In a prospective randomized and double-blind cross-over study, a new antiemetic regimen consisting of betamethasone (1 x 8 mg) and dixyrazine (a phenothiazine derivative) (4 x 10 mg) was compared with a standard high-dose metoclopramide (4 x 1 mg/kg) schedule for antiemetic treatment in doxorubicin and cisplatin chemotherapy. 100 consecutive patients (62 without prior experience of chemotherapy and 38 with prior experience) entered the study and were followed during 1-4 courses of chemotherapy. Effect and side effect parameters were recorded on questionnaires for patients and nurses using the visual analog scale for quantification. The correlation between the two ways of recording (self-scoring versus recording by nursing staff) was very high, both for effect variables (nausea and vomiting) and the adverse reactions (sedation and extrapyramidal reactions). The median number of courses per patient was 3.0 (range 1-4) and altogether 299 courses were studied. Full emetic protection was achieved in 58% with betamethasone-dixyrazine and in 34% with high-dose metoclopramide regardless of prior patient experience or the cytostatic agents administered. With doxorubicin regimens, betamethasone-dixyrazine gave full protection in 80% compared to 40% for metoclopramide. Cisplatin regimens were a greater challenge and protection against nausea and vomiting was achieved only in 27% with betamethasone-dixyrazine and in 18% with metoclopramide. Adverse reactions were a significant problem with metoclopramide: restlessness 33%, akathisia 19%, parkinsonism 16%, and acute dystonia 3%. Sedation was the same with the two regimens (80%).