CONTEXT: Dexamethasone is widely used to prevent postoperative nausea and vomiting (PONV) in pediatric tonsillectomy. OBJECTIVE: To assess whether dexamethasone dose-dependently reduces the risk of PONV at 24 hours after tonsillectomy. DESIGN, SETTING, AND PATIENTS: Randomized placebo-controlled trial conducted among 215 children undergoing elective tonsillectomy at a major public teaching hospital in Switzerland from February 2005 to December 2007. INTERVENTIONS: Children were randomly assigned to receive dexamethasone (0.05, 0.15, or 0.5 mg/kg) or placebo intravenously after induction of anesthesia. Acetaminophen-codeine and ibuprofen were given as postoperative analgesia. Follow-up continued until the 10th postoperative day. MAIN OUTCOME MEASURES: The primary end point was prevention of PONV at 24 hours; secondary end points were decrease in the need for ibuprofen at 24 hours and evaluation of adverse effects. RESULTS: At 24 hours, 24 of 54 participants who received placebo (44%; 95% confidence interval [CI], 31%-59%) had experienced PONV compared with 20 of 53 (38%; 95% CI, 25%-52%), 13 of 54 (24%; 95% CI, 13%-38%), and 6 of 52 (12%; 95% CI, 4%-23%) who received dexamethasone at 0.05, 0.15, and 0.5 mg/kg, respectively (P<.001 for linear trend). Children who received dexamethasone received significantly less ibuprofen. There were 26 postoperative bleeding episodes in 22 children. Two of 53 (4%; 95% CI, 0.5%-13%) children who received placebo had bleeding compared with 6 of 53 (11%; 95% CI, 4%-23%), 2 of 51 (4%; 95% CI, 0.5%-13%), and 12 of 50 (24%; 95% CI, 13%-38%) who received dexamethasone at 0.05, 0.15, and 0.5 mg/kg, respectively (P =.003). Dexamethasone, 0.5 mg/kg, was associated with the highest bleeding risk (adjusted relative risk, 6.80; 95% CI, 1.77-16.5). Eight children had to undergo emergency reoperation because of bleeding, all of whom had received dexamethasone. The trial was stopped early for safety reasons. CONCLUSION: In this study of children undergoing tonsillectomy, dexamethasone decreased the risk of PONV dose dependently but was associated with an increased risk of postoperative bleeding. TRIAL REGISTRATION: clinicaltrials.gov Identifier: NCT00403806.