Results of a randomized trial on antiemesis for cisplatin (CDDP) and non-CDDP chemotherapy-induced vomiting are reported. One hundred and sixty-three outpatients received 282 chemotherapy courses (141 with CDDP and 141 without CDDP). Patients were randomly assigned to receive either high-dose metoclopramide plus methylprednisolone (arm A) or the same drugs plus lorazepam (arm B). In both arms a high protection rate for vomiting was obtained, on the whole without statistically significant differences. Patients who received lorazepam had, however, significantly fewer nausea episodes during first day post-chemotherapy (p less than 0.05). Arm B was also superior in anxiety control during the first day of chemotherapy (p less than 0.01). Both regimens were significantly more effective in patients who had not been given chemotherapy previously (p less than 0.01). No differences in antiemetic protection were found between CDDP and non-CDDP courses. No significant differences were found in premonitory vomiting control between the two arms of the trial. Toxicity was very mild with both regimens, although sedation was significantly higher in arm B (p less than 0.001). We conclude that high-dose metoclopramide plus methylprednisolone is a highly effective combination for chemotherapy-induced nausea and vomiting, and that it is quite suitable for outpatient use. Lorazepam did not significantly increase the antiemetic potency of the combination, nor did it improve premonitory vomiting control, although it gave a better control of acute nausea and anxiety.