Uncertainty due to low-dose extrapolation: Modified BMD methodology for epidemiological data

Traditional environmental risk assessment methodologies, including benchmark dose (BMD) estimation, were originally developed to be used with animal toxicology data. We discuss some problems that can occur when toxicology-based methods are applied to human cohort studies, and we propose a new method for BMD risk estimation adapted to epidemiological data. Instead of extrapolating to zero-dose levels to estimate risk in an unexposed population, our proposed alternative involves specifying a nonzero background exposure level instead. To account for uncertainty in that level, we can average modified BMDs over a distribution of exposure levels. We apply our modified BMD method to analyze data from a cohort study designed to assess the effects of prenatal polychlorinated biphenyl (PCB) exposure on cognitive ability in young children. © 2013 John Wiley & Sons, Ltd