The authors evaluated a 3-week schedule of bevacizumab in patients with recurrent high-grade glioma (HGG). Patients received bevacizumab 15 mg/kg every 3 weeks and were evaluated every 6 weeks until tumor progression. Tissue correlates were used to quantify tumor content of vascular endothelial growth factor A (VEGFA) and vascular endothelial growth factor receptor-2 (VEGFR2). Of 61 patients who were treated (35 men and 26 women; median age, 52 years; age range, 21-78 years), 50 patients had glioblastoma multiforme (GBM), and 11 patients had anaplastic glioma (AG). The median number of previous chemotherapies was 2 (range, 1-5 previous chemotherapies), and 16 patients had received ≥3 previous chemotherapies. The median number of bevacizumab doses was 4 (range, 1-20 doses), and 45% of patients received >5 doses. The toxicities observed were primarily grade 1 and 2, and the most common were fatigue, hypertension, and headache. One grade 2 intratumoral bleed and 1 bowel perforation were reported. For patients with GBM, the 6-month progression-free survival rate was 25%, the median time to tumor progression was 10.8 weeks, and the median overall survival was 25.6 weeks. The best response included a partial response in 15 patients (24.5%) and stable disease in 31 patients (50.8%) patients; radiographic recurrence patterns included increased changes in fluid attenuation inversion recovery (24%) and multifocal recurrence (20%). The median survival after bevacizumab failure was 10 weeks. The ratio of tumor VEGFA/VEGFR2 was increased in patients aged >55 years; an increased VEGFA/VEGFR2 ratio was correlated nonsignificantly with decreased survival (P = .052). An every-3-week schedule of bevacizumab had antitumor activity and was relatively nontoxic for patients with recurrent HGG. The predictive value of VEGFA/VEGFR2 in tumor will require validation in a larger patient cohort.