Sodium thiosulfate (STS) provides protection against carboplatin-induced ototoxicity in an animal model. The purpose of this study was to determine the STS dose required for otoprotection, in patients with malignant brain tumors treated with carboplatin in conjunction with osmotic blood-brain barrier disruption. Twenty-nine patients received STS intravenously 2 hr after carboplatin. Doses were escalated from 4 g/m2 to 8, 12, 16 and 20 g/m2 on consecutive months. Audiologic assessment was performed at baseline and monthly. The audiograms were compared with those of 19 similarly treated historical control patients who did not receive STS. The incidence of ototoxicity in the historical control group of patients was 79% (15/19). This group had an average loss of 20.8 +/- 5.9 dB (n = 19) at 8 kHz after one treatment with carboplatin, whereas the STS treatment group lost only 3.7 +/- 2 dB (n = 15) after one treatment. This difference was statistically significant as assessed by Student's t test (P < .05). Furthermore, patients in the STS treatment group with excellent base-line hearing showed little change in hearing thresholds at 8 kHz after the second treatment (8.0 +/- 8.3 dB) (n = 5) compared with the historical control patients with excellent base-line hearing, (40.5 +/- 8.6 dB) (n = 11). Our data support that doses of 16 or 20 g/m2 of STS decrease carboplatin-induced hearing loss without central nervous system entry. Clinical demonstration of an otoprotective effect with a two-compartment system to prevent drug-induced hearing loss, while preserving central nervous system cytotoxicity, has not been reported previously.