The efficacy of secobarbital sodium plus chlorpromazine (SC) in the prevention of cisplatin induced emesis was compared to the combination of metoclopramide, diphenhydramine, and dexamethasone (MDD). Twenty-three patients were entered onto protocol. Eighteen were evaluable. Good to excellent antiemetic prophylaxis was obtained in 72% with MDD versus 17% with SC (P < 0.01). Sedation and anticholinergic side effects were more common with SC. Extrapyramidal reactions were more commonly seen with MDD. Significantly more patients preferred the combination of metoclopramide, diphenhydramine, and dexa-methasone (P < 0.05). Cancer 58:959-962, 1986. HE INTRODUCTION of cisplatin into the chemother-T apeutic armamentarium gave impetus to therapeutic attempts to control the associated severe nausea and vomiting. High-dose metoclopramide was the first truly effective antiemetic regimen for cisplatin. ' This regimen has been subsequently improved2 by the addition of di-phenhydramine to prevent extrapyramidal side effects, and dexamethasone to increase the antiemetic efficacy and decrease the frequency of diarrhea. Yet metoclo-pramide is not a perfect antiemetic because of extrapy-ramidal side effect^^,^ and the high cost.5 Sevin d aL6 demonstrated the antiemetic activity of the barbiturate-phenothiazine combination of pentobar-bital sodium plus prochlorperazine with dexamethasone in patients treated with cisplatin regimens. They observed no vomiting in 22/30 (73%) patients. We postulated that barbiturates might enhance the antiemetic effect of phe-nothiazines through their sedative potential. This study was undertaken to investigate the antiemetic efficacy of the relatively inexpensive regimen of secobar