Study designs of evaluations included in the review
Only randomised controlled trials (RCTs) were eligible for inclusion. The included study began with an open phase during which all patients received etanercept. Patients who responded positively to the treatment (based on pre-specified criteria) entered the double-blind phase of the trial; they were randomised to continue with etanercept or change to placebo. This was followed by an open-label phase during which etanercept was received.
Specific interventions included in the review
Comparisons of the anti-tumour necrosis factor (TNF) agent etanercept with any other agent, including placebo, were eligible for inclusion. The included study was a placebo comparison. The participants received 0.4 mg/kg etanercept, up to a maximum of 25 mg, by subcutaneous injection twice per week.
Participants included in the review
Studies with participants 18 years or under experiencing JIA (known formally as juvenile rheumatoid arthritis in the USA) were eligible for inclusion. The included trial had participants aged 4 to 17 years (mean: 10.5); 26 were male and 43 were female. They were unresponsive to treatment with non-steroidal anti-inflammatory drugs and to methotrexate at doses of at least 10 mg/m2 per week.
Outcomes assessed in the review
Only studies reporting clinical outcomes were eligible for inclusion. The primary outcome measure in the included study was disease flare in the 4-month period after entry into the double-blind phase of the trial. Disease flare was determined on the basis of 6 variables: the number of active joints; the number of joints with limited motion plus pain or tenderness or both; physician's global assessment of disease severity; patient or parent's global assessment of disease severity; Child Health Assessment Questionnaire functional ability score; and erythrocyte sedimentation rate score. Flare was defined as the patients being worse by at least 30% in 3 of the 6 measures and having a minimum of two active joints, although at least 30% improvement was allowed on one variable. Improvement at the end of the study was defined as at least 30% improvement in at least 3 of these outcome variables and a deterioration of more than 30% in no more than one outcome variable. Fifty per cent and 70% improvements were also measured.
How were decisions on the relevance of primary studies made?
Two reviewers independently assessed the identified articles and consensus was reached. Non-English language publications were screened by one reviewer.