Controlled-release (CR) formulations of oxycodone and morphine were compared in 45 patients with chronic cancer pain. The study was started with an open-label, randomised titration phase to achieve stable pain control for at least 48 h, followed by a double-blind, randomised, crossover phase in two periods, 3-6 days each. To blind the study using available tablet strengths, the dose ratio of oxycodone to morphine was set at 2:3. A daily telephone contact was maintained between the patient and the investigator. The patients were asked to assess pain intensity four times a day and acceptability of therapy twice daily, and to record possible adverse effects. Pharmacodynamic evaluations were performed at the end of each double-blind period. The patients were allowed to use escape analgesic (respective opioid as oral solution) as needed. Twenty-seven patients were evaluable for both safety and efficacy. Pain was well-controlled during both stable phases. When the period effect was taken into account the two opioids provided comparable analgesia. If the results of the two periods were combined, the patients consumed significantly more escape doses and the mean pain intensities were significantly greater with CR oxycodone. The total opioid consumption ratio of oxycodone to morphine was 2:3 when oxycodone was administered first, and 3:4 when oxycodone was administered after morphine. The total incidence of adverse experiences reported by the patients was similar, but significantly more vomiting occurred with morphine, whereas constipation was more common with oxycodone.