Foeldvari 2009.

Methods	Allocation: randomised Blinding: double‐blind Controlled: active comparator Centre: multicentre Arm: 2 arms, parallel groups
Participants	Inclusion criteria: children ≥ 9 kg, with pauciarticular of polyarticular course JRA, with or without systemic onset, according to ACR criteria; > 1 swollen joint with limited motion; parent global assessment ≥ 10 mm (100‐millimetre VAS) Exclusion criteria: active systemic manifestations; oral corticosteroid doses ≤ 0.2 mg/kg/day or 10 mg prednisone or methotrexate < 1 mg/kg/week Baseline characteristics N = 242 Age: 2 to 16 years Gender: male (71); female (171) Number randomised: intervention A (77); intervention B (82); control (83) Number completed: intervention A (67); intervention B (71); control (74) Setting and location: 17 centres worldwide
Interventions	Intervention group (N = 77): celecoxib 50 mg/5 mL oral suspension (target dose approximately 3 mg/kg twice daily) Intervention group (N = 82): celecoxib 100 mg/5 mL oral suspension (target dose approximately 6 mg/kg twice daily) Control group (N = 83): naproxen 125 mg/5 mL oral suspension (target dose approximately 7.5 mg/kg twice daily) Study duration: 12 weeks
Outcomes	Primary outcomes Time‐weighted average proportion of patients achieving ACR Pediatric 30 (at least 30% improvement in any 3 of 6 variables) Investigators' global assessment of disease activity (100‐millimetre VAS) Parent/patient's global assessment of overall well‐being (100‐millimetre VAS) Measure of physical functional ability (CHAQ: 0‐to‐3‐point scale) Number of joints with active arthritis Number of joints with limited range of motion Measure of inflammation (ESR) Secondary outcomes Change from baseline at each visit for the individual Juvenile Rheumatoid Arthritis score set measures Parent's assessment of child's arthritis pain (100‐millimetre VAS) as reported on the CHAQ Health‐related quality of life (Pediatric Quality of Life Inventory)
Notes	Sources of funding: editorial support funded by Pfizer
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "children were randomly assigned to 1 of 3 treatment groups in a 1:1:1 ratio ... randomized according to the allocation number provided by an interactive voice response system"
Allocation concealment (selection bias)	Unclear risk	Comment: Insufficient information
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: Insufficient information
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: Insufficient information
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: All participants were accounted for. Lost to follow‐up and withdrawals explained. However, authors do not report whether there were significant differences between completers and non‐completers.
Selective reporting (reporting bias)	High risk	Comment: Secondary outcome data not reported (e.g. Pediatric Quality of Life Inventory)
Size	Unclear risk	Comment: Total participants = 242 (between 50 and 200 per treatment arm)
Other bias	Low risk	Comment: No other potential sources of bias found.