AbbVie (NYSE: ABBV) reported that upadacitinib (Rinvoq) beat adalimumab (Humira) on the primary endpoint of a Phase IIIb/IV head-to-head trial in rheumatoid arthritis. The trial, specifically focused on patients who had already failed one TNF inhibitor, showed that nearly twice as many patients reached low disease activity on AbbVie's JAK inhibitor at 12 weeks.
The SELECT-SWITCH trial is a multicenter, randomized, double-blind, double-dummy, active comparator-controlled Phase IIIb/IV study enrolling 492 adults with moderate to severe rheumatoid arthritis on a stable background of methotrexate who had an inadequate response or intolerance to a single TNF inhibitor other than adalimumab.
At week 12, 43.3% of patients receiving Rinvoq (upadacitinib) achieved the primary endpoint of low disease activity, defined as a Disease Activity Score 28 C-reactive protein of 3.2 or below, compared with 22.4% of patients on Humira (adalimumab) (p<0.001). On the ranked secondary endpoint of remission, defined as DAS28-CRP below 2.6, 28.4% of upadacitinib-treated patients met the threshold versus 14.5% on adalimumab (p<0.001). Rates of serious adverse events were low and broadly balanced, at 2.0% for upadacitinib and 2.4% for adalimumab, with no adjudicated venous thromboembolism, major adverse cardiovascular events, or deaths observed in the 12-week period.
The question SELECT-SWITCH was designed to answer sits at a clinically contested juncture: what to do after a first TNF inhibitor fails. TNF inhibitors remain the most commonly prescribed first-line targeted therapy in rheumatoid arthritis, yet cycling to a second TNF inhibitor after initial failure is widespread in practice despite limited prospective evidence supporting that strategy over a mechanistic switch. Real-world data and retrospective analyses have consistently suggested that switching to a different mode of action produces better outcomes than TNF cycling, but direct randomized evidence comparing the two approaches has been absent — until now.
Upadacitinib is a selective JAK1 inhibitor that works intracellularly, suppressing cytokine-driven inflammatory signaling downstream of multiple receptor pathways simultaneously. That mechanistic breadth distinguishes it from adalimumab, a fully human monoclonal antibody that neutralizes tumor necrosis factor alpha at a single extracellular target. The hypothesis underlying SELECT-SWITCH is that patients who have already experienced inadequate response or intolerance to TNF blockade may derive more benefit from targeting a different node in the inflammatory cascade than from re-engaging the same pathway with a different agent.
The magnitude of the difference observed at week 12 is notable in context. Roughly one in five patients on adalimumab achieved low disease activity, compared to roughly two in five on upadacitinib — a separation that held across both the primary low disease activity endpoint and the remission threshold. Treatment guidelines from both the American College of Rheumatology and EULAR identify remission as the optimal treatment target in rheumatoid arthritis, and the data suggest that a JAK inhibitor vs TNF inhibitor strategy, at least in this post-TNF population, produces meaningfully different rates of reaching that goal within the first three months of therapy. Cross-trial comparisons are limited by differences in patient populations, background therapies, and outcome definitions, so these figures should be interpreted within the SELECT-SWITCH design rather than extrapolated broadly.
Upadacitinib already carries US FDA approval for moderate to severe rheumatoid arthritis in adults following inadequate response or intolerance to one or more TNF inhibitors, among several other approved indications spanning psoriatic arthritis, ankylosing spondylitis, ulcerative colitis, Crohn's disease, atopic dermatitis, and giant cell arteritis. SELECT-SWITCH does not alter the regulatory status of the drug, but it adds prospective randomized evidence to a treatment decision that clinicians have historically navigated using observational data and clinical judgment alone.