1991 Herbert E. Stokinger Lecture: Uncertainties in Risk Assessment

Abstract

Results obtained in animal toxicology studies are extrapolated to humans to set standards in the workplace and to evaluate the levels of contaminants found in the environment. Predictions are made about potential adverse effects in humans based on the toxicity observed in laboratory animals if human data are limited or nonexistent. This approach is an important first step. However, humans do not always respond to chemical exposures in the same way animals do, nor is it always true that proximity to a chemical substance necessarily means that the chemical is actually taken up and absorbed or that absorption is complete.

Most common chronic diseases of the general population are multifactorial in origin. Important factors that modulate chronic diseases are genetic make-up, chronic virus infections, improper nutrition, smoking, and drug and alcohol abuse. Exposure to trace amounts of synthetic chemicals found in the environment has, in most instances, little or no impact. However, certain occupational diseases where higher exposures occur form an exception, and it would be prudent to make a greater effort to reduce worker exposures to high levels of chemicals and to shift our focus to occupational rather than environmental health in the area of synthetic chemicals. As an example, the most recent data on 2, 3, 7, 8-tetrachlorodibenzo-p-dioxin are reviewed. These data show that the estimates of potential human health risks based on animal data have been exaggerated.

If chemicals cause cancer in rodents at doses where pronounced tissue damage is also present in the target organ, at lower doses where such tissue damage does not occur, the response for many chemicals would be less than linear or no tumors would develop. These biological factors are presently not considered in the United States for most cancer risk assessments that are performed.