The authors concluded that there was insufficient data available to evaluate the comparative effectiveness of DMARDs. However, there was evidence to suggest that DMARDs improved symptoms associated with juvenile idiopathic arthritis. Further trials were needed. Based on such diverse and limited evidence, the authors' conclusions seem appropriate, but potential for language and reviewer bias should be considered.

To compare the efficacy and safety of disease-modifying antirheumatic drugs with each other or versus conventional treatment (with or without methotrexate), in children with juvenile idiopathic arthritis.

MEDLINE (1966 to December 2010), EMBASE (1947 to December 2010) and Cochrane Database of Systematic Reviews were searched for relevant articles in English. Search terms were reported. Grey literature and conference abstracts from the 2008 and 2009 meetings of the American College of Rheumatology (ACR) and the Paediatric Academic Societies were also searched. Reference lists of review articles and included studies were manually searched for additional studies.

Eligible randomised controlled trials (RCTs), non-RCTs and cohort studies compared the efficacy of disease-modifying antirheumatic drugs (DMARDs; as specified in the review) with each other or versus conventional treatment, with or without methotrexate. Intervention duration had to be least three months. Study populations had to include 80% or more children aged 18 years or younger with juvenile rheumatoid arthritis (according to the current ACR definitions), juvenile idiopathic arthritis (International League of Associations for Rheumatology definitions) or juvenile chronic arthritis (European League Against Rheumatism criteria). Eligible conventional treatments were non-steroidal anti-inflammatory drugs (NSAIDs) or intra-articular corticosteroids. Efficacy outcomes included intermediate laboratory measures and long-term clinical outcomes, such as pain control, clinical remission, quality of life, growth, development, joint function, functional ability and mortality. Case series and case reports were also eligible to assess adverse events of the stated treatments in the specified population.

Included efficacy studies were conducted worldwide between 1979 and 2006 (where this was reported). Comparator studies compared six non-biologic DMARDs (azathioprine, leflunomide, methotrexate, hydroxychloroquine, penicillamine and sulfasalazine) and seven biologic DMARDs (abatacept, anakinra, tocilizumab, intravenous immunoglobin, adalimumab, etanercept and infliximab) compared with each other or conventional treatment with or without methotrexate. Patients received other arthritis drugs, including corticosteroids, NSAIDs, analgesics, opioids, slow-acting antirheumatic drugs and other DMARDs. There were significant differences in outcome measures and how these were reported.

The authors did not state how many reviewers screened studies for inclusion.

The strength of the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. The quality of RCTs, non-RCTs and comparative cohorts was assessed using published criteria, including items on randomisation and methods of allocation, drop-out rates, and use of appropriate outcome measures. Studies were rated as good quality (the least bias and results considered valid); fair quality (susceptible to some bias, but probably not sufficient to invalidate the results); or poor quality (significant bias that may invalidate the results).

One reviewer performed the quality assessment and a second reviewer checked the data. Discrepancies were resolved by consensus.

One reviewer extracted the efficacy outcome data to calculate risk ratios (RRs) and 95% confidence intervals (CIs), where possible. A second reviewer checked the data for accuracy. One reviewer extracted the rates of adverse events.

Where appropriate, risk ratios and confidence intervals were pooled using a fixed-effect model, or random-effects model where there was evidence of statistical heterogeneity. Where meta-analysis was not appropriate, data were presented as a narrative synthesis. Data were presented separately for biologic and non-biologic disease-modifying antirheumatic drugs (DMARDs). Statistical heterogeneity and inconsistency were assessed using the Cochran Q test and Ι² statistic.

DMARDs versus conventional treatment (18 studies, 20 publications; 1,532 patients): four RCTs were of good quality, four fair quality and two poor quality. Two open-label comparison studies were of poor quality, three randomised discontinuation studies were of good quality, two fair quality and one poor quality. Study follow-up ranged from three months to five years.

With the exception of infliximab (one fair quality RCT) the remaining biologic DMARDs showed inconsistent improvements in laboratory measures of inflammation and health status. Meta-analysis of four randomised discontinuation trials showed that patients who initially responded to a treatment and continued with DMARD were at greater risk of flare compared with conventional treatment (RR 0.48, 95% CI 0.36 to 0.63; Ι²=6%).

There was some evidence of improvement in laboratory measures of inflammation and health status with methotrexate (two good quality RCTs, one poor quality non-RCT) and sulfasalazine (one good quality RCT) compared with conventional therapy. Other non-biologic DMARDs generally showed non-statistically significant results compared with conventional treatment.

DMARDs versus DMARDs (five studies, six reports; 520 patients): two good quality RCTs, two poor quality RCTs and one poor quality open-label, non-RCT. Study follow-up ranged from six to 12 months.

There were no statistically significant differences in outcome measures between two biologic DMARDs; etanercept versus infliximab (one poor quality non-RCT).

Comparison of non-biologic DMARDs showed similar results between penicillamine versus hydroxychloroquine (one good quality RCT, one poor quality open-label RCT), sulfasalazine versus hydroxychloroquine (one poor quality RCT) and leflunomide versus methotrexate (one good quality RCT).

Adverse Events: Studies that reported adverse event data had methodological issues, but did not appear to show differences in rates of events between different DMARDs or between DMARDs and conventional treatment or placebo.

There was insufficient evidence to evaluate the efficacy, effectiveness, safety or adverse effects of treatment with DMARDS by type of juvenile idiopathic arthritis.

There was insufficient data available to evaluate the comparative effectiveness of DMARDs. However, there was some evidence to suggest that DMARDs improved symptoms associated with juvenile idiopathic arthritis. Further trials were needed to evaluate the effectiveness of DMARDS.

The review questions were clear and were supported by clearly defined inclusion criteria. Several relevant sources were searched for data, with attempts to locate unpublished data. The search was limited to papers in English, which meant language bias may have been introduced. Study quality and data extraction were performed in duplicate, but it was unclear whether this was true for study selection so reviewer error and bias could not be ruled out. Appropriate methods were used to assess comparative study quality, but most studies appeared to have some quality issues.

Studies varied considerably in terms of country of origin and study period. The authors acknowledged the clinical and methodological heterogeneity in studies, and the challenges in summarising such diverse data. The methods of syntheses seemed appropriate. Evidence for the different treatment comparisons was generally from one or two studies, and the sample sizes and study quality were generally limited. Based on such diverse and limited evidence, the authors' conclusions seem appropriate, but potential for language and reviewer bias should be considered.

Practice: The authors stated that the study populations were reflective of patients seen in clinical practice, and the intensity of biologic treatments was consistent with current practice. However, non-biologic DMARDs (with the exception of methotrexate) and some outcome measures used in the studies were generally used less often in practice. The authors stated that the timing of follow-up was appropriate to reflect clinically important benefits and harms.

Research: The authors stated that future RCTs and observational studies were needed to evaluate the effectiveness of DMARDs compared with each other and to conventional therapy, particularly long-term trials in populations representative of typical patients with juvenile idiopathic arthritis. Trials should standardise outcome measures and report definitions of adverse events. Patient-level meta-analyses may also be useful to address these issues. Trials investigating the effects of DMARDs on the specific health conditions associated with juvenile idiopathic arthritis should be assessed.

Agency for Healthcare Research and Quality.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.