Morphine and oxycodone were administered to ten patients suffering from severe cancer pain in a double-blind cross-over study. The patients titrated themselves pain-free, first intravenously, using a patient-controlled analgesia device, and then orally. Each titration phase lasted for 48 hours. Blood samples were drawn after 36 hr of each administration phase. The plasma levels of morphine, morphine-6- and morphine-3 glucoronides were determined with high performance liquid chromatography (HPLC), whereas the oxycodone samples were assayed with gas chromatography (GC). Twin samples were analyzed for plasma opioid activity with a radioreceptor assay (RRA) using 3H-dihydromorphine and 3H-naloxone as radioligands. Adequate analgesia was achieved with both morphine and oxycodone. About 30% more oxycodone was needed intravenously, whereas 25% less oxycodone than morphine was consumed orally. There was a good linear correlation between the morphine concentrations measured with HPLC and RRA. The mean morphine-6-glucuronide to morphine concentration ratio was 2.3 after intravenous and 4.6 after oral administration. Results from RRA indicate that oxycodone in vivo is a potent mu-agonist and that at least part of its analgesic action is mediated by active metabolites. In vitro morphine glucuronides enhanced morphine in displacing radioligands from the opioid receptors, thus suggesting their complex interactions in vivo.