[Antiemetic effect of ondansetron in recurrent chemotherapy with cisplatin].

Six hundred and twenty-eight courses of cisplatin containing chemotherapy were recorded in patients receiving up to three courses of cytotoxic drugs with ondansetron (OND) given as an antiemetic agent (32 mg as a single iv dose or as a continuous infusion). The sample size of patients decreases from one course to another due to phenomena which may or may not be related to the chemotherapy and the anti-emetic treatment or to the evolution of the cancer disease. For patients with incomplete response to ondansetron, withdrawals could be related to an insufficient antiemetic effect as is known with other antiemetic drugs. Conversely, it is not the case for patients who had a complete response. Therefore, to avoid any bias due to patient selection, the analysis is based on the probability of changes from complete to incomplete response or from complete to complete response between two subsequent courses (i and i + 1). The response to OND treatment for the course i + 1 depends not only on the efficacy during this course (i + 1) but also on the response during the prior course (i). A discrete time-dependent statistical model (Markov chain) was used to test the evolution of the probability of remaining in complete response. This probability was equal to 66% between the first and the second course, and to 88% between the second and the third one. The probability of remaining in complete response significantly increased during repeated courses (P = 0.001). These results show that in patients for whom OND treatment allows a complete antiemetic control during the first course of chemotherapy, the probability of remaining with no emetic episodes at all increases during the two subsequent courses.