Background: Conflicting results are reported for maintenance treatment with bevacizumab (bev) during chemotherapy free intervals (CFI) in metastatic colorectal cancer (mCRC). Methods: The objective was to compare the tumor control duration (TCD) by either bev maintenance (Arm A) or no treatment during CFI (Arm B) after induction chemotherapy (CT) (12 cycles of FOLFIRI + bev). CT was reintroduced at progression (8 cycles) and then a new CFI. The randomization was performed before induction CT. TCD was defined by the time between randomization and tumor progression during a CT sequence (Aparicio T et al, Dig Liver Dis, 2015). Per Protocol (PP) population was defined as patients (pts) with at least one CT reintroduction after first progression during CFI. We present the final analysis of the trial. Results: From March 2010 to July 2013, 491 pts were randomized. The median age was 64.6 years [range: 27 - 89], 64% of the pts were men, 93% of the pts were OMS 0-1 and 218 (44%) pts had a non-resected primary tumor. A progression during induction CT happen in 85 (17%) pts, 261 (53%) pts had at least one reintroduction, 107 (22%) pts had two and 51 (10%) pts had three or more re-introductions. Multivariate analysis in all pts revealed that WHO performance status ≥ 2, unresected primary tumor and BRAF mutation were associated with a shorter TCD. Unresected primary tumor and BRAF mutation were associated with a shorter overall survival (OS). Grade 3-4 toxicities were observed in 80% of the pts in arm A and 79% in arm B. Conclusions: The alternate irinotecan-based CT sequences with or without bev maintenance during CFI, revealed an impressive TCD. Bev maintenance monotherapy did not improved TCD, progression free survival (PFS) or OS. Clinical trial information: NCT00952029