[Too early to recommend universal screening for colorectal cancer]

Randomised controlled trials have shown that screening may significantly reduce mortality in colorectal cancer. In the present evaluation it is shown that the reported reduction of 15 per cent, or higher, is given in relative figures. With absolute figures the reduction is only 0.15 per cent or lower. The number of individuals needed to screen (NNT) biannually for 10 years to prevent one death in colorectal cancer is 1,173 (confidence interval 741-2,807). Side-effects, costs and ethical issues are seldom, or never, discussed. Postoperative follow-up may be regarded as the screening of a high-risk group. The reduced mortality seen in the meta-analysis of randomised controlled trials is here regarded as small and should also be weighed against the drawbacks of follow-up before any recommendation for more active surveillance is given.