Background: The optimal follow up after potentially curative resection of primary colorectal cancer is not known. Methods: In this National Institute of Health Research funded trial 1,211 patients with curatively treated primary colorectal cancer were randomised in a 2x2 trial design: Arm 1; Single CT chest/abdomen/pelvis (CT-CAP) at 12-18 months post treatment: Arm 2; Regular CEA measurement in family practice plus a single CT-CAP at 12-18 months: Arm 3; Intensive hospital based imaging (6 monthly CT-CAP for 2 years, annually for an additional 3 years, combined with colonoscopy at 2 years: Arm 4; Combination of Arms 2 and 3. All patients had full evaluation by CT-CAP after surgery or adjuvant chemotherapy if given and had a normal CEA prior to trial entry. All patients have a colonoscopy scheduled at 5 years. Recurrence is managed by a multi-discipline team which includes liver and lung surgeons. The primary end point is complete resection of recurrent disease with curative intent (89% power to detect a 6% difference between arms 1 and 4). Secondary end points are overall survival, recurrence free survival, quality of life and cost effectiveness of follow up. Results: This observational analysis includes all randomisation arms at a median follow up of 54 months. The mean age of the participants was 68.6 years, 61% were male and 97% ethnically white. 22% were in stage I, 48% in stage II and 30% in stage III. At follow up 85.6% of patients were alive without recurrence (stage I: 90.9%, stage II: 86.7%, stage III: 80.4%). Of the 174 recurrences 42 (24.1%) were liver only and 29 (16.1%) lung only. The remainder were loco-regional or at multiple sites. 72 (41.4%) of the 174 patients with recurrence had further surgery. In 38 (23%) this was potentially curative. Conclusions: If patients are fully staged by CT after treatment of the primary tumour and have a normal CEA the relapse rate is low. Only 38 (3.1%) of the 1,211 patients were suitable for potentially curative treatment and most patients with recurrence had incurable disease. These preliminary data militate against the cost effectiveness of very intensive follow up schedules. The final analysis is planned in 2013 to minimise the risk of lead-time bias.