Thesafety and efficacy of upadacitinib 15mg (UPA15) through week 216 was evaluated in patients with rheumatoid arthritis (RA) from the long-term extension (LTE) of the phase 3 SELECT-CHOICE study. Patients with RA refractory to biologic disease-modifying antirheumatic drugs (bDMARDs) were randomized to UPA15 or abatacept (ABA) for 24weeks. During the open-label LTE, patients on ABA switched to UPA15 at week 24, and those on UPA15 continued treatment. Thesafety and efficacy of continuous UPA15, and ABA to UPA15, are summarized through week 216. The LTE was comprised of 91.4% (n = 277/303) of patients that initially received UPA15, and 89.6% (n =277/309) that initially received ABA. Of patients on UPA15 in the LTE (n =547), 28.3% (n =155/547) discontinued the study drug by week 216. Relative to other adverse events of special interest, and largely consistent with previous findings at week 24, higher rates of serious infection, COVID-19, herpes zoster, and elevated creatine phosphokinase were reported, while rates of malignancy excluding nonmelanoma skin cancer (NMSC), NMSC, major adverse cardiovascular event (MACE), and venous thromboembolism (VTE) were low. Long-term safety data with UPA through week 216 aligned with previous observations and no new safety risks were identified, including in patients who switched from ABA to UPA15. Proportions of patients achieving 28-joint disease activity score based on C-reactive protein (DAS28[CRP]) < 2.6/ ≤ 3.2, clinical disease activity index (CDAI) and simple disease activity index (SDAI) low disease activity/remission, ≥ 20%/50%/70% improvement in theAmerican College of Rheumatology (ACR20/50/70) response criteria, and Boolean remission were maintained or improved with UPA15 through week 216. Improvements in theHealth Assessment Questionnaire-Disability Index (HAQ-DI), patient's assessment of pain, and Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) were also maintained or improved with UPA15 through week 216. Across all efficacy endpoints, similar results were observed in patients who switched from ABA to UPA15 versus continuous UPA15. Patients with an inadequate response to ≥ 1 prior tumor necrosis factor (TNF) inhibitor (UPA15: n =263/303, 86.8%; ABA to UPA15: n =273/309, 88.3%) showed similar responses to the total population. The long-term safety profile of UPA was consistent with previous findings and the broader RA clinical program. Compared to the primary analyses at week 24, efficacy responses were maintained or further improved with UPA15 through week 216 in patients with RA. Trial registration, ClinicalTrials.gov identifier: NCT03086343.