Impact of pharmacogenetics on toxicological studies. Statistical implications.

Genetic differences in response to chemicals are probably ubiquitous. The implications of such genetic differences for the design of toxicological experiments is reviewed. The inclusion of genetic differences as factors in factorial experimental designs is illustrated by an experiment to investigate differences between inbred strains of mice in the length of anaesthesia following the administration of pentobarbitone to control and phenobarbitone pre-treated animals. The effects of including genetic differences as a factor in the standard toxicological studies such as the long-term rodent carcinogenicity bioassay (LTRCB) are reviewed. Such designs are likely to result in more chemicals being identified as carcinogens based upon the criteria of a positive result in the LTRCB. However, they are unlikely to provide any improvement in the estimation of risk to the human population from low level exposures to the chemicals because of the limitations of the existing extrapolation methods. Information on the degree of genetic variability in response to chemical exposure should, though, help in the more qualitative biologically-based risk assessment approaches favoured by some toxicologists. The presence of genetic differences in response also provides potential models for investigation of the mechanisms underlying toxicological responses.