Determining acceptable risks: experimental and epidemiological issues.

The extent to which results of experimental toxicology studies can be used to assess risk for humans varies. Administering high doses of chemicals to experimental animals may produce indirect adverse metabolic and nutritional effects, and metabolic pathways may differ at high and low doses. The rate of metabolism at high and low doses and the metabolic pathways of chemicals in different species are important for establishing the reasons for differences in species responses. Animal species best suited to serve as a surrogate for humans vary for different chemicals. For realistic risk assessments, more detailed information on the toxicokinetics of chemicals in humans is needed. As data are developed to link early biochemical changes with future disease, it is important to determine their predictive value. Our present expectations of using various biomarkers to predict future health outcomes may be unrealistic. More baseline data are needed to determine which biomarkers are expressions of exposure and which are predictors of future disease. It is also unclear how lifestyle affects most biomarkers.