Quantitative human health risk assessment for 1,3-butadiene based upon ovarian effects in rodents

A case study was prepared for noncancer risk assessment of 1,3-butadiene (BD) based upon the ovarian atrophy effects in rodents with specific consideration of the guidelines described by NAS (2009). Ovarian toxicity has been identified in the past as a sensitive endpoint for BD, and serves as the basis for noncancer risk assessment by regulatory agencies. A meta-analysis was conducted in which the available dose-response data from rats and mice were normalized using an internal dose estimate (DEB in blood) that is causally related to ovarian toxicity. A time-to-response (multistage-Weibull) model was used to simultaneously fit the pooled rodent data sets with exposure durations ranging from 13 to 105 weeks. Human variation in ovarian follicle count was assumed to reflect variation in sensitivity to the adverse effects associated with follicle depletion (i.e., premature menopause). Information on follicle count in women was used in two ways: (1) the window of susceptibility (from birth to menopause) was defined as 49.6 years for women born with an average follicle count, 38.7 years for women born with a low follicle count, and 60.0 years for women born with a high follicle count; and (2) follicle count was assumed to reflect human susceptibility due to toxicodynamic factors. The multistage-Weibull model was used to predict dose-response curves for three scenarios (average, low, and high follicle counts at birth to generate reference concentration values ranging from 0.2 to 20 ppm). This case study illustrates how information on mode of action can be used to guide key decisions in the dose-response assessment with respect to identifying a dose measure, low-dose extrapolation method, background exposure, and sensitive subpopulations.

► Meta-analysis was conducted for ovarian toxicity in rodents exposed to 1,3-butadiene. ► Hemoglobin adduct data were used to estimate internal dose for diepoxide metabolite. ► The pooled rodent data set was described using a time-to-response model. ► Reference concentration values ranging from 0.2 to 20 ppm were derived. ► Information on mode of action, human variation, and background exposures were incorporated.