Eleventh Interim Report of the Subcommittee on Acute Exposure Guideline Levels (2004)

In 1991, the U.S. Environmental Protection Agency (EPA) and the Agency for Toxic Substances and Disease Registry (ATSDR) asked the National Research Council (NRC) to provide technical guidance for establishing community emergency exposure levels (CEELs) for extremely hazardous substances (EHSs) pursuant to the Superfund Amendments and Reauthorization Act of 1986. In response to that request, a subcommittee of the NRC Committee on Toxicology prepared a report titled Guidelines for Developing Community Emergency Exposure Levels for Hazardous Substances (NRC 1993). That report provides step-by-step guidance for the derivation of CEELs for EHSs.

In 1995, EPA, several other federal and state agencies, and several private organizations convened an advisory committee—the National Advisory Committee on Acute Exposure Guideline Levels (AEGLs) for Hazardous Substances (referred to as the NAC)—to develop, review, and approve AEGLs (similar to CEELs) for up to 400 EHSs. AEGLs developed by the NAC have a broad array of potential applications for federal, state, and local governments and for the private sector. AEGLs are needed for prevention and emergency response planning for potential releases of EHSs, either from accidents or as a result of terrorist activities.

The NRC convened the Subcommittee on Acute Exposure Guideline Levels to review the AEGL documents approved by the NAC. The subcommittee members were selected for their expertise in toxicology, pharmacology, medicine, industrial hygiene, biostatistics, risk assessment, and risk communication.

The charge to the subcommittee is to (1) review AEGLs developed by the NAC for scientific validity, completeness, and conformance to the NRC (1993) guidelines report, (2) identify priorities for research to fill data gaps, and (3) identify guidance issues that may require modification or further development based on the toxicological database for the chemicals reviewed.

This interim report presents the subcommittee’s comments concerning the NAC’s draft AEGL documents for 11 chemicals: chloromethyl methyl ether, jet-propulsion fuel 8, tetranitromethane, carbon monoxide, acetone cyanohydrin, monochloroacetic acid, phosphorus trichloride, phosphorus oxychloride, fluorine, cis-1, 2-dichloroethylene, trans-1, 2-dichloroethylene, and acrylic acid.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on chloromethyl methyl ether (CMME). The presentation was made by Sylvia Milanez of Oak Ridge National Laboratory. The subcommittee recommends a number of revisions. A revised draft would be reviewed by the subcommittee at its next meeting.

The subcommittee recommends that comments on the AEGL values that would be derived on the basis of the toxicity for bis-chloromethyl ether (BCME) be added to the document. The AEGLs for BCME are expected to be similar to those calculated for CMME, after adjusting for the level of contamination. The explanations for not deriving an AEGL-1 need to be consistent throughout the document. Because CMME is classified as a human carcinogen, more detail on the cancer risk assessment and how it impacts on the AEGL values should be included in the main text of the document. The subcommittee recommends that data from single dose studies, including those involving BCME, be examined in more detail prior to AEGL-2. The use of single exposure studies might lead to AEGL-2 values exceeding those calculated on the basis of carcinogenicity. The sensitivity of the cancer risk assessment to different BCME contamination levels should also be examined. It should also be noted that the International Agency for Research on Cancer (IARC) classifies CMME as a “possible human carcinogen,” category 2B.

The subcommittee was concerned about the use of a repeat exposure study (Drew et al. 1975) to derive the AEGL-2 values and suggests recalculating the AEGL-2 values using the single 7-hour exposure study in which exposures to CMME ranged from 0.7 to 9.5 ppm (Drew et al. 1975) and the single exposure studies with BCME. There was concern that the repeat exposure study might be inappropriate for the derivation of AEGL-2 values. At the concentration used to derive AEGL-2 values, 2 of 25 rats died, and death is an AEGL-3 effect. Three possible solutions were discussed:

1. The easiest solution, but perhaps not the best, is to make a clear statement of why the repeat exposure study was used as the basis for AEGL-2 despite the occurrence of this AEGL-3 effect (e.g., by arguing that in the view of the author there is no good single exposure study, so, this multiple exposure study was chosen). Because of the repeated exposures, 2 of 25 rats died at 1 ppm, but that can be disregarded because a single exposure study acceptable for deriving AEGL-3 showed a LC₀₁ (lethal concentration in one percent of the exposed animals) of 15 ppm (Drew et al. 1975).

2. An alternative solution is to use the lung edema (lung-to-body weight change) from the single exposure study (Drew et al. 1975) to derive AEGL-2 values if that effect can be

considered serious and long lasting. (At 12.5 ppm, there was no mortality and no increase in lung-to-body weight ratios.)

3. Use the Leong et al. (1975, 1981) studies that show that BCME at 10 ppb did not cause AEGL-2 effects in rats and mice.

Uncertainty factors (UFs) of 3 were used to account for both species differences (interspecies) and intraspecies variability in the derivation of the AEGL-2 and AEGL-3 values. The reasons given for using those values need to be expanded and strengthened to justify deviating from the default value of 10. For AEGL-2, the reason given for using an interspecies UF of 3 is that “the key study is a repeat-exposure; CMME is a local-acting irritant (hydrolyzes in situ) and metabolism is unlikely to play a role in its toxicity” (page 24, lines 31–32). For AEGL-3, the reason given is that “rat and hamster yielded similar LC₅₀ values in key study” (page 27, lines 30–31). Perhaps the intended meaning here is that the basis for choosing an interspecies UF of 3 was the same for both AEGL-2 and AEGL-3, but it does not come across that way. The reasons given should be consistent (if in fact the reasons are the same) and clearly stated. The basis for choosing the UF for AEGL-2 is especially troubling. That two animal species had similar LC₅₀ values is not an adequate basis for determining that humans will respond in a similar manner to those two species. The reasons given for choosing these UFs needs to be expanded and strengthened to justify deviating from the default UF of 10.

For intraspecies variability, the same reason is given for using a UF of 3 in AEGL-2 and in AEGL-3 values—response to an irritant gas hydrolyzed in situ “is not likely to vary greatly among humans” (page 28, line 23 and page 29, lines 1–2). The document has not established (with documentation) that response to an irritant gas is not likely to vary greatly among humans or that this is a sufficient basis for deviating from the default UF of 10. The reasons given for using these values need to be expanded and strengthened.

The reasons and justifications presented for why AEGL-1 values were not calculated are inconsistent and thus confusing. The reason given in the text is that “there were no inhalation exposure studies with technical grade CMME that produced end points consistent with the definition of AEGL-1” (page 23). This description differs from the statements in the Executive Summary, in the summary table in the Executive Summary (page vii), and on page 29. That toxicity occurs below the odor threshold might be a reason to consider as well.

A summary of the AEGLs values derived using cancer as a critical end point should be included in the text along with some analysis of the data. The text includes only a short statement that the calculation was done and that it did not impact the outcome. The risk level for each AEGL should be included along with a short discussion of its impact on the overall derivation of the AEGL values.

The discussion of cancer assessment is not clearly presented. The design of the Kuschner study is unclear; the risk is incorrectly formatted as “10x ⁻³”; the risk calculations are based on an assumption of 8% BCME, not on “pure” CMME, which needs to be explained better; it is unclear if the 6-fold modifier is appropriate for this substance; the NAC needs to determine if unit cancer risk is based on data from the 10-exposure study.

Page vii, line 25. Two out of how many rats?

Page vii, line 34. “AEGL-1 values were not derived” is repetitive (just stated above).

Page viii, lines 9–10. Eliminate “of small populations in limited geographic areas.”

Page 1, line 4. Add “which, however, is noted only at concentrations exceeding the lowest life-threatening concentrations” after “…with an irritating odor.”

Page 1, lines 7–8. Add citation for the statement “Acute exposure can lead to delayed fatal pulmonary edema.”

Page 1, lines 12–13. The t 1/2 for hydrolysis of CMME in water was extrapolated from that in aqueous isopropanol to be…?

Page 7, line 22. Typo: “form” should be “from.”

Page 12, lines 4–5. The exposure time for the Drew et al. (1975) study (30-day exposure study) is given in the original paper. The exposure time was 6-hours per day (see Table 5, page 65 of original Drew et al. [1975] paper).

Page 13, line 33. Give full terms instead of abbreviations.

Page 14, Section 3.2.2. The second paragraph describes a study where mice were exposed for 6 hours to CMME at 14.6 to 100 ppm and none of the mice died within 14 days. Isn’t this a lethality study (even though it was negative)? Why not include this study in the lethality section? See page 10, lines 40–41: “Inhalation of 100 ppm CMME can apparently be tolerated (i.e., death does not occur) by animals for several hours (Hake and Rowe 1963).”

Page 15, line 6. How much CMME?

Page 15, line 16. How much CMME?

Page 15, line 18. How much CMME?

Page 15, line 30. Explain “demeanor.”

Page 20, line 4. “CMME was degraded.” Within what time frame?

Page 20, lines 18–21. This does not appear to be fully plausible. Hence, it may be appropriate to write “It has been reported that the higher carcinogenicity” instead.

Page 20, line 39. The text states that “the study by Drew et al. (1975) indicated that there is not a great deal of variability between species.” Drew et al. found that there was little difference in response and variability between two species—rats and hamsters. That is not sufficient evidence that humans will respond in the same way. See comment above.

Page 21, Section 4.4.5, Neoplastic Potential of CMME by Other Routes of Exposure. In the past, we have typically not included data for routes of exposure other than inhalation. Is this meant to be supportive of the inhalation data?

Page 23, lines 1–3. Perhaps a better argument would be “because concentrations leading to AEGL-1 effects are higher than those already causing AEGL-2 effects.”

Page 23, lines 31–32. What was not specified?

Page 25, line 23. Carcinogenicity calculations are in Appendix C, not B.

Page 26, last paragraph. Shouldn’t this discussion mention that the median lifespan for animals exposed at both 6.9 and 9.5 ppm was 2 days? That is important.

Page 28, lines 31–43. Summary explains more details than were explained in the preceding full text. Move part of the Summary to page 24 and 27.

Page 30, Table 8, Extant Standards and Guidelines for CMME. Why not include and comment on the OSHA standard of 0.1 ppb for BCME?

Page 30, lines 23–24. Should read “at concentrations below those leading to irritation and below the odor detection”

Page C-2, Appendix C. Why was no UF used to account for intraspecies differences in the cancer risk assessment?

Page E-2, Category Plot for CMME. The scale for this graph is incorrect. The axis values below 1 should not be zeros.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on jet-propulsion fuel 8 (JP-8). The document was presented by Sylvia Talmage of Oak Ridge National Laboratory. The subcommittee recommends a number of revisions. The subcommittee will review the revised AEGLs draft at its next meeting.

JP-8 is significantly different from JP-4. JP-8 is essentially kerosene with additives, whereas JP-4 is a mixture of gasoline (65%) and kerosene (35%) with additives. Much of the toxicological data presented in the document for JP-4 is associated with the light-end fraction (gasoline). That discussion should be eliminated. A better comparison is with kerosene or Jet A and Jet A-1

Because JP-8 has a very low vapor pressure, high concentrations are probably associated with aerosols. The toxicology of the lighter-end vapor components to the full-mixture aerosols makes exposure generation and evaluation of the data very complex.

An interspecies uncertainty factor (UF) of 1 was used to derive the AEGL-2 values. The reason given for making that decision is that no adverse effects were observed and the exposures were repeated for up to 90 days (page 46, lines 22–23). This explanation is not convincing. Using an interspecies UF of 1 when relying on animal data is a major departure from past decisions as well as from the SOP. The subcommittee does not agree with the argument. If the NAC believes the UF of 1 is correct, the reasons for its use have to be clear and well stated. The current explanation is insufficient.

The use of the Alarie’s 10-fold reduction factor is troubling. This procedure needs to be more fully described. As written, its basis is unclear. How well-established and accepted is it among members of the scientific community? Is it appropriate to use it here? Furthermore, there is no discussion of whether this one-time conversion step takes into account variability among species, such as effects on children and the elderly. The explanation for why no UFs were used is simplistic and unclear to anyone unfamiliar with this procedure.

In the example used to support the AEGL-1 values derived from the Alarie 10-fold reduction factor, an interspecies UF of 1 was also used to derive AEGL-1. The justification given for that decision is that “no species differences were observed in multiple studies with rats and mice and the exposures were repeated” (page 45, lines 19–20). This is unconvincing. As stated above for AEGL-2, using an interspecies UF of 1 when relying on animal data is a major departure from past decisions as well as from the SOP. The subcommittee does not agree with the argument provided. If the NAC believes the UF of 1 is correct, the reasons for its use have to be clear and well stated. The current explanation is insufficient.

Page 8, line 37. The subcommittee questions “18% aromatic hydrocarbons,” specifically calling out benzene, ethylbenzene, toluene, and xylenes. Although those components might be present, there are probably more higher-ringed components and if so, that should also be taken into consideration. The distillation fraction of JP-8 should minimize the presence of benzene and low-boiling aromatic hydrocarbons.

Page 12, line 1. Reference to Table 2 implies that the table includes information on health effects. (“Olsen compared the health of 18 Air Force personnel exposed to jet fuels with 18 non-exposed subjects.”) There are no health effects data in this table. Take out the reference or be more clear about why you are referring the reader to this table.

Page 12, lines 9–10. It is stated here that subjects exposed to JP-8 had both “lower” and “higher” mean corpuscular hemoglobin. Which is correct?

Page 12, lines 15–16. Enzymes were elevated in both groups but there are no differences? Either explain or eliminate as unnecessary information.

Page 12, lines 26–30. This example—comparing effects following exposure to unleaded gasoline—is 60 years old (1943). Isn’t there a more recent study that could be used here?

Pages 13–14, Table 2. Add data on benzene, xylene, toluene reported as part of these monitoring studies. Why are those data not included? This comment assumes that these substances make up a significant percentage of JP-8.

Page 14, lines 16–21. Benzene is extremely difficult to separate from complex mixtures, especially in air, where the mixture is first trapped on an adsorptive media. This is especially true for short-term samples when quantities trapped are at or generally below the detection level. The appropriate analytical method is capillary gas chromatography with mass spectroscopy. Otherwise, identification by retention time leads to biased high results.

Page 15, lines 15–26. Which hydrocarbons did Carlton and Smith (2000) monitor?

Page 16, lines 1–3. Were any adverse health effects reported in Richie et al.?

Page 16, lines 10–28. What is the toxicological significance of the effect of JP-8 on postural sway Smith et al. reported? Given that this is apparently not a manifestation of acute CNS effects (lines 23–25), does it reflect residual neurological dysfunction? How much time typically elapsed between the last JP-8 exposure and testing?

Page 16, lines 17–18. Are benzene, toluene, and xylene exposures associated with postural sway? What about peripheral neuropathy?

Page 16, line 30. It should probably be noted here (and in similar instances) that the findings of McInturf et al. (2001) were only reported in an abstract.

Page 17, line 7. What is the origin of the 1,1,1-trichloroethane?

Page 17, lines 13–14. Are these common dilutions? Is 1:75 toxic?

Page 17, lines 23–24. Painters have a totally different exposure that includes ketones and aromatics, especially toluene and xylene.

Page 17, line 34. JP-4 is closer to aviation gasoline than JP-8.

Page 17, Section 2.8, Carcinogenicity. What about data on the carcinogenicity of kerosene? ATSDR (1998) summarizes several studies that are not mentioned here. Why not include these as part of the document? Also, what about benzene, a component of JP-8? Should there be some mention of the carcinogenicity of JP-8’s components?

Page 18, lines 21–22. How were aerosols measured? It seems strange to encounter aerosols during refueling. What is the aerosol generating process?

Page 18, lines 26–27. This statement is incorrect. Statistically significant associations were found between exposures and increased postural sway, particularly for the components benzene, toluene, and xylene (page 16, lines 17–18).

Page 19, line 13. It is surprising that the undiluted fuels did not produce eye irritation, because MacEwen and Vernot (1985) reported eye irritation in mice and rats exposed for 1 hour to JP-5 at 625 ppm. Other investigators have described eye irritation from exposure to JP-8.

Page 19, lines 14–5. Were the undiluted fuels placed onto abraded or unabraded skin? Were occlusive patches applied?

Page 20, lines 8–17. This is good methodology for testing and documenting the animal exposure.

Page 22, Table 3. The NAC should list and describe the studies on JP-8 first and then include the other jet fuels. The focus of the document is on JP-8, so data on that fuel should be discussed first in the text and summarized first in the tables. It is distracting to have to go through all the data on JP-4 before getting to the data on JP-8.

Page 27. Discussion on the toxicity of JP-4. Same comment as above. Start with studies on JP-8 and then include studies on other jet fuels.

Page 28, lines 25–37. This paragraph describes effects of exposure to JP-5. Switch this paragraph with the next paragraph so that the two JP-5 studies are described together and the JP-8 studies are discussed as a group.

Page 29, line 27. The upper bound of inhaled JP-8 concentration is said here to be 1,084 mg/m³, whereas in line 2 of the table on page 25 a value of 1,094 mg/m³ appears.

Page 30, lines 32–33. It is believed that aerosols can pass through charcoal because they are trapped in the air stream and do not have enough Brownian motion to hit the charcoal. It is not clear how the authors distinguished between aerosol and vapor, although the conclusion that the vapor contained more light components than the vapor/aerosol is not remarkable—it is exactly what should happen.

Page 31, lines 12–14. What is meant by “a saturation of the chemical in the nasal passages”?

Page 31, lines 15–18. n-nonane is C₉. What are they trying to say?

Page 31, lines 15–21. If the RD₅₀ decreases with increasing lipophilicity of alkanes (e.g., for heptane to octane), how do this document’s authors account for the inability of nonane, decane, and undecane to reduce respiratory rate by 50%?

Page 31, lines 22–34, and page 41, lines 3–5. Has the toxicological significance of the changes in protein expression been established, or even postulated? It would be a good idea to state that the observed changes are of uncertain significance, if that is the case.

Page 32, line 3. Give a brief explanation of BALF fluid analysis.

Page 34, lines 2–4. “The rats got stronger”?

Pages 35–36, Section 3.4, Immunotoxicity. Was consideration given to using the immune studies to derive the AEGL values (Robledo and Whitten 1998 or Harris 2001)? Could those studies be used to derive AEGL values?

Page 40, lines 12–13. This sentence does not make sense. It should be rewritten.

Page 40, lines 15–16. Explain why “the male rat nephropathy and resulting kidney cancer associated with exposure to jet fuels is not relevant to humans.” It is not obvious to everyone.

Page 41, lines 8–9. Although there are apparently no published absorption, distribution, metabolism, and excretion (ADME) studies of JP fuels, there have been investigations of the pharmacokinetics of combinations of ≥3 aromatics and/or long-chain aliphatics (Pedersen et al. 1984; Zahlsen et al. 1990, 1992; Lof et al. 1999). There have also been efforts to develop PBPK models for such mixtures (Tardif et al. 1997; Haddad et al. 2001).

Page 41, line 37, and page 42, lines 8–10. It would be worth mentioning the “time-honored” mechanism of hydrocarbon-induced CNS depression. Because hydrocarbons are lipophilic, they partition into and accumulate in neuronal membranes and myelin. The more lipophilic the hydrocarbon is (i.e., the higher its neuronal tissue:blood partition coefficient), the more potent a CNS depressant it is. The mere presence of hydrocarbons has generally been thought to disrupt the ability of the neuron to propagate an action potential and repolarize. Recent research has revealed that hydrocarbons might act by more specific mechanisms and might affect specific neurotransmitters and membrane receptors. Hypotheses and pertinent experimental results have been published by a number of researchers, including Mihic et al. (1994), Engelke et al. (1996), Cruz et al. (1998), and Balster (1998).

Page 42, lines 13–26. The potential relationship of the male-rat-specific nephropathy and nephrocarcinogenicity should be mentioned here.

Page 42, line 30. The phrase “to acute effects” should be inserted between “susceptibilities” and “usually.”

Page 43, lines 2–3 The meaning of the following is unclear “N-acetyltransferase should have no effect on toxicity based on the composition of JP-8.”

Page 43, lines 22–24. What is the basis for postulating that CNS depression is less dependent on duration of exposure than on exposure concentration? Duration of exposure is also a major determinant of dose until near steady-state or equilibrium is reached.

Page 44, lines 12, 23. It is not clear how aerosols are generated during refueling. Unless that can be explained, the subcommittee recommends deleting references to refueling. It could be an issue during cold starts and also during foam replacement when saturated foam is pulled from the fuel tanks through small openings.

Page 44, line 29. The phrase “exposures to JP-8 were generally low” is not germane to determining whether there are well-designed studies that identify reversible health problems relevant to AEGL-1 values. That phrase should be deleted from the text.

Page 45, lines 8–11 The 10-fold reduction factor of Alarie et al. (1981) essentially includes a 3-fold factor for both intraspecies and interspecies variability.

Page 46, lines 24–25 It would be worthwhile to point out that doses of volatile organic hydrocarbons (VOCs) absorbed systemically are considerably greater in mice and rats than in humans subjected to equivalent inhalation exposures. This is attributable to rodents’ relatively high respiratory rates, cardiac outputs, and blood:air partition coefficients. Therefore, no-effect levels for CNS depression in rodents are quite protective for humans.

Page 47, line 7. Omit the word “lethal.”

Page 47, lines 12–13, and 21–22: It is not accurate to state or imply that lethal JP-8 levels might be attained in confined spaces at high ambient temperatures. The highest vapor (only) concentration Wolfe et al. (1996) could generate was 3,700 mg/m³. A 4-hour exposure at that concentration was not lethal to any of a group of male rats. Most vapor levels measured within airplane fuel tanks are considerably lower. The only conceivable way a person might receive a lethal inhaled dose of JP-8 would be to inhale a very high aerosol concentration for a prolonged period of time. That is unlikely to occur, because the aerosol would be irritating to mucus membranes, and the individual would leave the situation.

Page 50, lines 36–38. The statement that “exposure to higher concentrations of jet fuels occurred only when personnel were wearing respirators, thus negating the inhalation exposure route” is unclear and potentially misleading. The issue here is that workers wore respirators and thus adverse health effects were not observed; therefore, information on adverse health

effects could not be derived from these studies. The statement is repeated several times in the text, but the point is never made clearly.

The subcommittee believes that the JP-4 data do not add substantially to the scientific arguments made in the document for JP-8.

Because AEGLs generally have not been previously based on RD₅₀ levels, the subcommittee recommends that the NAC provide a more detailed explanation of the rationale for deriving the AGEL-1 from the RD₅₀. Specifically, the data from the Alarie study that support dividing the RD₅₀ by a factor of 10 should be discussed. It could be explained that the Alarie data are consistent with human data showing no sensory irritation for other compounds, and a clear argument could be made for why no intraspecies UF was applied.

According to the NAC’s general rules for solvents, small animals receive a larger dose of direct acting compounds that affect the CNS than humans because of their more rapid respiration rate, which induces CNS effects. Therefore, the interspecies UF should be 2 or less; however, the intraspecies factor should be greater than 3.

An interspecies UF of 1 was used to derive AEGL-2. The reason given for that decision was that no adverse effects were observed and the exposures were repeated for up to 90 days (page 46, lines 22–23). This explanation is not convincing. Using an interspecies UF of 1 when relying on animal data is a major departure from past decisions as well as from the SOP. The subcommittee does not agree with the argument. If the NAC believes the UF of 1 is correct, the reasons for its use have to be clear and well stated. The current explanation is insufficient.

Further justification for the lack of time scaling for AEGL-2 (which is based on systemic toxicity) is also needed.

The AEGL-2 seems a little high if based on escape impairment rather than irreversible health effects. The RD₅₀ in mice was 2,876 mg/m³, and the recommended AEGL-2 was 1,100 mg/m³. This is a very high percentage of the RD₅₀, and it might be satisfactory for healthy military personnel, but AEGLs apply to the general population. A suggested range is one-fourth or one-fifth of the RD₅₀, unless there was data in asthmatics or more RD₅₀-type data. Therefore, the subcommittee recommends a value in the range of 500 to 700 mg/m³ for AEGL-2.

Page 47, line 9. It is unlikely oxygen deprivation is a possible cause of death. Assuming 15% oxygen is the limit for death from oxygen depletion (actually closer to 12%), that would require 15% oxygen, 60% nitrogen, and 25% JP-8. The JP-8 concentration would be 250,000 ppm. The victim would more likely be anesthetized to death.

Balster, R.L. 1998. Neural basis of inhalant abuse. Drug Alcohol Dep. 51:207–214.

Cruz, S.L., T.Mirshahi, and B.Thomas. 1998. Effects of the abused solvent toluene on recombinant N-methyl-D-aspartate and non-N-methyl-D-aspartate receptors expressed in Xenopus oocytes. J.Pharmacol. Exp. Therap. 286:334–340.

Engelke, M., H.Tahti, and L.Vaalavirta. 1996. Perturbation of artificial and biological membranes by organic compounds of aliphatic, alicyclic and aromatic structure. Toxicol. In Vitro 10:111–115.

Haddad, S., M.Beliveau, R.Tardif, and K.Krishnan. 2001. A PBPK modeling-based approach for interactions in health risk assessment of chemical mixtures. Toxicol. Sci. 63:125–131.

Lof, A., H.Lam, E.Gullstrand, G.Ostergaard, and O.Ladefoged. 1999. Distribution of dearomatised white spirit in brain, blood and fat tissue after repeated exposure of rats. Pharmacol. Toxicol. 85:92–97.

Mihic, S.J., S.J.McQuilkin, and E.I.Eger. 1994. Potentiation of γ-aminobutyric acid type A receptor-mediated chloride currents by novel halogenated compounds correlated with their abilities to induce general anesthesia. Mol. Pharmacol. 46:851–857.

Pedersen, L., A.Larsen, and K-H Cohr. 1984. Kinetics of white spirit in human fat and blood during short-term experimental exposure. Acta Pharmacol. Toxicol. 55:308–316.

Tardif, R., G.Charest-Tardif, J.Brodeur, and K.Krishnan. 1997. Physiologically based pharmacokinetic modeling of a ternary mixture of alkyl benzenes. Toxicol. Appl. Pharmacol. 144:120–134.

Zahlsen, K., A.Nilsen, I.Edie, and O.Nilsen. 1990. Accumulation and distribution of aliphatic (n-nonane), aromatic (1,2,4-trimethylbenzene), and naphthenic (1,2,4-triethylcyclohexane) hydrocarbons in the rat after repeated inhalation. Pharmacol. Toxicol. 67:436–440.

Zahlsen, K., I.Edie, A.Nilsen, and O.Nilsen. 1992. Inhalation kinetics of C6 to C10 aliphatic, aromatic and naphthenic hydrocarbons in rats after repeated exposures. Pharmacol. Toxicol. 71:144–149.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on tetranitromethane. The document was presented by Sylvia Milanez of Oak Ridge National

Laboratory. The document can be finalized after the subcommittee’s recommended revisions have been made appropriately.

Page vi, line 14, 16, and throughout the document. The use of the term “threshold” is not justified in this context. For lethality, NOEL is more appropriate.

Page 17, lines 23–34. The AEGL-2 was derived from a 5-ppm concentration in a multiple dose mouse study (6 hours per day, 5 days per week for 2 weeks). This is appropriate when no adequate single dose study is available. However, there is a single dose study (Kinkead et al. [page 4, line 4 from bottom]) in which rats exposed to TNM at 10 ppm for 4 hours lost weight the first 4 days and then recovered and had mild lung congestion. Those effects are not severe or long-lasting and hence are suitable for AEGL-2 derivation (at the next higher concentration of 15 ppm three often rats died). Corresponding changes should be made throughout the document.

Page 18, last two paragraphs, and page 19, first two paragraphs. The multiple exposure study (NTP 1990) was used as the basis for AEGL-3 (10 ppm is considered the highest concentration below lethality, but one of five mice died). The single exposure study of Kinkead et al. (page 18, last two paragraphs) from which a 4-hour LC₀₁ can be calculated (9.4 ppm for rats) is a more appropriate basis. Rats are clearly more sensitive than mice (4-hour LC₀₁ of 20.5 ppm), and experimentally, there are no deaths in rats at 10 ppm for 4 hours.

Page 10, lines 2–7. Some quantitative statement would be useful (in the crudest terms, strong or weak, or even numerical).

Page 11, line 32. “dog pups exposed up to 6.25 ppm” for 6 hours.

Page 14, lines 1–2 and lines 4–6. Repetitive.

At its January 28–30, 2004, meeting, the subcommittee reviewed the revised AEGL document on carbon monoxide. The document was presented by Peter Griem of FOBIG GmbH, Germany. The subcommittee recommends a number of revisions. The revised document will be reviewed by the subcommittee at its next meeting.

The subcommittee concurs with the AEGLs proposed by the NAC, but recommends several clarifications relating to (1) the fact that exposures exceeding the 8-hour AEGL-2 could be experienced in a number of settings, (2) the footnote to the AEGL summary table indicating why an AEGL-1 was not derived, and (3) the apparent inconsistency between the 253-ppm exposure concentration in the key study for AEGL-2, which led to 4.4% carboxyhemoglobin (COHb), and the 1-hour AEGL-2 of 83 ppm.

The changes made after the last review of this document have significantly improved the document. Most of the comments made by the subcommittee were addressed and incorporated into the current draft. The reply to the subcommittee’s comments provides detailed explanations for the changes made in the earlier draft. Only a few issues that need to be addressed remain, as discussed below.

1. The reason given for not recommending values for AEGL-1 is “high acute toxicity without warning signs” (pages ix-x, Summary Table) of carbon monoxide. That description is vague, and the basis for not deriving AEGL-1 values needs to be better explained. The term “without warning signs” is especially unclear.

2. Consider the following text to address the fact that exposures exceeding the 8-hour AEGL-2 could be experienced in a number of settings:

In addition, there are a number of instances in the document, especially in the section on AEGL-3, where data from specific studies do not seem consistent with the proposed AEGL values. Examples include the studies by Dahms et al. (1993) that found that people exposed for 1 hour to CO at 292 +/− 31 ppm were fine; Chiodi et al. (1941) found that four adults exposed repeatedly to CO levels as high as 1,500 to 3,500 ppm for 70 minutes showed no adverse effects. Henderson et al. (1921) exposed volunteers to CO at levels as high as 1,000 ppm for 60 minutes. Some symptoms were reported at the upper exposure levels (600 to 1,000 ppm), but nothing serious. No symptoms were reported at levels up to 500 ppm. The AEGL-3 for 1 hour is 330 ppm. The findings in these studies do not seem consistent with this value. Explain or clarify.

3. The concentration associated with approximately 4% COHb seems to be 253 ppm (1 hour exposure) in the studies of Allred et al. (1989, 1991). However, the document reports a 1 hour AEGL-2 of 83 ppm on the basis of the CFK model. The basis for this inconsistency should be verified (see page 92).

4. There is a detailed discussion on “other susceptible populations” for AEGL-2 (page 49, line 1744, to page 51, line 1807) that concludes that a UF of 1 is sufficient “because the values are based on observations in the most susceptible human subpopulations (patients with coronary artery disease).” This conclusion is not obvious to the reader. Make a more direct comparison between the COHb levels reported in the studies of other susceptible populations and those found in patients with coronary artery disease. That would drive the point home more clearly. The wording used in the discussion for AEGL-3 (page 53, lines 1879–1882) is quite good an could be duplicated here.

Page 4, Section 2, Human Toxicity Data, Table 2. The units in this table (% COHb) do not match the units used to derive the AEGL values, which makes it difficult to compare the effects presented in the table with the AEGL values. Can the values in the table be converted to ppm or mg/m³ so that the data in the table are useful for comparisons? The table would be more valuable if it were more user-friendly.

Page 5–6, Section 2.1, Acute Lethality. Four of the case studies in this section describe situations in which the victims survived, and thus they do not appear to be lethality studies.

Why are they included in this section? The case studies are Grace and Platt (1981) (two cases studies); Ebisuno et al (1972); and Marius-Nunez (1990).

Page 37, lines 1342–1344. The sentence is not clear; either some wording is missing or there is a typo (“and” should be “an”).

Page 38, line 1388. Typo: “the” should be “there.”

Page 39, lines 1404–1405. It is not clear how the references to WHO (1999a) and EPA (2000) show that the derived AEGL values are protective of these risk groups. Include a statement or something from these sources that supports the conclusion.

Page 39, lines 1408–1410. Same comment as above.

Page 48, lines 1684–1699. The data in this paragraph are not consistent with the updated (since the last draft) data included on page 39. Update the statistics in this entire paragraph to be consistent with page 39.

Page 49, line 1746. The COHb percentage at AEGL-2 given here (4.5–5.2%) differs from the value given in the summary table in the Executive Summary and in Table 19 in Appendix B (4.9–5.2%). Correct whichever is incorrect.

Page 49, line 1746. Insert “in Appendix B” following the reference to Table 19.

Page 50, line 1772. Insert “only” between CO and because.

Page 50, line 1776. Change sentence to read “disease patients is likely to protect other elderly people.”

Page 50, line 1778. Insert “in Appendix B” following the reference to Table 19.

Page 50, line 1799. The COHb percentage at AEGL-2 given here (4.5–5.2%) differs from the value given in the summary table in the Executive Summary and in Table 19 in Appendix B (4.9–5.2%). Correct whichever is incorrect.

Page 53, line 1884. Insert “in Appendix B” following the reference to Table 22. Also, it is not clear how Table 22 shows that exposure to the derived AEGL-3 values will result in COHb values of about 14–17% in adults. Shouldn’t this statement refer to Table 21?

Page 53, line 1908. Insert “in Appendix B” following the reference to Table 21.

Page 58, Table 17. Several of the values are not consistent. Although changes have been made in this table to reflect that concern, there remain a number of values that are not transparent (see WHO Air Quality Guidelines). The NAC should discuss any differences.

Page 74, Appendix A. Can the corresponding COHb levels be added to the list of AEGL values in lines 2428–2432?

Page 81, Table 19. Do not split the table into two pages.

Pages 86–88, Appendix C. This appendix on electrocardiography is very detailed and seems out of place in the document. The subcommittee suggests dropping it, but make sure there is a clear definition of “ST segment” in the text.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on acetone cyanohydrin. The document was presented by Peter Griem of FOBIG GmbH, Germany. The subcommittee recommends minor revisions to the document. A revised draft can be finalized if the recommended revisions are made appropriately.

The mechanism of action of cyanide toxicity is misrepresented and needs revision. There are substantial misconceptions about the mechanism of cyanide intoxication evident in the draft document. Tachypnea is a consequence of respiratory stimulation at the carotid and aortic chemoreceptors. Those cells respond as they would to reduced PO₂. All of the pharmacological actions of cyanide result from cyanide’s reversible complex with the ferric (+3) state of mitochondrial cytochrome c oxidase also known as ferrocytochrome c-oxygen oxidoreductase. This enzyme is also known as cytochrome aas, and it is the terminal oxidase in aerobic metabolism of all animals, plants, yeasts, and some bacteria. This enzyme is a heme-copper lipoprotein and cytochromes a and as are combined in the same large oligomeric protein molecule. Mammalian cytochrome c oxidase contains two molecules of heme A and two copper atoms. This helical protein also contains 820 amino acids. The integrity of the disulfide groups to maintain the 30% helix structure is essential to the oxidase mechanism.

The interaction of cytochrome c oxidase with cytochrome c was reviewed by Lemberg (1969). The reaction proceeds by first-order kinetics with respect to the concentration of cytochrome c (Smith et al. 1979). Once absorbed, cyanide complexes with many metal ions and interferes with the activities of at least 39 heme-zinc, -copper, and -disulfide enzymes (e.g., catalase, peroxidase) whose activities depend on either metals as cofactors or prosthetic groups (Dixon and Webb 1964). Cyanide also binds to non-hematin metal-containing enzymes, like tyrosinase, ascorbic acid oxidase, xanthine oxidase, amino acid oxidase, formic dehydrogenase, and various phosphates. The cyanide concentration required for cytochrome c oxidase inhibition is 2–6 orders of magnitude less than that required for inhibition of these other enzymes. Thus, it is the critical position of cytochrome c oxidase in aerobic metabolism that makes its inhibition felt earliest, such that the effects of hydrogen cyanide (HCN) on other enzyme systems have scant chance to appear (Rieders 1971). The oxidase-HCN (not CN¯) (Stannard and Horecker 1948; Gibson and Greenwood 1963) complex is dissociable, (Swinyard 1975).

Administration of NaNO₂ oxidizes a modest fraction of circulating hemoglobin (Hb-Fe⁺²) to methemoglobin (MetHb) (Hb-Fe⁺³). Each of the four ferric heme groups of a molecule of MetHb reacts with a molecule of CN¯. The concentration gradient favors formation of cyanmethemoglobin; dissociation of the cytochrome c oxidase HCN complex occurs and resumption of oxidative metabolism then follows

(Gosselin et al. 1976). If the complex were not dissociable, then nitrite (a standard cyanide antidote) would be completely ineffective in treatment of these victims.

The AEGLs for acetone cyanohydrin should be based on HCN unless additional justification can be provided. Exposure to acetone cyanohydrin is like exposure to HCN in that rapid and near complete dissociation of the nitrile occurs in the presence of moisture. The human, nonhuman primate, and other animal data for HCN are more robust than that for cyanohydrin. Because the mechanism for acetone cyanohydrin poisoning is identical to that for HCN, the AEGL values should be equivalent on a molar basis. AEGL-2 and -3 values were derived on the basis of decomposition to HCN. Because of the lack of relevant data on acetone cyanohydrin and the fact that in toxicological assessments the toxicities of the two substances are indistinguishable, this approach is reasonable. On page 17, Section 5.2, describe in brief the AEGL-1 for HCN (NRC, 2002, pages 248–250).

Unless rigorous data can be presented to show that the irritation associated with acetone exposure (reviewed by ACGIH); Am. J. Ind. Med. 31(5):558–569, 1997; Am. Ind. Hyg. Assoc. J. 58(10):704–712, 1997) contributes to the reduced AEGL-1 or that a combined acetone and nitrile exposure contributes to headache or systemic toxicity as a result of cyanohydrin contact, the AEGL values at each level and duration of exposure should be equivalent to those for HCN (NRC, 2002). In section 4.4.1, some mention should be made of the insignificant contribution of the acetone and its metabolic products (Casazza et al. 1984; Gentry et al. 2003; Kosugi et al. 1986) to the toxicity of acetone cyanohydrin. Depending on the rate of presentation of the cyanohydrin/HCN vapor, the concentrations of acetone cyanohydrin/HCN sufficient to cause headache (NRC, 2002, pages 247–250) or acetone-associated irritation will only be slightly less than those that cause serious cyanide consequences.

The AEGL-1 values in the document are based on an assessment of the effects of acetone cyanohydrin in rats exposed for 5 days per week. There is little basis for the AEGL-1 in that study. Red nasal discharge is not consistently seen in any of the Monsanto studies and, when present, is not always dose-responsive. In addition, control animals vary widely in terms of whether that end point is present or not. In light of the variability of the red nasal discharge in repeat studies, it seems a poor end point on which to base the AEGL-1. Furthermore, the repeat exposures used in the Monsanto studies are not appropriate for the derivation of AEGL-1 values.

In light of the significant limitations of the toxicological data, it is recommended that AEGL-1 values be derived from HCN values, as was done for AEGL-2 and -3.

Because absorbed cyanide causes intoxication and death in all aerobic species by precisely the same mechanism of action, what are the specific “toxicodynamic differences” that justify the application of an uncertainty factor (UF) of 3 for AEGL-1 (page vi, lines 141–142)?

The reliance on HCN toxicity is appropriate. Some notation should be made by the AEGL values that the total of both acetone cyanohydrin and hydrogen cyanide concentrations should be measured and considered.

The very rapid breakdown of acetone cyanohydrin with moisture would present some challenges in any accidental spill or release. Because acetone cyanohydrin breaks down so readily to HCN, and the toxicity is due to HCN, both materials are present in a mixture and the ratio of the two could be rapidly changing. Therefore, both materials would need to be tracked to give an indication of the risk. Some notation should be made to explain that the AEGL values are based on rapid decomposition to HCN, and that detectors need to measure the total of both acetone cyanohydrin and HCN.

Page vi, lines 138–142. The sentence about species differences makes no sense. What are the specific toxicodynamic differences referenced here?

Page vi, line 145, and page vii, lines 146–148. The sentence about species differences as written here makes no sense.

Page vi, lines 142–143. Describe the duration of exposure and the circumstance (controlled study? Anecdotal report?) surrounding the El Ghawabi et al. (1975) results. What is the level of confidence the reader can place in the 10-ppm value as a cyanide NOAEL for humans?

Page vii. Why is it important that the specific dissociation of acetone cyanohydrin to free cyanide and acetone is “accelerated by heat”? Does that mean the dissociation occurs faster or to a greater extent on a hot summer day than it would on a very cold day? Is that not true of all endothermic chemical reactions? Water and ethanol exert specific dissociative effects on cyanohydrins, rather than acting as mere diluents (Jones 1914; Steward and Fontana 1940).

Page vii, line 155 and 156. Insert “it” to read “it was taken into account.

Page vii, line 161–165. Explain why the lack of respiratory distress is a part of this discussion. As written, it is unclear.

Page vii, lines 170–171. The sentence should be rewritten to indicate that, once absorbed, a dose of acetone cyanohydrin behaves in a manner identical to that of its molar equivalent in absorbed free cyanide.

Page vii, lines 158–159 and lines 169–170. The statement that “hydrogen cyanide has a far higher vapor pressure than acetone cyanohydrin” is not necessarily true. Because of its 9.2 pKa (Andregg 1957), cyanide is present as the nonvolatile CN¯ under alkaline conditions. Only when cyanide is present as the protonated species under neutral to acidic conditions is the associated HCN volatile. Thus, the vapor pressure argument presented here has no relation to the AEGL derivation. Acetone cyanohydrin behaves as its molar equivalent in cyanide under physiologic conditions.

Page 2, Section 2.2. Although it is not directly relevant to airborne acetone cyanohydrin, there is a wealth of data on the toxicity of acetone cyanohydrin ingested as a consequence of its liberation from linamarin (a cyanogenic glycoside found in cassava and other plant foodstuffs) (Conn 1979, 1981). Linamarin is the common name given to a molecule composed of glucose and acetone cyanohydrin.

Section 2.2. Sunderman and Kincaid (1953) described a number of nonfatal cases in plant operators who had contact with acetone cyanohydrin while packing pumps leading from storage tanks. It is not clear whether the three men described in Rohm and Haas (1969) (page 2, lines 264–266) are the same as those discussed by Sunderman and Kincaid (1953).

Section 2.2. This section should include accounts of the 19-year-old man who was exposed to 30–40 ml acetone cyanohydrin for 40–60 minutes (Lang and Stintzy 1960) and of the two cases described by Zeller et al. (1969).

Page 6, line 350. The prostration seen in these animals is not the result of simple mucous membrane irritation. This observation is associated with acute cyanide intoxication.

Page 10, lines 457–458. It is appropriate to summarize the discussion of the HCN carcinogenic potential from the NRC (2002) cyanide AEGL document.

Page 10, line 443. Were these “statistically significant differences” reflected in reduced or increased numbers of implantation sites and numbers of corpora lutea? What are the “post-implantation” loses (e.g., resorptions, late fetal death) that were measured here?

Page 12, line 519. It is preferable that the document cite the original reports (Casazza et al. 1984; Kosugi et al. 1986) rather than the ATSDR review.

Page 16, Table 4. Replace commas with decimals.

Page 17, line 699, and page 19, lines 770–771. Shkodich (1966) published the odor threshold for acetone cyanohydrin in water (0.06 mg/L). However, the odor would necessarily be the consequence of a mixed presentation of the HCN and cyanohydrin levels in air.

Casazza, J.P., M.E.Felver, and R.L.Veech. 1984. The metabolism of acetone in the rat. J. Biol. Chem. 259:2312–236.

Conn, E.E. 1979. Cyanogenic glycosides. Int. Rev. Biochem. 27:21.

Conn, E.E. 1981. Unwanted biological substances in foods: Cyanogenic glycosides. P. 105 in Impact of Toxicology on Food Processing, J.C.Ayers and J.C.Kirschman, eds. Westport, CT: AVI Publications.

Dixon, M., and E.G.Webb. 1964. Enzymes. Pp. 337–340.

Gentry, P.R., T.R.Covington, M.J.Clewell, and M.E.Anderson. 2003. Application of a physiologically-based pharmacokinetic model for reference dose and reference concentration estimation for acetone. J.Toxicol. Environ. Health 66A:2209–2225.

Gibson, Q.H., and C.Greenwood. 1963. Reactions of cytochrome oxidase with oxygen and carbon monoxide. Biochem. J. 86(3):541–554.

ITI (International Technical Information Institute). 1977. Toxic and Hazardous Industrial Chemicals, Safety Manual for Handling and Disposal with Toxicity and Hazard Data. Kenkyusho, Tokyo: Kaigai Gijutsu Shiryo.

Jones, W.J. 1914. The interaction between hydrogen cyanide and aldehydes and ketones in dilute solutions. J. Chem. Soc. 105:1560–1564.

Kosugi, K., V.Chandramouli, K.Kumaran, et al. 1986. Determinants in the pathways followed by the carbons of acetone in their conversion to glucose. J. Biol. Chem. 261:13179–13181.

Lang, J., and F.Stintzy. 1960. Un cas d’intoxication lente a l’acide cyanhydrique par l’acetone-cyanohydrine [in French]. Arch. Mal. Prof. Med. Hyg. Travail 21:652–657.

Lemberg, M.R. 1969. Cytochrome oxidase. Phys. Rev. 49(1):48–121.

Rieders, F. 1971. Noxious gases and vapors. I. Carbon monoxide cyanides, methemoglobin and sulfhemoglobin. Pp. 1180–1205 in Drill’s Pharmacology in Medicine, 4th Ed, J. Dipalma, ed.

Shkodich, P.E. 1966. Eksperimental ‘noe obosnovanie predel’ no dopustimoi kontsentratsi atsentontsiangidrina v vode vodoemov [in Russian]. Gig. Sanit. 31(3):335–341.

Smith, L., H.C.Davies, and M.E.Nava. 1979. Kinetics of reaction of cytochrome c with cytochrome c oxidase. Pp. 293–304 in Cytochrome Oxidase, Developments in Biochemistry, Vol. 5, T.E.King et al., eds. Amsterdam: Elsevier/North Holland.

Stannard, J.N., and B.L.Horecker. 1948. The in vitro inhibition of cytochrome oxidase by azide and cyanide. J. Biol. Chem. 172:599–608.

Stewart, T.D., and B.J.Fontana. 1940. Effect of salvation upon the dissociation of acetone cyanohydrin. J. Am. Chem. Soc. 67:3281–3285.

Swinyard, E.A. 1975. Noxious gases and vapors. Pp. 900–911 in The Pharmacological Basis of Therapeutics, 5^th Ed, L.S.Goodman and A.Oilman, eds. New York: Macmillan.

Zeller, H., H.T.Hofman, A.M.Thiess, and W.Hey. 1969. Zur toxizitat der nitrile. Tierexperimentelle untersuchungsergebnisse und Werksarztliche Erfahrungen in 15 Jahren [in German]. Zbl. Arbeitsmed. 19(18):225–238.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on monochloroacetic acid (MCA). The document was presented by Peter Griem of FOBIG GmbH, Germany. The subcommittee recommends the following revisions. A revised draft should be reviewed by the subcommittee at its next meeting.

The high water solubility and low vapor pressure of this chemical together raise concerns about the relevance of the derived of AEGLs for MCA. It is unclear in what it is found (vapor or aerosol), and the document does not provide enough information about the extent of its production and use in the United States or worldwide. MCA is also a by-product of the chlorination of drinking water. It occurs along with other chloroacetic acids and trihalomethanes; however, there is no mention or reference to those studies or sources. That should be rectified.

The adequacy of the Dow study is questionable given that the nominal concentration and analytical concentration vary by more than a factor of 10. Furthermore, the end points from the Dow study, “eye squint” and “some lethargy,” are not appropriate for deriving AEGL-2 values.

The justification for the intraspecies UF of 3 is weak. The document states (page 41, line 1413) that the UF is justified by the small variability in target enzymes. However, that would only be true for systemic effects, and not for local effects.

The conversion of volume to quantity of MCA in the document should account for density (1.37 g/ml). Using the density data, the dose would be 2.94 mg/kg/d (by oral exposure). The comparison of that number with doses calculated based on AEGL values only raises concerns and does not provide support for the proposed AEGL-2 values. In making such comparisons, an oral absorption of 100% is assumed, and that is questionable; the corresponding paragraphs (lines 865–878, on page 22) should be deleted.

The wording “the enzymes inhibited by MCA are evolutionarily highly conserved” is not scientifically correct, because it implies possible explanation for the low variability of those enzymes. It should be rephrased as “‘the enzymes inhibited by MCA do not vary considerably within and between species” on pages vii (lines 150 and 154) and 41 (line 1314).

Page 2, line 216. Acedemia—metabolic or respiratory acidosis?

Page 5, Section 3.1.1. It should be stated that this study would be ethically unacceptable nowadays.

Page 9, line 1410. How did the investigators measure “eye squint” and “slight lethargy”? Are those the same as “amblyopia” and “slight drowsiness”?

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on phosphorus trichloride (PCl₃). The document was presented by Robert Young of Oak Ridge National Laboratory. The subcommittee recommends the following revisions. A revised draft should be reviewed by the subcommittee at its next meeting.

This is a well-written, concise document, although the literature on which it is based is limited, both in a qualitative and a quantitative sense. The key study is Weeks (1964), which has its limitations, but it is the best available study.

The major shortcomings of the Weeks (1964) study are (1) that there is mixed exposure to vapor and aerosol and (2) that whole-body exposures were applied, not nose-only. That means that skin absorption might have taken place in addition to pulmonary absorption.

It is remarkable that ammonia neutralization decreased signs of irritation but did not decrease the mortality rate. That might suggest a systemic effect of PCl₃, in addition to a local irritating effect on the pulmonary epithelium.

Executive Summary. In what chemical processes and in what quantities is PCl₃ being used? Is it an irritant per se, or does it work through HCl and phosphoric acid formation?

Page 1, line 5. Is “glyophosphate” the same as glyphosate?

Page 1, line 7. Insert “or” after “hydrogen chloride.”

Section I. Insert the quantities produced and consumed each year, mode of transport, number of facilities handling PCl₃ in the United States, Western Europe, and Canada.

Page 3, line 31. Increased lethal dose might mean that the liver is also a target organ. Was the organic source of the enzyme specified?

Page 4, lines 9–14. On the basis of the uncertainty, why include this information at all?

Page 4, line 22. What was the analytical method, and what was the LOD (limit of detection)?

Page 4, lines 32–34. The subcommittee recommends changing the wording “and, therefore, results in a lack of confidence in the reported results.”

Page 4, line 37. Who was the source of the personal communication?

Page 5, line 16–19. Genotoxicity and carcinogenicity cannot be excluded because of the epithelial damage.

Summary section. A likelihood of other organ damage (LD increase) in addition to mucosal irritation should be mentioned.

Page 7, lines 41–42. Delete.

Page 8, line 16. Squamous metaplasia is an indication that genotoxicity and carcinogenicity might occur.

Page 9, line 28. There appears to be a latency period.

Page 11, line 3. From previous information, aren’t the exposures for 6 hours at concentrations of 0.71 ppm and for 1 hour at 1.8–5.4 ppm?

Page 11, line 26. What is the basis for the assertion that rodents appear to be more sensitive than humans?

Page 11, lines 28–31. It is not clear how the description of the effects supports the UF selected.

Page 11, lines 36–39. This rationale is not clear.

Page 12, line 14. In addition, it is unclear whether the toxic effects were due to PCl₃ itself or to its hydrolysis products.

Page 13, line 3. Exposed to something for 6 hours/day.

Derivation of AEGL-3. Why not use the more sensitive species to derive the AEGL, rather than the less sensitive?

Page 16. The AEGL values indicate a fairly steep dose-response curve. Is that really warranted by the available data set? Also, the 8-hour AEGL-1 is higher than the TLV.

Page 35, line 12. Replace “for may irritant” with “for many irritant.” Fix typo throughout report.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on phosphorus oxychloride (POCl₃). The document was presented by Robert Young of Oak Ridge National Laboratory. The subcommittee recommends some minor revisions. The document can be finalized if the recommended revision is made appropriately.

This is a well-written, concise document, although the literature on which it is based is limited, both in a qualitative and a quantitative sense. The key study is Weeks (1964), which has its limitations, but it is the best available.

The major comment concerns the use of an uncertainty factor (UF) of 10 for interspecies variation; it is recommended that the UF be lowered to 3.

The major shortcomings of the Weeks (1964) study are (1) that there is mixed exposure to vapor and aerosol and (2) that whole-body exposures were applied, not nose-only. That means that skin absorption might have taken place in addition to pulmonary absorption.

It is remarkable that ammonia neutralization decreased signs of irritation but did not decrease the mortality rate. That may suggest a systemic effect of POCl₃, in addition to a local irritating effect on the pulmonary epithelium.

Suggested phrasing for the use of UF of 3 instead of 10: “Although the animal species used (female rats) appeared to be more sensitive than humans, the species difference was small and the toxic effects appear to be direct and not metabolism-dependent. An interspecies UF of 3 is recommended because of the paucity of data.”

Executive Summary. 1) In what chemical processes and in what quantities is POCl₃ being used? 2) Apparently its use is much more widespread than PCl₃. 3) In the fourth paragraph, “mucosal” is preferred to “epithelial,” because skin effects are not predominant.

Section I. Insert the quantities produced and consumed each year, mode of transport, number of facilities handling PCl₃ in the United States, Western Europe, and Canada.

Page 3, lines 4–12. The lack of effect on PFT is inconsistent with the reported symptoms. Some comments on the time frame between symptom reporting and actual respiratory testing should be included if that information is available from the source.

Page 5, line 6. What do the authors mean by “porphyrin secretions”? That should be “hemorrhagic exsudate” or something else, because they probably did not measure porphyrin.

Page 5, line 32. It should be mentioned that these experiments would not be ethically acceptable nowadays, but the data are used in this document because they are relevant. That does not mean that the subcommittee agrees with the ethics of the experiment.

Page 8, lines 3–4. The LC₅₀ values do not seem to make sense, because it seems that HCl should have a lower LC₅₀ than phosphoric acid.

Page 8, lines 28–29. The similarity between PCl₃ and POCl₃ favors the use of the same UFs for both substances.

Page 11, lines 19–23. In reality, there is no difference between values of 48 and 52.

Page 11, lines 32–35. It is not clear why the mechanism of toxicity stated justifies a UF of 3.

Page 13. The AEGL-3 is about 2.5 times the TLV. That suggests a relatively tight dose-response curve. Is this justified by the available data?

Page 21. Replace “for may irritant” with “for many irritant.” A total UF of 10 will be sufficient; it leads to AEGL-3s comparable to those for

A copy of the full paper in Italian by C.Sassi (Occupational poisoning by phosphorous oxychloride, Med. Lavoro., 45(1954):171–177) should be requested through the Italian Toxicology Society and sent to the subcommittee as soon as it becomes available. It will then be decided whether it should act as additional key study or replace the Weeks et al. (1964) study and whether that would affect the proposed AEGL values.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on fluorine. The document was presented by Sylvia Talmage of Oak Ridge National Laboratory.

The subcommittee recommends the following revisions. A revised draft will be reviewed by the subcommittee at its next meeting.

Inconsistencies are evident regarding the database uncertainty factor (UF). The sentence on page 20, line 23, summarizes Section 3.3, and that section should be edited to conform to the conclusion. Unless data can be cited to show that systemic fluoride toxicity results from inhaled fluorine gas, the value of references to systemic uptake is not clear. In this context, it might be worthwhile to review the Biological Exposure Indices (BEI) documentation for various fluorides. As the document is constructed, reference to the Heindel et al. (1996) study (e.g., page 21, lines 22–23) is clearly irrelevant. Why is a database UF of 2 assigned (page 28, line 16)? What specific data on pulmonary edema and irritation are missing that warrant assignment of that value? The data support an AEGL-1 of 3.4 ppm at all time points (page 28, lines 25–30); however, it is noteworthy that the TLV for an 8-hour fluorine gas exposure is limited to 1 ppm and that there is a 2-ppm STEL for four 15-minute exposures per shift.

It is not clear whether the evidence put forward to support the NAC conclusion that repeated fluorine gas exposures result in adaptation is correct. Given the inherent variability in LC₅₀ values, how is it that 15-minute LC₅₀ values of 315 ppm and 350 ppm (page 26, lines 10–11) are not biologically identical? It appears that these values are within the limits of experimental error or variability commonly encountered in performance of these assays. It appears that these values are toxicologically identical.

For AEGL-2, how can the NAC state that the interspecies factor should be less than the intraspecies factor? (interspecies factor=1, intraspecies factor=3). Are we more like rats than each other?

Page 6, lines 15–16. Provide authoritative reference to substantiate the conclusion that “at low concentrations [does this apply specifically to elemental fluorine?] there is accommodation to irritant gases.” What is the definition of “low”? Does this include the AEGL-2 range of exposures or does this relate only to the AEGL-1 range or lower? It appears that this statement is derived from the observations of Keplinger (1969).

Page 8, line 16. The production data cited are 10 years old. What are current quantities of fluorine gas produced in the United States and Western Europe?

Page 8, line 20. What is the source of the anhydrous hydrogen fluoride?

Page 8, lines 30–31. If no deaths have occurred, it is speculation that death might occur from pulmonary edema?

Page 10, line 21. What is the significance of 0.1 ppm? The next sentence says the subjects were exposed to concentrations ranging from 0.3 to 1.4 ppm.

Page 12, lines 1–10. Were these mongrel dogs or beagles or something else?

Page 12, line 15. Insert the gender and strain of rat, if available.

Page 14, line 7. Insert gender and strain of mice, if available.

Page 14, line 18. Insert gender if available.

Page 16, line 18. Delete. The phrase “hyperperfusion and shock” is speculation as written.

Page 16, lines 18–25. Kidney and liver damage seems significant.

Page 17, line 4. Define the nature of the specific renal and hepatic “gross changes.”

Page 18, lines 1, 3, 4, 6, 10, 11. Describe the organ and type of “irritation” listed.

Page 18, lines 16–23. How can repeated exposures result in less damage than a single exposure? There was an initial exposure in all cases.

Page 20, lines 13–36, and page 11, lines 1–3 and 8–11. Delete. An ATSDR review of ingested fluoride salts has no bearing on the AEGL derivation unless the NAC can cite data to support the similar disposition of the quantities of systematically available fluoride ion after inhalation and ingestion. Ingestion data are irrelevant in the present context inasmuch as mammals die after acute fluorine gas exposure as a result of acute pulmonary edema and consequent respiratory failure, not as a result of the actions of systemic fluoride.

Page 20, line 42. It is proper to specify the Ames tester strains used here.

Page 21, lines 6–12. If the NAC insists that chronic ingestion of fluoride ion in food or water is relevant to AEGL development for fluorine gas, it is incumbent upon the NAC to cite data concerning systemic uptake of fluoride after fluorine gas exposure in relation to the circulating and or target-tissue (e.g., bone, teeth) fluoride concentrations associated with systemic fluorine toxicity. In this case, the Biological Exposure Indices (BEI) documentation might be of some utility.

Page 24, lines 1–2. At some points the document states that fluorine gas is rapidly hydrolyzed. However, this sections states it will remain in saturated water vapor for over an hour. Is it hydrolyzed to HF? One other location states fluorine oxidizes water.

Page 24, lines 3–6. Delete and close up to line 8.

Page 24, line 26. Explain whether the “relative toxicities of HF>HCl>HBr” are associated with nasal pathology alone or does that include death and/or pulmonary edema as well?

Page 24, line 37. This is a sentence fragment. Rewrite the text to complete the bromine and chlorine comparison.

Page 25, lines 17–19. Do the Machle and Evans (1940) results pertain to increased asthmatic episode frequency or intensity or both?

Page 25, lines 29–34. Doesn’t this section argue against the intraspecies factor of 1?

There is no C-t graph. Why is it only based on the LC₅₀? Is the mechanism the same? The subcommittee recommends that n=1 and n=3 are more appropriate. Also, the number of significant figures of 3 (1.77) seems like a stretch of the quality of the data. The subcommittee needs to see the data plotted.

Page 25, line 33. Re-order the concentrations from lowest to highest.

Page 25, line 39. How is it that similar LC₅₀ or LD₅₀ values across species suggest “a common cause of death”? Delete the phrase unless substantiating information can be brought to bear on this question.

Page 26, line 16. The word “levels” is usually reserved in toxicology for the “levels” of a material in human or animal food. Use the word “concentration” in the present context.

Page 28, lines 12–14. It is not clear how the Lyon (1962) report brings rigorous empirical data to bear on the validity of the 10-ppm 15-minute NOAEL.

Page 28, line 16. What data are missing that have direct relevance to the derivation of fluorine gas AEGLs for humans and that justify the database UF of 2? Surely the authors cannot be reducing the AEGL by 50% on the basis of the acknowledged “questionable relevance” of the human fluoride dust and/or HF experience (page 11, lines 19–23). Considered and specific justification is necessary for so drastic a reduction to the permissible once-in-a-lifetime exposure (to less than 2 ppm for 10 minutes) when the empirical data from multiple sources clearly identified a 10-ppm NOAEL (more than 5-fold the proposed AEGL-1) (see Keplinger and Suissa [1968] and Belles [1965]). As to the animal data, the conclusion presented on page 29 (lines 14–16) does not support the introduction of this additional UF.

Page 28, line 40. AEGL-2 should be AEGL-1. This is the definition for AEGL-1.

Page 29, line 24. Explain the specific basis for the “modifying factor of 2.”

Page 30, line 10. Delete the editorialization.

The subcommittee recommends that the NAC use the 25-ppm concentration that elicited slight to moderate discomfort. That fits the AEGL-1 definition.

The subcommittee recommends that the NAC use a higher concentration. The concentration used actually fits the definition for AEGL-1.

Same as for AEGL-1 and -2. The AEGL-3 definition includes death; reversible effects and labored breathing do not fit the definition.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on acrylic acid. The document was presented by Peter Griem of FOBIG GmbH, Germany. The subcommittee recommends the following revisions. A revised draft should be reviewed by the subcommittee at its next meeting.

The subcommittee is uncomfortable with NAC citing the Renshaw personal communication. It has not been provided to the subcommittee for review.

What is the likely form of an exposure to the general public? It would seem that even if an aerosol was formed, it would quickly convert to vapor because of the relatively high vapor pressure of acrylic acid. Basing the AEGL on the aerosol requires further information. Is acrylic acid generally heated in the plant? It definitely is not when transported. The subcommittee postulates aerosols will rapidly evaporate to the vapor state. If that is the case, an AEGL derived on the basis of the vapor is more applicable.

Provide current production quantities, number of U.S. and Western European facilities, and information on how it is handled and how it is transported. What is the opportunity for widespread release and dispersal of acrylic acid?

Substantial difficulties remain with the draft technical support document. There remains considerable confusion between the notion of odor perception and irritation, and there is a failure to account rigorously for smoking (Cometto-Muiz and Cain 1982), gender (Cometto-Muniz and Noriega 1985), age (Stevens et al. 1982; Stevens and Cain 1986), and physiologic inurement on the part of the individuals surveyed. The NAC failed to distinguish between the objectionable odor associated with materials and frank irritation. It is important to take into account the quantitative differences between eye irritation threshold, nasal pungency, and odor threshold—the former of which often lays “orders of magnitude above odor thresholds” (Cometto-Muniz

and Cain 1995). Just as important, “persons with normal olfaction provide rather unreliable estimates of a nasal trigeminal threshold” and “it seemed necessary to rely strictly on the data from anosmic subjects who would have no distraction from accompanying odor” (Cometto-Muniz and Cain 1995). Controlled human quantitative structure-activity relationships exist for nasal pungency thresholds; those data are at present most robust for aliphatic alcohols, acetates, ketones, and carboxylic acids (e.g., Hau et al. 1999; Abraham et al. 1998; Cometto-Muniz and Cain 1993). In the present context, a current literature search might reveal similar studies with acrylates. Nevertheless, there seems to be no compelling reason that those data cannot be collected for a series of acrylates and acrylic acid before AEGLs for acrylic acid are adopted.

What is the rodent RD₅₀ for this material? How do AEGL values derived from the rodent RD₅₀ compare with those derived from the anecdotal report of Renshaw (1988)?

Page x, line 146. Does the statement here refer to ocular, mucous membrane, upper respiratory tract, or deep lung irritation or does it refer to all of those organ systems? Were the results obtained from controlled laboratory inhalation or instillation studies? Was the Renshaw (1988) report the result of simple anecdotal complaints? Describe here whether the people were male or female adults, smokers or nonsmokers, what the age range was, and whether the subjects were naïve or were workers accustomed to elevated concentrations of acrylic acid in workplace air. Were these complaints of irritation (what kind?) from people subjected to a single (preferably once-in-a-lifetime) exposure or from people who had a history of repeated occupational acrylic acid exposures (describe duration and range of concentrations routinely encountered) in workplace air?

Page x, lines 147–152. The conclusions reached here relate to repeated inhalation exposure. Because AEGL values apply only to once-in-a-lifetime exposures, the many details from subchronic studies can only be used as supporting evidence, at best. As written, it is not clear whether these changes in rabbit, mouse, and nonhuman primates were seen after the first 6 hours (the first encounter) of exposure to airborne acrylic acid.

Page x, line 165. Describe the cell system and whether clastogenicity was observed at only cytotoxic concentrations in vitro.

Page x, line 171. As written, it is not clear why the irritation “threshold” is so broad (4.5–23 ppm) (Renshaw 1988). It appears that the acknowledged difficulties Renshaw had in reporting duration and concentration (page x, lines 174–175) render these data anecdotal and of such low confidence that they can only fill a supporting role in the present circumstance.

Page xi, line 197 and page xiii, line 263. Specify species, gender, and age of the “monkeys” used in the Harkema (1997, 2001) assays.

Page xi, lines 208–211. As presented, the text does not follow. The AEGL-2 is said (line 197) to be derived on the basis of the results of nonhuman primate inhalation studies conducted by

Harkema and associate, yet rats are mentioned at line 198 and the rat deposition data of Frederick et al. (1988) entered into identification of the interspecies UF? If the AEGL is based on the primate data, why is it necessary to re-work the rodent nasal end point here?

Page xi, lines 216–217. Cite the specific SOP section that supports this conclusion. What data are available to justify the NAC generalization that “For local effects [what kind? Anesthesia?], the toxicokinetic differences [what kind? Rate of absorption?] between individuals are usually much smaller [by what factor?] when compared to systemic effects [like what?].” Delete the generalization unless specifics for acrylic acid can be included.

Page xi, lines 217–220. How does the mode of action of acrylic acid differ so greatly among humans as to support a 3-fold reduction in the moderate primate olfactory changes after a continuous 3-hour inhalation exposure to acrylic acid? How were the changes “moderate”? Does NAC have data to support the development of nasal pathology in as little as 10 minutes after similar exposures?

Page xi, line 220 and page xii, lines 221–223. What is the justification for applying an n derived from acrylic acid aerosol studies of mortality in rats to the vapor study results collected in primates by Harkema and coworkers?

Page xiii, line 263. Why are the rat data referenced here when the AEGL-2, was derived from the Harkema nonhuman primate data (page ix, line 196)?

Page 4, line 33. Why would whole-body exposure have a greater toxic effect than nose-only? Include the explanation presented at the last meeting.

Page 22, lines 31–32. Is maternal toxicity different than other types of toxicity? The effects noted seem to be common to all species regardless of gender or pregnancy. If body-weight gain is the issue, emphasize that and state that eye and nose irritation also occurred. The reference to maternal toxicity was removed in Section 3.2.2. That is not the explanation provided in the comment-response section.

Section 4.2. Is this the common mechanism of a simple irritant? Restated, is this a common irritant or is there another mechanism occurring?

It is an unusual approach to calibrate a total hydrocarbon analyzer to an infrared instrument. Generally, standards are generated and the total hydrocarbon analyzer response is measured for a known concentration. The cited approach introduces potentially large sources of error.

Pages 34–35, AEGL-2. It is stated that histopathological changes consistently are a more sensitive toxicological response than clinical signs of irritation. That statement is supported by comparison of these effects in three species (rabbits, rats, mice). If that is the case, are we being protective enough by using irritation response in humans as the basis for the AEGL-1?

The monkey studies (Rohm and Haas 1995; Harkema et al. 1997; Harekma 2001) should be included in Table 10, in the Summary (Section 3.6, page 27), and in the mechanisms

discussion (Section 4.2, page 30). The human eye irritation data should be included in Figure 3.

It should be emphasized that the present AEGL-1 value of 1.5 ppm is based on an evaluation of a wide range of occupational concentrations and provide a stronger rationale for the choice of 4.5 ppm as the starting point for the derivation of the AEGL-1. More details on the 11 human subjects should be given, including their inurement status. The RD₅₀ should be included in the document, and its implications for the AEGL-1 should be considered. The AEGL-1 should also be discussed in relation to the TLV.

The monkey study, which represents the most suitable animal model for human risk assessment, seems most appropriate for the derivation of AEGL-2 values. Clearer explanations of the use of an interspecies UF of 1 and an intraspecies UF of 3 also need to be provided. The rodent data provide useful supporting information, leading to similar AEGL-2 values. Time-scaling should be based on default levels of n=1 and n=3, because n=1.8 is based on whole-body aerosol exposures.

The vapor data appears to be more relevant than the aerosol data. If the AEGL-3 is based on the vapor data (using the highest NOAEL from all of the available studies), the default time-scaling values of n=1 and 3 should be used.

Abraham, M.H., R.Kumarsingh, J.E.Cometto-Muniz, and W.S.Cain. 1998. An algorithm for nasal pungency thresholds in man. Arch. Toxicol. 72:227–232.

Cometto-Muniz, J.E., and W.S.Cain. 1982. Perception of nasal pungency in smokers and nonsmokers. Physiol. Behav. 29:727–731.

Cometto-Muniz, J.E., and W.S.Cain. 1993. Efficacy of volatile organic compounds in evoking nasal pungency and odor. Arch. Environ. Health 48(5):309–314.

Cometto-Muniz, I.E., and W.S.Cain. 1995. Relative sensitivity of the ocular trigeminal, nasal trigeminal and olfactory systems to airborne chemicals. Chem. Sense 20(2):191–198.

Cometto-Muniz, J.E., and G.Noriega. 1985. Gender differences in the perception of pungency. Physiol. Behav. 34:385–389.

Hau, K.M., D.W.Connell, and B.J.Richardson. 1999. Quantitative structure-activity relationships for nasal pungency thresholds of volatile organic chemicals. Toxicol. Sci. 47:93–98.

Stevens, J.S., and W.S.Cain. 1986. Aging and the perception of nasal irritation. Physiol. Behav. 37:323–323–328.

Stevens, J.C., A.Plantinga, and W.S.Cain. 1982. Reduction of odor and nasal pungency associated with aging. Neurobiol. Aging 3:125–132.

At its January 28–30, 2004, meeting, the subcommittee reviewed the AEGL document on cis-1,2-dichloroethylene (cis-1,2-DCE) and trans-1,2-dichoroethylene (trans-1,2-DCE). The document was presented by Cheryl Bast of Oak Ridge National Laboratory. The subcommittee recommends the following revisions. A revised draft should be reviewed by the subcommittee at its next meeting.

The use of a modifying factor (MF) of 2 should be justified on the basis of irritation-potency differences.

Narcosis is one mechanism for which there is documented evidence of species similarity in response and in critical lipid concentration necessary to produce such effects (see ecotoxicology literature by Lynn McCarty and others). The toxicodynamic component of the uncertainty factor (UF) should therefore be 1. It is acceptable to use a total factor of 10 to account for pharmacokinetic diferences within and between species. However, the authors should stay away from back-calculating the UF, even though some cross-checking is desirable.

The application of UF of 10 is acceptable for the above reasons. As pointed out earlier, back-calculating the UF is not recommended. Further, it may not be appropriate to say “twice as toxic” (Page 27, line 14), because the difference is likely the result of twice the exposure concentration needed to yield the same critical lipid concentration responsible for narcosis-mediated lethality.

Page 8, lines 22–24. Why not cite newer editions (12th or 13th) of the Merck Index?

Page 10, lines 21 and 22. The two human guinea pigs inhaled 275–2,200 ppm, not mg/m³.

Page 11, lines 5, 6, and 20. If six often rats succumb to cis-1,2-DCE at 13,500 ppm, wouldn’t the LC₅₀ be lower than 13,700 ppm?

Page 11, Table 3. Include a footnote that all of the deaths in exposed rats occurred during exposure to the chemical.

Page 15, line 24, and page 26, line 5. The Freundt et al. (1977) study results are inconsistent with those of the GLP study by Kelly (1999). The morphological changes described by Freundt et al. are of very questionable toxicological significance and are probably undeserving of the extent of coverage they are given on page 15, line 24, and on page 16, line 5. It should be noted here that morphological changes were seen in just one of six rats (16%) exposed at 200 ppm versus five of 48 of the controls (10%).

Page 16, line 43, and page 17, line 26. It would be worthwhile to summarize the subacute and subchronic oral toxicity studies by McCauley et al. (1995) and the NTP (2002) study. Extremely high daily doses of the cis and trans isomers had very little effect on rats or mice other than increased liver weights at the highest doses. These findings of an apparent lack of systemic toxicity support those of inhalation studies conducted by Kelly (1996) in rats.

Page 19, lines 1–12. Can cis-1,2-DCE be characterized as a weak, moderate, or strong mutagen on the basis of these findings?

Page 19, lines 32–34. It is stated that the 4-hour LC₅₀ from Kelly (1999) for cis-1,2-DCE is 13,700 ppm and that the 4-hour lethality NOAEL is 12,100 ppm. Is the lethality dose-response curve actually that steep? Specify in the text that the NOAELs are for 4-hour exposures.

Page 19, line 39. Substitute the word “potent” for “toxic.”

Page 19, line 42, and page 20, lines 1–2. Loss of equilibrium in cats and mice is not equivalent to dizziness in humans. Intoxication/inebriation in humans would be closer.

Page 20, lines 11–14. The “pathological changes” reported by Freundt et al. (1977) should be discounted. The apparent morphological changes they described (e.g., hyperemia, alveolar septum distension, fibrous swelling, poorly maintained cardiac muscle striations), with the exception of fatty liver change, are vague, nonspecific, and of questionable toxicological significance. The findings of Kelly (1998, 1999)—no histological alterations in heart, liver, kidneys, and lungs after subchronic exposures—should be relied on instead.

Page 20, lines 19–30 It is stated in lines 21 and 22 that 1,2-DCE has a “relatively high affinity for blood.” That statement is inaccurate. Its blood:air partition coefficients and those of other lipophilic chlorinated aliphatic hydrocarbons (halocarbons) are relatively low because of their limited solubility in blood (which is primarily aqueous). Relatively water-soluble hydrocarbons (e.g., ethanol, acetone) have much higher blood:air partition coefficients than 1,2-DCE.

Page 20, lines 35–36. Barton et al. (1995), Lilly et al. (1998), and Hanioka et al. (1998) have published the results of more recent suicide enzyme inhibition studies of 1,2-DCE. Lilly et al. (1998) found the trans isomer to be more potent in that regard in male rats. However, Hanioka et al. (1998), observed that the isomers’ inhibitory effects were isozyme-specific and limited to male rats.

Page 21, lines 11–13. Why is the V_max for the cis isomer lower than that for the trans isomer if cis is more rapidly metabolized?

Page 21, lines 22–32. It should be pointed out that CYP2E1-catalyzed oxidation of 1,2-DCE to an epoxide, 2,2-dichloroacetaldehyde, and 2,2-dichloroethanol represents metabolic activation. Each of these metabolites is cytotoxic. Collectively, they are likely responsible for the hepatic centrilobular fatty degeneration 1,2-DCE causes. The more rapid and extensive metabolism of the cis isomer and the more extensive production of dichloroethanol and its unstable predecessors from cis are consistent with this isomer’s greater ability to affect the liver (Kelly 1999). The delayed deaths in cats exposed during the investigation by Lehmann and Schmidt-Kehl (1936) might have been due to hepatotoxicity.

Page 21, lines 24–25. The paper by Gargas et al. (1990) should be cited here. 1,2-DCE-induced enzyme inhibition and resynthesis was also discussed for in a later publication by this research group (i.e., Lilly et al. 1998).

Page 21, lines 34–39. The work of Eger et al. (2001) should be cited rather than that of Anon (1988a,b). Eger et al. (2001) found cis-1,2-DCE to be a more potent anesthetic in rats than trans-1,2-DCE. The authors felt that this supported the hypothesis that 1,2-DCE and other lipophilic anesthetics act by specific receptor interactions as opposed to simple partitioning into neuronal membrane lipids.

Page 21, line 6. Insert the word “metabolic” at the end of this line.

Page 21, line 24. This line should read “inhibition of the metabolism of other cytochrome P-450 substrates by 1,2-dichloroethylene.”

Page 21, line 42. Substitute “cis” for “trans.”

Page 22, line 2. Tables 7 and 8 should be Tables 8 and 9. Cite the source of these data (i.e., Lehman and Schmidt 1936).

Page 22, line 6. Tables 4 and 6 should be Tables 5 and 7.

Page 22, lines 11–13. The 6-hour LC₅₀ of 21,723 ppm for trans-l,2-DCE might be of questionable accuracy. Was a complete translation of the report by Gradiski et al. (1978) available so that their study could be evaluated? Lehmann and Schmidt-Kehl (1936) found that all mice inhaling trans-1,2-DCE at 18,750 ppm for 102 minutes and at 20,000 ppm for 95 minutes died. It would be anticipated that the LC₅₀s for mice would be significantly lower than those for rats. Mice will achieve higher blood and brain levels of 1,2-DCE because of their higher blood:air partition coefficient, pulmonary (blood) perfusion rate, alveolar ventilation rate, and 1,2-DCE metabolic rate.

Page 22, lines 23–29. The results of Freundt and Macholz (1978) should be included here, namely that 1,2-DCE prolonged hexobarbital sleeping time and zoxazolamine paralysis time. These findings are a good illustration of the ability of 1,2-DCE to inhibit the P-450-catalyzed detoxification of certain chemicals.

Page 22, lines 33–38. It does not appear that the use of Cⁿ×t=k for time-scaling is warranted in deriving the 4- and 8-hour AEGL-2. One might anticipate that blood (and tissue) profiles for 1,2-DCE would resemble those of other well-metabolized halocarbons (that blood 1,2-DCE levels would increase rapidly upon initiation of an inhalation exposure and soon approach near-steady-state). Although no time-course data were located, Filser and

Bolt (1979) estimated that cis- and trans-1,2-DCE would attain near-steady-state within about 2 hours in rats inhaling 1,2-DCE at 100 ppm. If that is the case, blood levels and the extent of CNS depression would only increase modestly with increasing duration of exposure once near-steady-state was reached. This phenomenon is exhibited by trichloroethylene, another well-metabolized halocarbon (Bruckner et al. 2004). An apparent plateau for anesthetic effects is cited as the rationale for keeping the 10-, 30-, and 60-minute AEGL-2 values constant. 1,2-DCE is one of the few halocarbons that exhibits suicide enzyme inhibition. Metabolic inhibition would result in a decrease in the systemic uptake of inhaled 1,2-DCE. Lilly et al. (1998) observed virtually complete metabolic inhibition in rats inhaling trans-1,2-DCE at 10 ppm.

Page 23, lines 34–37. It is not necessary to apply a modifying factor (MF) of 2 for cis-1,2-DCE. Its action as a mild direct irritant has nothing to do with its increased potency as a CNS depressant and hepatotoxicant. The same AEGL-1 values should apply to the cis and trans isomers.

Page 25, lines 1–3. Hurtt et al. (1993) do not state when narcosis first became evident in the pregnant rats during the 6-hour exposures. It seems quite likely that narcosis ensued within the first hour or two and became somewhat more pronounced during the last 4 hours (judging from the aforementioned pharmacokinetic calculations of Filser and Bolt [1979]). Have any data been published showing the magnitude of CNS depression as a function of time of inhalation of a fixed vapor concentration?

Page 26, lines 16–18. It should be stated here that Kelly saw no histopathological changes in the liver, heart, kidneys, or lungs of any rats in the LC₅₀ study.

Page 26, line 19. Give the 20,000–114,000-mg/m³ values in ppm so that they can be more readily compared with the values in lines 1–3.

Page 26, line 28. Insert “up to 4,000 ppm” between “to” and “trans” at the end of the line.

Page 26, lines 33 and 34. A vapor level of trans-1,2-DCE at 12,300 ppm—a NOEL for death of rats—was chosen as the basis for 4- and 8-hour AEGL-3 values. The cat (Table 7) is apparently more susceptible to 1,2-DCE lethality, although the cats were exposed to cis-1,2-DCE. Justify the use of the rat data.

Barton, H.A., J.R.Creech, C.S.Godin, G.M.Randall, and C.S.Seckel. 1995. Chloroethylene mixtures: Pharmacokinetic modeling and in vitro metabolism of vinyl chloride, trichloroethylene, and trans-1,2-dichloroethylene in rat. Toxicol. Appl. Pharmacol. 130:237–247.

Bruckner, J.V., D.A.Keys, and J.W.Fisher. 2004. The acute exposure guideline level (AEGL) program: Applications of physiologically based pharmacokinetic modeling. J. Toxicol. Environ. Health, Part A, 67:621–634.

de long, R.H., and E.I.Eger. 1975. AD₅₀ and AD₉₅ values of common inhalation anesthetics in man. Anesthesiology 42:384–389.

Eger, E.I., M.J.Halsey, D.D.Koblin, M.J.Laster, P.Ionescu, K.Konigsberger, R.Fan, B.V. Nguyen, and T.Hudlicky. 2001. The convulsant and anesthetic properties of cis-trans isomers of 1,2-dichlorohexafluorocyclobutane and 1,2-dichloroethylene. Anesth. Analg. 93:922–927.

Gregory, G.A., E.I.Eger, and E.S.Munson. 1969. The relationship between age and halothane requirement in man. Anesthesiology 30:488–491.

Hanioka, N., H.Jinno, T.Nishimura, and M.Ando. 1998. Changes in hepatic cytochrome P450 enzymes by cis- and trans-1,2-dichloroethylenes in rat. Xenobiotica 28:41–51.

Lilly, P.D., J.R.Thorton-Manning, M.L.Gargas, H.J.Clewell, and M.E.Andersen. 1998. Kinetic characterization of CYP2E1 inhibition in vivo and in vitro by the chloroethylenes. Arch. Toxicol. 72:609–621.

McCauley, P.T., M.Robinson, F.B.Daniel, and G.R.Olson. 1995. The effects of subacute and subchronic oral exposure to cis-1,2-dichloroethylene in Sprague-Dawley rats. Drug Chem. Toxicol. 18:171–184.

Stevens, W.C. et al. 1975. Minimum alveolar concentrations (MAC) of isoflurane with and without nitrous oxide in patients of various ages. Anesthesiology 42:197–200.

This page intentionally left blank.