Reflections on the assessment of the toxicity of "dioxins" for humans, using data from experimental and epidemiological studies.

"Dioxins" (polyhalogenated dibenzo-p-dioxins and dibenzofurans, PHDDs/PHDFs) have gained considerable scientific interest, and (unjustified or justified) also received tremendous political attention. The data pool available on sources, distribution in the environment, kinetics in animals and humans, and on biological and toxic actions in various species (including humans), is one of the largest among all environmental organic substances; but quality and predictive power of the data on possible effects in humans vary widely, from adequate to not acceptable. This fact is often ignorantly or perhaps even deliberately disregarded, and such divergent data are frequently given the same weight in attempted risk assessments. It must be stressed that the quality of the toxicological data on most "environmental compounds" in general is far below today's standards required for preclinical and clinical data on medicinal substances. The crucial question is whether humans constitute an especially vulnerable species for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) or the other PHDDs/PHDFs. Since ample data on body burdens and some results of clinical and epidemiological studies on TCDD have become available, answering this question may now be attempted on the basis of a comparison of animal and human data. Quality of the data and the predictive power of the methods used must be considered, dose-response relationships must be critically evaluated, and body burdens achieved in humans and experimental animals must be taken into account. Pitfalls in attempts to extrapolate data from animal studies to humans and limitations of conclusions to be drawn from epidemiological data on humans are discussed in this presentation.