The efficacy of diflunisal in cancer pain was evaluated and compared with dipyrone. Diflunisal was given at the dosage of 500 mg perorally twice a day, and dipyrone was given at the dosage of 500 mg perorally three times a day. Duration of each treatment was 7 days; after a 12-hour wash-out period, patients were given the other drug for another 7 days. A total of 50 patients were enrolled in the study. Pain intensity was assessed by 10-point visual analog scale (VAS). Patients who had a VAS score higher than 5 were included. A total of 47 patients were evaluable. Initial VAS score was a mean of 8.57+/-1.33. Diflunisal reduced the pain score by a mean of 4.65+/-3.10, whereas dipyrone reduced the pain score by a mean of 3.25 < or = 2.85 (p < 0.001). Patients were also analyzed in three subgroups according to the presence of nonmetastatic, metastatic, and bone metastatic diseases. In each of these subgroups, diflunisal reduced the pain score more than dipyrone; however, the difference was statistically significant only in patients who had bone metastasis. Adverse reactions were rare and acceptable with both drugs. Diflusinal is superior to dipyrone at this dosage and schedule in the treatment of moderate to severe cancer pain.