Background: Trials with anti-PD1 in rGBM have shown limited efficacy. VEGF is highly up regulated proangiogenic growth factor in GBM contributing to tumor-associated immunosuppression. Preclinical data suggests a potential dose effect of anti-VEGF therapy on immunomodulation. Hence, a combination of anti-PD1 and anti-VEGF may be a promising approach in rGBM. Methods: 90 patients with first-recurrent GBM were randomized (1:1) to nivolumab (240 mg IV Q2 weeks) with bevacizumab at standard (10 mg/kg; Arm A) or at low dose (3 mg/kg; Arm B) IV Q2 weeks. Stratification included extent of resection, age, performance status and MGMT methylation status. Single cell RNA sequencing with CITE-seq was used to analyze blood samples from pre- and 8 weeks post-treatment among 8 responders and 8 non-responders. Progression-free survival (PFS) and overall survival (OS) were compared between two arms. Results: 90 patients (Median age 60.6 years ranged 27.4-86.4, 67.8% male, median KPS 80) were enrolled between May 2018 and Jan 2020. Patients were followed in median 7.7 months (Range 0.7, 28.2). 35 of 88 patients were MGMT methylated (2 indeterminate). Overall OS was not significantly different between arm A and arm B (1 year: 41.1 vs 37.7%, p = 0.14), while OS was better for arm A in age > 60 (At 1-year: 46.2% vs 23.8%; Median: 10.6 vs 5.9 months; P = 0.046). OS was no different in the two arms for age ≤ 60 years (At 1-year: 35.6% vs 56.4; Median 8.0 vs 12.4 months; P = 0.90). Single cell RNA sequencing with CITE-seq was used to analyze blood samples from 16 patients, baseline and 8 weeks post treatment. Standard dose bevacizumab treated patients had decreased myeloid derived suppressor cells and an inflammatory response gene signature at 8 weeks. Most frequent toxicities ( > 20%) included fatigue (45.6%), proteinuria (34.4 %), diarrhea (28.9%), hypertension (23.3%) and lipase increase (21.1%). Toxicities in grade 3-4 were hypertension (7.8%), fatigue (5.6) and other non-neurological toxicities including DVT, PE, infection, and abnormal liver function. Conclusions: Overall PFS and OS rates appear similar for nivolumab with either standard or low-dose bevacizumab compared to historical benchmarks of bevacizumab monotherapy. Nivolumab with standard bevacizumab may benefit older but not younger patients. Ongoing response evaluation and immunocorrelative data will be presented. Clinical trial information: NCT03452579