The authors concluded that tight control treatment regimes significantly improved clinical outcomes compared to usual care in patients with early rheumatoid arthritis. Findings suggested that using a protocol for treatment adjustment may be beneficial. These conclusions were supported by the evidence, but a limited search and incomplete reporting of review methods made it difficult to judge their reliability.

To compare tight control strategies with or without protocolised treatment adjustments versus usual care in patients with rheumatoid arthritis.

PubMed and The Cochrane Library were searched from 1995 to August 2009 for studies reported in English. Search terms were reported. Reference lists of all selected articles were screened. Articles were tracked using the related articles function in PubMed.

Randomised controlled trials (RCTs) and clinical trials were eligible if they compared tight control strategies (defined as the frequent assessment of inflammation to guide treatment changes and keep inflammation at an agreed low level) with usual care (defined as decisions about treatment left to physician and patients and reflecting daily clinical practice) in patients with rheumatoid arthritis diagnosed according to American College of Rheumatology 1987 revised criteria. Prescribed treatment regimens had to at least include disease-modifying antirheumatic drugs (DMARDs), anti-TNF-alpha agents or glucocorticoids. Studies had to assess the disease activity score (DAS), DAS-28, or all separate components of the DAS-28 as a primary of secondary outcome. Studies were excluded if they evaluated only safety outcomes and if they compared in-patient and day care teams.

The included studies evaluated treatment regimens that contained conventional DMARDs, methotrexate, sulfasalazine (SSZ) and hydroxychloroquine (HCQ) (the last two acronyms were not defined in the paper). Most studies did not use biological agents; one study did use these drugs in a minority of patients. All studies permitted use of prednisone. Most patients took nonsteroidal anti-inflammatory drugs.

Half of the included studies evaluated use of a protocol for treatment adjustment in the tight control arm; most of these studies were in DMARD–naïve patients with early rheumatoid arthritis. Studies that did not use a protocol for treatment adjustment were all in patients with a mix of early and established rheumatoid arthritis. Mean age of patients ranged from 51 to 60 years. Most patients were female. Between 42% and 80% of participants were rheumatoid factor positive. Mean baseline DAS-28 scores ranged from moderate (between 3.2 and 5.1) to high (>5.1). Mean duration of rheumatoid arthritis ranged from 0.4 to 11 years. Most studies were set in The Netherlands.

Data were reported for the percentage of patients in remission, with joint damage, increased physical function and toxicity.

Two reviewers independently selected studies. Disagreements were resolved by consensus.

Validity was assessed using concealment of allocation, blinding of outcome assessment and treatment of withdrawals. The authors did not state how many reviewers assessed validity.

Findings reported as DAS scores were transformed into DAS-28 scores. Mean changes and standard deviations of the DAS-28 from baseline to study completion were extracted or calculated for each treatment group and used to calculate the mean difference. Authors were contacted if required for missing data.

The authors did not state how many reviewers extracted data.

Pooled weighted mean differences (WMD) and 95% confidence intervals (CI) were calculated using a fixed-effect model in the absence of significant heterogeneity (I^2_<30%); otherwise a random-effects model was used. Heterogeneity was assessed using Q and I².

Sensitivity analysis was conducted to explore the influence of randomisation (random versus non-randomised studies). The effect of use of a protocol for treatment adjustment in the tight control arm was examined using subgroup analysis.

Six studies were included (n=1,740 participants): four RCTs (n=1,041) and two nonrandomised controlled studies (n=699). Sample size ranged from 110 to 435. Study duration ranged from 12 to 24 months. The authors stated that all studies met criteria for concealment of allocation, blinding of outcome assessment and treatment of withdrawals. Drop-out rates ranged from 4% to 39%.

At one-year follow-up, tight control regimens were associated with a statistically significant improvement in DAS-28 scores compared to usual care for all studies combined (WMD random-effects model 0.59, 95% CI 0.22 to 0.97; significant heterogeneity I²=80%) and for RCTs (WMD random-effects model 0.65, 95% CI 0.10 to 1.21; level of heterogeneity not reported).

Tight control regimens that incorporated a protocol for treatment adjustment were associated with a statistically significant improvement in DAS-28 scores compared to usual care (WMD random-effects model 0.97, 95% CI 0.64 to 1.30; three studies, significant heterogeneity I²=61%). The authors stated that heterogeneity was mostly due to one study with large efficacy and a small number of patients.

Tight control regimens that did not incorporate a protocol for treatment adjustment were associated with a small statistically significant improvement in DAS-28 scores (WMD fixed-effect model 0.25, 95% CI 0.03 to 0.46; three studies, no significant heterogeneity).

Tight control treatment regimens significantly improved clinical outcomes compared to usual care in patients with early rheumatoid arthritis. Findings suggested but did not prove that tight control regimens that used a protocol for treatment adjustment were more beneficial than regimens without a protocol.

The review question was clearly stated. Inclusion criteria were appropriately defined. Limiting the search to English-language reports listed in two databases plus references risked publication and language biases and may have resulted in the omission of studies. Relevant information was provided about the included studies. Study validity was assessed. Methods used during study selection minimised risks of reviewer errors and bias; whether similar steps were taken during data extraction and validity assessment was unclear. In the main meta-analysis, data from studies of diverse design were pooled and this may not have been appropriate. RCTs were also analysed separately and results were similar. Heterogeneity was assessed. Subgroup analyses were conducted.

The authors’ conclusions were supported by the evidence, but the limited search, incomplete reporting of review methods, evidence based on a small number of studies and some evidence of statistical heterogeneity between studies made it difficult to judge their reliability.

Practice: The authors stated that tight control treatment strategies that incorporated monitoring of disease activity and a protocol for treatment adjustment should be used for patients with rheumatoid arthritis. They also stated that the advantages of tight control shown for patients with early rheumatoid arthritis may not generalise to patients with established rheumatoid arthritis.

Research: The authors stated that there was a need for research to determine the optimal protocol for treatment adjustment for patients with rheumatoid arthritis.

One reviewer was funded by a grant from Wyeth Pharmaceuticals, The Netherlands.

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.