Withdraw DMARDS after remission? Save

The possibility of withdrawing DMARDs after patients achieve remission has been in our minds for a while. Yet when our patients ask whether it is a good idea to taper or stop their DMARD when they are doing well, most of us don’t have a black or white answer for them. The arrival on the market of several biosimilars and very reassuring long term safety data with bioDMARDs also provides encouraging prospects for longer term exposure.

ARCTIC REWIND might be the evidence we were looking for. This study was a randomized, multicenter, open-label, clinical trial randomizing 2:1:1 to stable csDMARDs, half-dose csDMARDs, or half-dose csDMARDs for one year followed by withdrawal. 

The trial enrolled 160 RA patients in sustained remission for ≥1 year and the primary endpoint was the absence of disease activity flare at 3 years. The 1-year results of the ARCTIC REWIND trial was published in the JAMA in 2021. For memory, at 1 year, treatment with half-dose vs stable-dose csDMARDs resulted in disease flares in 25% vs 6% over 12 months; for which the authors rejected the non-inferiority of this strategy versus stable-dose. The results of this study at 3 years were presented by the authors in San Diego (abstract 2543). 89% patients completed the 3 years follow-up. 80 % remained flare-free in the stable group, compared with 60 % in the half-dose group and 38 % in the tapering to withdrawal group. The adjusted hazard ratio for flare was 4.3 in the withdrawal group. 75% of those who flared at half-dose or after withdrawal recovered remission. However, in the withdrawal group, 18% more bioDMARDs and 50% more glucocorticoids were prescribed. The half dose group was exposed to significantly fewer other treatments, and it might be the reason this specific group presented with higher % of > or equal to 3 unit changes in Sharp van der Heijde scores suggesting radiographic progression over time as opposed to the other groups.

Now one question remains: can we say the same about bioDMARDS?

The same authors also presented the 3 years outcomes of the randomized, multicenter, open-label, clinical trial including 99 patients (abstract L07) in sustained remission for 12 months on stable TNFi therapy, with no swollen joints at inclusion. Of 99 randomized patients, only 25% (remained flare-free in the tapering TNFi compared to 85% in the stable TNFi group. Withdrawing TNFi  corresponded to a hazard ratio for flare of 9.4. Although 81% of the patients would recover remission at the next visit, the tapering TNFi group had lower Boolean remission rates. AE/serious AE occurred in 81% in the stable dose group and 21% in the withdrawal group.

To my opinion, what these studies mainly highlight, is the absence of a ‘one size fits all’ tapering/withdrawal strategy; and the need for stratification biomarkers in order to predict outcome after withdrawal. I would suggest a stratification into 3 groups (and not 2): those who will flare but recover response; those who will flare and not recover response (in whom we should absolutely avoid to withdraw); and, those who will achieve sustained remission.

In that respect, translational science has been very helpful by providing insights into why patients in remission relapse after bioDMARD withdrawal; and as often, the truth lies in the synovial tissue. Alivernini et al. described a tissue specific myeloid cell cluster responsible for relapse in RA patients after drug withdrawal. This is not ready for clinical applications yet but should give interesting insights into further stratification strategies for drugs tapering/withdrawal in RA.