Abstract Methylnaltrexone, a peripherally-acting quaternary opioid antagonist, is an investigational treatment for opioid-induced constipation in patients with advanced illness. This randomized, parallel-group, repeated dose, dose-ranging trial included a double-blind phase for one week followed by an open-label phase for a maximum of three weeks. Opioid-treated patients with advanced illness who met criteria for opioid-induced constipation despite laxative therapy were potentially eligible. Double-blind treatment occurred on Days 1, 3, and 5; open-label therapy could be administered as often as every other day. The initial dose range of 1 mg, 5 mg, or 12.5 mg was extended by adding a 20 mg group during the study while still maintaining the double blind; the initial open-label dose of 5 mg could be titrated. The primary outcome was a laxation response within four hours after the first dose. Thirty-three patients received at least one dose of methylnaltrexone. Only one of 10 patients (10%) who received the 1 mg dose experienced laxation within four hours of dosing. The median time to laxation was >48 hours for the 1 mg dose group, compared to 1.26 hours for all patients receiving ≥5 mg ( P = 0.0003). There was no apparent dose–response above 5 mg. Most adverse events were related to the gastrointestinal system, were mild, and did not lead to discontinuation. In conclusion, methylnaltrexone relieved opioid-induced constipation at doses ≥5 mg in patients with advanced illness, and did not reduce analgesia or cause opioid withdrawal symptoms.