A historically controlled phase II study was undertaken to investigate the efficacy and toxicity of a postoperative treatment consisting of high-dose continuous tamoxifen, carboplatin and radiotherapy in patients with newly diagnosed glioblastoma. Between 1995 and 1998, 50 patients with newly diagnosed glioblastomas underwent surgery and were subsequently treated with 200 mg day(-1) tamoxifen continuously, 3 cycles of carboplatin (300 mg m(-2)), and radiotherapy. Survival data for a historical control group were calculated from respective prognostic indices and were obtained from studies with comparable patient populations treated with operation and radiotherapy only. In our study, the median time to tumor progression was 30 weeks and the median survival time (MST) 55 weeks (95% confidence interval: 46-63 weeks). The MST of the control group (48 weeks) showed to be within this interval. In addition to already known prognostic factors in malignant gliomas (age, Karnofsky performance score, extent of tumor resection), the gender (females lived longer than males, p = 0.0025) showed to influence survival. Serious side effects (thrombosis, pulmonary embolism) occurred in 6 patients. A high incidence of multifocal tumor recurrences (33%), which might be related to study-treatment, was observed. In conclusion, the combined therapy failed to demonstrate a higher efficacy than standard treatment for glioblastoma patients.