Article Text

Extended report
Discontinuing treatment in patients with rheumatoid arthritis in sustained clinical remission: exploratory analyses from the BeSt study
  1. N B Klarenbeek1,
  2. S M van der Kooij1,
  3. M Güler-Yüksel1,
  4. J H L M van Groenendael2,
  5. K H Han3,
  6. P J S M Kerstens4,
  7. T W J Huizinga1,
  8. B A C Dijkmans4,5,
  9. C F Allaart1
  1. 1Leiden University Medical Center, Leiden, The Netherlands
  2. 2Sint Franciscus Hospital, Roosendaal, The Netherlands
  3. 3Maasstad Hospital, Rotterdam, The Netherlands
  4. 4Jan van Breemen Institute, Amsterdam, The Netherlands
  5. 5VU Medical Center, Amsterdam, The Netherlands
  1. Correspondence to N B Klarenbeek, Leiden University Medical Center, P O Box 9600, 2300 RC Leiden, The Netherlands; n.b.klarenbeek{at}


Objectives To determine the relapse rate after discontinuing treatment in patients with rheumatoid arthritis (RA) in sustained clinical remission, to identify predictors of a relapse and to evaluate treatment response after restarting treatment.

Methods Five-year data from the BeSt study were used, in which 508 patients with recent-onset RA were randomised into four dynamic treatment strategies, aiming at a disease activity score (DAS) ≤2.4. When DAS was <1.6 for ≥6 months, the last disease-modifying antirheumatic drug (DMARD) was tapered and discontinued. If DAS increased to ≥1.6, the last DMARD was immediately reintroduced.

Results During a 5-year period, 115/508 patients (23%) achieved drug-free remission. Of these, 53 patients (46%) restarted treatment because the DAS was ≥1.6 after a median of 5 months, 59 patients (51%) remained in drug-free remission for a median duration of 23 months and 3 (3%) were lost to follow-up. In those who restarted treatment, mean (SD) DAS increased from 1.13 (0.73) at remission before tapering to 2.18 (0.65) at restart, reflecting an increase in all four components of DAS. Multivariable predictors for restarting treatment were anti-cyclic citrullinated peptide (anti-CCP), last DMARD sulfasalazine, low baseline Health Assessment Questionnaire score and high mean DAS until remission. Of the 53 patients who restarted treatment, 39 (74%) again achieved remission 3–6 months after the restart. The median (IQR) damage progression in those who restarted treatment during the year of DAS increase was 0 (0–1) Sharp-van der Heijde units.

Conclusion During 5 years DAS steered treatment, nearly 25% of patients with RA achieved drug-free remission; 46% restarted DMARD monotherapy because of a relapse, the majority of whom again achieved clinical remission within 3–6 months without showing radiological progression during the relapse.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


As a result of the increasing percentage of patients achieving remission with the introduction of early intensive goal-steered therapy,1 more often the dilemma is whether a patient with rheumatoid arthritis (RA) in prolonged remission could discontinue disease-modifying antirheumatic drugs (DMARDs) or whether treatment should be continued. Stopping DMARDs while keeping remission would be beneficial with respect to adverse events and costs. On the other hand, discontinuation of DMARDs could contribute to a relapse of disease with potential harmful consequences.

Historically, the question whether DMARD therapy can be discontinued has often been asked. Several small studies from the 1970s/1980s assessed the need for long-term DMARD maintenance therapy in patients treated according to the pyramid ‘go low go slow’ approach.2,,8 These studies reported high relapse rates after discontinuation of DMARDs (range 58–100%). Later, ten Wolde et al performed a randomised clinical trial in 285 patients with long-standing inactive RA9 10 and reported flare rates of 38% in the placebo (discontinuation) group versus 22% in the group continuing DMARD therapy. Data on withdrawal of DMARDs in patients with established disease have recently been summarised in a meta-analysis.11

Since the approach in RA has shifted towards early intensive goal-steered treatment, the discontinuation of DMARDs has rarely been studied.12 13 There are few data on the safety of withdrawing DMARDs, the chance of flares and the response to reintroduction of DMARDs.10 In addition, it is unclear whether a flare after a drug-free remission episode can be predicted. With prediction, subgroups of patients might be identified in which medication can be safely withdrawn compared with other subgroups in whom treatment should be continued.

The protocol of the BeSt study, a randomised clinical trial in recent-onset RA, allowed discontinuation of DMARDs in patients in prolonged remission under strict control of disease activity.14 The objectives of this analysis were: (1) to assess the flare rate in patients in drug-free remission; (2) to describe the severity of relapse; (3) to identify predictors for relapse; and (4) to assess the response to reintroduction of DMARDs.


Study design and patients

Five-year data of the BeSt study were used. Details of the design have been described elsewhere.15 In summary, 508 patients with recent-onset active RA according to the 1987 RA classification criteria (disease duration <2 years)16 were randomised into four treatment strategies: sequential monotherapy (n=126), step-up combination therapy (n=121), initial combination therapy with prednisone (n=133) and initial combination therapy with methotrexate and the tumour necrosis factor α inhibitor infliximab (MTX+IFX, n=128). Treatment was adjusted based on 3-monthly disease activity score (DAS) measurements (disease activity score 44 joints, DAS44), aiming at DAS ≤2.4.17 18 If DAS was >2.4, the next step of the protocol was taken. If DAS was ≤2.4 for ≥6 months, treatment was tapered to the maintenance dose. For details on treatment steps per arm see figure in online supplement.

Discontinuation of DMARD therapy

Remission was defined as a DAS <1.6.19 Two years after inclusion, the protocol allowed tapering and discontinuation of the last DMARD if patients fulfilled the following conditions: (1) DMARDs in maintenance dose according to the protocol; and (2) clinical remission (defined as DAS <1.6) for ≥6 months.

For MTX monotherapy, sulfasalazine (SSA) monotherapy, leflunomide and intramuscular gold, the maintenance doses were 10 mg/week, 2000 mg/day, 10 mg every other day and 50 mg every other week, respectively. All combinations were first tapered to MTX monotherapy which was then tapered to 10 mg/week, with the exception of the COBRA combination (MTX, SSA, prednisone)20 which was tapered to SSA 2000 mg/day as maintenance dose, and the combination azathioprine and prednisolone which was tapered to azathioprine 2 mg/kg/day.

Discontinuation of the last DMARD occurred by tapering MTX with 2.5 mg/4 weeks and SSA with 500 mg/4 weeks; maintenance doses of leflunomide, gold and azathioprine were simply discontinued.

Restart of DMARD therapy

If the DAS was ≥1.6, the last tapered DMARD was immediately restarted in maintenance dose and could not be discontinued twice. Patients who remained drug-free until year 5 will be referred to as ‘sustained drug-free remission patients (SDFR)’ whereas patients who restarted treatment before 5 years of follow-up will be called ‘restarters’.

Statistical analysis

The software program SPSS version 17.0 was used for all statistical analyses.

Among the patients who achieved drug-free remission, variables associated with restarting treatment were identified by univariable logistic regression using characteristics from baseline and from the last visit before tapering the last DMARD to 0. Multivariable logistic regression with univariable logistic determinants (p<0.10) was used to identify independent predictors for restart using a backward selection procedure (p<0.05). Subsequently, the multivariable predictors from the backward procedure were entered into a new logistic regression model. Variables that were not associated with restarting treatment in the univariable logistic regression were then added one by one to assess whether they had additional predictive value.


After 5 years 115 of the 508 patients (23%) achieved drug-free remission with no significant differences between the groups (p=0.20, table 1). Of these, 53 (46%) restarted treatment after a median (IQR) period of 5 (2–16) months. Fifty-nine (51%) remained in drug-free remission with a median (IQR) duration of 23 (15–25 months) at 5 years and 3 (3%) were lost to follow-up. Details of the treatment steps at which drug-free remission was reached are given in table 2.

Table 1

Overview of the distribution of patients who achieved drug-free remission during 5 years of follow-up in the four treatment groups

Table 2

Treatment details of the 115 patients who discontinued all DMARDs due to sustained clinical remission

In the restarters, mean (SD) DAS increased from 1.13 (0.73) at remission before tapering to 2.18 (0.65) at restart, reflecting an increase in all four components of DAS: median (IQR) erythrocyte sedimentation rate increased from 7 (5–14) to 19 (7–28), swollen joint count from 0 (0–0) to 2.5 (0–4), Ritchie articular index21 from 0 (0–1) to 3 (1–6) and visual analogue scale (VAS) global health from 15 (2–22) to 28 (15–55). In 38/53 patients (72%) the highest disease activity during the flare was ≤2.4 (low disease activity), in 12/53 (23%) it was 2.4–3.7 (moderate disease activity) and in 3/53 (6%) it was >3.7 (high disease activity).

The presence of anti-cyclic citrullinated peptide (anti-CCP2), rheumatoid factor, lower VAS global health and lower health assessment questionnaire (HAQ) at baseline were univariably associated with restarting treatment (table 3). Furthermore, the baseline Sharp-van der Heijde score (SHS) tended to be higher in those who restarted treatment than in SDFR patients (NS). None of the characteristics at the time of remission were associated with restarting treatment. In the multivariable analysis, the presence of anti-CCP was the strongest independent predictor for restart, followed by a higher mean DAS until remission, lower baseline HAQ and SSA as last DMARD (table 4). Both restarters and SDFR patients had good functional ability during their remission (table 3), comparable to HAQ scores of the general population (age- and sex matched: median (IQR) HAQ 0.20 (0.15–0.34) and 0.18 (0.10–0.49) for restarters versus SDFR, respectively).22

Table 3

Univariable analysis: potential variables associated with a flare during a drug-free remission period

Table 4

Multivariable predictors for a flare during a drug-free remission period

Of the 53 restarters, 25 (47%) again achieved clinical remission within 3 months after restarting treatment with the last used DMARD in maintenance dose and another 14 (26%) within 6 months. Eleven patients (21%) achieved a DAS ≤2.4, one patient (2%) did not achieve a DAS ≤2.4 and two patients (4%) were lost to follow-up. The large majority of patients did not show joint damage progression in the first year after discontinuation of DMARDs, with a median (IQR) SHS progression of 0 (0–1) units in the restarters during the year of DAS increase compared with 0 (0–0) in SDFR patients in the first year completely drug-free (p=0.44, Mann–Whitney U test, figure 1).

Figure 1

Cumulative probability plot showing radiological joint damage progression in the first year after stopping the last disease-modifying antirheumatic drug in patients who remained in clinical remission (circles) versus patients who restarted treatment (triangles). The dashed line represents clinically relevant progression. SHS, Sharp-van der Heijde score.


We implemented discontinuation of DMARDs early in the course of the disease in patients with RA in clinical remission treated according to an early, aggressive and dynamic treatment approach. Twenty-three percent of patients could discontinue their DMARDs because of remission during ≥6 months: 51% of them remained in drug-free remission (median 23 months) and 46% restarted treatment. The majority of these restarters again achieved clinical remission within 3–6 months after restarting DMARDs without suffering joint damage progression.

Earlier reported relapse rates were comparable or higher (38–100%)2,,9 11 13 than the 46% relapse rate in our study. The largest well-designed study by ten Wolde et al9 in patients with established RA showed a relapse rate in the same range as the rate we found in our early RA cohort (38% vs 46%). Direct comparison of these relapse rates is difficult owing to different definitions of relapse/remission and differences in patient populations. In the older studies,2,,8 11 the majority of patients had longstanding disease with significant joint damage and poor functional ability. We report discontinuation of DMARDs early in the disease course with the key advantage that patients in drug-free remission had limited joint damage and enjoyed a functional ability comparable to an age- and sex-matched healthy reference population.22

Tanaka et al recently published the RRR study on discontinuation of infliximab in patients with RA after attaining low disease activity.23 As in our study, medication was withdrawn and reintroduced at a predefined cut-off and the occurrence of a flare was registered. This study differs from the BeSt study in several ways. First, in the RRR study, infliximab was discontinued while methotrexate was continued whereas, in BeSt study, all antirheumatic treatment was withdrawn. Second, in the RRR study, infliximab was discontinued and reintroduced at the low disease activity cut-off point. No clinical remission was required. Furthermore, the disease duration at inclusion was higher in the RRR study (5.9 vs 0.4 years). Despite these differences in study design, the observed chance of a flare was remarkably comparable (45% versus 46%) and, as we found, the majority of patients responded well to reintroduction of treatment after a relapse.

The presence of CCP2 antibodies is one of the strongest known predictors for a worse disease course in RA.24 In line with this, anti-CCP-positive patients have a lower chance of achieving drug-free remission.14 In addition, we found that among the patients achieving drug-free remission, the presence of anti-CCP was the strongest predictor for the occurrence of a flare. Nevertheless, 30% of the patients in sustained drug-free remission were anti-CCP positive, indicating that, even in anti-CCP positive-patients, successful drug-free remission is possible.

Surprisingly, low HAQ at baseline was predictive for restarting treatment in the univariable and the multivariable analysis. The univariable results of VAS general health pointed in the same direction. This suggests that if patients are able to improve more in HAQ—that is, if patients have gained more (mean HAQ improvement 1.14 vs 0.83 in SDFR vs restarters)—they have a higher chance of retaining remission after discontinuation of drugs.

Patients who discontinued MTX maintenance therapy had a higher chance of retaining remission than patients who discontinued SSA. Although the patient numbers are low, these results may suggest that SSA is less potent in inducing sustained remission after discontinuation than MTX. Additional research is needed to confirm this finding.

We hypothesised that patients with a DAS just below the cut-off score of 1.6 might have a higher chance of relapse than patients with a lower DAS. We therefore assessed whether the level of inflammation at the time of remission, as measured with the DAS and its components, was predictive for a flare. There appeared to be no association between inflammation measures and the risk of a flare, indicating that the ‘depth’ of the DAS remission is not useful in predicting whether the remission will be maintained after discontinuation of DMARDs. ‘Deeper’ remission is not ‘truer’ remission in that sense.

During the relapse the duration and severity of a higher level of disease activity seems limited. DMARDs in a low maintenance dose were restarted immediately if DAS rose to ≥1.6 and the large majority again achieved clinical remission within 3–6 months. During the flare the DAS increased to a low disease activity level in 72% of patients; few experienced high disease activity during the flare. The temporarily higher DAS level might have contributed to the slightly higher (non-significant) joint damage progression in the restarters than in the SDFR patients. Another explanation could be that the restarters had less favourable characteristics than the SDFR patients, including a higher percentage of anti-CCP-positive patients, leading to a higher risk of progression.24 This is supported by the observation that, before discontinuing DMARDs, restarters already had more joint damage than patients retaining remission (median SHS 5.0 vs 1.5).

Since DMARDs were stopped in all patients in prolonged remission, it is unknown what the flare rate would have been if treatment had been continued. Ten Wolde et al found that, in patients who continued treatment, the flare rate was also considerable (22%) but significantly lower than in patients discontinuing treatment (38%).9 Being aware of the differences in patient population, these findings suggest that part of the flares we observed could have happened even if DMARDs were continued.

In summary, in 23% of patients with recent-onset RA, DMARDs could be discontinued because of prolonged clinical remission. Based on strict criteria, almost half of them had to restart treatment. The presence of anti-CCP was the strongest predictor for restarting treatment. The large majority of patients who lost remission remained in low disease activity, regained clinical remission shortly after reintroduction of low-dose monotherapy and showed no joint damage in the year of the restart. We therefore propose that, under continued tight control, discontinuation of the last DMARD can be considered in patients in stable clinical remission. The final decision whether or not to withdraw treatment in an individual patient should be made by the physician and patient together, carefully weighing the advantages and disadvantages.


The authors thank all patients and the following rheumatologists (other than the authors) who participated in the Foundation for Applied Rheumatology Research (all locations are in The Netherlands): W M de Beus (Medical Center Haaglanden, The Hague); C Bijkerk (Reinier de Graaf Gasthuis, Delft); M H W de Bois and G Collée (Medical Center Haaglanden, The Hague); J A P M Ewals (Haga Hospital, The Hague); A H Gerards (Vlietland Hospital, Schiedam); R J Goekoop (Haga Hospital, The Hague); B A M Grillet (ZorgSaam Hospital, Terneuzen); J M W Hazes (Erasmus Medical Center, Rotterdam); H M J Hulsmans (Haga Hospital, The Hague); M H de Jager (Albert Schweitzer Hospital, Dordrecht); J M de Jonge-Bok (retired); M V van Krugten (Admiraal de Ruyter Hospital, Vlissingen); H van der Leeden (retired); W F Lems (VU Medical Center, Amsterdam); M F van Lieshout-Zuidema (Spaarne Hospital, Hoofddorp); A Linssen (retired); P A H M van der Lubbe (Vlietland Hospital, Schiedam); C Mallée (Kennemer Gasthuis, Haarlem); T H E Molenaar (Groene Hart Hospital, Gouda); M van Oosterhout (Groene Hart Hospital, Gouda); H C van Paassen (Sint Franciscus Gasthuis, Rotterdam); A J Peeters (Reinier de Graaf Gasthuis, Delft); H K Markusse (deceased); H K Ronday (Haga Hospital, the Hague); D van Schaardenburg (VU Medical Center, Amsterdam and Jan van Breemen Institute, Amsterdam); A A Schouffoer (Groene Hart Hospital, Gouda); P E H Seys (Lievensberg Hospital, Bergen op Zoom); R M van Soesbergen (retired); P B J de Sonnaville (Admiraal de Ruyter Hospital, Goes); I Speyer (Bronovo Hospital, The Hague); K S S Steen (Kennemer Gasthuis, Haarlem); J Ph Terwiel (retired); A E Voskuyl (VU Medical Center, Amsterdam); M L Westedt (Bronovo Hospital, The Hague); S ten Wolde (Kennemer Gasthuis, Haarlem); J M G W Wouters (Sint Franciscus Gasthuis, Rotterdam); D van Zeben (Sint Franciscus Gasthuis, Rotterdam). We would also like to thank all other rheumatologists and trainee rheumatologists who enrolled patients in this study, and all research nurses for their contributions.


Supplementary materials

  • Web Only Data

    Files in this Data Supplement:


  • Funding This study was funded by a grant from the Dutch College of Health Insurances (College Voor Zorgverzekeringen) with additional funding provided by Scheringh-Plough BV and Centocor. The authors, not the sponsors, were responsible for the study design, collection, analyses and interpretation of all data, the writing of the article and the decision to publish.

  • Competing interests TWJH, BACD and CFA received speakers' fees from various pharmaceutical companies (less than US$5000 per year).

  • Ethics approval This study was conducted with the approval of the medical ethical committees of the participating hospitals. All patients gave written informed consent.

  • Provenance and peer review Not commissioned; externally peer reviewed.