Isoprene cancer risk and the time pattern of dose administration.

Most statistical risk assessment models assume that equal, measured on scale such as mg/kg/day, create equal tumor risks. This equivalent dose metric (EDM) hypothesis allows risks to be extrapolated from high concentrations to low-concentrations and from and species, sex, and strain to another, since it implies that all administered dose histories corresponding to the same total dose create the same risk. This paper tests the EDM hypothesis using data on tumor rates in B6C3F mice administered isoprene via inhalation. Its major conclusion is that the EDM hypothesis does not hold for isoprene. For example, it appears that exposure concentration has a greater impact on tumor rates than weeks of exposure. To predict tumor probabilities, the time pattern of dose administration must be considered. The asymmetric effects of concentration, hours-per-day, and number of days of exposure on tumor risks imply that complex dynamic risk models may be needed to accurately describe dose-time-response relations. The traditional concept of a dose-response relation as a static curve relating a numerical summary of dose to a numerical summary of response probability is probably not predictively useful for chemicals such as isoprene, and extrapolations of risk based on the EDM hypothesis could be misleading for such chemicals.