Background:Pulmonary manifestations such as interstitial lung disease (ILD) and chronic obstructive pulmonary disease (COPD) are frequent extra-articular features that carry a poor prognosis in Rheumatoid Arthritis (RA). Little data is available on how RA patients (pts) with pulmonary disease are managed in real-world settings.Objectives:To assess treatment patterns and DMARD discontinuation in RA patients with comorbid lung disease in comparison with other RA patients.Methods:The study included RA Patients enrolled in the Forward Databank with ≥1 year observation after 2000 initiating a DMARD. Forward is a large longitudinal rheumatic disease registry in the US. RA patients’ diagnoses were rheumatologist-confirmed, and every 6 months participants completed comprehensive questionnaires regarding symptoms, disease outcomes, medications, and clinical events. Lung disease (LD+) was defined as at least one of the following: emphysema, asthma, bronchitis, COPD, pleural effusion, fibrosis of the lung, “RA lung”, or ILD, the later classified by ICD9 codes (England 2019). DMARDs were categorized hierarchically into four groups: csDMARDs, TNFi and NTNFi (bDMARDs), and tsDMARDs. Percentage of patients who initiated different DMARDs were reported for pts with LD+/LD-. Discontinuation was analyzed by Kaplan Meier (KM) curves, log-ranks tests, and Cox regression models using time-varying covariates. Best models were created using backward selection models (10% probability of removal) and pre-defined clinical models.Results:Of the 21,525 eligible RA patients, 13.8% had LD+ at the time they initiated a DMARD (follow-up: 69,597 pt-yrs (median 1.9 yrs/pt)). LD+ patients tended to have more severe RA outcomes and comorbidities. MTX-monotherapy (48% vs 44%, p<0.001) and NTNFi were initiated more frequently in LD+ pts with lower use of TNFi (Figure). DMARD discontinuation rates were higher among LD+ patients for all DMARD groups, but KM curves were only significantly different for csDMARDs and TNFi. Different HRs for LD+ were found depending on the model used ranging from 1.18 to 1.28, and all models revealed an increased risk of discontinuation for LD+ patients. Compared to csDMARDs, TNFi were more often discontinued (Table). Other variables associated with an increased risk of discontinuation included: HAQ, Rheumatoid Disease (RD) comorbidity index, pain, prior bDMARDs, and csDMARDs.Conclusion:Different DMARD treatment patterns were found for LD+ patients, who tended to initiate more csDMARD and NTNFi and less likely to initiate a TNFi. LD+ patients were at a higher risk of discontinuation irrespectively of the DMARD treatment, but with greater risk for TNF users.References:[1]England BR, et al. Arth Care Res. doi:10.1002/acr.24043.Figure.DMARD treatment initiators by disease groupTable .Cox models for DMARD discontinuation by stepwise (removal probability 10%) and clinical models including DMARD treatment.Model of DMARD persistence*Model 1- Stepwise-Without drugsModel 2 – StepwiseModel 3 - ClinicalLD+ vs LD–1.181.281.20(1.08 - 1.29)(1.13 - 1.45)(1.08 - 1.34)TNF vs csDmard1.321.22(1.08 - 1.63)(1.04 - 1.44)NTNF vs csDmard1.131.13(0.83 - 1.52)(0.90 - 1.41)tsDmard vs csDmard1.301.02(0.65 - 2.60)(0.64 - 1.62)*Best models searched/Clinical adjusted for LD+/LD-, DMARDs, age, sex, education, HAQ disability, RD comorbidity index, smoking, pain, glucocorticoids, year of entry, prior bDMARDs and csDMARDs counts and MRC breath scale.Disclosure of Interests:Sofia Pedro: None declared, Ted Mikuls Grant/research support from: Horizon Therapeutics, BMS, Consultant of: Pfizer, Joe Zhuo Shareholder of: Bristol-Myers Squibb, Employee of: Bristol-Myers Squibb, Kaleb Michaud: None declared
Epidemiology and treatment patterns of rheumatoid arthritis in a large cohort of Arab patients
