Distributions of individual susceptibility among humans for toxic effects. How much protection does the traditional tenfold factor provide for what fraction of which kinds of chemicals and effects?

A significant data base has been assembled on human variability in parameters representing a series of steps in the pathway from external exposure to the production of biological responses: contact rate (e.g., breathing rates/body weight, fish consumption/body weight); uptake or absorption (mg/kg)/intake or contact rate; general systemic availability net of first pass elimination and dilution; systemic elimination or half-life; active site availability/general systemic availability; physiological parameter change/active site availability; functional reserve capacity--change in baseline physiological parameter needed to pass a criterion of abnormal function or exhibit a response. This paper discusses the current results of analyzing these data to derive estimates for distributions of human susceptibility to different routes of exposure and types of adverse effects. The degree of protection is tentatively evaluated by projecting the incidences of effects that would be expected for a tenfold lowering of exposure from a 5% incidence level if the population distribution of susceptibility were truly log-normal out to the extreme tails, and if the populations, chemicals, and responses that gave rise to the underlying data were representative of the cases to which traditional uncertainty factor is applied. The results indicate that, acting by itself, a tenfold reduction in dose from a 5% effect level is associated with effect incidences ranging from slightly less than one in ten thousand, for a median chemical/response, to a few per thousand, for chemicals and responses that have greater human interindividual variability than 19 out of 20 typical chemicals/responses. In practice, for many of the cases where the traditional tenfold factor is applied, additional protection is provided by other uncertainty factors. Nevertheless, the results generate some reason for concern that current application of traditional safety or uncertainty factor approaches may allow appreciable incidences of responses in some cases.