Remifentanil is a short-acting, esterase-metabolized opioid analgesic. This study compared the abuse liability of remifentanil with that of fentanyl and placebo in a randomized, double-blind, crossover study. Twelve recreational users of opioids received increasing doses of remifentanil (0.6, 1.2, 1.8, 2.4, and 3.0 microg/kg), fentanyl (0.4, 0.8, 1.3, 2.0, 3.0, and 4.5 microg/kg) and placebo via an intravenous infusion pump. Subjective measures (Cole/Addiction Research Center Inventory [ARCI] scales and visual analog scale [VAS] items such as "High" and "Good Effects") and physiologic variables (blood pressure, O2 saturation, pupil diameter) were recorded. For each measure, the differences from baseline were reduced to an area under the response curve (AUC) and a peak, and each subject's response to the maximum tolerable dose for each of the two active drug classes and mean response to several placebo infusions were entered into a 12 x 3 analysis of variance. All differences in drug versus placebo effects were significant. Although a majority of the peak effects that were measurable within 4 minutes after drug infusion reflected greater remifentanil effects, only one, High VAS, was statistically significant. In contrast, observations that could only be made > or = 5 minutes after drug infusion predominantly indicated significantly greater fentanyl peak effects, including High VAS, Liking VAS, Good Effects VAS, and Cole/ARCI Abuse Potential. Fentanyl AUCs were generally significantly larger than the corresponding remifentanil AUCs. A drug abuser seeking longer-lasting drug effects might select fentanyl over remifentanil, but these data do not completely rule out remifentanil abuse by some individuals with access to both the drug and the infusion equipment or by those who prefer briefer, repeated effects.