Despite treatment, emesis remains a major problem with cisplatin (CDDP) chemotherapy. Reasons for variability in antiemetic response among patients and in subsequent cycles are largely unknown and toxicity is sometimes severe. We have, therefore, carried out a multicenter, double-blind randomized trial comparing a combination of high-dose metoclopramide (MTC) (1 mg/kg x 4) and methylprednisolone (P) (treatment A) with a shorter but higher single-dose schedule of metoclopramide (3 mg/kg x 2) combined with dexamethasone (DEX) and diphenhydramine (DIP) to prevent extrapyramidal reactions (treatment B). Three hundred sixty-seven consecutive patients treated with various chemotherapy combinations containing CDDP were studied. Complete protection from vomiting/nausea was, at first cycle, 72.5%/79.5% with treatment B and 55.8%/65.1% with treatment A, a statistically significant difference (P less than .002/P less than .005). In subsequent cycles, protection from emesis significantly decreased with no difference between the two treatments. Multifactorial analysis shows that women, younger patients, outpatients, and patients who experienced emesis in previous cycles were at higher risk of suffering nausea and/or vomiting. Both regimens were well tolerated, but patients treated with treatment B had significantly less extrapyramidal reactions (1.7%/6.1%, P = .053). Treatment B is preferred due to its greater efficacy and lower incidence of extrapyramidal reactions. Trials on antiemetic therapy should take into account the important variables able to influence the efficacy of treatment. There is still a need for improving prevention of emesis in CDDP-treated patients.