Article Text
Abstract
Objectives To perform a systematic literature review of effective strategies for the treatment of rheumatoid arthritis (RA).
Methods As part of a European League Against Rheumatism (EULAR) Task Force investigation, a literature search was carried out from January 1962 until February 2009 in PubMed/Ovid Embase/Cochrane and EULAR/American College of Rheumatism (ACR)) abstracts (2007/2008) for studies with a treatment strategy adjusted to target a predefined outcome. Articles were systematically reviewed and clinical outcome, physical function and structural damage were compared between intensive and less intensive strategies. The results were evaluated by an expert panel to consolidate evidence on treatment strategies in RA.
Results The search identified two different kinds of treatment strategies: strategies in which the reason for treatment adjustment differed between the study arms (‘steering strategies’, n=13) and strategies in which all trial arms used the same clinical outcome to adjust treatment with different pharmacological treatments (‘medication strategies’, n=7). Both intensive steering strategies and intensive medication strategies resulted in better outcome than less intensive strategies in patients with early active RA.
Conclusion Intensive steering strategies and intensive medication strategies produce a better clinical outcome, improved physical function and less structural damage than conventional steering or treatment. Proof in favour of any steering method is lacking and the best medication sequence is still not known.
Statistics from Altmetric.com
Introduction
The treatment of rheumatoid arthritis (RA) has undergone tremendous changes. Most clinical trials compare medical treatment according to rigid protocols. These trials are useful to assess efficacy of drugs, but do not give information on the order in which treatment should be given, nor when and how treatment regimens should be adjusted. Studies examining the effect of treatment strategies were designed.1,–,3 Our systematic literature search focused on these strategic trials. The aim of the search was to define what medication sequence or adjustment methods should be started or modified and how patients should be monitored to obtain low disease activity or remission, good physical function and low progression of joint damage most effectively in patients with RA.
Methods
Systematic literature search
An international steering committee consisting of 29 experts defined the scope of an extensive literature search in order to give recommendations for the treatment of RA.4 A subgroup of this steering committee (see author list) focused on strategy trials.
In collaboration with a librarian the project fellow (RK) searched Pubmed, Ovid Embase and Cochrane from January 1962 until February 2009 and in European League Against Rheumatism (EULAR) and American College of Rheumatism (ACR) abstracts of 2007–8. The search was performed using an extended combination of search strings (see online supplementary table 1). Inclusion criteria were RA, adults and medical treatment. Publications without an abstract or in a language other than English were excluded. Trials were included when they met the following definition of strategy:
A strategic trial is a clinical trial of any treatment of RA in which at least one arm consists of medication adjustment according to protocol, based on clinical outcomes aiming at a specific target.
Relevant articles were selected in a three-step approach. First, titles and abstracts of identified references were screened to exclude articles that did not deal with the topic of interest. Second, the full paper of selected articles was reviewed. Lastly, references of the finally selected articles and relevant reviews were hand searched for additional relevant publications. EULAR and ACR abstracts were searched by hand and with a simplified search string.
Data extraction
A standard form was used to extract study design and outcome data on three domains: clinical efficacy, functional and structural outcome (online supplementary table 2). Since the published data were heterogeneous it was not possible to make comparisons with summarising statistics. However, the studies most often compared an intensive strategy with a conventional strategy, which could be used as the main comparison between the trials. To make this comparison, data were extracted at approximately 2 years of follow-up.
Results
The search retrieved 6464 papers and abstracts for further evaluation (online supplementary figure 1), resulting in a selection of 17 comparative strategy trials from 23 published articles and three abstracts.5,–,30 Extension of the search to the inclusion of non-comparative trials resulted in three additional papers.31,–,33
Close study of the literature disclosed two different types of strategy studies: (1) steering studies—same treatment but different target (eg, Disease Activity Score (DAS)-steered versus clinical judgment) or different follow-up (eg, routine outpatient vs intensive follow-up); (2) medication studies—same target (eg, low disease activity) and strategy (eg, adjustment every 3 months) with different treatments (online supplementary figure 2). Since these differences provide answers to different questions, the retrieved literature is analysed using this dichotomy.
Steering strategies
Thirteen trials studied the effect of steering methods.5,–,15 ,31,–,33 Seven of these were randomised controlled trials (RCTs) comparing intensive with conventional steering.5,–,10 ,14 ,15 Three non-randomised controlled trials were included.11,–,13 Finally, three non-comparative studies investigating the effect of intensive steering in daily practice were included.31,–,33 Detailed information is provided in table 1 (baseline characteristics and treatment target) and table 3 (outcome) and online supplementary table 3 (design).
Three of the five RCTs, TICORA (TIght COntrol of Rheumatoid Arthritis)5 CAMERA (Computer Assisted Management for Early Rheumatoid Arthritis)6 and Fransen et al,7 showed significantly better results for the intensively treated arm than for the conventionally treated arm on clinical outcome in patients with early RA. TICORA also observed a difference in progression of structural damage. Such a difference was not found by CAMERA while Fransen et al did not assess progression of structural damage.
The other two RCTs found no difference: BROSG (British Rheumatoid Outcome Study Group),8 ,9 which in contrast to the RCTs mentioned above, examined patients with established RA; van Hulst et al10 found that since the doctors were not obliged to adjust treatment in their study, the frequency of treatment adjustments was similar in both arms, resulting in no difference in clinical outcome.
All non-randomised studies used historical data as a comparison. Two of these studies concluded in favour of the intensive treatment arms. Allaart et al11 observed better clinical outcomes for the DAS44-steered patients than for conventionally treated patients. Remarkably, the latter had less severe joint damage. The authors attributed this effect to the differences in DAS44 at baseline. Stenger et al12 observed significantly lower area under the curve values of C-reactive protein and progression of joint damage in the high-risk intensive treatment group versus the high-risk historic control group. This difference was not seen in the low-risk group. Finally, van der Woude et al13 did not find a difference in remission rate between BeSt-patients and Emotional Approach Coping Scale-patients. However, the studies by Allaart et al and van de Woude et al were not randomised.
RCTs analysing the best target to use did not come to definitive conclusions.14 ,15 The three non-comparative trials observed significant benefits on all three outcome measures (structure, function, signs and symptoms), concluding that strategy was effective in daily practice.31,–,33
Medication strategies
Seven comparative trials investigated patients with different medication sequences but with similar treatment goals and adjustment criteria.16 ,21 ,22 ,27,–,30 Six trials were RCTs16 ,21 ,22 ,27,–,29 and one had a prospective cohort design.30 Detailed information is provided in table 2 (baseline characteristics and treatment target) and table 3 (outcome) and supplementary table 3 (design).
Finnish Rheumatoid Arthritis Combination Therapy (FIN-RACo)16,–,20 and BeSt22,–,26 found significantly more rapid improvement in clinical and functional outcome in the combination groups than in the initial monotherapy groups. Although, the difference in clinical response decreased over time, a difference in structural damage favouring combination therapy was still seen after 5 years in BeSt and FIN-RACo. The difference in structural damage developed during the first period of the study with a relatively constant difference over time. Adding infliximab in the first 6 months to FIN-RACo combination therapy, carried out by Neo Rheumatoid Arthritis Combination Therapy (Neo-RACo)21 resulted in higher remission rates and more patients without radiological damage than adding placebo.
GUEPARD (GUérir la PolyArthrite Rhumatoide Débutante (cure early RA))27 found the same faster response of DAS for the initial combination group of methotrexate (MTX) plus adalimumab versus MTX only, resulting in a higher DAS area under the curve for the latter. However, unique in this study was the quick addition of adalimumab to MTX after 3 months if the target was not reached, which apparently led to similar functional and radiological outcomes in both groups.
The small study by Saunders et al28 found neither a difference in clinical outcome nor in physical functional or structural damage in the comparison of step-up versus initial triple therapy. Ferraccioli et al29 observed better clinical response in the two intensively treated arms than in the arm treated with monotherapy. The difference in clinical outcome between the two intensively treated arms was inconclusive. Finally, Verschueren et al30 observed quicker clinical response in the intensively treated group, although they had higher disease activity at baseline.
Discussion
The conclusions of the steering and medication trials are remarkably consistent: Five trials showed significant benefits for intensive steering on clinical outcome, physical function and structural damage.5,–,8 ,12 ,13 The trials that found no differences studied patients with established RA, and one of the trials did not demand treatment adjustment.8 ,10 These results lead to the conclusion that patients with early active RA benefit from active steering. Proof in late disease is lacking. Additionally, protocol-based adjustment appears to be required. Regardless of the steering method used, active steering resulted in better outcomes, suggesting that the method is less important than the active steering itself. Almost all successful studies used a composite index as steering method. Based on these data in early RA, a prompt start and active steering of medication is advisable.
The medication sequence used, the allowance of prednisone use and the lack of appropriate comparisons across strategies make it impossible to prioritise a particular sequence. The results of FIN-RACo16 and BeSt22 favoured initial combination therapy. However, the combination treatment of BeSt consisted of prednisone or antitumour necrosis factor (anti-TNF), which does not answer the question as to whether a combination of conventional disease-modifying antirheumatic drugs (DMARDs) is better than initial monotherapy. FIN-RACo compared sequential monotherapy with combination therapy. Saunders et al28 compared step-up therapy with combination therapy and found no difference between the treatment arms. Although this could be caused by a difference in sulfasalazine dose, it suggests that rather than starting with combination therapy, it was the prompt adjustment of treatment for non-optimal response that made the difference. This is underlined by the conclusion of GUEPARD, that initial anti-TNF does not give better outcomes than prompt adjustment to anti-TNF when MTX treatment fails after 3 months.
Conclusion
Patients with early active disease may benefit substantially from an immediate start of treatment, active steering and prompt adjustment of treatment intensity. Proof in favour of any steering method is lacking and the best medication sequence is still not known. The benefits of intensive treatment of patients with established RA should be further explored.
Acknowledgments
J W Schoones, Walaeus Library, Leiden University Medical Centre, The Netherlands, participated in the elaboration of the systematic search strategy. All participants of the EULAR Task Force, who guided the search and reviewed the conclusions.
References
Supplementary materials
Web Only Data ard.2009.126748
Files in this Data Supplement:
Footnotes
RK and MS contributed equally.
-
Competing interests None. Francis Berenbaum was the Handling Editor.
-
Provenance and peer review Not commissioned; externally peer reviewed.