An experimental design for judging synergism on consideration to endocrine disruptor animal experiments

This paper investigates an appropriate statistical design for an animal experiment to evaluate synergism of two test chemicals. It assumes a certain number of animals are divided into groups, each of which is treated with a combination of dose levels of two chemicals. A design is identified by the set of group size for each combination of doses, including the case where the dose of either one chemical is zero. The power of t-test to detect synergism by positive surplus of response on a simultaneous administration group from the additivity plane composed of the responses on single administration groups is adopted as the criterion for the appropriate design. The applicable design is investigated for the application to real cases of endocrine disrupter study conducted at the National Institute of Health Sciences of Japan. It revealed that the dose level of the simultaneous administration group should be located inside or on the boundary of a triangular region and that the total number of animals should be the same as those for single administration groups. Copyright © 2008 John Wiley & Sons, Ltd.