The optimum duration of first-line treatment with chemotherapy in combination with bevacizumab in patients with metastatic colorectal cancer is unknown. The CAIRO3 study was designed to determine the efficacy of maintenance treatment with capecitabine plus bevacizumab versus observation. In this open-label, phase 3, randomised controlled trial, we recruited patients in 64 hospitals in the Netherlands. We included patients older than 18 years with previously untreated metastatic colorectal cancer, with stable disease or better after induction treatment with six 3-weekly cycles of capecitabine, oxaliplatin, and bevacizumab (CAPOX-B), WHO performance status of 0 or 1, and adequate bone marrow, liver, and renal function. Patients were randomly assigned (1:1) to either maintenance treatment with capecitabine and bevacizumab (maintenance group) or observation (observation group). Randomisation was done centrally by minimisation, with stratification according to previous adjuvant chemotherapy, response to induction treatment, WHO performance status, serum lactate dehydrogenase concentration, and treatment centre. Both patients and investigators were aware of treatment assignment. We assessed disease status every 9 weeks. On first progression (defined as PFS1), patients in both groups were to receive the induction regimen of CAPOX-B until second progression (PFS2), which was the study's primary endpoint. All endpoints were calculated from the time of randomisation. Analyses were done by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00442637. Between May 30, 2007, and Oct 15, 2012, we randomly assigned 558 patients to either the maintenance group (n=279) or the observation group (n=279). Median follow-up was 48 months (IQR 36-57). The primary endpoint of median PFS2 was significantly improved in patients on maintenance treatment, and was 8·5 months in the observation group and 11·7 months in the maintenance group (HR 0·67, 95% CI 0·56-0·81, p<0·0001). This difference remained significant when any treatment after PFS1 was considered. Maintenance treatment was well tolerated, although the incidence of hand-foot syndrome was increased (64 [23%] patients with hand-foot skin reaction during maintenance). The global quality of life did not deteriorate during maintenance treatment and was clinically not different between treatment groups. Maintenance treatment with capecitabine plus bevacizumab after six cycles of CAPOX-B in patients with metastatic colorectal cancer is effective and does not compromise quality of life. Dutch Colorectal Cancer Group (DCCG). The DCCG received financial support for the study from the Commissie Klinische Studies (CKS) of the Dutch Cancer Foundation (KWF), Roche, and Sanofi-Aventis.