OBJECTIVE: Bevacizumab has been reported not to improve survival in patients with glioblastoma in addition to radiotherapy-temozolomide. The Japanese Ministry of Health, Labour and Welfare has approved BEV for newly-diagnosed and recurrent malignant glioma on June 14, 2013. Therefore, we analyze retrospectively clinical outcomes of glioblastoma patients treated with bevacizumab-containing and non-bevacizumab-containing regimens in our institution. SUBJECTS AND METHODS: Between 2006 and 2013, 94 patients with newly-diagnosed glioblastoma were treated with radiotherapy with concomitant and adjuvant temozolomide. To patients with tumor progression, we applied ICE (ifosfamide, carboplatin, and etoposide), CPT-11, etoposide, bevacizumab and bevacizumab-containing regimen with temozolomide, ICE, and CPT-11. RESULTS: 43 patients have been treated with bevacizumab-containing regimens (bevacizumab group) and 51 patients with non-bevacizumab-containing regimens (non-bevacizumab group). Median survival times in bevacizumab group and non-bevacizumab group were 21.7 and 17.2 months, respectively (p = 0.45). Recurrences were identified in 41 patients of bevacizumab group and 38 patients of non-bevacizumab group. In analysis of patients with recurrent glioblastoma, median survival time in bevacizumab group (21.7 months) was significantly longer than in non-bevacizumab group (14.2 months) (p = 0.031). CONCLUSION: Bevacizumab-containing regimens showed significant survival benefit in the patients with recurrence of glioblastoma and temozolomide failure, compared with non-bevacizumab-containing regimens.