A second reviewer indicated that a choice of 5% extra risk over a 10%
extra risk as the
BMR was only briefly presented in the Toxicological Review. This reviewer suggested that if
the effect is moderately adverse as stated, meaning that it is neither very severe and it occurs at
only the highest exposure dose, then a higher level of extra risk would be tolerated. The
reviewer did not see a scientific justification for using a 5% extra risk, instead of a 10% extra
risk which is the standard approach.
A third reviewer indicated the selection of a 5% BMR may not be scientifically justified.
This reviewer stated that the lack of an adequate database could require the use of a lower
percentage extra risk as the BMR (i.e., 0%). For these reasons, this reviewer recommended
using the standard default benchmark of 10% extra risk.
A fourth reviewer indicated the choice of 5% BMR was arbitrary in view of the
limitations of the data base. This reviewer stated additional explanation should be represented in
the text. A fifth reviewer indicated that, because of the limited dose-response data, BMD
modeling at any BMR was not warranted. This reviewer indicated the only feasible approach
was a NOAEL/LOAEL approach for determining the point of departure for deriving the RfD.
Response: The selection of a BMR depends, in part, on the relative severity of the critical effect
and whether there are sufficient data to predict the shape of the dose-response curve at low
doses. The reviewers commented on both of these issues.
Nephropathy was observed in approximately half of the animals in the high dose group,
and these lesions were associated with a significant decrease in survival. Lesions observed at
intermediate dose were deemed to be on a continuum leading to severe nephropathy. For this
reason, the effects at this dose were considered irreversible and of a substantial nature to warrant
the use of a lower benchmark response. The BMR at 10% extra risk is provided for comparison
purposes.
The other issue of concern is the suitability of the data for low-dose extrapolation. The
dose-response information for chemical-related nephropathy in male mice contains two data
points: the high dose group with lesions in 32% of the animals and the middle dose group with
lesions in 3% of the animals. The incidence of nephropathy in the middle dose group was not
statistically significant, but the histomorphology and severity of the lesions indicates that they
were not spontaneous in origin. Moreover, these data provide an increased level of confidence
in the BMD model predictions for effects in the low dose range.
A-12
For this assessment, a BMR of 5% extra risk was used to derive
the RfD because it was
determined that nephropathy was a substantially severe adverse effect and because the data
supported modeling to this effect level (i.e., a response was measured near this effect level).
Additional text discussing the selection of BMR was added to Section 5.1.2 and to Appendix B.
D1) Are the choices of uncertainty factors transparently and objectively described?
Comment: The reviewers generally agreed with the choice of uncertainty factors (UFs) and
largely indicated that the description of their selection was transparent. One reviewer thought
that it should be emphasized that the application of a 10-fold UF is a default EPA policy.
Another reviewer felt that the rationale for the choice of UFs should contain more detail. For
example, the evidence of interspecies differences could be summarized. A third reviewer noted
that describing the application of the data base deficiency UF is especially difficult because,
unlike the other UFs, the rationale is often chemical-specific.
Response: EPA’s practice is to examine all of the relevant health effect data and apply default
assumptions when the data are insufficient or there are data gaps. The application of 10-fold
UFs in situations where data are lacking is a standard EPA practice based on empirically derived
data (U.S. EPA, 2002). An explanation for each uncertainty factor is provided below (in
response to Question D2) and in Section 5.1.3 of the Toxicological Review. Additional
information about the application of UFs can be found in A Review of the Reference Dose and
Reference Concentration Processes (U.S. EPA, 2002) available at:
http://www.epa.gov/iris/RFD_FINAL[1].pdf.
Comment: The reviewers generally agreed with the proposed UF values. One reviewer
expressed concern about the 10-fold UF for interspecies variability and the threefold UF data
base deficiency. This reviewer stated that the available information about factors that contribute
to intraspecies susceptibility is limited and inconsistent but ultimately concluded that it was
reasonable to retain the 10-fold UF. Regarding the threefold UF for data base deficiencies, this
reviewer thought that it was inappropriate to consider limitations in the data base for areas
unrelated to the critical effect of nephropathy (i.e., neurotoxicity). A second reviewer agreed
that neurotoxicity data, if they were available, would not be likely to affect the RfD and
therefore should not be used as a justification for this UF. This reviewer thought that other
factors, such as the lack of a two-generation reproductive toxicity study, might justify the use of
A-13