Nutritional epidemiological studies in cancer prevention: what went wrong, and how to move forwards.

It has been almost 30 years since Doll and Peto suggested that most of the differences in cancer rates could be attributed to the environment and behavioral factors including diet. Since then epidemiological studies have reported that individuals consuming large amounts of fruits and vegetables have a reduced risk of cancer. From this evidence, large randomized trials of long duration were designed to test the hypothesis that some micronutrients contained in plant foods decrease cancer risk. Despite the promising experimental and epidemiological data, most randomized controlled trials failed to confirm a protective role for single or combined elements in cancer prevention. The results from randomized trials of micronutrients for cancer prevention have been mixed. Some trials did not demonstrate chemopreventive efficacy for their primary endpoints but showed statistically significant reductions in secondary outcomes, whereas others showed unexpected harmful effects. On the basis of these findings and reflections of what went wrong it is important to find alternative approaches to move forwards in cancer prevention. Clearly, the evidence is not encouraging for further preventive trials on cancer in healthy populations. However, if we identify high-risk individuals using models that include genetic polymorphisms, or in the future, MicroRNA profiles, prevention trials could be designed to target only these groups