The treatment of recurrent glioblastoma (GBM) remains challenging notwithstanding the recent approval of bevacizumab for this indication. Bendamustine has a bifunctional mechanism of action including alkylation, penetrates the CNS and does not show cross resistance to other alkylator chemotherapies. In a single institution phase 2 trial, patients with recurrent GBM were treated with bendamustine (100 mg/m(2)/day administered intravenously for two consecutive days every 4 weeks). The primary study endpoint was 6-month progression free survival (PFS-6). An interim analysis for futility was conducted according to a Simon two-stage minimax design. Complete blood counts were obtained bimonthly, clinical evaluations and brain imaging every month for the first cycle and bimonthly thereafter. Treatment responses were based upon MacDonald criteria. Sixteen patients were enrolled (nine men; seven women), with a median age of 53 years (range 36-68) and a median Karnofsky performance status of 90 (range 70-100). Nine patients were treated at first relapse and seven at second relapse (five patients were bevacizumab failures). A total of 25 cycles of bendamustine were administered with a median of 1 (range 1-6). Bendamustine-related toxicity was seen in eight patients; lymphopenia in seven (5 grade 3; 2 Grade 4), thrombocytopenia in two (1 Grade 3; 1 Grade 4), and neutropenia in one (1 Grade 3). Fourteen patients have died due to disease progression, two patients are alive and on alternative therapies. Only one patient was progression-free at 6 months, triggering the stopping rule for futility. Bendamustine was reasonably well tolerated but failed to meet the study criteria for activity in adults with recurrent GBM.