On the basis of experimental data showing the efficacy of glutathione (GSH) as a protective agent on cisplatin-induced neurotoxicity and the clinical evidence of the low incidence of neurotoxicity in high-dose cisplatin + GSH treated patients we evaluated the neuroprotective effect of GSH in a randomized phase II study. Thirty-three patients with relapsed ovarian cancer after a disease-free interval of at least 1 year and a cumulative dose of prior cisplatin ranging 450-650 mg m-2 were randomized to receive cisplatin 50 mg m-2 weekly +/- 2.5 g GSH for 9 consecutive weeks. Clinical and instrumental neurologic and otologic evaluations were made at the baseline and at the end of the study. Overall response rate in 31 evaluable patients was: 9/15 in group A and 12/16 in group B, including 4/15 vs 7/16 complete responses. The administered dose intensity of cisplatin was higher in the GSH treated patients (100% dose intensity was received by 56% vs 27%). A trend in terms of neuroprotection was detected in the GSH treated group, and no major difference was observed in the other toxicities between the two groups. It is concluded that possible benefit can be expected from the concomitant administration of GSH and cisplatin in patients at high risk of developing neurotoxicity, without decreasing the anti-tumor activity.