The first issue to consider is how long patients are able to remain on DMARDs, particularly if they continue to need them. Most patients who start DMARD monotherapy do not stay on treatment with single DMARDs indefinitely [2]. More than half the patients who start treatment with a single DMARD have had to stop treatment after 2–3 years. Retention rates differ across different DMARDs. Patients stay longer on methotrexate than other DMARDs. The probability of continuing methotrexate for 5 years approaches 80 %. For other DMARDs retention rates are lower. They may be no more than 30 % after 2 years treatment. Such low retention rates of patients starting DMARDs mean it is crucial to consider carefully the benefits and risks of discontinuing DMARDs in patients in whom therapy is controlling RA without causing adverse effects.

There have been many attempts to stop DMARDs when patients have achieved good clinical responses [3]. This has included a substantial number of clinical trials of treatment withdrawal. Most of these trials are historical. They mainly studied DMARD withdrawal in patients with RA who were in remission or were achieving good clinical responses. They followed patients for as long as 2 years. Most involved classical DMARDs such as gold and penicillamine, though these are not widely used today.

About one fifth of patients who continued to take DMARDs flared. In contrast about two fifth of patients who stopped DMARDs had a flare. When patients who had stopped DMARDs restarted treatment most regained control of their disease. Only a few patients did not benefit from resuming DMARDs.

One way to give intensive treatment is to start several DMARDs at the same time [4]. They can then be stepped down to DMARD monotherapy. This approach is particularly favoured in early RA. Several trials have evaluated using two conventional DMARDs, usually methotrexate and sulfasalazine, with prednisolone. Treatment is then stepped down after 6 months or longer to DMARD monotherapy. Patients benefit from the use of intensive treatment without the need to take combinations of drugs indefinitely. Overall trials in early and established RA show that step-down combination therapy is effective and has sustained benefits. Remaining on one anchor DMARD reduces subsequent flares. The optimal maintenance DMARD therapy has not been identified in these trials. However, most experts believe methotrexate is the best anchor drug to continue as monotherapy.

Several historical case series looked at reducing the frequency of DMARD administration [5]. Reducing methotrexate from weekly to fortnightly did not appear to cause more flares. The same was seen when reducing the frequency of giving penicillamine, though clearly the relevance of findings with this historical DMARD to current practice is uncertain.

There is some evidence that when patients are receiving biological treatment with infliximab the dose of methotrexate can be tapered without an increased risk of flares.

Despite these positive findings the long-term goal of “drug-free” remission remains elusive. It is only rarely achieved. About 10 % of patients in early arthritis cohorts who require DMARDs can eventually achieve sustained drug free remissions when treatments are stopped.

Detailed analyses of early rheumatoid cohorts have identified several predictors of drug-free remissions [6]. These are symptom duration, rheumatoid factor positivity and presence of the HLA–DRB1 shared epitope alleles. Rheumatoid factor positivity is the strongest predictor and seropositive patients are far less likely to be able to withdraw treatment than seronegative patients.

After reviewing the available evidence, expert groups have different perspectives about discontinuing DMARDs. There appears to be no overall consensus. UK guidelines from the National Institute for Health and Care Excellence (NICE) recommend that if RA is stable, DMARD doses should be cautiously reduced, returning promptly to disease controlling doses if there are any indications of a flare [7]. EULAR European guidelines are more guarded about DMARD tapering. They recommend that in sustained long-term remission cautious titration of synthetic DMARD dose may be considered [8]. By contrast American College of Rheumatology guidelines do not comment on DMARD withdrawal [9].