Tumour incidence reductions in carcinogenicity bioassays

Most experimental evidence for chemopreventive activity comes from studies in which animals are administered the suspect chemopreventive agent before, after or simultaneously with a carcinogen predicted to increase tumour incidence in a particular tissue. Such experimental designs have inherent mechanistic assumptions. Evidence of chemopreventive function is also theoretically derivable from carcinogenicity assays, which in principle involve exposure to a single agent in addition to all other factors experienced by the control group. In these assays, exposure is normally for a very large proportion of the animal's lifespan, so they could indicate chemopreventive activity with much greater facility than could human trials, in which exposure for 10% of the lifespan would be considered to be a long study. Site-specific tumour incidence reductions do occur in carcinogenicity assays, often in the same experiments in which increases occur at other sites. These reductions could be due to a carcinogenic response at another site which increases mortality within the group. They could also be secondary to a noncarcinogenic toxic effect that results in either a lifespan reduction (thereby reducing the chance of tumour development) or a reduced food intake (this being the clearest example of an experimental manipulation resulting in reduction in tumour incidence). Nevertheless, if chemoprevention is possible in humans, then it should also be experimentally demonstrable at nontoxic dose levels. Studies involving experimentally reduced tumour incidence are reviewed here, with the aim of describing special requirements for their evaluation and their significance in the field of chemoprevention.