Experimental studies have indicated that N-methyl-D-aspartate (NMDA) receptor antagonists may be effective analgesics in a wide variety of chronic pain states. The mechanism is presumed to be related to decreased firing of dorsal horn neurons after constant repeated C-fiber stimulation. Dextromethorphan (DM), a potent NMDA antagonist with a good safety profile, may be a promising agent for the treatment of persistent pain. An open-label randomized trial was designed to examine the effects of combining DM with NSAIDs, dextropropoxyphene, or morphine in cancer patients with pain. Patients who required a change in the step of the World Health Organization's (WHO) analgesic ladder because of a pain level of 4 or more on a numerical pain scale were randomly allocated to receive DM 30 mg three times a day (30 patients) or conventional treatment (30 patients). There were 20 patients randomized for each step of the analgesic ladder. Pain mechanisms, pain intensity (numerical 0-10 scale), symptoms more frequently present in advanced cancer patients or associated with opioid therapy (graded on a 0-3 scale--not at all, slight, a lot, awful), opioid escalation index, days on opioid treatment, and adverse effects were recorded. After 2 days 75%, 80%, and 100% of patients treated with DM in steps 1, 2, and 3, respectively, required conventional treatment, because adequate pain relief had not been achieved (pain scale > 4). Failure of this treatment was equally observed in neuropathic pain or nociceptive pain syndromes. Four patients treated with DM who did not require the conventional treatment immediately did require this change after some days, due to poor pain control. A highly significant reduction in pain was observed in patients directly treated with the conventional treatment in all the three steps of the analgesic ladder. No significant analgesic effects could be found when DM at this dose was combined with NSAIDs, dextropropoxyphene, or morphine.