AEGL-1

The comment that response to sensory irritants is not expected to vary greatly among individuals (page 41, lines 13-18) is not correct. There is often wide variability in responses to such chemicals. The wording used to justify the use of an uncertainty factor of 3 for intraspecies variability should be modified to explain the reason for departing from the default value of 10 (see Standing Operating Procedures [NRC 2001, pages 87-88).

AEGL-2

The committee approved the derivation of the AEGL-2 values for epichlorohydrin.

AEGL-3

The National Institute for Public Health and the Environment (RIVM) in The Netherlands recently published a report (Ruijten 2009) in which the lethality studies used in the draft TSD as the basis of the AEGL-3 values for epichlorohydrin are considered more “uncertain” and “not qualitatively superior” to data from an unpublished study by Kimmerle (1967), which was cited in the RIVM report but not included in AEGL TSD. Those data include a 1-h NOAEL for lethality, plus lethal concentrations. The RIVM evaluation (using the ten Berge software) produced values in good agreement with reported data. This information might be useful in refining the derivation of the AEGL-3. The analysis in the RIVM report should be explicitly considered in the draft TSD with rationale given for accepting or rejecting it.

Confidence in the 8-h AEGL-3 value would be increased if alternative PODs were evaluated (that is PODs other than a no-effect level in a lifetime exposure study with total uncertainty factor of 1) in a manner similar to that done with the 6-h exposure data in the Laskin et al. (1980) study. See, for instance, the studies referenced in Ruijten (2009) or those listed in the Centers for Disease Control and Prevention’s Registry of Toxic Effects of Chemical Substances (http://www.cdc.gov/nioshrtecs/TX4AC4a0.html).

Responses to Previous Comments

In general, the responses were clear, and the changes made to the draft TSD were appropriate. Exceptions are noted in the following three items.

Page 15, lines 29-30 and 34-35: A previous comment dealt with chromosome aberrations and whether they are long-lived effects. The specific citation has been addressed, but disparate statements remain in these lines. An easy resolution might be to change the wording of line 35 to read “long-term clinical effects” or some similar phrasing.

Page 26, lines 7-10: Although the rewritten section on rat kidney damage (or the lack thereof) is responsive to previous comments, the wording is ambiguous. Is the conclusion that the observed changes are not toxicologically significant attributable to the authors of the study (Robinson et al. 1995) or to the professional judgment of the author(s) of the draft TSD? These two sentences should be rewritten to clarify this point.

Page 38, lines 28-29: The second sentence here begins “All three compounds,” but the previous sentence merely lists the synonyms for epichlorohydrin, a single compound (unless the two optical enantiomers are counted separately). The first sentence should be reworded to emphasize and make explicit the point made here: “Structurally, epichlorohydrin can be related to either ethylene oxide or propylene oxide (i.e., either as chloromethyl ethylene oxide or as chlorinated propylene oxide).” The chemical nomenclature using “oxirane” can be discussed in the Introduction, if desired, but it adds nothing to this section.

Potential Carcinogenicity

Page 7, lines 41-46: The statement regarding the cancer unit risk for epichlorohydrin should disclose that it is for risks at the 1 in 10,000 level (10^-4), and should state that this is the level of risk most relevant for emergency exposure and response purposes (see Standing Operating Procedures, Section 2.8.4).

Page 13, lines 6-8: “The U.S. Environmental Protection Agency considered the human data to be inadequate for evaluating the carcinogenicity of epichlorohydrin (U.S. EPA 2006).” The sentence, although correct with regard to the human evidence, does not capture EPA’s full evaluation. EPA’s Integrated Risk Information System (IRIS) database shows that as of 1992 epichlorohydrin has been classified as a B2 carcinogen (probable human carcinogen, on the basis of sufficient evidence of carcinogenicity in animals). This statement should be added to the TSD.

Page 13, lines 21-22: The description of the International Agency for Research on Cancer (IARC) review of epichlorohydrin, as with the EPA IRIS citation above, is technically correct but substantively incomplete. The TSD should state that IARC categorized epichlorohydrin as “probably carcinogenic to humans (Group 2A).”

It is also noteworthy that the TLV for epichlorohydrin carries a carcinogen classification of A3 (confirmed animal carcinogen with unknown relevance to humans [since 1997]).

Page 58, Appendix B: The equation and calculations of the virtually safe dose and subsequent calculations should be checked for accuracy. The relevant AEGL values for noncancer effects should be presented in the table reporting the values associated with cancer risks of 10^-4, 10^-5, and 10^-6 to facilitate comparisons. It would be preferable to structure the table in the traditional format of presenting AEGL values: The exposure durations should be the column headings, and the AEGL values and cancer risk levels should constitute the row designations; entries should be sorted by ascending level of risk using the 1-h values.

Odor Issues

Section 2.2.1, page 10: Another useful reference for odor thresholds is Ruth (1986).

Page 10, lines 19-20, and PAGE 53, lines 26-27: Van Doorn et al. (2002) is cited as the source for determining the level of distinct odor awareness, but the reference is to an unpublished report with no other source information. As cited, this reference is of limited use. A 2009 version of this report (Ruijten et al. 2009) has since been published, so the updated citation should be used.

Use of Tables to Present Data

Tables are important in examining and making comparisons of the data, especially when comparing the exposure conditions and results across several studies. Many studies (or useful data points) are discussed in the text but are not included in existing tables. For example, Table 3 does not include the studies by Freuder and Leake (1941) described on page 19 of the TSD or the Mobay Chemical Corporation (1983) study described on page 20.

Table 7 summarizes nonlethal effects, but mixes both single and repeat exposure studies. The data presentation would be easier to follow if these were separated into two tables. These tables should summarize all components of the studies regardless of where they are discussed in the text, unless good reason exists for excluding them (which should be clearly stated). Examples of places where one paper is discussed in multiple sections include the following:

• The results of the UCC (1983) study with rats are described on page 19, with monkeys (acute and repeated exposures) on page 25, and with dogs on page 31.
• The Mobay Chemical Corporation (1983) 5-day exposure studies with rats are described on page 19, with mice and guinea pigs on page 21, with rabbits on page 22, and with cats on page 23.
• The John et al. (1983) study with rabbits is described on page 22 and with rats on page 28.
• The Gage (1959) study with rats is described on page 28 and with rabbits on page 31.

Sometimes errors are introduced as summary tables are constructed, or useful data points are overlooked. For example,

• In Table 3, the approximate lethal concentration for the dog in the UCC (1983) study (430 ppm) was excluded.
• In Table 7, the Quast et al. rat study is cited as 1979b but is described on page 28 as 1979a.
• In Table 7, the Quast et al. (1979b) mouse study described on page 29 should include a note that there were 9 exposure days during the 12-day study period.

Concern about the utility of aerosol exposure data (or the comparability of vapor data with aerosol data) should not lead to exclusion of the data; it can be presented in a separate table(s) if there are several studies or appropriately annotated in existing tables.

General Comments

Page 42, line 33: “Studies using aerosols should not be used to derive AEGL values.” It is unclear why aerosol data were excluded from consideration. The AEGL Standing Operating Procedures appear to support the use of such data. For example, the procedures state, “Therefore, no dosimetry adjustments have been made to date by the NAC-AEGL Committee for attaining human-equivalent doses in the development of AEGLs for gases, vapors, and aerosols” (p. 57). The implication is that AEGLs might be developed for aerosols. The procedures also state, “The determination of susceptibility entails the presence of observable changes in biochemical or physiologic processes reflecting dose-response relationships unique to a chemical (e.g., sulfur dioxide) or class of chemicals (e.g., acid aerosols)” (p. 81).

The TSD for titanium tetrachloride (discussed elsewhere in this report) involves an aerosol exposure; a second example is the metal phosphides (NRC 2008). A number of chemicals can occur as a vapor, an aerosol, or both (Perez and Soderholm 1991).