We undertook a meta-analysis of randomized controlled trials to summarize the efficacy of anti-obesity drugs in reducing BMI and improving health in children and adolescents. Data sources included Medline, Embase, the Cochrane controlled trials register and other registers of controlled trials, together with reference lists of identified articles. All data sources were searched from January 1996 to July 2008. We searched for double blind randomized placebo controlled trials of approved anti-obesity drugs used in children and adolescents (age < 20) with primary obesity for > or = 6 months. Six trials, 4 of sibutramine (total patients = 686) and 2 of orlistat (n = 573) met inclusion criteria. No trials of rimonabant were identified. Compared with placebo, sibutramine together with behavioural support reduced BMI by 2.20 kg/m(2) (95% CI: 1.57 to 2.83) and orlistat together with behavioural support reduced BMI by 0.83 kg/m(2) (95% CI 0.47 to 1.19). Sibutramine improved waist circumference, triglycerides and high density lipoprotein (HDL)-cholesterol, but raised systolic and diastolic blood pressure and pulse. Orlistat increased rates of gastrointestinal side-effects. We conclude that sibutramine in adolescents produces clinically meaningful reductions in BMI and waist circumference of approximately 0.63 SD, with improvements in cardiometabolic risk. Orlistat modestly reduces BMI (effect size approximately 0.24 SD) with a high prevalence of gastrointestinal adverse effects.