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.
1 ~ ntroduction In light of the task before the U. S . Environmental Protection Agency (EPA) of regulating exposure to toxic chemicals, its Office of Drinking Water asked the Safe Drinking Water Committee of the National Research Council's Board on Environmental Studies and Toxicology to hold a small workshop, with each participant addressing some aspect of the methodology for assessing the risk associated with exposure to mixtures of chemicals found in drinking water. This report is the product of that workshop, held in October 1987 in Washington, D.C., and of the deliberations of the Subcommittee on Mixtures in a followup meeting. This chapter briefly describes the background of the workshop, defines concepts and terms, and suggests ways of grouping chemicals for estimating their combined risk. BACKGROUND OF THE STUDY More than 6 million chemicals have been listed and given identifying numbers by the Chemical Abstracts Service of the American Chemical So- ciety. Most have not been adequately tested for toxicity (NRC, 1984) either individually or in combinations. Some 67,000 of them are registered with federal regulatory agencies for use as industrial chemicals; as pesticide, food, drug, and cosmetic ingredients; or for other commercial purposes. Industrial discharges or nonpoint discharges, such as runoff from hazardous-waste sites or agricultural application, might cause many of those chemicals to appear in surface water or groundwater and hence in drinking water. One consumer advocacy group (Center for Study of Responsive Law) has compiled a list 101
102 DRINKING WATER AND HEALTH of some 2,000 contaminants in drinking water as detected in surveys con- ducted since 1974 (Duff Conacher and Associates, 19881. Contaminants and their concentrations vary with site, time, and temperature, and many (in- cluding the by-products of disinfection) have not been characterized or even identified (NRC, 19871. Because of the variability of drinking water com- position and because the relatively low concentrations of the chemical con- taminants in water would require large lifetime studies to reveal long-term effects, regulatory authorities and theoretical scientists have attempted to model the effects of mixtures by using the results of tests of the individual components of the mixtures, often for shorter periods at higher doses (Bingham and Morris, 1988; NRC, 19881. The National Toxicology Program has ini- tiated short-term and subchronic studies of a mixture of 25 groundwater contaminants at concentrations actually encountered (Yang and Rauckman, 19871; however, almost nothing is known about how chemicals interact when they are ingested by humans as mixtures or with substances from other sources, including medications. The potential for interactions that could have adverse health consequences must be considered in any assessment of the quality of drinking water. A recent study showed a statistically significant association between the ingestion of chlorinated surface water and human bladder cancer (Cantor et al., 1987~. Although the specific components re- sponsible for that association remain unidentified, the by-products of chlorine disinfection are currently the prime suspects. CONCEPTS AN D DEFI N ITIONS Exposure to two or more chemicals simultaneously can produce interac- tions that qualitatively or quantitatively differ from biologic responses that would be predicted from the actions of the individual chemicals separately (Murphy, 1980; NRC, 1980, 19881. When the response is greater than that predicted on the basis of adding the separate responses, the interaction is said to be synergistic. When the response is less than that predicted on the basis of additivity, the interaction is said to be antagonistic. There is a contrast between dose additivity and response additivity that needs to be addressed. If two (apparently different) toxic materials lead to the same type and severity of toxic effect, they might be considered as one material in the effect they produce. When there is no threshold and the dose- response curve is essentially linear (at least across some modest range of doses), a response-additive model is reasonable. Under such circumstances, a response-additivity model and a dose-additivity model will give the same answer for the same range of doses. However, if some minimum dose must be reached before toxicity is man- ifest (i.e., if there is a threshold), then the response-additivity model can be
Introduction 103 misleading. At subthreshold doses, each of the materials in a mixture will produce zero response separately. A response-additivity model would predict that the sum of any number of zero responses will be a zero response. However, if the sum of the doses is above the threshold, then a response can occur and lead to an appearance of synergism i.e., greater than response additivity. For a more technical definition and more detailed models of synergism, see Elashoff et al. (1987) and Fears et al. (1988, 19891. For computational purposes for distinguishing between possible additivity and multiplicativity of responses, a dose that leads to a relative risk of 1.01 or less is called a "low" dose. That does not, in any way, imply that such a dose is "acceptable" or "safe." For example, if the overall age-adjusted mortality from cancer in 1 year in the United States is 200 x 10-s, a dose leading to a 1% increase would imply an excess mortality of 2 x 10-s, which is generally considered to be unacceptable, although the relative risk, (202 x 10-/200 x 10-s), equals 1 .01, an increase (as defined) of "only" 1%. Exposure to two materials, each at such a dose, would, if results were strictly response-additive, produce a relative risk of 1.02 (i.e., 1 + 0.01 + 0.011. If results were multiplicative, the relative risk would be (1.011~1 .01) = 1.0201 implying an excess over an additive risk that is extremely un- likely ever to be measured or even measurable. Synergistic interactions between chemicals have been suspected of causing health effects in humans that could not be predicted by simply adding the expected effects of the component chemicals. One such case was the incident of mass organophosphorus insecticide poisoning among field workers in Pakistan in 1976 (Baker et al., 19781; two of the pesticide formulations contained contaminants, which could well have increased the toxicity attrib- uted to the designated active pesticidal ingredient, malathion, by inhibiting its detoxification (see Chapter 41. The present approach for regulating organic chemicals in drinking water (EPA, 1987) is to establish a maximum contaminant level goal (MCLG) and a maximum contaminant level (MCL) for each organic compound, except that trihalomethanes (THMs) as a class are regulated by a single MCL (EPA, 1979~. That exception is based on the "potential" carcinogenicity of chlo- roform in humans and the similarities of chloroform to less-studied THMs; chloroform is assumed to be representative of a class (the THMs) that is ubiquitous in treated drinking water in the United States and whose members' concentrations can be reduced simultaneously. All other EPA standards are established after the toxicologic data, treatment capabilities, and occurrence data are interpreted and evaluated for each chemical. The single-chemical approach is scientifically appealing, but it could pose major problems, be- cause it ignores both the possibility of interaction and the presence of many unidentified chemicals (NRC, 1987) in treated drinking water.
104 DRINKING WATER AND HEALTH GROUPING CHEMICALS FOR ESTIMATION OF COMBINED RISK Methods for assessing health risks associated with mixtures have not changed substantially in recent years (EPA, 1985, 1986; Murphy, 1980; NRC, 1980, 19881. Lack of toxicologic information and the complexity of a mixture can impede and complicate the application of any method, and considerable reliance is still placed on knowledge of the toxicity of individual chemicals in approaching the regulation of mixtures. Because of the substantial backlog in testing and regulation, regulatory agencies need to explore carefully some ways to group chemicals so as to facilitate their control in the absence of complete toxicologic information. Grouping could be used to establish prior- ities for testing, to formulate rules for testing, or to develop standards for allowable concentrations of contaminants in drinking water. The subcommittee did not consider either the complete universe of clas- sifications that might be devised or the regulatory consequences of imple- menting standards for classification. It did, however, review options and suggested the following four types of grouping for consideration by EPA: 1. Contaminants can be grouped on the basis of their being carcinogenic. According to the currently preferred dose-extrapolation models (EPA, 1986; NRC, 1987), the risk of one end point associated with exposure to a mixture of carcinogens at low concentrations can be theoretically approximated as the sum of risks associated with the individual carcinogens; i.e., additivity of re- sponse or risk is usually assumed for carcinogens associated with relative risks of less than 1.01. However, the subcommittee recognizes that this assumption of low-dose additivity of response does not have much empirical foundation. Rather, it rests on theoretical considerations and observations from limited epidemiologic studies, and it might not apply for carcinogens at doses yielding high relative risks or when alternative dose-extrapolation models are considered. For exposures at higher concentrations, synergistic interactions appear to occur in humans exposed to combinations of several kinds of agents such as cigarette smoke, asbestos, and alcoholic beverages (NRC, 19881. The assumption of low-dose additivity needs to be carefully assessed in future research. 2. Systemic contaminants that have similar toxic end points, such as those resulting in specific organ toxicity or peripheral nerve damage, can be grouped and treated as having additive effects under most conditions. A general description of this approach is given in Chapter 3, and the anticholinesterases, which have similar toxic consequences, are examined in detail as a biolog- ically based class in Chapter 5. Materials that are assumed to have thresholds for response require special attention to the biologic mechanisms leading to a toxic response. As indicated earlier, where the mechanisms of toxicity of two or more toxicants are the same, combining below-threshold doses (i.e., doses that produce a zero response) could lead to an above-threshold dose,
Introduction 105 (i.e., a dose that produces a response). Thus, dose additivity should be considered for materials that yield the same toxic end point. Furthermore, in some cases in which toxicity information is limited and exposure concen- trations are high, an uncertainty factor could be applied to accommodate the possibility of synergism; this is discussed in Chapter 3. 3. Dissimilar chemical compounds might be physiologically transformed into similar metabolites with similar reactivity and stability. For the purposes of assessing combined risk, chemicals can be grouped on the basis of this kind of similarity. A caution to keep in mind is that materials that appear to have similar metabolites might nonetheless at times have different toxic end points (see Chapter 54. 4. Contaminants can be grouped according to structural similarity, which might imply similar biologic responses. In grouping chemical mixtures by whatever method, a "toxic-equiva- lence" approach can be considered assigning numerical potency values to individual mixture components that are representatives of specific classes, estimating potencies of other class members that are present relative to those of the appropriate representative chemicals, and then summing the products of the relative potencies and concentrations of all the chemicals present across all end points. Risks associated with exposure to polycyclic aromatic hydro- carbons (Clement Associates, 1988) and the chlorinated dibenzo-p-dioxins and dibenzofurans (Berlin and Barnes, 1987) have been estimated by this method. Again, the concept of dose additivity is inherent in the consideration of toxic equivalents and relative potency; this approach implies that one material is operationally a dilution (in effect) of the other material. Volatile halogenated hydrocarbons including carbon tetrachloride, te- trachloroethylene, trichloroethylene, and 1,2-dichloroethane are frequently found in drinking water, and several could be placed into more than one of the above groups (e.g., into the group of carcinogens or into a group of chlorinated compounds). These substances have similar physical and chem- ical properties, they are metabolized in the liver, and similar methods are used in treating drinking water to reduce their concentrations. The highest recommended concentrations for these compounds are all set within a rather narrow range (2-5 ~g/liter), because their parallel toxicologic properties are similar and it is feasible to control them together (EPA, 19871. The chemicals could be evaluated together with an additivity formula or with a single standard for a mixture of them. Combining the carcinogenic potentials of such chemicals is discussed in Chapter 7. STRUCTU R E OF TH E R EPORT Chapter 2 explains the importance of pharmacokinetics to an estimation of the health risks associated with multiple-chemical exposures. Appendix
106 DRINKING WATER AND HEALTH A gives examples of the use of physiologically based pharmacokinetics to determine the risk to humans of exposure to volatile synthetic organic com- pounds by extrapolating data from inhalation studies performed on laboratory animals. The current EPA approach for assessing health risks related to noncarcinogens is reviewed in Chapter 3, which suggests modifications of the approach. A simple illustrative mathematical model that supports some of the modifications is given in Appendix B. Appendix C discusses dose addit~vity and response additivity. Chapter 4 considers issues of exposure, with emphasis on the organophosphates, carbamates, and volatile organic compounds. Chapter 5 reviews the biologic mechanisms of and interactions among anticholinesterases. Chapter 6 shows how the assumptions inherent in EPA's risk assessment methods for carcinogens can be used to combine the estimated risks associated with individual carcinogenic components in a mixture. Although the workshop led to an affirmation of current methods for the risk assessment of mixtures in drinking water, attempts at developing a firmer empirical base and reevaluation should continue. Chapter 7 rec- ommends research to facilitate further improvement. Because there are so many chemicals and so many possible mixtures to which humans could be exposed, the absence of toxicity data might result in human exposure to chemicals or mixtures that are not being studied by regulatory agencies. More research is needed on the scientific basis for grouping chemicals for testing and regulation. Specific research proposals are given in Chapters 2-6 and summarized in Chapter 7. REFERENCES Baker, E. L., Jr., M. Warren, M. Zack, R. D. Dobbin. J. W. Miles. S. Miller, L. Alderman, and W. R. Teeters. 1978. Epidemic malathion poisoning in Pakistan malaria workers. Lancet 1 (8054):31-34. Bellin, J. S., and D. G. Barnes. 1987. Interim Procedures for Estimating Risk Associated with Exposures to Mixtures of Chlorinated Dibenzo-p-dioxins and Dibenzofuran (CDDs and CDF). U.S. Environmental Protection Agency Report No. EPA/625/3-87/012. Washington, D.C.: Risk Assessment Forum, U.S. Environmental Protection Agency. 27 pp. + appen- dixes. gingham, E., and S. Morris. 1988. Complex mixtures and multiple agent interactions: The issues and their significance. Fund. Appl. Toxicol. 10:549-552. Cantor, K. P., R. Hoover, P. Hartge, T. J. Mason, D. T. Silverman, R. Altman, D. F. Austin, M. A. Child, C. R. Key, L. D. Marrett, M. H. Myers, A. S. Narayana, L. I. Levin, J. W. Sullivan, G. M. Swanson, D. B. Thomas, and D. W. West. 1987. Bladder cancer, drinking water source. and tap water consumption: A case-control study. J. Natl. Cancer Inst. 79: 1269- 1279. Clement Associates. 1988. Comparative Potency Approach for Estimating the Cancer Risk Associated With Exposure to Mixtures of Polycyclic Aromatic Hydrocarbons. Interim final report to the U.S. Environmental Protection Agency, Office of Research and Development. April 1, 1988. Contract No. 68-02-4403. Vienna, Va.: Clement Associates, Inc.
Introduction 107 Duff Conacher and Associates. 1988. Troubled Waters on Tap: Organic Chemicals in Public Drinking Water Systems and the Failure of Regulation. Washington, D.C.: Center for Study of Responsive Law. 120 pp. Elashoff, R. M., T. R. Fears, and M. A. Schneiderman. 1987. The statistical analysis of a carcinogen mixture experiment. I. Liver carcinogens. J. Natl. Cancer Inst. 79:509-526. EPA (U.S. Environmental Protection Agency). 1979. National interim primary drinking water regulations; control of trihalomethanes in drinking water. Final Rule. Fed. Regist. 44(231):68624. EPA (U.S. Environmental Protection Agency). 1985. Proposed guidelines for the health risk assessment of chemical mixtures. Fed. Regist. 50(6):1170-1176. EPA (U.S. Environmental Protection Agency). 1986. Guidelines for the health risk assessment of chemical mixtures. [FRL-2984-2. ] Fed. Regist. 51 (185)34014-34025. EPA (U.S. Environmental Protection Agency). 1987. National primary drinking water regu- lations; Synthetic organic chemicals; Monitoring for unregulated contaminants. Fed. Regist. 52(130):25690-25717. Fears, T. R., R. M. Elashoff, and M. A. Schneiderman. 1988. The statistical analysis of a carcinogen mixture experiment. II. Carcinogens with different target organs, N-methyl-N- nitro-N-nitrosoguanidine, N-butyl-N-(4-hydroxybutyl)nitrosamine, dipentylnitrosamine and nitrilotriacetic acid. Toxicol. Indus. Health 4:221-255. Fears, T. R., R. M. Elashoff, and M. A. Schneiderman. 1989. The statistical analysis of a carcinogen mixture experiment. III. Carcinogens with different target systems, aflatoxin B 1, N-butyl-N-(4-hydroxybutyl)nitrosamine, lead acetate, and thiouracil. Toxicol. Ind. Health 5: 1-23. Murphy, S. D. 1980. Assessment of the potential for toxic interactions among environmental pollutants. Pp. 277-294 in The Principles and Methods in Modern Toxicology, C. L. Galli, S. D. Murphy, and R. Paoletti, eds. New York: Elsevier/North-Holland Biomedical Press. NRC (National Research Council). 1980. Principles of Toxicological Interaction Associated with Multichemical Exposures. Washington, D.C.: National Academy Press. 213 pp. NRC (National Research Council). 1984. Toxicity Testing: Strategies to Determine Needs and Priorities. Washington, D.C.: National Academy Press. 382 pp. NRC (National Research Council). 1987. Drinking Water and Health. Vol. 7. Disinfectants and Disinfectant By-Products. Washington, D.C.: National Academy Press. 207 pp. NRC (National Research Council). 1988. Complex Mixtures: Methods for In Vivo Toxicity Testing. Washington, D.C.: National Academy Press. 227 pp. Yang, R. S. H. and E. J. Rauckman. 1987. Toxicologicial studies of chemical mixtures of environmental concern at the National Toxicology Program: Health effects of groundwater contaminants. Toxicology 47: 15-34.
