Sublingual buprenorphine, a long-acting, partial mu-opioid agonist, is as effective as methadone in the treatment of heroin dependence, with a better safety profile due to its antagonist activity. However, the safety of therapeutic doses (8 to 16 mg) that might be diverted for intravenous (i.v.) use has not been demonstrated. To evaluate the safety and possible ceiling effects of buprenorphine administered i.v. to experienced opioid users, buprenorphine was administered to 6 nondependent opioid abusers residing on a research unit; the doses tested, in separate sessions, were 12 mg buprenorphine sublingual, i.v./sublingual placebo, and escalating i.v. buprenorphine (2, 4, 8, 12, and 16 mg). Physiologic and subjective measures were collected for 72 hours post-drug administration. Buprenorphine minimally but significantly increased systolic blood pressure. Changes in heart rate or oxygen saturation among the 7 drug conditions were not statistically significant. The mean maximum decrease in oxygen saturation from baseline was greatest for the 8-mg i.v. dose. Buprenorphine produced positive mood effects, although with substantial variability among participants. Onset and peak effects occurred earlier following i.v. administration: peak i.v. effects occurred between 0.25 and 3 hours; peak sublingual effects occurred at 3 to 7 hours. Duration of effects varied among the outcome measures. The dose-response curves were flat for most parameters, particularly subjective measures. Side effects were mild except in one participant who experienced severe nausea and vomiting after the 12-mg i.v. dose. Buprenorphine appears to have a ceiling for cardiorespiratory and subjective effects and a high safety margin even when taken by the i.v. route.