To determine the antitumor efficacy and safety profile of temozolomide in patients with malignant astrocytoma at first relapse. This open-label, multicenter, phase II trial enrolled 162 patients (intent-to-treat [ITT] population). After central histologic review, 111 patients were confirmed to have had an anaplastic astrocytoma (AA) or anaplastic mixed oligoastrocytoma. Chemotherapy-naive patients were treated with temozolomide 200 mg/m(2)/d. Patients previously treated with chemotherapy received temozolomide 150 mg/m(2)/d; the dose could be increased to 200 mg/m(2)/d in the absence of grade 3/4 toxicity. Therapy was administered orally on the first 5 days of a 28-day cycle. Progression-free survival (PFS) at 6 months, the primary protocol end point, was 46% (95% confidence interval, 38% to 54%). The median PFS was 5.4 months, and PFS at 12 months was 24%. The median overall survival was 13.6 months, and the 6- and 12-month survival rates were 75% and 56%, respectively. The objective response rate determined by independent central review of gadolinium-enhanced magnetic resonance imaging scans of the ITT population was 35% (8% complete response [CR], 27% partial response [PR]), with an additional 26% of patients with stable disease (SD). The median PFS for patients with SD was 4.4 months, with 33% progression-free at 6 months. Maintenance of progression-free status and objectively assessed response (CR/PR/SD) were both associated with health-related quality-of-life (HQL) benefits. Adverse events were mild to moderate, with hematologic side effects occurring in less than 10% of patients. Temozolomide demonstrated good single-agent activity, an acceptable safety profile, and documented HQL benefits in patients with recurrent AA.