van Leeuwen's Response

W, Connolly R. Concerns with the use of a toxic equivalency factor (TEF) approach for risk assessment of "dioxin-like" compounds. In our paper, "Toxic Equivalency Factors (TEFs) for PCBs, PCDDs, PCDFs for Humans and Wildlife" (1), we described the results of a World Health Organization (WHO) working group that evaluated the existing TEFs for human risk assessment and derived consensus TEFs for fish and birds. Starr et al. comment on the approach taken for the derivation of TEFs as described in our paper (1) and criticize the inadequate characterization of the uncertainty of the TEFs. They state that the following sentence is the only uncertainty characterization: ... it is unlikely for the use of this additive model to result in a great deal of error in predicting the concentrations of TCDD TEQs or responses at environmentally relevant levels due to nonaddi-tive interactions. In addition, they request explicit quantitative estimates on the uncertainty in each REP value, each TEF, and on the deviation from parallelism for the different end points. Their first statement is incorrect. In the paper we clearly stated that the TEFs that were derived are "an order of magnitude estimate." This is a clear illustration of the overall uncertainty in TEF values based on the differences in outcomes of the different end points and the variation in available data for the different congeners. In addition , Starr et al. misinterpreted the sentence quoted above. This sentence is based on the opinion of Van den Berg et al. (1) that the use of an additive model in the TEF approach, in contrast to including nonad-ditive (synergistic or antagonistic) effects, does not result "in a great deal of error." Providing a quantitative estimate of the uncertainty of the individual REPs, as requested by Starr et al., is often not possible. Uncertainty in no-observed-(adverse)-effect levels [or lowest-observed-(adverse)-effect lev-els] or EC50 is usually not given in the studies used by the WHO working group. Therefore, a more qualitative, tiered approach was chosen to select the REP values in which we had greatest confidence, and not because we believe they are "without error." This weighted procedure is clearly outlined in the original paper. For those scientists who want to address the variation in REP values in more detail, the database containing all the information that was used in the derivation of TEFs is available on request. In their comments , Starr et al. …