Studies on the monoclonal VEGF-A antibody bevacizumab gave raise to questions regarding the lack of an overall survival benefit, the optimal timing in the disease course and potential combination and salvage therapies. We retrospectively assessed survival, radiological progression type on bevacizumab and efficacy of salvage therapies in 42 patients with recurrent malignant gliomas who received bevacizumab and nitrosourea sequentially. 15 patients received bevacizumab followed by nitrosourea at progression and 27 patients vice versa. Time to treatment failure, defined as time from initiation of one to failure of the other treatment, was similar in both groups (9.6 vs. 9.2 months, log rank p = 0.19). Progression-free survival on nitrosoureas was comparable in both groups, while progression-free survival on bevacizumab was longer in the group receiving bevacizumab first (5.3 vs. 4.1 months, log rank p = 0.03). Survival times were similar for patients with grade III (n = 9) and grade IV (n = 33) tumors. Progression-free survival on bevacizumab for patients developing contrast-enhancing T1 progression was longer than for patients who displayed a non-enhancing T2 progression. However, post-progression survival times after bevacizumab failure were not different. Earlier treatment with bevacizumab was not associated with better outcome in this series. The fact that earlier as compared to later bevacizumab treatment does not result in a different time to treatment failure highlights the challenge for first-line or recurrence trials with bevacizumab to demonstrate an overall survival benefit if crossover of bevacizumab-naïve patients after progression occurs.