Background: The optimal use of bevacizumab (bev) in recurrent glioblastoma (GBM) remains uncertain including the choice between monotherapy or combination therapy as well as the role of continuing bev beyond disease progression. Methods: This was a sequential stratified two part randomized phase II study. Eligibility criteria included: recurrent GBM after radiotherapy and temozolomide, no other chemotherapy for GBM, ECOG PS 0-2. The primary objective (Part 1) was to determine the effect of bev plus carboplatin versus bev alone on 6 month progression-free survival (6PFS) using modified RANO criteria. Bev was given 2-weekly 10mg/kg, carboplatin 4-weekly AUC5. On progression, patients (pts) able to continue treatment were randomized to continue or cease bev (Part 2). Secondary endpoints included response rate (RR); cognitive function; quality of life; toxicity and overall survival (OS). Results: 122 pts (median age 55) were enrolled from 18 Australian sites. 87 (71.3%) were PS 0-1. The current median follow up is 14.7 months with median on-treatment time 3.7 months. 6PFS was 26% (combination) versus 24% (monotherapy) (HR 0.96, 95%CI (0.66, 1.39), p = 0.82). RR (CR+PR) was 15% versus 13%. Median OS was 6.9 versus 6.4 months (HR 1.08, 95%CI (0.74, 1.59), p = 0.68). There were 2 treatment (bev) related deaths in the combination arm (one GI perforation, one CNS hemorrhage). To date, overall incidence of bev-related AEs is similar to prior literature with 10 (8.3%) venous thromboembolic events and 5 (4.2%) hemorrhages (all grades) reported. There were 3 episodes of G3-4 neutropenia, all in the combination arm (5%) and 9 episodes of G3-4 thrombocytopenia. As of January 14, 2013, 47 pts have continued on to Part 2. Data for bev beyond progression is not yet available. Conclusions: In this study of patients with recurrent GBM, the addition of carboplatin to bev did not result in additional clinical benefit compared to bev monotherapy. This large multicentre population-based study demonstrated that clinical outcomes in patients with recurrent GBM treated with bev were inferior to previously reported studies. Ongoing follow-up of patients on bev beyond progression, and novel secondary and exploratory endpoints continues. Clinical trial information: ACTRN12610000915055.