by testing at lower doses has any relevance to those other target sites. For example, if an HDT were selected to be below a dose that caused a physiologic change in the liver but not the lung, one could miss a carcinogenic effect on the lung at a dose that does not alter lung physiology.
Use MTD testing as part of an overall testing strategy.
The animal bioassay that uses the MTD is one part of a complete program for identifying human carcinogens. It generally is conducted after some indication that a substance merits examination—e.g., information that a chemical has a structural similarity to a known carcinogen, results of a test for mutagenicity, or a suggestion that there will be extensive human exposure to the substance. It is then used as a screening technique to separate carcinogens from noncarcinogens it can be followed by tests to determine mode of action, pharmacokinetics, and applicability of results to the human experience. The workshop participants and the committee discussed two-ways to use the MTD test in a complete program. They are described below.
Use a two-track system that comprises full testing and limited testing.
In this option, chemicals would be placed into two-tracks for testing. A small number of selected chemicals would be subjected to rigorous testing (the full-testing track). All the remaining chemicals would be subjected to less rigorous testing (the limited-testing track). The option is based on three premises:
A large amount of additional information might be needed to assist in understanding the importance of positive results at the MTD.
It might not be feasible to collect the additional information for all the chemicals whose regulation is appropriate.
Without accompanying information on mechanisms or results at low doses, animal bioassay results alone (i.e., without parallel data on mechanisms and dose-response relationships) do not add greatly to our ability to make regulatory decisions because of the uncertainty about the human implications of positive results in animal bioassays.
Chemicals could be chosen for full testing on the basis of expected human exposures, importance in commerce, structural similarity to a known carcinogen, or results of mutagenicity tests—i.e., in much the same way that chemicals are currently chosen for testing. The method of Gaylor (1989), basing a preliminary estimate of the carcinogenic potency of a chemical on its MTD, could also be used to select chemicals for full or limited testing. As described earlier, given the relationship between the carcinogenic potency of chemicals and their MTDs, Gaylor pointed out that a preliminary estimate of the dose corresponding to a carcinogenic risk of one in a million human lifetimes could be found by dividing the MTD by 380,000. If human exposures were unlikely to be greater than the quotient, the chemical would not be assigned to full testing on the grounds that, even if it were a carcinogen, human risk would be unlikely to be greater than one in a million per lifetime. (A different divisor could be selected if a level of safety different from one in a million per lifetime were required.)
When a class of structurally similar chemicals that are thought likely to have similar mechanisms of action is being considered, it might be a good use of resources to test fully only a small number (perhaps only one) of representative chemicals in the class and to evaluate the others in the class for relative potency on the basis of data from short-term, less-expensive studies.
Chemicals chosen for full testing would be tested in a standard bioassay that used the MTD and an array of doses below the MTD. If a chemical were positive, additional testing would be performed as needed to clarify the dose-response relationship and to improve the predictive value of the positive findings for humans.
Chemicals chosen for limited testing would be considered for regulation without testing in a long-term cancer bioassay. Regulatory decisions for these chemicals would be based on more limited data, such as estimates of the MTD from short-term studies, mutagenicity information, and other data that can be gathered much more quickly and cheaply than results from a lifetime bioassay.