-
Views
-
Cite
Cite
Barbara A. Wetmore, John F. Wambaugh, Stephen S. Ferguson, Mark A. Sochaski, Daniel M. Rotroff, Kimberly Freeman, Harvey J. Clewell, David J. Dix, Melvin E. Andersen, Keith A. Houck, Brittany Allen, Richard S. Judson, Reetu Singh, Robert J. Kavlock, Ann M. Richard, Russell S. Thomas, Response to “Accurate Risk-Based Chemical Screening Relies on Robust Exposure Estimates”, Toxicological Sciences, Volume 128, Issue 1, July 2012, Pages 297–299, https://doi.org/10.1093/toxsci/kfs144
Close - Share Icon Share
Extract
We thank Rudel and Perovich for their interest in our recent article integrating human dosimetry and exposure information with high-throughput in vitro toxicity screening data. To a large extent, we agree with their main premise that more robust exposure estimates are needed for accurate risk-based chemical screening. However, we would like to respond to their concerns as we believe they have either over- or misinterpreted some of the conclusions in the study.
First, Rudel and Perovich expressed concern with the methods used to estimate human exposure and, in particular, with our use of the U.S. Centers for Disease Control and Prevention’s (CDC) Fourth National Report on Human Exposure to Environmental Chemicals (NHANES) (CDC, 2010) to derive some of the values. In our study, the vast majority of exposure estimates (178 of the 182 estimates) came from the Reregistration Eligibility Documents (REDs). Only four of the exposure estimates came from the NHANES data. For the four NHANES chemicals, the demographic with the highest 95th percentile value for the chemical in question was selected to provide a conservative estimate of exposure. In other words, it was not just the 95th percentile of the general U.S. population as Rudel and Perovich imply, but the 95th percentile of the most highly exposed demographic subpopulation that was selected. This decision was stated in the Methods section (p. 161). Granted, the 95th percentile of the most highly exposed subpopulation may still not capture the most “extreme variability in the high end of the exposure distributions,” but it is still a conservative estimate of human exposure and one with precedence in the risk assessment community. According to the Presidential/ Congressional Commission on Risk Assessment and Risk Management, “if a single value representing a population’s or subpopulation’s exposure is required, such as for priority setting, a point in the upper end of the distribution should be used, such as the 95th percentile” (PCRARM, 1997).
Comments