Stop Arthritis Patients’ DMARDs Ahead of COVID Jabs? Trial Brings Clarity

Patients with inflammatory arthritis treated with biologic or other targeted drugs saw no increase in immune response to COVID-19 vaccines when the drugs were temporarily stopped in a randomized trial, researchers said.

With 840 arthritis patients assigned to either a 2-week hold for targeted disease-modifying antirheumatic drugs (DMARDs) or to continue them on their usual schedule, increases in antibodies against the virus’s spike protein were virtually identical 6 weeks after vaccination (geometric mean fold rise ratio 0.96, 95% CI 0.36-2.56), according to Jeffrey R. Curtis, MD, MS, MPH, of the University of Alabama at Birmingham, and colleagues.

Meanwhile, the group reported in JAMA Internal Medicine, odds of developing a new arthritis flare more than doubled during follow-up (OR 2.27, 95% CI 1.41-3.65).

These results should put paid to suggestions that targeted DMARDs — agents that suppress autoimmune activity, often including antibody production, and not always selectively — should be withheld temporarily when COVID vaccinations are planned so as to boost the shots’ immunogenicity.

As Curtis and colleagues put it, “Given the increased flare risk and absence of humoral immunogenicity benefit, these findings do not support routine temporary interruption for COVID-19 vaccine optimization.”

The researchers noted that, especially in the weeks after COVID vaccines were broadly launched in early 2021, it was suggested that, to maximize their effectiveness in patients on immunosuppressant drugs — an important goal given the potentially lethal outcomes from COVID-19 that remained common — those drugs be stopped at least for a while. These considerations prompted the investigators in late 2021 to mount a formal trial that would quantify the benefits and risks of targeted DMARD holidays for inflammatory arthritis patients getting COVID shots.

Called COVER, the trial enrolled 602 patients with rheumatoid arthritis, 198 with psoriatic arthritis, and 40 with some form of spondyloarthritis. They were randomized in equal numbers to the 2-week DMARD hold or to continue as usual, and thus the patients were not blinded. Data analysis began in 2024.

Change from baseline in levels of immunoglobulin G antibodies against the SARS-CoV-2 spike protein was the trial’s primary outcome measure. The chief secondary outcome was disease worsening or overt flares, which was determined in two ways: patients’ responses when asked simply whether they had a flare or worsened disease, and on the more detailed and quantitative Outcome Measures in Rheumatology (OMERACT) Rheumatoid Arthritis Flare Questionnaire. Changes of 10 points or more on this instrument were defined as clinically meaningful disease worsening.

About 160 patients were on one of four tumor necrosis factor (TNF) inhibitors; some 330 were taking a Janus kinase (JAK) inhibitor, including 186 on upadacitinib (Rinvoq); and about 150 were receiving interleukin-17 inhibitors such as secukinumab (Cosentyx). Additionally, 185 were on abatacept (Orencia), which blocks T-cell activation.

Patients assigned to the hold strategy saw antibody titers rise an average 4.69-fold, versus a 4.86-fold mean increase among those who continued their DMARDs. Neither the timing nor the level of response when stratified by DMARD medication type differed between groups.

As seen in earlier studies, however, patients using methotrexate, JAK inhibitors, or abatacept in both arms had muted immune responses when compared with other DMARD classes. That led Curtis and colleagues to write that “vaccination status should ideally be optimized before initiating biologic therapy when feasible”; as well, they suggested, those likely to need repeat vaccinations may benefit from shorter-acting DMARDs.

The investigators didn’t report absolute rates for responses to the yes-or-no question about flares. They did, however, for 10-point changes in OMERACT score: 21.1% in the hold arm, versus 9.1% with uninterrupted DMARDs (difference 12.0 percentage points, 95% CI 5.9-18.1). Most events occurred in the first 2 weeks post-vaccination and resolved before the 6-week follow-up. Curtis and colleagues wrote that these results “indicated that the disease perturbation associated with holding was real and validated … and self-limited.” The timing might suggest that the flares were some sort of vaccine side effect, but the researchers didn’t address this possibility.

Limitations included a substantial (31%) loss to follow-up in the analysis; only 509 of the 840 patients completed the OMERACT assessment. The authors also pointed to the 2-week hold as perhaps insufficient for an adequate washout of some DMARDs. The 3-year period for enrolling and vaccinating patients meant that some received different vaccine formulations than others, and their previous exposures to SARS-CoV-2 may have involved different strains and thus differing underlying immunity. And the researchers didn’t attempt to stratify analyses by diagnosis.