Background: Most patients with Crohn's disease need an intestinal resection, but a majority will subsequently experience disease recurrence and require further surgery. This study aimed to identify the optimal strategy to prevent postoperative disease recurrence. Methods: In this randomised trial, consecutive patients from 17 centres in Australia and New Zealand undergoing intestinal resection of all macroscopic Crohn's disease, with an endoscopically accessible anastomosis, received 3 months of metronidazole therapy. Patients at high risk of recurrence also received a thiopurine, or adalimumab if they were intolerant to thiopurines. Patients were randomly assigned to parallel groups: colonoscopy at 6 months (active care) or no colonoscopy (standard care). We used computer-generated block randomisation to allocate patients in each centre to active or standard care in a 2:1 ratio. For endoscopic recurrence (Rutgeerts score ≥i2) at 6 months, patients stepped-up to thiopurine, fortnightly adalimumab with thiopurine, or weekly adalimumab. The primary endpoint was endoscopic recurrence at 18 months. Patients and treating physicians were aware of the patient's study group and treatment, but central reading of the endoscopic findings was undertaken blind to the study group and treatment. Analysis included all patients who received at least one dose of study drug. This trial is registered with ClinicalTrials.gov, number NCT00989560. Findings Between Oct 13, 2009, and Sept 28, 2011, 174 (83% high risk across both active and standard care groups) patients were enrolled and received at least one dose of study drug. Of 122 patients in the active care group, 47 (39%) stepped-up treatment. At 18 months, endoscopic recurrence occurred in 60 (49%) patients in the active care group and 35 (67%) patients in the standard care group (p=0·03). Complete mucosal normality was maintained in 27 (22%) of 122 patients in the active care group versus four (8%) in the standard care group (p=0·03). In the active care arm, of those with 6 months recurrence who stepped up treatment, 18 (38%) of 47 patients were in remission 12 months later; conversely, of those in remission at 6 months who did not change therapy recurrence occurred in 31 (41%) of 75 patients 12 months later. Smoking (odds ratio [OR] 2·4, 95% CI 1·2-4·8, p=0·02) and the presence of two or more clinical risk factors including smoking (OR 2·8, 95% CI 1·01-7·7, p=0·05) increased the risk of endoscopic recurrence. The incidence and type of adverse and severe adverse events did not differ significantly between patients in the active care and standard care groups (100 [82%] of 122 vs 45 [87%] of 52; p=0·51) and (33 [27%] of 122 vs 18 [35%] of 52; p=0·36), respectively. Interpretation Treatment according to clinical risk of recurrence, with early colonoscopy and treatment step-up for recurrence, is better than conventional drug therapy alone for prevention of postoperative Crohn's disease recurrence. Selective immune suppression, adjusted for early recurrence, rather than routine use, leads to disease control in most patients. Clinical risk factors predict recurrence, but patients at low risk also need monitoring. Early remission does not preclude the need for ongoing monitoring.