%) ILD diagnosis. Use of conditionally recommended first‐line therapies in the ACR/CHEST guideline (mycophenolate, azathioprine, rituximab, and cyclophosphamide) increased after ILD diagnosis, but the overall frequency was low (range 1.2% to 12.0%). Methotrexate use decreased (47.1% to 30.4%), whereas leflunomide, tumor necrosis factor inhibitors, and abatacept use increased after ILD diagnosis. Reduced forced vital capacity, anti–cyclic citrullinated peptide/rheumatoid factor seropositivity, ILD diagnosis between 2012 and 2021 (vs 2006 and 2011), and higher pulmonary visit frequency were associated with guideline recommended medication initiation while older age at RA‐ILD diagnosis, and higher comorbidity burden were associated with less frequent initiation. Conclusion Most ACR/CHEST first‐line recommended therapies for RA‐ILD treatment were infrequently used among VA patients with RA‐ILD between 2006 and 2021. These findings illustrate real‐world use of RA‐ILD therapies and suggest that treatment decisions in RA‐ILD are guided by many factors in addition to ILD severity.