This study investigated the analgesic effects of three intravenous bolus doses of hydromorphone (10, 20, 40 micrograms/kg) on experimental pain measures in normal humans. Ten healthy male volunteers participated in four study sessions, one for each of the hydromorphone doses as well as a placebo (saline). They received the four treatments in counterbalanced order under double-blind conditions and with study days at least 1 week apart. During each session subjects underwent repeated electrical tooth pulp stimulation at intensities sufficient to elicit a rating of 'strong pain' before drug administration. Subjective pain reports (PRs) and dental evoked potential amplitude measures (EPs) served as analgesic effect indicators. We observed dose-dependent analgesia as measured by both PR (P = 0.009) and EP (P = 0.017). Area under the PR versus time curve as well as the EP versus time curve decreased in a log dose-dependent fashion. Although the peak effect was poorly defined, the onset of analgesia was rapid, within 5 min, and maximum analgesic effect was seen between 10 and 20 min after maximum plasma hydromorphone concentration. However, within sessions we found a poor correspondence between hydromorphone plasma concentration and effect. Compared to pain report data from other human studies done in our laboratory, hydromorphone has a shorter time to peak effect compared to morphine, and overall, hydromorphone hydrochloride is approximately five times as potent as morphine sulfate on a milligram basis.