Background Juvenile idiopathic arthritis (JIA) is the more common chronic arthritis of childhood, and can lead to significant long-term morbidity, including physical disability. Treatment of the disease has greatly improved short and medium-term outcomes, and its effectiveness has been often tested with trials using placebo as comparator. However, until now the real magnitude of placebo effect has not been quantified in children with JIA.
Objectives To estimate the placebo effect in JIA through a meta-analysis on phase III trials which compares the active compound versus placebo using different study designs (parallel or withdrawal).
Methods The systematic literature search was carried out from June to December 2012 using publications retrieved in Medline, Cochrane Controlled Trials, ClinicalTrials.gov registers. The study was developed according to the PRISMA statement. For parallel design studies the outcome was evaluated as a single one-dimensional (OD) variable or, when available, as a composite score (CS); withdrawal studies were evaluated only by CS.
Results A total of 23 trials were included in the final meta-analysis. In the seven OD studies, 38% (95% CI: 32-43%) of patients responded to placebo. Considering CS parallel trials, the rate of response to placebo was high in studies enrolling different JIA categories (38%; 95% CI: 33-43%) and lower in studies with only systemic included (13%; 95% CI: 6-20%). In withdrawal design trials, the percentages of placebo patients who flared during the double blind phase in the placebo arm was lower in the studies recruiting various JIA subgroups (62%; 95% CI: 50-74%) than trials with only a population of systemic subjects (67%; 95% CI: 37-97%) with substantial evidence of heterogeneity.
Conclusions A sizable number of patients seems to benefit from placebo treatment. The placebo effect is mainly influenced by the study design and the rate of systemic patients enrolled.
References Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009 July 21;6(7):e1000097.
Disclosure of Interest None Declared