The purpose of this article was to study the effectiveness of a prospective follow-up programme in patients after curative surgery for colorectal cancer. Of the initial 151 selected patients, 61 (40%) developed a recurrence in whom only six cases (10%) of potentially curable recurrent lesions were detected. The first clues to recurrence in the 61 patients were history or physical examination in 49%, a rising CEA in 29% and a positive imaging finding in 10%, being difficult to decide which test first signalled a recurrent cancer in an additional 11%. Endoscopy and CEA determinations were the most rewarding investigations. CEA was a sensitive means of identifying disseminated recurrent disease and liver metastases compared with liver function tests or liver ultrasound every 3 months. Endoscopy was useful in the diagnosis of local recurrences. However no follow-up test was capable of detecting recurrent colorectal cancer when it might still have been curable. As a direct result of this follow-up programme 15 patients (23%) underwent re-exploration. No symptomatic patients were candidates for curative re-operation. Of the asymptomatic patients six (four colonic and two rectal cancers) (19.5%) were re-resected for cure. Only three of these were alive and without evidence of disease, 40, 43 and 69 months later so that long term survivors after curative re-resection represent only 5% of all patients with recurrences (7.2% of the recurrent colonic cancer and 3% of the rectal cancer). Our follow-up programme did not permit us to alter the incidence of disseminated recurrent disease, and the effectiveness of the curative re-resection represents an increase of only 1.3% in the global 5-year survival rates for colorectal cancer. Our study does not demonstrate any great value of 'classical' postoperative follow-up programme.