Abstract

Logically, the choice of any ultimate optimum therapy requires, as well as comparison of the survival outcomes, a comparison of both subjective and objective toxicities in terms of incidence, degree of severity, and duration. Frequently such detail is not collected in large studies. Both cisplatin and paclitaxel are effective but neurotoxic drugs for ovarian cancer. The optimum choice is further complicated in that carboplatin is a possible alternative for cisplatin, being less neurotoxic but having greater hematologic toxicity. Similarly, 3-h and 24-h infusion schedules of paclitaxel have different incidences in opposite directions of hematologic and neurologic toxicities. One hundred fifty two eligible Canadian patients entered in a European-Canadian study that compared paclitaxel-cisplatin (PT, 79) patients with cyclophosphamide-cisplatin (PC, 73 patients) had both subjective and objective neurotoxicity data collected from treatment initiation to disease progression. Incidence, degree, and duration (compared in an analogous way to remission durations) of neurotoxicity were compared in the two arms to quantify the additional paclitaxel toxicity. No significant differences were found for motor toxicity, motor impairment, hearing impairment, or insomnia. For sensory changes during treatment, toxicity (all grades, 91% vs. 49%; grade 3 or higher, 29% vs. 3%) incidence, subjective impairment (a little or more, 89% vs. 40%; lots, 54% vs. 11%) incidence, and toxicity duration (all grades only), and impairment durations (both degrees) were all worse for PT. During follow-up, only the incidence of all-grade sensory toxicity was worse and this was not reflected by any other parameters. We conclude that paclitaxel adds considerably, but only temporarily, to the sensoy neurotoxicity of cisplatin.