Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 21 2 Assessment of Health Risks from the Use of Flame Retardants THE assessment of health risks from the use of flame retardants (FRs) will be central to the U.S. Consumer Product Safety Commission's (CPSC) determination of whether substances likely to be used as FRs in residential upholstered furniture would present a hazard to consumers. Risk estimates derived from this process are often vital in making informed and balanced decisions to effectively manage risks and to communicate about the significance of known or anticipated risks. The risk assessment process enables the systematic evaluation of data and the quantitative presentation of complex information. This chapter provides an overview of the risk assessment process and describes procedures applied for interpretation of data on physical and chemical properties, toxicity, and exposure to estimate human health risks from exposures to FRs. Many documents are available to the reader to obtain more detailed information about risk assessment procedures and their applicability; some include: EPA (1986, 1992a, b, 1996, 1999a), NRC (1977, 1980, 1983, 1994), Tardiff and Rodricks (1987), and WHO (1978, 1984, 1986, 1999). The basic principles guiding the assessment of a substance's toxicity are outlined in the Guidelines for Carcinogen Risk Assessment (EPA 1999b); Chemical Carcinogens: A Review of the Science and Its Associated Principles (OSTP 1985), Guidelines for Developmental Toxicity Risk Assessment (EPA 1991a).
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 22 THE SUBCOMMITTEE'S RISK ASSESSMENT PROCESS Risk assessment is the process of quantitatively determining the likelihood of adverse effects resulting from exposures to FR chemicals. The risk assessment paradigm used in this report and widely acknowledged by the U.S. scientific community to be the current standard used in industrial risk management and public policy settings was first articulated by the U.S. National Research Council in its report, Risk Assessment in the Federal Government: Managing the Process (NRC 1983). In a sequel entitled Science and Judgment in Risk Assessment (NRC 1994), the National Research Council defined major elements in the estimation of health risks and placed the process within contexts of managing such risks. Among the elements elaborated in detail are (1) the values and limitations of default assumptions and the circumstances under which they should be replaced with empirical findings; (2) the importance of understanding the mode of toxic action of a compound to provide increased accuracy in defining the presence of hazards to human health; (3) the role of variability in human response as a basis for deciding the degree of health protection for susceptible subpopulations; and (4) aggregation of exposures and risks from all sources of exposure so as to provide realistic estimates of risk to specific groups of individuals. In this report, health risks are characterized for potential effects in the body after repeated and prolonged exposures to FRs. Although acute toxicity information was reviewed, it was not used in developing risk assessments because chronic exposure data are more relevant. The four basic steps of risk assessment process applied to FRs are hazard identification, dose-response assessment, exposure assessment, and risk characterization. Each is described below. Hazard Identification In the hazard identification step, a determination was made of causal relationships that exist between exposure to an FR and its adverse health effects. It involves gathering and critically evaluating toxicity data on the types of health effects that might be produced by an FR and on the conditions of exposure under which adverse effects are produced. Such an evaluation requires the development of the weight of all the evidence related to the toxicity of an FR. Toxicity data are derived from observations of humans (epidemiological studies, clinical findings, and case reports), from investigations of laboratory animals (most often rodents), and from in vitro studies.
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 23 Observations of Humans Epidemiological studies provide the most relevant kind of information for hazard identification, because they involve observations of humans, not laboratory animals. This obvious and substantial advantage is offset by various difficulties associated with obtaining and interpreting epidemiological findings. Rarely are convincing causal relationships identified with a single or even a few studies. Epidemiologists usually weigh the results from several studies, ideally involving different populations and investigative methods, to determine the degree of consistency between exposures and responses among them, including the dose-response relationships. Few human data were found for the FRs reviewed in this report. Studies of Laboratory Animals When human studies are unavailable or unsuitable, risk assessments for FRs are based on findings in studies on laboratory animals (usually rodents). An advantage of animal studies is that they can be controlled, so establishing causation (assuming that the experiments are well conducted) is not in general difficult. Another advantage is that animals can be used to collect toxicity information on chemicals before being marketed, whereas epidemiological data can be collected only after human exposure has occurred. Indeed, many countries require that some classes of chemicals (e.g., pesticides, food additives, and drugs) be subjected to toxicity testing in animals to demonstrate appropriate safety before the product can be marketed. Other advantages of animal studies include the facts that: â¢ A quantitative relationship between exposure (or dose) and extent of toxic response can be established. â¢ Animals and animal tissues (including some that would be inaccessible in humans) can be closely examined by toxicologists and pathologists, so the full range of toxic effects produced by a chemical can be identified and, in some instances, the progression to physiologic impairment can be characterized. The reversibility of adverse effects once exposure ceases can also be studied. Animal data reviewed included studies of neurotoxicity, immunotoxicity, reproductive and developmental toxicity, organ toxicity, dermal and pulmonary toxicity, and other local and systemic effects. Toxicokinetic studies were also
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 24 reviewed to understand the absorption, distribution, metabolism, and excretion of the FRs. In Vitro Studies The subcommittee also evaluated the result of in vitro studies. These studies are inexpensive and can be done in a relatively short time. The use of in vitro tests is increasing because they reduce the need to use experimental animals. These studies also provide some insight into possible mode of action of a chemical. They also provide information on structure-activity relationships (SAR). The SAR information can be useful when the data on chemical under consideration are inadequate or nonexistent. The results of these studies are important in hazard identification and can provide supporting data when the animal data are inadequate or absent. Discussion The subcommittee's hazard-identification step describes the types of toxic responses, if any, that can be caused by the FR under review, the weight of the primary and supporting evidence, the scientific merits of the data, and their value and reliability for estimating human toxicity under defined conditions of exposure. The weight of evidence analyses for the individual FRs takes into account replication, reproducibility, and concordance of results, as well as the degree of correspondence between observations in experimental animals and expected responses in humans for a given form of toxicity. These concepts are described in detail elsewhere (Tardiff and Rodricks 1987). The product of the hazard identification leads only to a statement about the toxic properties of an FR that may occur in humans. It does not reveal whether the FR poses a risk for specific populations with specific exposure circumstances. That determination requires three additional analytic steps: (1) evaluation of toxic potency of the FRs by examining their dose-response characteristics, (2) characterization of the nature and magnitude of human exposure, and (3) characterization of risk by combining the information on magnitude of exposure with dose-response relationships. Toxic Potency or Dose-Response Assessment Toxic potency or dose-response assessment information is used to determine the quantitative relationship between increases in the dose of an FR and changes in magnitude of the incidence and/or severity of adverse health effects.
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 25 The slope of the dose-response curve in the low-dose range, combined with the increment in exposure (dose), provides a quantitative estimate of the increase in incidence or severity of some adverse effect within an exposed population group. Dose-response relationships are often grouped into two classes based on two distinct modes of toxicity: (1) adverse effects expected to have a nonlinear dose-response relationships (sometimes referred to as âbiological thresholdâ) and (2) those likely to have a linear (i.e., having no biological threshold) dose-response in the low- dose range. A threshold for a particular toxic effect is defined as the dose or dose rate below which the adverse effect attributable to the specific agent is unlikely to occur (Brown 1987). For the evaluation of FRs, toxic effects other than cancer are considered to have a nonlinear dose-response, whereas carcinogenic responses induced by genotoxic carcinogens are treated as having a linear dose-response. Non-genotoxic carcinogens or tumor promoters may have a nonlinear dose response. For the carcinogenic FRs evaluated in this report, the data are insufficient to determine whether dose responses are linear or nonlinear. The subcommittee assumed that they have linear dose responses because this assumption is conservative or health protective. For some FRs, more than one dose-response relationship may exist, depending on conditions of exposure and a variety of responses (e.g., cancer, birth defects, kidney damage, etc.). The subcommittee's process of evaluating relevant dose-response relationships takes into account diverse information about the body's ability to generate metabolites that are more toxic than the parent compound, its ability to detoxify potentially toxic compounds or metabolites, and differences between the mechanisms of toxicity in test organisms and humans. Exposure circumstances (e.g., duration, frequency, and route) have considerable impact on toxic potency. Dose-response assessment includes the process of extrapolating adverse effects observed in experimental animal organism from high to low doses; it also includes extrapolating data from animals to humans. To perform such extrapolations, two fundamentally different approaches are used, one for carcinogenic responses and another for all other forms of toxicity. It should be noted that these procedures used by the subcommittee relate specifically to toxicity resulting from repeated and prolonged exposures, as much as a full lifetime in duration. The subcommittee's approach for determining noncancer and cancer potency of FR chemicals is described below. Determination of Toxic Potency for Noncancer Effects For all types of noncancer effects, the subcommittee's procedure used to evaluate the dose-response involves identifying the lowest-observed-adverse-
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 26 effect level (LOAEL) and the no-observed-adverse-effect level (NOAEL). The LOAEL is the lowest dose at which a statistically or biologically significant increase (when compared with a control group) in an observable adverse (although at times mild, reversible) effect has been reported. The NOAEL is the highest exposure level below the LOAEL at which no such significant increase is observed in the frequency or severity of an adverse effect when compared with a control group. Conceptually, NOAELs are more conservative than LOAELs as a starting point for extrapolating findings to low doses experienced by humans. The NOAEL-uncertainty-factor (UF) approach was used to identify reference dose (RfD) or reference concentration (RfC). An RfD is defined as an estimate (with uncertainty spanning an order of magnitude or greater) of daily oral or dermal doses that are unlikely to have deleterious effects during a life span. An RfC for an inhalation exposure is similarly defined. Statistical deficiencies and improvements in the NOAEL based uncertainty-factor (UF) approach to derive RfDs and RfCs have been described by several investigators (Crump 1984, Kimmel and Gaylor 1988, Chen and Kodell 1989, Gaylor 1989, Kodell et al. 1991). Uncertainty Factors Once a NOAEL is estimated from experimental data or observations of humans, an RfD is obtained by dividing it by one to several uncertainty factors (UFs). The degree of confidence in the data used to derive the NOAEL determines the magnitude of UFs. The major UFs, some or all of which may be applied for derivations of RfDs or RfCs, are described below: Interspecies Extrapolation Despite physiological similarities among mammalian species, laboratory animals are not human beings, thereby providing a clear disadvantage in estimating possible adverse human health effects from exposure to FRs. Based on both toxicological principles and empirical observations, reasons exist to support the hypothesis that many forms of biological responses, including toxic responses, can be extrapolated across mammalian species, including humans. One of the most important reasons for species differences in response to chemical exposures is that toxicity is often a function of the metabolism of a chemical. This is also applicable to FRs. Differences in metabolic handling of a chemical among animal species, or even among strains of the same species,
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 27 are not uncommon. Such differences can account for toxicity differences. In most cases, because information on a chemical's metabolic profile in humans is lacking (and often unobtainable), identifying the animal species and toxic response most likely to predict the human response accurately is generally not feasible. Therefore, by convention, one assumes that, in the absence of clear evidence that a particular toxic response is not relevant to humans, any compound-related adverse effects found in laboratory animals tested in properly designed studies are potentially predictive of response in at least some humans. This subcommittee agrees with this convention and also adopted that approach. A UF of 10 is used for FRs when data from laboratory animals are extrapolated to humans; if the toxic potency of a compound is known to be similar in humans and experimental animals, then a factor less than 10 is used. Intraspecies Extrapolation Tests conducted in a homogeneous laboratory animal population (or even a small human group) do not account fully for the heterogeneous human population. Individuals vary considerably in their susceptibility to chemical insult due to their genetic make-up, lifestyle factors (e.g., nutrition, smoking), age, hormonal status (e.g., pregnancy), immune system integrity, and pre-existing illness. The subcommittee used a UF of 10 to account for intraspecies differences. If data indicate the existence of a narrow range of susceptibilities, a UF of less than 10 was used. Therefore, an understanding of ranges of metabolic differences is of substantial value in determining the most suitable UF. Even when a NOAEL is based on human data, those data are likely to have been collected in exposure circumstances that differ appreciably from the exposure circumstances of interest. Consequently, there will still be uncertainty in such a NOAEL, and it may also be adjusted with UFs to produce an RfD or RfC. Extrapolation from Subchronic to Chronic Exposures In some data sets for FRs, the toxicity data used to derive an RfD or RfC is of a duration of less than a lifetime. The UF for extrapolation from subchronic to chronic exposure ranges from one to 10. Selection of a specific value depends in part on the quality of the data and on expectations for cumulative toxicity and accumulation of the substance. For instance, several short-term studies of varying durations may indicate that toxicity of a FR is or is not cumulative. Likewise, an understanding of the chemistry of a FR may assist in
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 28 deciding whether extending durations of exposure might yield even lower NOAELs. Extrapolation from LOAEL to NOAEL At times, toxicity studies reveal only a LOAEL and no NOAEL. Under these circumstances, the subcommittee used a UF of between one and 10. Selection of a UF is dependent largely on the quality of the entire toxicity data set. A factor of 10 was applied, unless the evidence indicates that lesser values could be used confidently. Route-to-Route Extrapolation When the toxicity information on an FR is obtained from one route of administration (e.g., ingestion) yet individuals are exposed by another route (e.g., dermal or inhalation), the resulting differences in toxic potency are taken into account by applying a UF of 10; if evidence indicates little or no such variability, a factor of less than 10 is used. Reliance on toxicokinetic data, elucidating differences in absorption rates and metabolic pathways by different routes of exposure, is useful in selecting an appropriate value of the UF. Adjustment to Account for Poor Quality of the Database The quality of toxicity studies vary considerably. The subcommittee considered the quality of data in the derivation of an RfD or RfC. Some deficiencies include the use of small numbers of animals and the lack of replication of results. To account for these deficiencies, the subcommittee used UFs between one and 10 and the actual values used varied with the subcommittee's evaluation of the adequacy and quality of data. Overall Uncertainty Factor The RfD is obtained from the adjusted NOAEL by dividing it by an overall UF that is equal to the product of all the UFs discussed in the preceding paragraphs. For example, a composite UF of 100 is applied when the NOAEL is derived from chronic toxicity studies (typically 2-year studies) that are considered to be of high quality and when the purpose is to protect members of the
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 29 general population who could be exposed daily for a full lifetime (10 to account for interspecies differences and 10 to account for intraspecies differences). Carcinogenic Potency Carcinogenic potency of chemicals in humans can be estimated for doses far below those in the range of observations only by the use of appropriate mathematical models that extend dose-response curves (NRC 1983; NRC 1994; EPA 1998). Because such dose-response functions cannot be determined empirically, the actual shapes of such dose-response curves at the lowest dose are unknown, and must currently be hypothesized on the basis of assumptions about biological processes. For most carcinogens, it is believed that the probability for the occurrence of a cancer increases linearly with dose at sufficiently low doses, so that only certain mathematical models (nonthreshold models) are used to predict carcinogenic responses at low doses. In the case of nonthreshold chemicals, it is assumed that there is no dose (except zero dose) that corresponds to zero risk of injury. In theoretical terms, any dose of a carcinogen results in an incremental increase in the risk of cancer (NRC 1983). If an FR causes cancer in laboratory animals, the slope of the dose-response curve is used as the unit to describe carcinogenic potency and is called the cancer potency factor (also technically designated as q1*) (Crump 1996). Many models have been developed to assess the effects of low doses of carcinogens. The most frequently used cancer risk dose-response model is the linearized multistage model (Crump 1996), and it was used for FRs that produced tumors in rodents; this model is used because it provides conservative cancer estimates. A major limitation with low-dose extrapolation models is that they all often fit the data from animal bioassays equally well, and it is not possible to determine their validity based on goodness of fit. Each model may fit experimental data equally well, but they are not all equally plausible biologically. The dose-response curves derived from different models diverge substantially in the dose range of interest (NRC 1983). Therefore, low-dose extrapolation is more than a curve-fitting process, and considerations of biological plausibility of the models must be taken into account before choosing the most suitable model for a particular set of data. Mathematically fitting the multistage model to experimental results in laboratory animals allows estimation of the low-dose slope of a dose-response curve, together with an upper 95% confidence limit (UCL) on that slope. These are often referred to as cancer potency factors. These cancer potency factors are extrapolated to humans, usually by incorporating various
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 30 other assumptions such as constant exposure for a lifetime of 70 years and a suitable dose metric for the extrapolation. Through the use of the UCL, and by the nature of the animal-to-human extrapolation used in practice, it is possible that the cancer potency factors estimates obtained will overestimate risk. Such potency factors are, therefore, suited more to standard setting than to defining actual risks to a specific population. Quantitative estimate of risk is obtained by multiplying cancer potency factors by lifetime average dose rates. Actual risks are unlikely to be higher than such quantitative estimates, are likely to be lower, and may be zero. The benchmark dose (BD) offers a variant to estimate the cancer potency of a carcinogenic FR or other compound (EPA 1996, Crump 1984). When applied to a carcinogen data set, the BD process first uses the linearized multi-stage model to define the dose-response curve in the observed response range, and to estimate a dose that is likely to produce an increment in response rate of approximately 10%. However, rather than applying UFs to the dose representing the 10% response rate, a straight line is interpolated to the origin of the dose versus incremental-response curve from the lower confidence limit of the BD at 10% incremental response, and the slope of this line is used as the cancer potency factor. The extrapolation to humans then proceeds as for the linearized multi-stage model already discussed. It is likely that linear extrapolation to zero from the lower confidence limit of the BD at 10% incremental response will be close to cancer risk estimates based on q1* (Cogliano et al. 1988). Exposure1 Assessment for Flame Retardants This third step in the risk assessment process is used to describe the nature (e.g., distribution of age, sex, and unique conditions such as pregnancy, preexisting illness, and lifestyle) and size of the various populations exposed and the magnitude and duration of their exposures. The assessment might include past, current, and prospective exposures. For FRs, prospective exposures in residential settings are estimated; no attempt has been made to reconstruct previous 1Exposure is defined as the opportunity for a dose, such as concentrations in food, air, or water, and is generally reported in units such as ppm or ppb or mg/L, etc. Dose is defined as the amount received by the body of the target organism (e.g., humans) or a target organ (e.g., the liver or kidneys); it is generally reported in units of weight of the substance (e.g., mg or Âµg) per body weight or body weight. Inhalation doses are expressed as concentrations in air either as ppm or as mg/m3. Doses resulting from skin contact are generally described as concentrations per unit of surface area (e.g., mg/meter2); however, when the toxicity from skin contact is systemic, the dose is also described in the traditional unit of mg/kg-day.
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 31 exposures. Human exposure to FRs in residential furniture fabric was assumed to possibly occur via skin contact, ingestion (specifically for infants or children who might chew on the fabric), inhalation of particles generated during abrasion of the surface fibers, and inhalation vapors off-gassing from treated fabric. An ideal exposure assessment for FRs in furniture fabrics, would include the following steps: 1. determining concentrations of the chemicals at the surface of the fabric, their degree of binding, and their rate of disappearance; 2. estimating the amount of an FR that may come into contact with skin (characterizing the surface area of contact is necessary) and that may be ingested or inhaled, and contact with internal surfaces (gut and living membranes); 3. characterizing human behaviors that directly bear on external dose (e.g., frequency of contact, years of contact, etc.); and 4. establishing or estimating the rate of membrane penetration (absorption), and evaluating the pharmacokinetics of the compound to ascertain doses to target tissues, which might be some distance from the point of entry into the body. These four steps were simplified by assuming worst-case scenarios. For example, individuals were assumed to be exposed to 100% of the compound of interest for a plausible period of time over plausibly large areas of their body. The approach taken results in upper-bound estimates of risk that have utility in screening compounds âthey allow elimination from concern of compounds that are not a risk. In Chapter 3, the details of this process are presented; and for each FR, estimates of doses are described within their respective chapters. Risk Characterization of Flame Retardants This final step of a risk assessment process involves integration of data and analyses from the other three steps of risk assessment to determine the likelihood that groups of individuals may not experience any of the various forms of toxicity associated with a chemical under its known or anticipated conditions of exposure. This step includes estimations of risk to individuals and population groups and a full exposition of the uncertainties associated with the conclusions. Scientific knowledge is usually incomplete, so the reliance upon inferences about risk is inevitable. A well-constructed risk assessment relies on inferences that are most strongly supported by general scientific understanding and, to the extent feasible, do not include assumptions derived solely from risk management or policy directives.
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 32 Risk Characterization of Noncarcinogenic Effects For adverse health effects other than cancer, the subcommittee used a âhazard indexâ approach; it is the ratio of an actual or estimated dose to RfD or RfC. This approach was used by the subcommittee to judge whether a particular exposure is unlikely to present a noncancer risk. For each FR, a hazard index was calculated for noncancer toxicity by dividing the estimated human dose by the estimated RfD or RfC. A hazard index of less than 1 was deemed to provide an adequate margin of safety; a hazard index greater than 1 was considered to possibly pose concern for noncancer effects. The inability to draw more definitive conclusions for hazard indices greater than 1 arises from four factors: 1. The conservative approach taken to estimate dosesâthe estimated doses are certainly much higher than may occur in practice, because the methodology used in developing these estimates was deliberately chosen to be ultraconservative. 2. The conservative approach taken to estimate the RfDsâthe estimated RfDs are certainly substantially lower than doses that might cause health effects, because the methodology used in developing these estimates was deliberately chosen to be conservative. 3. The interaction between the 2 factors discussed in the previous two paragraphs. For any health effect to occur in any individual, the dose for that individual must be higher than the threshold for that same individual. The estimates of dose, however, are for the highest exposed individuals, while the RfD is estimated for the most sensitive individuals, and these two will coincide only with low probability. 4. FRs are to be used to prevent or reduce a known riskâthat of fires caused by ignition of FR-treated upholstered furniture. Even if FRs were to produce adverse effects, the net effect of using them might nevertheless be a reduction in risks. Determination of the acceptability of the various risks requires consideration of the trade-offs involved. The subcommittee's charge did not include evaluation of trade-offs, and it did not attempt to make such evaluation. Risk Characterization of Carcinogenic Effects The approach used by the subcommittee to characterize risks from exposure to carcinogenic FRs involves extrapolation of observations of cancer in animals at relatively high doses to much lower doses anticipated in residential settings. The risks for a specified carcinogen in a defined set of circumstances are esti
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 33 mated by multiplying the cancer potency factor by the various measures of dose. The risk is expressed as a probability, for instance, as the proportion of individuals, among all individuals exposed to a cancer-causing agent, that might develop (or die from) cancer attributable to that agent over a specified time, usually a lifetime of 70 years. A specific example would be a lifetime risk of less than one in a million, meaning that chance exists that less than one excess cancer may occur among a million identically exposed persons. The subcommittee's upper limits of cancer estimates reflect the chance that cancer may occur and not that they must inevitably occur; however, because of limitations in knowledge about the processes of cancer causation, it is also possible that the risk may be zero and that no excess cancers above the background levels would ensue from a specified exposure. REFERENCES Brown, C.C. 1987. Approaches to interspecies dose extrapolation. Pp. 237â268 in Toxic Substances and Human Risk, Principles of Data Interpretation. R.G.Tardiff, and J. V.Rodricks, eds. New York: Plenum Press. Chen, J.J., and R.L.Kodell. 1989. Quantitative risk assessment for teratological effects. J. Am. Stat. Assoc. 84:966â971. Cogliano, V.J., A.M.Koppikar, J.M.Conis, S.C.Gibson, J.S.Gift, et al. 1988. Methodology for Evaluating Potential Carcinogenicity in Support of Reportable Quantity Adjustments Pursuant to CERCLA Section 102. Office of Health and Environmental Assessment, U.S. Environmental Protection Agency, Washington, DC. Crump, K.S. 1984. A new method for determining allowable daily intakes. Fundam. Appl. Toxicol. 4(5):854â871. Crump, K.S. 1996. The linearized multistage model and the future of quantitative risk assessment Hum. Exp. Toxicol. 15(10):787â798. EPA (U.S. Environmental Protection Agency). 1986. Guidelines for Carcinogen Risk Assessment. Fed. Regist. 51 (September 24):33992â 34003. EPA (U.S. Environmental Protection Agency). 1989. Exposure Factors Handbook. EPA/600/8â89/043. Exposure Assessment Group, Office of Health and Environmental Assessment, U.S. Environmental Protection Agency, Washington, DC. EPA (U.S. Environmental Protection Agency). 1991a. Guidelines for Developmental Toxicity Risk Assessment. Fed. Regist. 56 (Dec.5):63798â63826.EPA600FR91001. National Service Center for Environmental Publications, Cincinnati, OH. EPA (U.S. Environmental Protection Agency). 1992a. Guidelines for Exposure Assessment, Fed. Regist. 57(104):22888â22938. May 29, 1992. EPA (U.S. Environmental Protection Agency). 1992b. Draft Report: A Cross-species Scaling Factor for Carcinogen Risk Assessment Based on Equivalence of mg/kg 3/4/Day. Fed. Regist. 57(109):24152â24173(June 5). EPA (U.S. Environmental Protection Agency). 1996. Proposed Guidelines for Carcinogen Risk Assessment. EPA/600/P-92/003C. Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC.
ASSESSMENT OF HEALTH RISKS FROM THE USE OF FLAME RETARDANTS 34 EPA (U.S. Environmental Protection Agency). December 1998. Draft Revisions to the Proposed Guidelines for Carcinogen Risk Assessment. Office of Research and Development, U.S. Environmental Protection Agency, Washington, DC. EPA (U.S. Environmental Protection Agency). 1999a. Guidance for Conducting Health Risk Assessment of Chemical MixturesâExternal Scientific Peer Review Draft. NCEA-C-0148. [Online]. Available: http://www.epa.gov/nceawww1/mixtures.htm Office of Research and DevelopmentâNCEA, U.S. Environmental Protection Agency, Washington, DC. EPA (U.S. Environmental Protection Agency). 1999b. Guidelines for Carcinogen Risk Assessment. Review Draft. NCEA-F-0644, Risk Assessment Forum, U.S. EPA, Washington,DC (July 1999). [Online]. Available: http://www.epa.gov/nceawww1/raf/crasab.htm Gaylor, D.W. 1989. Quantitative risk analysis for quantal reproductive and developmental effects. Environ. Health Perspect. 79:243â246. Kimmel, C.A., and D.W.Gaylor. 1988. Issues in qualitative and quantitative risk analysis for developmental toxicology. Risk Anal. 8(1): 15â 20. Kodell, R.L., R.B.Howe, J.J.Chen, and D.W.Gaylor. 1991. Mathematical modeling of reproductive and developmental toxic effects for quantitative risk assessment. Risk Anal. 11(4):583â590. NRC (National Research Council). 1977. Chemical contaminants: Safety and risk assessment. Pp. 19â62 in Drinking Water and Health. Washington, DC: National Academy Press. NRC (National Research Council). 1980. Problems of risk estimation. Pp. 25â65 in Drinking Water and Health, Vol. 3. Washington, DC.: National Academy Press. NRC (National Research Council). 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press. NRC (National Research Council). 1994. Science and Judgment in Risk Assessment. Washington, DC: National Academy Press. OSTP (Office of Science and Technology). 1985. Chemical Carcinogens: A Review of the Science and Its Associated Principles. Fed. Regist. 50:10372â10442. Tardiff, R.G., and J.V.Rodricks, eds. 1987. Toxic Substances and Human Risk: Principles of Data Interpretation. New York: Plenum Press. WHO (World Health Organization). 1978. Principles and Methods for Evaluating the Toxicity of Chemicals. Environmental Health Criteria, Vol. 6. Geneva: World Health Organization. WHO (World Health Organization). 1984. Principles for Evaluating Health Risks to Progeny Associated with Exposure to Chemicals During Pregnancy. Environmental Health Criteria No. 30. Geneva: World Health Organization. WHO (World Health Organization). 1986. Principles and Methods for the Assessment of Neurotoxicity Associated with Exposure to Chemicals. Environmental Health Criteria No. 60. Geneva: World Health Organization. WHO (World Health Organization). 1999. Principles for the Assessment of Risks to Human Health from Exposure to Chemicals. Environmental Health Criteria No. 210. Geneva: World Health Organization.