Multiple DMARD discontinuations in rheumatoid arthritis: how often and in what patients? Results from a national clinical RA register

Objectives Multiple discontinuations of biologic (b-) or targeted synthetic (ts-) disease-modifying antirheumatic drugs (DMARDs) may indicate difficult-to-treat disease. We aimed to assess the occurrence of b/tsDMARD discontinuations in patients with rheumatoid arthritis (RA), specifically how this varies by definition, across patient subsets and over time.Methods Observational cohort study using data from the Swedish Rheumatology Quality Register on patients diagnosed with RA between 2010 and 2019. We identified three populations: (1) newly diagnosed (N=17 780), (2) initiating a first-ever DMARD (N=18 038) and (3) initiating a first-ever b/tsDMARD (N=8075). In each, we assessed the proportions and characteristics of patients fulfilling either of seven alternative DMARD discontinuation criteria (each encompassing a unique combination of number and type(s) of b/tsDMARD).Results At 4.5 years of follow-up, 10% in populations (1) and (2), and 25% in (3), had discontinued ≥2 b/tsDMARDs with different modes of action. The proportions meeting each of the other six definitions ranged from 0.3% to 10% in (1) and (2), and 1% to 25% in (3). Regardless of definition or time, the characteristics of discontinuers across populations (1) through (3) remained largely similar.Conclusions Applying treatment discontinuation-based definitions to an unselected RA population identifies widely varying proportions of patients with largely similar characteristics. Quantitatively, treatment-based definitions, follow-up time and study population must be clearly specified; qualitatively, the specific definition may be less critical.