• Support Care Cancer. 1994 May 1; 2 (3): 171-6.

AbstractIn comparative trials on antiemetic efficacy of different regimens, the positive correlation between the probabilities of vomiting and of nausea could hide some confounding effect. Our work seeks to detect such effects. The data from two large studies on prevention of cisplatin-induced emesis were re-analyzed using two multifactorial logistic models. The first study compared ondansetron + dexamethasone (Ond+Dex) with metoclopramide+dexamethasone+diphenhydramine (Mtc+Dex+Dip), the second compared Mtc+Dex+Dip with Mtc+prednisolone. A group of 267 patients in the first study and 343 in the second, all evaluable for clinical efficacy, were followed for three subsequent cycles of chemotherapy. The antiemetic regimen administered and the complete protection from acute vomiting and nausea were recorded. In the first study (cycle 1), after adjustment for the presence/absence of vomiting, the two therapies were no longer found significantly different on complete protection from nausea: the greater efficacy of Ond+Dex in preventing nausea was due to a confounding effect. Instead, in the second study, the greater efficacy of Mtc/Dex/Dip in preventing both nausea and vomiting was confirmed. The results indicate that, when a correlation between two responses is detected, multifactorial analyses should be performed to identify the possible presence of some confounding effect. The proof that the presence/absence of vomiting is a confounding factor for the relationship between the different efficacy of the two antiemetic regimens for complete protection from nausea, highlighting the same efficacy of the two therapies in preventing nausea, supports the hypothesis of the existence of two kinds of nausea, one independent of vomiting, the other concomitant with it.

30 keywords...

hide…