2.
Derivation of AEGL Values

2.1 CHARACTERIZATION OF AEGLs

Three tiers of AEGL severity levels represent short-term exposure values that are a threshold for specific biologic effects for the general public and are applicable to specified exposure durations of 10 min, 30 min, 1 h, 4 h, and 8 h. The values for these specified durations are maximum airborne concentrations above which there is an increasing likelihood of the adverse effects associated with the respective AEGL tiers. Therefore, to avoid the onset of these adverse effects, the values should not be exceeded during the specified exposure durations. Three tiers of AEGLs distinguished by varying degrees of severity of toxic effects are developed for each of the five exposure durations. Ten-minute AEGLs for the four chemicals included in the first publication of AEGLs by the National Research Council (NRC 2000b) will be developed at a future date.

Under the authority of the Federal Advisory Committee Act (FACA) P.L. 92–463 of 1972, the National Advisory Committee for Acute Exposure Guideline Levels for Hazardous Substances (NAC/AEGL Committee) has been established to identify, review, and interpret relevant toxicologic and other scientific data and develop AEGLs for high-priority acutely toxic chemicals.

AEGLs represent threshold exposure limits for the general public and are applicable to emergency exposure periods ranging from 10 min to 8 h. AEGL-2 and AEGL-3, and AEGL-1 values as appropriate, will be developed for each of five exposure periods (10 and 30 min, 1 h, 4 h, and 8 h) and will be distinguished by varying degrees of severity of toxic effects. It is believed that the recommended exposure levels are applicable to the general population



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals 2. Derivation of AEGL Values 2.1 CHARACTERIZATION OF AEGLs Three tiers of AEGL severity levels represent short-term exposure values that are a threshold for specific biologic effects for the general public and are applicable to specified exposure durations of 10 min, 30 min, 1 h, 4 h, and 8 h. The values for these specified durations are maximum airborne concentrations above which there is an increasing likelihood of the adverse effects associated with the respective AEGL tiers. Therefore, to avoid the onset of these adverse effects, the values should not be exceeded during the specified exposure durations. Three tiers of AEGLs distinguished by varying degrees of severity of toxic effects are developed for each of the five exposure durations. Ten-minute AEGLs for the four chemicals included in the first publication of AEGLs by the National Research Council (NRC 2000b) will be developed at a future date. Under the authority of the Federal Advisory Committee Act (FACA) P.L. 92–463 of 1972, the National Advisory Committee for Acute Exposure Guideline Levels for Hazardous Substances (NAC/AEGL Committee) has been established to identify, review, and interpret relevant toxicologic and other scientific data and develop AEGLs for high-priority acutely toxic chemicals. AEGLs represent threshold exposure limits for the general public and are applicable to emergency exposure periods ranging from 10 min to 8 h. AEGL-2 and AEGL-3, and AEGL-1 values as appropriate, will be developed for each of five exposure periods (10 and 30 min, 1 h, 4 h, and 8 h) and will be distinguished by varying degrees of severity of toxic effects. It is believed that the recommended exposure levels are applicable to the general population

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals including infants and children, and other individuals who may be susceptible. The three AEGLs have been defined as follows: AEGL-1 is the airborne concentration (expressed as parts per million or milligrams per cubic meter (ppm or mg/m3)) of a substance above which it is predicted that the general population, including susceptible individuals, could experience notable discomfort, irritation, or certain asymptomatic nonsensory effects. However, the effects are not disabling and are transient and reversible upon cessation of exposure. AEGL-2 is the airborne concentration (expressed as ppm or mg/m3) of a substance above which it is predicted that the general population, including susceptible individuals, could experience irreversible or other serious, long-lasting adverse health effects or an impaired ability to escape. AEGL-3 is the airborne concentration (expressed as ppm or mg/m3) of a substance above which it is predicted that the general population, including susceptible individuals, could experience life-threatening health effects or death. Airborne concentrations below the AEGL-1 represent exposure levels that can produce mild and progressively increasing but transient and nondisabling odor, taste, and sensory irritation or certain asymptomatic, nonsensory effects. With increasing airborne concentrations above each AEGL, there is a progressive increase in the likelihood of occurrence and the severity of effects described for each corresponding AEGL. Although the AEGL values represent threshold levels for the general public, including susceptible subpopulations, such as infants, children, the elderly, persons with asthma, and those with other illnesses, it is recognized that individuals, subject to unique or idiosyncratic responses, could experience the effects described at concentrations below the corresponding AEGL. 2.2 EMPIRICAL TOXICOLOGIC ENDPOINTS AND METHODS FOR DETERMINING EXPOSURE CONCENTRATIONS USED TO DERIVE AEGLs 1, 2, AND 3 The selection of the biologic endpoints that serve as the thresholds for each of the AEGL severity levels are based on the definitions for the community emergency exposure levels (CEELs) that were published in the NRC guidelines for developing short-term exposure limits (NRC 1993a). The AEGLs address the same defined population as the NRC CEELs. The NRC definitions of the three CEEL tiers have been modified slightly by the NAC/AEGL Committee to improve only the clarity of description of the

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals threshold levels. Hence, the defined threshold levels for CEELs and AEGLs are the same. The NRC guidelines describe CEELs as exposure limits applicable to emergency exposures to hazardous substances (NRC 1993a). The NRC guidance states the CEELs must be set low enough to protect most of the population that may be exposed, including those with increased susceptibilities, such as children, pregnant women, persons with asthma, and persons with other specific illnesses (NRC 1993a). The NRC definition of CEELs for each of the three tiers of adverse health effects states that the adverse effects for each CEEL tier are not likely to occur below that level for a specified exposure duration but are increasingly likely to occur at concentrations above that level in a general population, including susceptible individuals. For this reason, the NRC also refers to the CEELs as threshold levels (NRC 1993a). Because the data and methodologies used to derive AEGLs or any other short-term exposure limits are not sufficiently precise to make a distinction between a ceiling value and a threshold value, no distinction has been made with respect to AEGL values. No fine line can be drawn to precisely differentiate between a ceiling level, which represents the highest exposure concentration for which an effect is unlikely to occur, and a threshold level, which represents the lowest exposure concentration for the likelihood of onset of a given set of effects. Hence, AEGLs are not true effect levels. Rather, they are considered threshold levels that represent an estimated point of transition and reflect the best efforts to establish quantitatively a demarcation between one defined set of symptoms or adverse effects and another defined set of symptoms or adverse effects. Therefore, in the development of AEGLs, the NAC/AEGL Committee selects the highest exposure level from animal or human data where the effects used to define a given AEGL tier are not observed. 2.2.1 Selection of the Highest Exposure Level at Which the Effects That Define an AEGL Are Not Observed Traditionally, when setting acceptable (typically considered “safe”) levels of exposure, the risk assessor will select the highest experimental exposure that does not cause an adverse effect (no-observed-adverse-effect level (NOAEL)) in an experiment that demonstrated a graded exposure response from no effect to adverse effects. In standard risk-assessment practice (NRC 1993a), the exposure level identified as the NOAEL would then be divided by appropriate uncertainty factors and modifying factors to derive an acceptable exposure level for humans. However, there are a number of limitations in this

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals method. It does not consider the number of animals used in the experiment and the associated statistical uncertainty around the experimental exposure level chosen. It does not consider the slope of the exposure-response relationship and subjects the risk assessor to using the possibly arbitrarily selected exposure levels chosen in the face of an unknown exposure-response relationship. Under some conditions, especially when only a small number of animals are exposed per exposure, the NOAEL could be a level associated with significant adverse health effects (Leisenring and Ryan 1992). In recent years, Crump and Howe (1984), Barnes et al. (1995), the U.S. Environmental Protection Agency (EPA 1995a), Faustman et al. (1994), Gaylor et al. (1998, 1999), and Fowles et al. (1999) addressed these problems by using the concept of analyzing all the data to estimate statistically a benchmark concentration (BMC). The BMC is a statistical estimate of an exposure that will cause a specified incidence of a defined adverse health effect. The BMC is commonly defined as the 95% lower confidence limit (LCL) on the exposure level causing a specified level of response (typically 1% to 10%). This exposure level is intended to replace the NOAEL and is used like the NOAEL when setting acceptable exposure levels. The BMC method has a number of advantages over the traditional NOAEL approach. The BMC is derived from a statistical analysis of the exposure-response relationship and is not subject to investigator selection of exposure levels. It is a reflection of the exposure-response curve. Although the number of animals used in a study will affect the NOAEL and BMC estimates, the BMC, when compared with the maximum likelihood estimate (MLE), will explicitly reflect the variability in the study and the uncertainty around the number of subjects. The greater the variability and uncertainty, the greater the difference between the BMC and the MLE. The BMC calculation allows for the statistical estimation of a BMC in the absence of an empirical NOAEL. The data most relevant to the development of AEGL-3 values and most amenable to a BMC analysis are inhalation LC50 (lethal concentration for 50% of the animals) data. Fowles et al. (1999) analyzed 120 inhalation animal lethality data sets by using the BMC method. The analysis provides the basis for the application of the BMC approach used by the NAC/AEGL Committee in the development of AEGL values. BMCs (95% LCL) and MLEs were developed for the 1%, 5%, and 10% response levels using log probit and Weibull models. Species tested included rats, mice, guinea pigs, hamsters, rabbits, and dogs. Exposure times ranged from 5 min to 8 h. Each data set consisted of at least four data points. The BMC and MLE values were compared with the empirical NOAEL (highest exposure that did not cause death in the experiment) and LOAEL (lowest exposure that killed at least one

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals animal). The curve generated by the statistical models was subjected to a chisquared goodness of fit test (p>0.05). For statistical and data presentation reasons, 100 studies were analyzed with the probit analysis and 93 with the Weibull model. Most of the studies reported NOAELs (81/100 considered for the probit analysis and 74/93 considered for the Weibull analysis). The BMCs were generally lower than the NOAELs when analyzed with either statistical estimate. The mean NOAEL/BMC ratios for the 1%, 5%, and 10% response were 1.60, 1.16, and 0.99, respectively, when using a probit analysis, and 3.59, 1.59, and 1.17, respectively, when using the Weibull analysis. It is interesting to note that comparable means from a Weibull analysis of developmental toxicity data were considerably greater, the developmental toxicity means of the NOAEL/BMC ratios were 29, 5.9, and 2.9 (Allen et al. 1994). The proportion of times that the NOAEL exceeded the BMC for the 1%, 5%, and 10% response was 89%, 65%, and 42%, respectively, for the probit analysis and 95%, 80%, and 54%, respectively, for the Weibull analysis. In all cases, the LOAEL/BMC ratio exceeded 1 for the probit and Weibull analysis of the 1% and 5% response but not always for the 10% response (99%). For this reason, the BMC10 may be too high a response rate to use to predict a NOAEL. In contrast, the corresponding 1% and 5% response ratios were always greater than 1. The ratios of the MLE/BMC were not great, ranging from a mean of 1.39 for a probit analysis of the 10% response to 3.02 for a Weibull analysis of the 10% response. It is important to note that when using the probit analysis, the LOAEL/MLE ratios were equal to or greater than 1 in 99%, 94%, and 71% of the cases for the 1%, 5%, and 10% responses, respectively. The MLE would probably be protective at the 1% response level but not for the 5% and 10% response levels. Similar numbers of 99%, 97%, and 76% were observed for the Weibull analysis. The BMC approach can provide a more refined assessment of the prediction of the empirical NOAEL. It must be emphasized that even the empirical NOAEL may represent a response level that is not detected. When 5 to 10 animals are used in an experiment, a 10–20% response can be missed (Leisenring and Ryan 1992) and even a BMC10 is similar to a LOAEL with dichotomized data (Gaylor 1996). It is expected that the BMC is less than the empirical LOAEL. In the Fowles et al. (1999) analysis of the data, the BMC05 and BMC01 values were always below the empirical LOAEL for the studies analyzed. The probit analysis of the data by Fowles et al. (1999) provided a better fit with the data as measured by the “chi-squared goodness-of-fit test, mean width of confidence intervals, and number of data sets amenable to analysis by the model.”

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals It is interesting to note that the BMC05 is very close to the MLE01 in the Fowles et al. (1999) evaluation of inhalation acute toxicity data. Through 1999, the NAC/AEGL Committee has used the MLE01 to estimate the highest exposure at which lethality is not likely to be observed in a typical acute exposure study. Given the analysis by Fowles et al. (1999) and for the above reasons, the NAC/AEGL Committee will generally use the BMC05 (lower 95% confidence limit (LCL) of the exposure required to produce a 5% response to exposure to chemicals) in the future for this estimate, although the MLE01 will also be calculated and considered. This approach incorporates the uncertainties due to the number of animals used in an experiment and the experimental variability observed; it utilizes all the data and the slope of the exposure-response curve and provides for a reasonable estimate of a predicted experimental NOAEL. In all cases, the MLE and BMC at specific response levels will be considered when setting AEGL values. Statistical models in addition to the log-probit will also be considered. Since goodness-of-fit tests consider an average fit, they may not be valid predictors of the fit in the low-exposure region of interest. In this case, the output of the different models will be plotted and compared visually with the experimental data in selection of the most appropriate model. It should be emphasized that these methods will generally be considered for an acute lethal endpoint. Their use to set AEGL-1 and AEGL-2 values will be considered on a chemical-by-chemical basis. Different endpoints may require the use of different data sets in different or the same species, a different benchmark dose approach, or identification of a different response level. These factors will be considered for specific chemicals and toxicologic endpoints. The preferred approach will be to use the BMC approach to identify the highest exposure at which the toxicologic effects used to define an AEGL tier were not observed. If the data are insufficient for a meaningful statistical analysis to use that approach, then the level will be determined empirically from experimental data. 2.2.2 Selection of Health-Effect Endpoints for AEGL-1, AEGL-2, and AEGL-3 In addition to the working definitions of the three AEGL tiers, this section includes a summary of the specific biologic endpoints used to establish the AEGL values for individual chemicals. Also included are general principles for selection of AEGL health-effect endpoints that have been derived from the committee’s selections on a chemical-by-chemical basis. Since ideal data sets

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals for certain chemicals are not available, extrapolation methods and the committee’s scientific judgment are often used to establish threshold values. In the absence of adequate data, the committee can decide that no AEGL value be established. The basis for this decision is the failure to achieve a minimum two-thirds majority of a quorum of the committee that is in favor of establishing a value or a formal decision by two-thirds of the committee not to establish a value. Under ideal circumstances, the specific health effects would be identified that determine each of the AEGLs. A search of the published literature for data on the chemical would be performed, and AEGLs would be generated from those data. However, data relating exposure and effect do not always follow an ideal paradigm and may lead to apparent inconsistences in the use of endpoints to set AEGLs. The general principles laid down in the NRC (1993a) guidance for evaluating data and selecting appropriate health effects, combined with professional judgment, are used to establish AEGLs. From the evaluations of the first four chemicals in the subcommittee’s first full report (NRC 2000b) and experience with data sets on chemicals currently under review, the following refinements to the NRC guidelines have been adopted by the NAC/AEGL Committee to set AEGLs. For reasons discussed earlier in this chapter, the NAC/AEGL Committee generally selects the highest experimental concentration that does not elicit the symptoms or effects defined by the AEGL tier in question. This concentration represents the starting point for AEGL development. In instances in which appropriate data are available, the BMC method may be considered and used to select the AEGL endpoints. 2.2.2.1 AEGL-1 Endpoints The NRC (1993a) guidelines discuss the definition of the CEEL-1 (AEGL-1) endpoint on pages 10, 12, and 21 of that report. Above the AEGL-1 value, discomfort becomes increasingly likely. Below the AEGL-1 value (detectability), “Exposure insufficient to cause discomfort or adverse health effects may be perceived nevertheless by means of smell, taste, or sensations (mild sensory irritation) that are not uncomfortable. The awareness of exposure may lead to anxiety and complaints and constitutes what is termed here detectability.” (NRC 1993a, p. 21). Thus, below the AEGL-1 values, there may be specific effects, such as the perception of a disagreeable odor, taste, or other sensations (mild sensory irritation). In some people, that exposure level could result in mild lacrimation or coughing. Since there is a continuum in which it is difficult to judge the appearance of “discomfort” in animal studies and human experiences, the

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals NAC/AEGL Committee has used its best judgment on a case-by-case basis to arrive at appropriate and reasonable AEGL-1 values. One additional factor to consider is that the three tiers of AEGL values “provide much more information than a single value because the series indicates the slope of the dose-response curve” (NRC 1993a). If an accident occurs and people smell or otherwise “detect” a chemical, the extent of the concentration range between AEGL-1 and AEGL-2 values provides useful information and insight into the estimated margin of safety between a level of detection or mild sensory irritation (AEGL-1) and a level that may impair escape or lead to a serious long-term or irreversible health effect (AEGL-2). In cases in which the biologic criteria for the AEGL-1 value would be close to, or exceed, the AEGL-2 value, the conclusion is reached that it is “not recommended” (NR) to develop AEGL-1 values. In these cases, “detectability” by itself would indicate that a serious situation exists. In instances in which the AEGL-1 value approaches or exceeds the AEGL-2 value, it may erroneously be believed that people experiencing mild irritation are not at risk when in fact they have already been exposed to extremely hazardous or possibly lethal concentrations. Since a comparison of the AEGL-1 and AEGL-2 values indicates the slope of the dose-response curve that may be of value in emergency response, planning, or prevention, the NAC/AEGL Committee also attempts to establish AEGL-1 endpoints for adverse effects that are asymptomatic or nonsensory. Examples of such effects include significant (measurable) levels of methemoglobin, elevated blood enzyme levels, or other biologic markers related to exposure to a specific chemical. By establishing an AEGL-1 value in these instances, important information on the toxicologic behavior of a specific chemical is available to emergency responders and planners. The following criteria have been used by the NAC/AEGL Committee to select endpoints for use in setting the AEGL-1 values. 2.2.2.1.1 No Value Established—AEGL-1 Is Close to or Exceeds AEGL-2 State what aspects of the chemical toxicity profile make it inadvisable to recommend an AEGL-1 value. For example, the AEGL-1 value is not established, because levels that are “detectable” are close to, or exceed, an AEGL-2 value. These materials have poor warning properties. 2.2.2.1.2 No Value Established—Insufficient Data Insufficient data were available to establish AEGL-1.

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals 2.2.2.1.3 Highest Experimental Exposure Without an AEGL-1 Effect State the species, effect, concentration, and exposure time to cause the effect. Describe the toxicologic endpoint of concern. For example, the highest experimental exposure levels that did not cause (a) sensory irritation, (b) altered pulmonary function, and (c) narcosis in humans have been used to set AEGL-1 values. 2.2.2.1.4 Effect Level for a Response State the species, effect, concentration, and exposure time to cause the effect. Describe the toxicologic endpoint of concern. For example, levels for odor detection in humans, mild sensory irritation, asymptomatic or nonsensory effects, such as methemoglobin formation (22%) and altered pulmonary function (transient changes in clinically insignificant pulmonary functions of a susceptible individual), have been used as AEGL-1 endpoints. 2.2.2.2 AEGL-2 Endpoints NRC (1993a) discussed the CEEL-2 (AEGL-2) definition on pages 10, 12, and 21 of its report. The AEGL-2 has been defined as the threshold between reversible effects that cause discomfort and serious or irreversible health effects or effects that impair escape. Above the AEGL-2 value, there is an increasing likelihood that people may become disabled or are increasingly likely to experience serious or irreversible health effects. “The term disability is used here to indicate the situation where persons will require assistance or where the effects of exposure will be more severe or prolonged without assistance” (NRC 1993a, p. 21). In developing AEGL-2 values, the NAC/AEGL Committee estimates a NOAEL for serious or irreversible effects or effects that impair escape. It must be emphasized that reversible clinical toxicity may be observed below the AEGL-2 value. If minor reversible effects are seen at one level of exposure and disabling effects at a higher exposure, the former is used to set the AEGL-2 value. If the exposure associated with disabling effects cannot be determined from experimental data, then the highest level causing reversible effects and discomfort may be used to set the AEGL-2 value.

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals The following criteria have been used by the NAC/AEGL Committee to select endpoints for use in setting the AEGL-2 values. 2.2.2.2.1 Highest Experimental Exposure Without an AEGL-2 Effect State the species, effect, concentration, and exposure time to cause the effect. Describe the toxicologic endpoint of concern. The highest experimental exposure levels that did not cause decreased hematocrit, kidney pathology, behavioral changes, or lethality (effects observed at higher exposures were above the definition for AEGL-2) have been used as the basis for determining AEGL-2 values. 2.2.2.2.2 Effect Level for a Toxic Response That Was Not Incapacitating or Not Irreversible State the species, effect, concentration, and exposure time to cause the effect. Describe the toxicologic endpoint of concern. For example, overt ocular and/or respiratory tract irritation, dyspnea, pulmonary function changes, provocation of asthma episodes, pathology (respiratory tract), mild narcosis, and methemoglobin formation (approximately 40%) have been used as a basis for AEGL-2 values. 2.2.2.2.3 A Fraction of the AEGL-3 Value State the rationale for using a fraction of the AEGL-3. State why the specific fraction chosen is scientifically justified. In the absence of specific data used to determine an AEGL-2 value, one-third of the AEGL-3 value has been used to establish the AEGL-2 value. This approach can only be used if the data indicate a steep exposure-based relationship based on data for effects below the AEGL-2 value and lethal-effect value. 2.2.2.3 AEGL-3 Endpoints NRC (1993a) discussed the CEEL-3 (AEGL-3) definition on pages 10, 12, and 21 of its report. The AEGL-3 tier has been defined as the threshold exposure level between serious long-lasting or irreversible effects or effects

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals that impair escape and death or life-threatening effects. Above the AEGL-3, there is an increasing likelihood of death or life-threatening effects occurring. In determining AEGL-3 values, the NAC/AEGL Committee defined the highest exposure level that does not cause death or life-threatening effects. It must be emphasized that severe toxicity will be observed at levels exceeding the AEGL-3. In cases in which data to determine the highest exposure level that does not cause life-threatening effects are not available, levels that cause severe toxicity without producing death have been used. The following criteria have been used by the NAC/AEGL Committee to date to select endpoints for use in setting the AEGL-3 values. 2.2.2.3.1 Highest Exposure Level That Does Not Cause Lethality—Experimentally Observed Threshold (AEGL-3 NOAEL) State the species, effect, concentration, and exposure time to cause the effect. Describe the toxicologic endpoint of concern. When experimental lethality data are insufficient to determine statistically a BMC, the highest experimental exposure that did not cause lethality in an experiment in which death was observed was used to set the AEGL-3 value. 2.2.2.3.2 Highest Exposure Level That Does Not Cause Lethality—Estimated Lethality Threshold—One-Third of the LC50 State the species, effect, concentration, and exposure time to cause the effect. Describe the toxicologic endpoint of concern. If an exposure that does not produce death is estimated by dividing an LC50 value by 3 (or some other divisor), then give the slope of the exposure response curve or enough data points to support the division by 3 (or some other divisor). When experimental lethality data have been insufficient to determine statistically an LC01 value, but an LC50 value was determined and all exposure levels caused lethality, a fraction of the LC50 value is used to estimate the threshold for lethality. In all cases, the exposure-response curve was steep, and the LC50 value was divided by 3. Fowles et al. (1999) analyzed 120 published inhalation animal lethality data sets using the BMC method. Their analyses of inhalation toxicity experiments revealed that for many chemicals the ratio between the LC50 and the experimentally observed nonlethal level

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals levels in the range of 10−4 to 10−6 were generally considered acceptable risk levels (NRC 1986). However, the document states, “The role of short-term exposures in producing cancer is not clear…. [A]ny exposure to a carcinogen has the potential to add to the probability of carcinogenic effects…[but] the effects of long or repeated exposures could greatly overshadow brief exposures (up to 24 h).” Additionally, the NRC report states, “The assumption that the carcinogenic response is directly proportional to total dose is likely not to hold for all materials and all tissues that these materials affect.” However, these concerns not withstanding, the NRC set SPEGL values based on the carcinogenic risk-assessment method previously mentioned for hydrazine, methyl hydrazine, and 1,1-dimethyl hydrazine (NRC 1985). In each case, the theoretical excess cancer risk level used was 10−4, and the derived values were determined to be lower than corresponding airborne concentrations that were estimated to cause acute toxicity. SPEGL values for exposure periods of less than 24 h of other known or suspect human carcinogens were not based on carcinogenicity. These chemicals included benzene, trichloroethylene, ethylene oxide, and lithium chromate. The National Aeronautics and Space Administration (NASA) requested that the NRC develop spacecraft maximum allowable concentrations (SMACs) for space-station contaminants. The NRC published guidelines for the development of short-term and long-term SMACs (NRC 1992a). Short-term SMACs refer to concentrations of airborne substances that will not compromise the performance of specific tasks during emergency conditions lasting up to 24 h. Because of NASA’s concern for the health, safety, and functional abilities of space crews, SMACs for exposure from 1 to 24 h should not cause serious or permanent effects but may cause reversible effects that do not impair judgment or interfere with proper responses to emergencies. The long-term SMACs are designed to prevent deterioration in space-crew performance with continuous exposure for up to 180 days. The guidelines for determining SMACs for carcinogens recommend the methods proposed by Kodell et. al. (1987) based on the linear multistage model. The level of theoretical excess risk used in the computation was 10−4. The guidelines suggest extrapolations of long-term (often lifetime) exposures to shorter durations, such as 1, 30, or 180 days, and refer to a single-day exposure as “the case of near instantaneous exposure.” Further, the guidance states, “It must be remembered that extrapolation from a daily lifetime exposure level and conversion to an instantaneous exposure level using…[equations presented]…is an extreme case and is valid only under the assumptions underlying the multistage theory of carcinogenesis.” A review of the first three volumes of published SMACs (35 chemicals), including 10 known or suspected carcinogens, indicated that an assessment of excess risk for less than a 24-h exposure period was conducted on only 1 of the 10 carcinogenic

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals substances. Carcinogenic assessments for excess risk were conducted on all 10 chemicals for 24 h, as well as 7, 30, and 180 days. The reasons provided in the NRC report for not undertaking a risk assessment on carcinogenic substances for exposure periods of less than 24 h included the following: (1) “Data not considered applicable to the exposure time (1 hr)”; (2) “Extrapolation to 1 hour exposure duration produces unacceptable uncertainty in the values”; and (3) “The NRC model was not used to calculate acceptable concentrations for exposures shorter than 24 hours” (NRC 1992a). As stated previously, to date no U.S. federal or state regulatory agency has promulgated or established regulatory limits for single short-term (less than 24 h) exposures based on carcinogenic properties. 2.8.3 Scientific Basis for Credible Theoretical Excess Carcinogenic Risk Assessments for Single Exposures of 8 Hours or Less The NRC guidance (NRC 1993a) explains that AEGLs can be developed using carcinogenic risk-assessment methods for exposure durations of 1 to 8 h provided adequate data are available. However, the guidance states that risk assessments on chemical carcinogenicity in humans should be based on all relevant data and embody sound biologic and statistical principles. While some of the substances may be considered known human carcinogens, most of the information is based on animal testing information. Additionally, since the mode of action for animal carcinogens are not always the same with respect to biologic properties among animal species or strains and humans, a weight-of-evidence evaluation must be carried out on a case-by-case basis. The weight-of-evidence evaluation considers comparative metabolic disposition, dose-dependent pharmacokinetic parameters, routes of exposure, mechanisms of action, and organ or species differences in response in animals and humans. Uncertainties regarding lifetime theoretical excess carcinogenic risk assessments increase as shorter durations of a single exposure are considered. Most of these concerns stem from the reliance of both conclusions of carcinogenicity and quantitative assessments on long-term exposures in humans in occupational settings or in test animals. Thus, calculations for short-term risks require substantial extrapolation. At the same time, there are special concerns and unresolved issues regarding short exposures that will require more relevant data before they can be resolved. As evidenced from the actual application of these guidelines, COT was reluctant in most cases to develop quantitative carcinogenic risk assessments for less than 24-h exposures in the development of SMACs. To better understand the empirical database for single exposures, EPA

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals funded a study for the AEGL Program (Dr. Edward Calabrese of the University of Massachusetts was given a contract) to review the published literature and assess the circumstances during which a single exposure of short duration may be associated with a confirmed increase in carcinogenic response. This effort, referred to as the Single Exposure Carcinogen Database, represents a computerized summary that may assist the NAC/AEGL Committee’s assessment of whether a single exposure to a particular chemical under consideration for AEGL development might cause tumor development following a one-time inhalation exposure. The data base is designed to contain numerous parameters important to tumor outcome and/or quality of the studies conducted. The database will contain approximately 5,500 “studies” or data sets involving approximately 500 chemicals from nearly 2,000 references. Although a summary of the Single Exposure Carcinogen Database Project has been presented to the NAC/AEGL Committee, at the present time it is not known whether the data available on single exposures of carcinogenic substances will be sufficient to justify their use in the development of AEGL values. A preliminary review of the database indicates that only a limited number of short-term carcinogenesis bioassays conducted by the inhalation route are available. Hence, route-to-route extrapolations would be required in a manner that would not substantially weaken the conclusions reached for certain substances using standard EPA or NRC procedures if the toxicant is likely to cause tumors at a site other than the port of entry. If the substance causes tumors at the site of application or port of entry in oral or parenteral protocols, extrapolation to the inhalation route of exposure becomes problematic. For this reason, the NAC/AEGL Committee in most cases will rely on data from long-term animal studies as the basis for the quantitative cancer risk assessments for short-term exposures of 8 h or less. It is anticipated that work on the Single Carcinogen Database may be completed in 2001. The Single Carcinogen Database may prove to be useful in obtaining some important information for AEGL development. The database shows that single exposure to various chemical classes, using various species and strains of animals, can result in tumor formation. Furthermore, chemicals can be selected from the database for which there is dose-response information. Data and information from positive responses of the chemical in the database could be compared between the single-dose study and the long-term study. 2.8.4 Practical Issues of Using Quantitative, Carcinogenic Risk Assessments for Developing AEGLs In addition to fundamental scientific issues regarding carcinogenic risk assessments in the development of AEGL values, there are important practical

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals issues to be considered by emergency planners and responders regarding AEGL values based on possible carcinogenic risk. The theoretical excess cancer risk for a lifetime exposure to known or suspect human carcinogens considered safe and protective of the public health ranges from 10−6 to 10−4 for EPA and most other U.S. federal regulatory agencies (EPA 1991). The AEGL values, however, are designed for emergency planning for, response to, and prevention of accidental releases from chemical accidents. Thus, theoretical excess cancer risk may be accumulated in 30 min or in a few hours. In addition to the individual risk of 10−6 to 10−4, one should also consider a measure of population based risk. Experts in the chemical accident field indicate that the typical U.S. population at risk during most accidental chemical releases is in the range of 1,000 to 5,000 persons. The actual number of persons exposed depends on many factors, such as population density, quantity released, release rate, prevailing wind direction and velocity, terrain, and ambient temperature to name a few. Therefore, a population-based risk range of 10−6 to 10−4, assuming a credible carcinogenic assessment can be made, approaches zero for a population of 1,000 to 5,000 or higher. The consideration of population-based risks using assessment methodologies designed for individual risks has precedent in EPA assessments of new industrial chemicals under TSCA (Toxic Substance Control Act) Section 5 and pesticide chemicals under FIFRA (Federal Insecticide Fungicide and Rodenticide Act). Implementation of emergency-response procedures based on theoretical excess risk values of 10−6 to 10−4 may be problematical. For example, if such values were used, they would be based on an anticipated increased cancer risk of 10−6 to 10−4, a policy consistent with EPA’s acceptable cancer risk for lifetime exposures to known or suspect human carcinogens. However, the public health and safety risks associated with evacuation and other response measures might pose greater risks of injury or perhaps death. Thus, setting AEGL values based on uncertain theoretical cancer risk estimates might lead to response measures that increase actual or total risk for the exposed population. 2.8.5 Current Approach of the NAC/AEGL Committee to Assessing Potential Single Exposure Carcinogenic Risks On the basis of the discussions and considerations presented in the earlier sections of this chapter on cancer risk assessment, the NAC/AEGL Committee has developed no AEGL values based on carcinogenicity. In view of the great uncertainty of the assumptions used in extrapolating from lifetime exposures to 8 h or less, the paucity of single-exposure inhalation data, the relatively

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals small populations involved, and the potential risks associated with evacuations and other response measures, the NAC/AEGL Committee does not believe their use in setting AEGL values is justifiable at the present time. However, the NAC/AEGL Committee will continue to identify and evaluate carcinogenic data during the development of AEGLs on a chemical-by-chemical basis. The scientific parameters used in this analysis are presented later in this section. In those cases in which, in the judgment of the committee, it is appropriate, risk assessments for 10−4, 10−5, and 10−6 levels of cancer risk will be conducted. It is believed that information on known or suspect human carcinogens should be provided to emergency planners and responders and made available to the public even when such information is not used to set AEGL values. Therefore, the NAC/AEGL Committee will continue to provide data and information on the carcinogenic properties of chemicals in the TSDs and, in instances in which the appropriate data are available, develop quantitative cancer risk assessments at risk levels of 10−4, 10−5, and 10−6 in accordance with the NRC guidance (NRC 1993a). The NAC/AEGL Committee will attempt to limit potential cancer risk to 10−4 or less when there are scientifically credible data to support the risk estimates when based on a single exposure. When substantial and convincing scientific data become available that clearly establish a relationship between a single short-term inhalation exposure to a chemical and the onset of tumors that are likely to occur in humans, the carcinogenic risk in the development of the appropriate AEGL values will be given appropriate weight-of-evidence considerations. 2.8.5.1 Evaluation of Carcinogenicity Data The evaluation of the carcinogenic potential of a chemical exposure in humans must be based on analyses of all relevant data. Human epidemiologic and clinical studies, as well as accidental-exposure reports are considered and used to evaluate the carcinogenic potential of a substance. In the absence of human data, long-term bioassay data from controlled animal studies are used to derive theoretical excess carcinogenic risk estimates for exposed humans. The selection of data for estimating risk is based on the species and strain considered to resemble the human response most closely to provide the most accurate estimates. Data suggestive of a single exposure inducing a carcinogenic response, including related mechanistic data that support such a possibility, are considered. Weight should be given to those studies most relevant to estimating effects in humans on a case-by-case basis. Data for assessing the strength of conclusions drawn from controlled animal studies should include information

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals on comparative metabolic pathways, dose-dependent pharmacokinetic parameters, mode of exposure, mechanisms of action, and organ or species differences in response. In general, the NAC/AEGL Committee will follow a weight-of-evidence approach in the evaluation of carcinogenicity that is consistent with the availability and biologic variability of the data and its relationship to the likelihood of effects in humans. 2.8.5.2 Methodology Used for Assessing the Carcinogenic Risk of a Single Exposure Guidance published by the NRC (1993a) states that the setting of AEGLs (CEELs) should involve linear low-dose extrapolation from an upper confidence limit on excess risk for genotoxic carcinogens and for carcinogens with mechanisms of action that are not well understood. More specifically, the NRC guidance suggests an approach utilizing the methods proposed by Kodell et al. (1987) based on multistage models. Although the NRC guidance states that multistage models could be useful for setting AEGL values, the guidance acknowledges that sufficient information may not be available to postulate the total number of stages in the cancer process and the stages that are dose-related. In these instances, the NRC guidance recommends the use of the time-weighted-average dose where the instantaneous dose D at time t0 is assumed to be the equivalent of the lifetime excess carcinogenic risk as daily dose D up to time t. This equivalence is expressed by the equation D=d×t. As shown by Kodell et al. (1987), the actual risk will not exceed the number of stages in the model (k). In instances in which multistage models can be used and prudence dictates conservatism, the NRC guidance suggests reducing the approximation of D by an adjustment factor of 2 to 6, depending on the number of assumed stages in the multistage model used. To date, the NAC/AEGL Committee has evaluated excess theoretical risk at levels of 10−4, 10−5, and 10−6 for a one-time exposure to known or suspect human carcinogens by determining the total cumulative lifetime dose and applying Haber’s law for exposure periods ranging from 30 min to 8 h. The resultant doses are then divided by an adjustment factor to account for the multistage nature of carcinogens (see Appendix H). 2.8.5.2.1 Determination of an Adjustment Factor for the Dose-Dependent Stage of Carcinogenesis There is an extensive body of literature that deals with the concept of

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals malignant tumor development, progression of an initiated cell through of successive stages, and quantitative carcinogenic risk assessment. Two references, Crump and Howe (1984) and Kodell et al. (1987), are cited in the NRC (1993a) publication. The concept has been further discussed (Goddard et al 1995; Murdoch et al. 1992; Murdoch and Krewski 1988; Bogen 1989). This process is referred to as a cell kinetic multistage model. There are several published variations of the basic tenants in the model. If only one or more stages are dose-dependent and exposure is concentrated in the dose-dependent stage, it is possible to underestimate risk when the risk is based on lifetime exposure. For example, if the first stage is dose-dependent, and there is a single exposure to an infant, the probability of cancer induction is maximized, because the entire lifetime of the individual is available for progression through the remaining stages in the development of the cancer. If the same dose were given to an elderly person, the probability of inducing cancer approaches zero, because there is insufficient time remaining in the life of that individual for the initiated cell to progress through the subsequent stages to a malignant cancer. Kodell et al. (1987) demonstrated that the underestimation of risk that is based on a lifetime of exposure will not exceed the number of stages in the multistage model. For this reason, the NRC (1986) recommends dividing the risk assessment based on the lifetime exposure by a factor between 2 and 6 to account for the number of stages in the multistage model applicable to the particular chemical of concern. In addition to the multistage model, there have been a number of publications investigating the two-stage birth-death-mutation model (Morrison 1987; Chen et al. 1988; Murdoch and Krewski 1988; Moolgavkar and Luebeck 1990; Murdoch et al. 1992; Goddard et al. 1995). This model is similar to the multistage model. However, the impact of the number of stem cells at the time of chemical exposure is considered as well as the net growth rate of cells that have undergone the first stage of initiation. If the first-stage initiating event creates a cell that has a net growth rate greater than that of the stem cell, then the risk of that initiating event will be greater than it would be if the initiated cell grew at the same relative rate as the stem cell. In this case, exposure early in life results in a greater risk than exposure late in life. Conversely, exposure to a promoter (effects only the second stage) late in life will be more effective than early exposure, because relatively more initiated cells are present. If this stage is the only stage affected by the chemical, this situation is the same as that in the two stages of the multistage model. However, if the net growth rate of the initiated cells is 10 times the stem-cell rate, the relative effectiveness of exposure late in life could be 10-fold (Murdoch and Krewski 1988). Exposure to promoters between the first- and second-stage event can have an

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals impact by increasing the net growth rate of initiated cells over that of stem cells. For maximum effectiveness, exposure to promoters (generally considered to be a nongenotoxic event) must encompass multiple challenge (Chen et al. 1988; Murdoch and Krewski 1988). Thus, the cancer risk associated with a single exposure to a promoter should not be greater than that predicted for multiple exposures, and no correction to the estimated risk has to be made in this case. A major impact upon the risk assessment of the two-stage model comes from carcinogen exposure during the first stage in which the initiation event creates a cell with a greater net growth rate than the stem-cell rate. Modelers have considered a number of scenarios in which the net growth rate of initiated cells varies from −10 to +10. The greatest increase in risk in the two-stage model occurs when the first stage is dose-dependent and the initiating event creates a cell with a net growth rate of +10. In that case, the increased risk is 10-fold (Murdoch and Krewski 1988; Murdoch et al. 1992; Goddard et al. 1995). Data are lacking on the biologic plausibility of the maximum value for the net growth rate of initiated cells (Murdoch et al 1992). Major data requirements for the two-stage birth-death-mutation model include the number of stem cells at different times of the life cycle, their rates of division and differentiation, and their response to chemical exposure in terms of cell division and mutation rate. This information is also needed for the initiated-cell populations (Moolgavkar and Luebeck 1990). Because of these data gaps, the projections made for the two-stage model remain more speculative than those for the multistage model in which there is general agreement that the number of stages should not exceed six. For the above reasons, the linearized multistage model is used as a default when estimating risks for short-term exposures from lifetime carcinogenesis bioassays. In all of the above referenced publications on the multistage model, the maximum number of stages modeled was six. AEGL values are applicable to humans in all stages of life, so the maximum risk to an infant must be taken into consideration. In this case, the concentration based on a lifetime exposure study is divided by 6 unless there is evidence that the chemical is a late-stage carcinogen or operates by mechanisms different from those assumed in development of the linearized multistage model. As a first approximation, the NAC/AEGL Committee will use the divisor of 6 in agreement with the NRC (1993a) guidance on the development of short-term exposure limits, which states that a factor of 6 represents a conservative adjustment factor for a near-instantaneous exposure.

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals 2.8.5.3 Summary of Cancer-Assessment Methodology Used by the NAC/AEGL Committee The EPA q1* values listed on the Integrated Risk Information System (IRIS) or the GLOBAL86-generated slope-factor values (Howe et al. 1986) are used to compute lifetime theoretical excess carcinogenic risk levels. These values are based on EPA (1986) guidance. The EPA (1996a) proposed methodology will be considered in the future. These values are used to compute the concentration for a single exposure for the time periods of interest. As discussed in the beginning of this section, these values are typically divided by 6 to account for early exposure to a carcinogen in which the first stage is dose-dependent or for late exposure to a carcinogen in which the last stage is dose-dependent. If there is information about the number of stages required for development of the cancer or the stage that is dose-dependent, the divisor will be modified accordingly. An example of a carcinogenicity assessment is given in Appendix H. The cancer evaluation includes a weight-of-evidence discussion, which considers the following factors: Less evidence of carcinogenicity from a short-term exposure. No evidence for human carcinogenicity (may or may not lend support to cancer induction from a single exposure but an important consideration). Lifetime or long-term exposure necessary to elicit cancer. Positive response only at very high doses. Neoplasia appears reversible (when treatment is discontinued). Appears to be a “threshold” carcinogen. Weak or absent mutagenic response in multiple in vivo and in vitro test systems. Greater evidence of carcinogenicity from a short-term exposure. Confirmed human carcinogen (may or may not lend support to cancer induction from a single exposure but an important consideration). Short time to tumor. Evidence for cancer from one to a few exposures. Positive response at low doses. Complete carcinogen. Irreversible (when treatment is discontinued). Potent mutagen in multiple in vivo and in vitro test systems.

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals 2.9 GUIDELINES AND CRITERIA FOR MISCELLANEOUS PROCEDURES AND METHODS 2.9.1 Mathematical Rounding of AEGL Values Given the uncertainties involved in generating AEGL values, it could be argued that only one significant figure should be used. However, because of a number of considerations discussed below, AEGL numbers are rounded to two significant figures (e.g., 1.5, 23, or 0.35). The value 7.35 is rounded to 7.4. Trivial differences in numbers can give large differences in practice if only one significant figure is used. For example, values of 14.9 and 15.1 would yield AEGL values of 10 and 20, respectively. This is a 2-fold difference for a very small difference in computed AEGL values. Values of 18, 14, 11, and 6 ppm for 30 min, 1 h, 4 h, and 8 h would give values of 10, 10, 10, and 20 ppm, respectively, for the time points. As these numbers are often used in exposure models to make risk-management decisions, the use of two significant figures allows for a more reasonable progression when different exposure scenarios are considered. Two significant figures may seem overly precise when values less than 1 ppm are presented, because those levels may be difficult to quantify to that level of precision. However, the AEGL-2 values will often be used to compare with ambient air-dispersion modeling projections for planning purposes. In this case, the use of two vs one significant figure can have substantial practical impact. Other rounding schemes may be used on a case-by-case basis with justification. 2.9.2 Multiplication of UFs The NAC/AEGL Committee often multiplies two UFs of 3. Since the value 3 represents the geometric mean of 10 and 1, the actual number is 3.16. Therefore, the product of two different UFs is not 3×3 but 3.16×3.16, which equals 10. For simplicity’s sake, 3.16×10 is represented by 30. 2.9.3 Conversion Between Parts per Million and Milligrams per Cubic Meter Expressing the airborne concentration of a chemical in parts per million represents a volume-by-volume approach, and milligrams per cubic meter represents a mass-by-volume approach to quantifying the concentration. Be-

OCR for page 34
Standing Operating Procedures for Developing Acute Exposure Guideline Levels for Hazardous Chemicals cause a change of temperature with little or no change in pressure or a change in pressure with little or no change in temperature will result in changes in the volume of the air with no change in the mass of the chemical dispersed in the air, the airborne concentration of the chemical expressed as milligrams per cubic meter can vary at different elevations above sea level and at different temperatures at the same elevation. Airborne concentrations expressed as parts per million represent a volume-by-volume comparison and therefore do not change, regardless of changes in elevation (pressure) or temperature. AEGLs are expressed in parts per million. However, many inhalation studies on toxicity report the chemical concentrations in milligrams per cubic meter. In deriving AEGL values, it is assumed that the concentrations reported were measured at normal temperature and pressure (i.e., 25° C or 298° K and 760 mm Hg). The NAC/AEGL Committee uses this assumption in all cases in which concentrations in milligrams per cubic meter are converted from parts per million, or data in parts per million are converted to milligrams per cubic meter. The effect of elevation above sea level is approximately 15% when comparing mass-by-volume (milligrams per cubic meter) in New York City and Denver, Colorado. Although elevation may not be a major consideration, the effect of converting parts per million to milligrams per cubic meter at various pressures (elevations) and temperatures can be made by using the following equation: mg/m3 at Pa and Ta=(ppm)×((MW)/(24.45×(760 mm Hg)×Ta/(Pa×(298°K)))), where Pa=the absolute pressure (in mm Hg) at actual conditions. Ta=the absolute temperature (in °K) at actual conditions. MW=molecular weight.