Cancer pain treatment following the World Health Organization guidelines is effective and feasible. However, the evidence supporting the use of opioids for mild to moderate pain on the second step of the analgesic ladder is widely discussed. The present evaluation compares the efficacy and safety of high doses of oral tramadol (> or = 300 mg/d) with low doses of oral morphine (< or = 60 mg/d). Patients were included in this nonblinded and nonrandomized study if the combination of a nonopioid analgesic and up to 250 mg/d of oral tramadol was inadequate. 810 patients received oral tramadol for a total of 23,497 days, and 848 patients received oral morphine for a total of 24,695 days. The average dose of tramadol was 428 +/- 101 mg/d (range 300-600 mg/d); the average dose of morphine was 42 +/- 13 mg/d (range 10-60 mg/d). Additional nonopioid analgesics were given on more than 95% of days. Antiemetics, laxatives, neuroleptics, and steroids were prescribed significantly more frequently in the morphine group; the use of other adjuvants was similar in both groups. The mean pain intensity on a 0-100 numerical rating scale (NRS) was 27 +/- 21 (95% CI 26-29) in the tramadol and 26 +/- 20 (95% CI 24-27) in the morphine group (NS). The analgesic efficacy was good in 74% and 78%, satisfactory in 10% and 7%, and inadequate in 16% and 15% of patients receiving tramadol and morphine, respectively (NS). Constipation, neuropsychological symptoms, and pruritus were observed significantly more frequently with low-dose morphine; other symptoms had similar frequencies in both groups. These data suggest that tramadol can be used for the treatment of cancer pain, when nonopioids alone are not effective. High doses of tramadol are effective and safe.