Early characterization of difficult-to-treat rheumatoid arthritis by suboptimal initial management: a multicentre cohort study

Objectives To demonstrate that unsuccessful treatment optimization in early disease is associated with difficult-to-treat RA (D2T-RA). Methods In this retrospective multicentre cohort study conducted from 09/2021-03/2022, we enrolled individuals fulfilling the 2010 ACR/EULAR RA criteria diagnosed 2000-2019. The outcome was D2T-RA by the EULAR definition. We used robust regression to examine the associations with delay, dose, duration of methotrexate and discontinuation of glucocorticoids. We tested through multinomial regression which factors were associated with persistent inflammatory refractory RA (PIRRA) or non-inflammatory refractory RA (NIRRA). Sensitivity analysis included a case-control study matching the year of diagnosis. Results We enrolled 48 D2T-RA patients and 145 non-D2T-RA controls. Methotrexate was started within 3 months in 16.7% of D2T-RA vs 33.1% of non-D2T-RA (P = 0.011). Adequate duration of methotrexate was obtained in significantly fewer D2T-RA patients (70.8% vs 85.5%). Glucocorticoids were continued beyond 6 months in a higher proportion of D2T-RA patients (70.8% vs 33.8%, P < 0.001). In multiple regression, treatment delay beyond 3 months (OR 0.3; 95% CI 0.1, 0.9) and non-discontinuation of glucocorticoids after 6 months (OR 4.6; 95% CI 2.2, 9.5) were significantly associated with D2T-RA. Treatment delay was significantly associated with PIRRA only, while non-discontinuation of glucocorticoids was significantly associated with PIRRA and NIRRA. Results were replicated in sensitivity analyses. Conclusion Failure to start methotrexate within 3 months and not being off glucocorticoids within 6 months are early predictive features of D2T-RA.