The Predictive Value of Rodent Carcinogenicity Tests in the Evaluation of Human Risks

In the last fifteen years various authors have estimated that from about 70% to 90% of all human cancer cases can be attributed to environmental causes (4, 19, 20, 61). These percentages were based on two main considerations: (a) a certain number of environmental factors have been identified in the etiology of human cancer, and (b) significant variations in the incidence of cancer in different population groups in different countries suggest that environmental risk factors play an important role. The variation in different countries in the incidence of cancers that occur at the most common target organs fluctuates betweeen fourfold and more than one hundredfold (10, 19). It was therefore proposed that the lowest observed incidence rates could be taken as a baseline and that any increase over this baseline could be attributed to environmental factors, of which only a few so far have been identified. Ignoring, or disregarding, such considerations, it has been simplistically inferred that 90% of all cancer cases could now be prevented; this is certainly not what the above-men­ tioned authors meant to imply. The euphoria so engendered has not yet been dispelled, despite the words of caution that some authors have tried to interject, in an attempt to define environment and the role of environmen­ tal factors (21). An assessment of the number of cancers that could, in fact, be prevented would obviously be of paramount importance; however, it would be difficult