Six interaction profiles for simple mixtures

Chemosphere 53 (2003) 183–197 www.elsevier.com/locate/chemosphere Six interaction profiles for simple mixtures Hana R. Pohl *, Nickolette Roney, Sharon Wilbur, Hugh Hansen, Christopher T. De Rosa Agency for Toxic Substances and Disease Registry (ATSDR), US Department of Health and Human Services, Div. of Toxicology, Mailstop E-29, 1600 Clifton Road, Atlanta, Georgia 30333, USA Received 12 April 2002; received in revised form 26 February 2003; accepted 13 April 2003 Abstract The Agency for Toxic Substances and Disease Registry (ATSDR) has a program for chemical mixtures that encompasses research on chemical mixtures toxicity, health risk assessment, and development of innovative computational methods. ATSDR prepared a guidance document that instructs users on how to conduct health risk assessment on chemical mixtures (Guidance Manual for the Assessment of Joint Toxic Action of Chemical Mixtures). ATSDR also developed six interaction profiles for chemical mixtures. Two profiles were developed for persistent environmental chemicals that are often found in contaminated fish and also can be detected in human breast milk. The mixture included chlorinated dibenzo-p-dioxins, hexachlorobenzene, dichlorodiphenyl dichloroethane, methyl mercury, and polychlorinated biphenyls. Two profiles each were developed for mixtures of metals and mixtures of volatile organic chemicals (VOCs) that are frequently found at hazardous waste sites. The two metal profiles dealt with (a) lead, manganese, zinc, and copper; and (b) arsenic, cadmium, chromium, and lead; the two VOCs mixtures dealt with (a) 1,1,1-trichloroethane, 1,1-dichloroethane, trichloroethylene, and tetrachloroethylene; and (b) benzene, ethylbenzene, toluene, and xylenes (BTEX). Weight-of-evidence methodology was used to assess the joint toxic action for most of the mixtures. Physiologically based pharmacokinetic modeling was used for BTEX. In most cases, a target-organ toxicity dose modification of the hazard index approach is recommended for conducting exposure-based assessments of noncancer health hazards.
2003 Elsevier Ltd. All rights reserved. Keywords: Chemical mixtures; Risk assessment; Weight-of-evidence 1. Introduction To carry out US congressional mandates, the Agency for Toxic Substances and Disease Registry (ATSDR) developed and coordinated a program for chemical mixtures. The program includes trend analysis to identify the mixtures of real-life concern, in vivo and in vitro toxicological testing of mixtures, qualitative modeling of joint action, and methodological developments including computational modeling. * Corresponding author. Tel.: +1-404-498-0160; fax: +1-404498-0094. E-mail address: [email protected] (H.R. Pohl). The qualitative modeling of joint action is used for the health assessment of specific mixtures related to ATSDRs programs. ATSDRs Guidance Manual for the Assessment of Joint Toxic Action of Chemical Mixtures (ATSDR, 2002) describes the methodology for mixtures risk assessment that is applied in practice in documents called interaction profiles. Interaction profiles. The interaction profiles have three purposes: • to evaluate data (if available) on health hazards, and their dose–response relationships, from exposure to simple mixtures; • to evaluate data on the joint toxic actions of components of the mixture; and 0045-6535/03/$ - see front matter
2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0045-6535(03)00436-3 184 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 • to make recommendations for exposure-based assessments of the potential impact of joint toxic action of the mixture on public health. The interaction profiles evaluate mixtures that are of special interest to environmental public health. These evaluations are based on results reported in the scientific literature. The interaction profiles are peer reviewed to ensure accuracy of the data presented and validity of conclusions. Interaction profiles provide environmental health scientists with ATSDRs evaluation of whether interactions occur among the chemical components in the mixture and the types of interactions to be expected. Interaction profiles also make recommendations on how to incorporate concerns about the expected interactions or additivity into the public health assessment of the contaminated site. Interaction profiles provide the results of experimental and theoretical studies from current literature, an assessment of toxic interactions, and generalized rules that might be used inferentially for other related exposure scenarios. The relevance of interactions data and approaches to public health is an integral part of the profile. The primary users of interaction profiles are expected to be health assessors who need to know whether interactions are expected to occur among components of a chemical mixture and how to incorporate concerns regarding additivity and interactions into the public health assessment of the site. The concern in terms of potential impact on public health is whether additivity and toxicological interactions might increase the health hazard above what would be expected from an assessment of each component of the mixture singly. Health assessors will be able to use information on specific mixtures from interaction profiles together with site-specific exposure data and the methodology described in the Guidance Manual for the Assessment of Joint Toxic Action of Chemical Mixtures (ATSDR, 2002) to produce sitespecific assessments for mixtures. ATSDR has finalized six interaction profiles. Two profiles were written on persistent environmental chemicals that are often found in contaminated fish and can be also detected in human breast milk. Two profiles each were written for mixtures of metals and mixtures of volatile organic chemicals (VOCs) frequently found at hazardous waste sites. As described in ATSDRs guidance document (ATSDR, 2002), a chemical mixture can be evaluated using one of several options. Some of the basic methods were used previously by ATSDR and other agencies to derive health guidance values for specific mixtures such as chlorinated dibenzo-p-dioxins (CDDs), polychlorinated biphenyls (PCBs), jet fuels, and automotive gasoline (Pohl et al., 1997). The assessment can be based on literature data available for the mixture of concern (whole mixture or original mixture), on a similar mixture, or on the mixtures components. Tools used to evaluate the components include the hazard index (HI), target-organ toxicity dose (TTD) modification to the HI method, weight-of-evidence (WOE) modification to the HI method, toxic equivalency and relative potency, total cancer risk, and computational methods such as physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) and quantitative structure–activity relationships modeling. 2. Methods Most of these interaction profiles used the WOE method to evaluate the respective mixtures. The WOE approach for assessing toxicological interactions in chemical mixtures was developed by Mumtaz and Durkin (1992) and further explained by Mumtaz et al. (1994). In this approach, a mixture is broken down into binary pairs and a determination is made on the interaction potential of each pair of chemicals. This determination is called a binary weight-of-evidence (BINWOE) determination. More specifically, a BINWOE is a qualitative judgement, based on empirical observations and mechanistic considerations, that categorizes the most plausible nature of any potential influence of one compound on the toxicity of another compound for a given exposure scenario. When two chemicals are in a mixture, their interactions must be examined in two directions: the most likely effect of the first chemical on the second, then the effect of the second chemical on the first. With the WOE approach, these interaction determinations are based on an evaluation of the available information on the metabolism, health effects, and other pertinent data available in the literature on the chemicals. First, the direction of interaction is predicted (greater than additive, less than additive, or additive), then a classification is assigned on the basis of the mechanistic understanding and toxicological significance of the interaction determination (Table 1). In many instances the recommended approach for the exposure-based assessment of joint toxic action of a mixture is to use the HI method with the TTD modification and qualitative WOE method to assess the potential consequences of additive and interactive joint action of the components of the mixture. These methods are to be applied only under circumstances involving significant exposure to the mixture, i.e., only if hazard quotients for two or more of the components equal or exceed 0.1. Hazard quotients are the ratios of exposure estimates to noncancer health guideline values, such as minimal risk levels (MRLs). If only one or if none of the components have a hazard quotient that equals or exceeds 0.1, then no further assessment of the joint toxic action is needed because additivity and/or interactions H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 185 Table 1 WOE scheme for the qualitative assessment of chemical interactions Determine if the interaction of the mixture is additive (¼), greater than additive (>), or less than additive (<) Classification of mechanistic understanding III. Direct mechanistic data The mechanism(s) or possible mechanism(s) by which the interactions could occur has been well characterized and leads to an unambiguous interpretation of the direction of the interaction III. Mechanistic data on related compounds The mechanism(s) by which the interaction could occur is not well characterized for the compounds of concern, but structure/activity relationships, either quantitative or informal, can be used to infer the likely mechanisms and the direction of the interaction III. Inadequate or ambiguous mechanistic data The mechanism(s) by which the interactions could occur has not been well characterized, or information on the mechanism(s) does not clearly indicate the direction that the interaction will have Classification of toxicological significance A. The toxicological significance of the interaction has been directly demonstrated B. The toxicological significance of the interaction can be inferred or has been demonstrated in related compounds C. The toxicological significance of the interaction is unclear Modifiers 1. Anticipated exposure duration and sequence 2. A different exposure duration and sequence a. In vivo data b. In vitro data i. The anticipated route of exposure ii. A different route of exposure Adapted from Mumtaz and Durkin (1992). are unlikely to result in significant health hazard. The exposure-based assessment of potential health hazard should be used in conjunction with biomedical judgment, community-specific health outcome data, and community health concerns to assess the degree of public health hazard (ATSDR, 1992, 2001a). 3. Interaction profiles 3.1. Interaction profiles for persistent chemicals CDDs, hexachlorobenzene, dichlorodiphenyl dichloroethane (p; p0 -DDE), methyl mercury, and PCBs occur with high frequency in water, sediment, and fish, including fish from the North American Great Lakes and the Baltic Sea. The chemicals occur, to varying degrees, in other dietary components including meat, milk, and dairy products, and can be detected in human milk. One reason for selecting these five chemicals or chemical groups is the fair amount of overlap in the wide range of end points or organs that these chemicals affect in humans and animals. This overlap leads to a concern that after exposure to mixtures of the five chemicals in breast milk or other food sources, all five might jointly act to produce altered neurological development, suppression of immune competence, or cancer, and that three of the chemicals (CDDs, p; p0 -DDE, and PCBs) might jointly act to alter development of reproductive organs. ATSDR prepared two documents on this particular mixture of chemicals. The first interaction profile presented available information on possible health effects related to exposure via consumption of contaminated fish (ATSDR, 2001a); the second dealt with developmental impacts of the intake of contaminated breast milk (ATSDR, 2001b). The ‘‘fish’’ profile. Several studies are examining reproductive or physical developmental end points and neurodevelopmental or neurobehavioral end points in people whose diets include North American Great Lakes fish contaminated with complex mixtures of chemicals of concern and other chemicals (most notably chlorinated dibenzofurans and chlorinated pesticides such as mirex and chlordane) (ATSDR, 2001c). Studies of animals exposed to Great Lakes fish in their diet include • a two-generation study of reproductive, immunological, neurobehavioral, and neurochemical end points in rats fed diets containing varying amounts of similarly contaminated Great Lakes fish (Feeley et al., 1998; Seegal et al., 1998; Seegal, 1999; Stewart et al., 2000); 186 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 • a two-generation reproductive study of minks fed diets containing varying amounts of Great Lakes fish (Restum et al., 1998); and • other studies that examined neurobehavioral end points in rats fed diets with Great Lakes fish (Hertzler, 1990; Daly, 1991). Although these studies provided valuable information, they cannot be used for the assessment of this five-component mixture because of the confounding contamination of fish with other chemicals. The ‘‘breast milk’’ profile. PCBs, p; p0 -DDE, and hexachlorobenzene were selected for this profile because they have been detected with very high frequency in breast milk in several studies: • a recent study of women residing in the US Great Lakes region (Kostyniak et al., 1999), • a study of the general population in North Carolina (Rogan et al., 1986a), and • a Canadian study of the general population that included the Great Lakes basin (Newsome et al., 1995). Because CDDs and methyl mercury have been detected in Great Lakes fish (ATSDR, 2001b), it is also expected that breast milk of fish-eating populations in the US Great Lakes region might contain these chemicals. Whereas recent US monitoring studies have not focused on the presence of CDDs and methyl mercury in breast milk, CDDs have been detected in earlier US studies, as well as in studies in the Netherlands, Canada, Germany, New Zealand, Japan, and Russia (Pohl and Hibbs, 1996), and methyl mercury has been detected in breast milk samples from Japan, Germany, and Sweden (Abadin et al., 1997). In addition, elevated levels of PCBs and mercury were detected in samples of breast milk from mothers living in the North Atlantic Faroe Islands, where the seafood diet includes pilot whale meat and blubber (Grandjean et al., 1995a). Results from a North Carolina study (Rogan et al., 1986a,b, 1987; Gladen et al., 1988; Gladen and Rogan, 1991) and a Dutch study (Koopman-Esseboom et al., 1994a,b, 1996; Huisman et al., 1995a,b; Patandin et al., 1998, 1999a,b) of breast-fed children provide some evidence that exposure to mixtures of biopersistent chemicals in human breast milk at exposure levels in the upper range of background levels or exposure during gestation via placental transfer might be associated with mild neurodevelopmental delays in some children. Again, although data on combinations of some components of the mixture exist, these human studies have too many confounding factors. Both profiles. The assessment of joint toxic action was based on laboratory data. The WOE analysis (Table 2) indicated that only a limited amount of evidence is available to support the possible existence of greater- than-additive or less-than-additive joint actions of a few pairs of the components: • hexachlorobenzene potentiation of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (a most-studied representative of CDDs) reduction of body and thymus weights (Li et al., 1989); • PCB antagonism of TCDD immunotoxicity and TCDD developmental toxicity (Bannister et al., 1987; Davis and Safe, 1989); and • synergism between PCBs and methyl mercury in disrupting regulation of brain levels of dopamine that might influence neurological function and development (Birnbaum et al., 1985; Haake et al., 1987). For the remaining pairs, additive joint action at shared targets of toxicity is either supported by data (for a few pairs) or is recommended as a public health protective assumption because of the lack of adequate data to assess joint toxic action. In general, overlapping targets of toxicity for these five-components provide strong support for the plausibility of joint toxic action, but there is a notable lack of studies to characterize the modes of joint toxic action. Because no direct data are available to characterize health hazards (and dose–response relationships) from the five-component mixture, component-based approaches that assume additive joint toxic action are recommended for exposure-based assessments of possible noncancer or cancer health hazards from oral exposure to the whole mixture of CDDs, hexachlorobenzene, p; p0 -DDE, methyl mercury, and PCBs. A TTD modification of the HI approach is recommended for conducting exposure-based assessments of noncancer health hazards. Alternatively stated, HIs are computed on an organ-specific basis, assuming that target-organ toxicities are biologically independent. TTDs have been derived for several toxicity targets for each of the components, including TTDs for hepatic, endocrine, immunological, reproductive, developmental, and neurological effects (ATSDR, 2002). For assessment of cancer risks from joint toxic action of the mixture, a similar component-based approach is recommended that involves multiplication of intakes of the chemical components by US Environmental Protection Agency cancer slope factors and summation of the resultant risk estimates. In summary, this assessment of the joint toxicity can be used in real-life situations, especially in relation to contaminated food sources such as fish from polluted waters. It should be noted, however, that from the public health viewpoint the benefits of breastfeeding appear to outweigh the risks for most people with lowlevel environmental exposures (Pohl and Hibbs, 1996; Rogan, 1996; Abadin et al., 1997; Pohl and Tylenda, 2000). 187 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 Table 2 Matrix of BINWOE determinations for repeated simultaneous oral exposure to persistent chemicals On toxicity of -TCDD Effect of TCDD Hexachlorobenzene p; p0 -DDE Methyl mercury PCBs ? Antiandrogenic effects ¼ IIIC Other effects ? Immune suppression ¼ IIIBbii Other effects ? Body and organ weight changes, decreased retinoids in liver ¼ IIIC Other effects ? ? ? ? ? ? Hexachlorobenzene Body and thymus weight > IIIA2aii Other effects ? p; p0 -DDE Antiandrogenic effects ¼ IIIC Other effects ? ? Methyl mercury ? Liver effects ¼ IIIC Other effects ? ? PCBs Immune suppression < IIIB2aii Developmental < IIIC2ai Body and organ weight changes, decreased retinoids in liver ¼ IIIC Other effects ? ? ? Neurological effects > IIICb Reproductive performance ? Porphyria ¼ IIIB Other effects ? Neurological effects > IIICb Reproductive performance ? Porphyria ¼ IIIB Other effects ? See Table 1 for definition of classifications and modifiers. BINWOE: binary weight-of-evidence; p; p0 -DDE: dichlorodiphenyl dichloroethane; PCBs: polychlorinated biphenyls; TCDD: tetrachlorodibenzo-p-dioxin. 3.2. Interaction profile for 1,1,1-trichloroethane, 1,1dichloroethane, trichloroethylene, and tetrachloroethylene These VOCs were chosen for the development of an interaction profile because of their frequent occurrence at hazardous waste sites. An analysis of ATSDR public health assessments of 1608 National Priority List (NPL) hazardous waste sites indicated that a mixture of these VOCs was found at 210 sites and was the most frequently occurring mixture of four VOCs. Contaminated media at sites with this mixture included groundwater (186 of 210 sites), soil (45 of 210 sites), and air (25 of 210 sites). The most frequent completed exposure pathway for VOCs at these sites involved private well water. Completed exposure pathways involving municipal water or air contaminated with VOCs were less frequent. VOCs were unimportant in most completed exposure pathways involving soil. Acute or repeated inhalation exposure to any of these chemicals starting at concentrations as low as 20–100 ppm is expected to produce neurological impairment as the result of the parent chemicals (and a metabolite in the case of trichloroethylene) acting on components of neuronal membranes (ATSDR, 1990, 1995a, 1997a,b). Animal studies also provide evidence that repeated inhalation exposure at high exposure levels (>100–500 ppm) can damage liver and kidney tissue and produce cancer due to the formation of reactive metabolites. The weight of this evidence varies for the four chemicals. Neurological impairment forms the basis for ATSDRs inhalation MRLs for these chemicals. Acute oral exposure to trichloroethylene or tetrachloroethylene during pregnancy is also thought to present a hazard to the neurological development of offspring, and these effects form the basis of the oral MRLs for these chemicals. Only limited WOE exists to show that inhalation or oral 188 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 exposures to these chemicals might present significant cancer risks to humans. In the absence of pertinent data on neurological responses to a mixture of all four chemicals and PBPK models that predict the direction and magnitude of interactions among the four chemicals, health hazards from inhalation or oral exposure to mixtures of these chemicals might best be assessed by a components-based approach such as the HI approach (ATSDR, 2002). Such an approach requires judgments concerning the presence or absence of interactions affecting the response of the shared apparent critical target-organ, the nervous system, and other shared targets (the liver and kidney). Mechanistic data and interactions data were evaluated to determine how pairs of these chemicals might jointly act in producing nervous system, liver, and kidney effects. To characterize the overall potential for interactions among 1,1,1-trichloroethane, 1,1-dichloroethane, trichloroethylene, and tetrachloroethylene, BINWOE determinations were derived using the classification scheme described by ATSDR (2002). Each of the BINWOEs for nervous system effects determined an additive joint action with data quality factors of ‘‘II’’ for mechanistic understanding (reflecting moderate mechanistic understanding) and ‘‘C’’ for toxicologic significance to (reflecting lack of studies designed to test the hypothesis of joint additive actions on the nervous system) (Table 3). BINWOE determinations were also made for noncancer and cancer effects in the liver and kidney (Table 3). On the basis of animal study results, each of the chemicals is expected to produce noncancer and cancer effects in the liver and/or kidney via reactive metabolites formed under high-exposure chronic conditions. BINWOE determinations for liver and kidney end points were made in anticipation of public health concerns that greater-than-additive interactions might occur and might cause liver and kidney effects. The analysis of the available data, however, provides no indication that this type of interaction might occur. On the basis of plausibility from mechanistic understanding or limited evidence from rat studies examining joint action on liver or kidney end points, additive joint action was determined in 11 of the 12 BINWOEs (Stacey, 1989). The 12th BINWOE, for the effect of tetrachloroethylene on trichloroethylene, was determined as a less-than-additive joint action (i.e., tetrachloroethylene might antagonize trichloroethylene-induced liver and kidney effects by inhibiting the formation of trichloroacetic acid from trichloroethylene). This determination was made on the basis of in vivo evidence that tetrachloroethylene inhibits the metabolism of trichloroethylene in humans under occupational exposure conditions (Seiji et al., 1989), and that tetrachloroethylene and trichloroethylene acted in a less-than-additive manner in causing hepatic and renal peroxisomal proliferation in orally exposed rats (Goldsworthy and Popp, 1987). The use of a TTD approach did not appear to be warranted because neurological effects are the basis for all of ATSDRs MRLs for these chemicals, regardless of exposure duration or route, and no evidence exists for greater-than-additive interactions in the liver and kidney, which also are shared targets of the chemicals. Thus, to conduct exposure-based assessments of possible health hazards from exposure to mixtures of these chemicals, a component-based HI approach that assumes additive joint toxic action and uses ATSDR MRLs based on neurological impairment is recommended. No evidence exists to indicate that greater-thanadditive interactions would cause liver and kidney effects to occur at exposure levels lower than those influencing the nervous system. Table 3 Matrix of BINWOE determinations for simultaneous exposure to chemicals of concern On toxicity of 1,1,1-Trichloroethane Effect of 1,1,1-Trichloroethane 1,1-Dichloroethane Trichloroethylene Tetrachloroethylene Nervous system ¼ IIC Liver and kidney ¼ IIC Nervous system ¼ IIC Liver and kidney ¼ IIB Nervous system ¼ IIC Liver and kidney ¼ IIB Nervous system ¼ IIC Liver and kidney ¼ IIC Nervous system ¼ IIC Liver and kidney ¼ IIC 1,1-Dichloroethane Nervous system ¼ IIC Liver and kidney ¼ IIC Trichloroethylene Nervous system ¼ IIC Liver and kidney ¼ IIB Nervous system ¼ IIC Liver and kidney ¼ IIC Tetrachloroethylene Nervous system ¼ IIC Liver and kidney ¼ IIB Nervous system ¼ IIC Liver and kidney ¼ IIC See Table 1 for definition of classifications and modifiers. BINWOE: binary weight-of-evidence. Nervous system ¼ IIC Liver and kidney ¼ IIB Nervous system ¼ IIC Liver and kidney < IIB H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 3.3. Interaction profile for benzene, ethylbenzene, toluene, and xylenes Benzene, ethylbenzene, toluene, and xylenes (BTEX) frequently occur together at hazardous waste sites (ATSDR, 2001d). Combinations of these chemicals are among the most frequently found binary mixtures in completed exposure pathways at hazardous waste sites. Media contaminated with BTEX include air, water, and soil. When contaminated groundwater is used as household water, volatilization of BTEX into indoor air can occur. In addition, contamination of groundwater and subsurface soil can result in migration of these chemicals into basements as soil gas. Public health risks from exposure to BTEX might best be assessed by an approach that considers both the mechanism and toxic consequences of the joint action of the whole mixture, particularly the presence or absence of interactions affecting the responses of the critical target-organs. Toxicokinetic studies in humans and animals indicate that these chemicals are well absorbed; distribute to lipid-rich and highly vascular tissues such as the brain, bone marrow, and body fat because of their lipophilicity; and are rapidly eliminated from the body (ATSDR, 1995b, 1997c, 1999, 2000). Metabolism of the four chemicals is dose dependent and is generally extensive at dose levels that do not saturate the first metabolic step of each compound, which involves cytochrome P-450-dependent mixed function oxidases. The predominant cytochrome P-450 isozyme involved in the metabolism of each chemical is CYP2E1. All four chemicals can produce neurological impairment via parent-compound-induced physical and chemical changes in nervous system membranes. Exposure to benzene can also cause hematological effects including aplastic anemia, with subsequent manifestation of acute myelogenous leukemia, via the action of reactive metabolites. The HI method is recommended for assessing the joint neurotoxic hazard of BTEX because the HI approach is most appropriately applied to mixture components that cause the same effect by the same mechanism of action (ATSDR, 2002). Joint toxic action data on the whole BTEX mixture are lacking, and information pertaining to toxic interactions among the BTEX components is essentially limited to data on a few binary mixtures of the chemicals. However, predictions from PBPK modeling studies––when used in conjunction with mechanistic, interaction, and toxicity information on the components––provide a sufficient basis for assessing the joint toxic action of the whole mixture in humans. A PBPK model was developed that predicts blood levels of the four chemicals in rats (Haddad et al., 1999). Similar PBPK models have also been developed for binary, ternary, and quinary mixtures of BTEX components in 189 humans as well as rats (Purcell et al., 1990; Tardif et al., 1993a,b, 1995, 1997; Haddad et al., 1999, 2000). On the basis of PBPK studies as well as in vitro and in vivo metabolism and toxicity studies for some of the binary component mixtures, competitive metabolic inhibition is the most plausible mechanism of interaction among the BTEX components at relatively high doses. However, PBPK model predictions also clearly indicate that metabolic interactions are probably negligible at 620 parts per million (ppm) of each component, which implies that environmental exposures to BTEX are well below the threshold for interactions. Therefore, because of the apparent lack of competitive metabolic interactions in BTEX mixtures below approximately 20 ppm of each component, it is plausible that joint neurotoxic actions among the chemicals will be additive at environmental levels of exposure. Exposure to higher concentrations of BTEX components (i.e., above the threshold for metabolic inhibition) would be expected to lead to greater-than-additive increases in blood levels of parent compounds and, consequently, increased concern for neurotoxicity. However, it is unclear whether the PBPK model descriptions are adequate for predicting interactions from inhalation of BTEX mixtures above approximately 200 ppm of each component, or if they are appropriate for oral exposures. Studies that directly examined the joint toxic action of BTEX chemicals on the nervous system (Toftgard and Nilsen, 1981, 1982; Frantık et al., 1988; Korsak et al., 1988, 1992; Dudek et al., 1990; Frantik and Vodickova, 1995) generally support the predictions of the PBPK studies that joint action is expected to be additive at BTEX concentrations below approximately 20 ppm of each component. In summary, on the basis of evidence from PBPK and neurotoxicity studies supporting the plausibility of additive joint action at the shared target of toxicity at relatively low levels of exposure, the HI approach is recommended for assessing possible neurotoxic health hazards from environmental exposures to BTEX. HI is a conservative approach for assessing BTEX because • the MRLs and guidance values on which it is based are protective in nature, • the data indicate that greater-than-additive interactions are unlikely at component doses that would not otherwise be overtly toxic, and • the neurotoxicity of the mixture would be decreased, not increased, if interactions were less than additive at low levels of exposure. The PBPK studies facilitate the use of the HI method to assess the joint neurotoxic action of BTEX by providing an estimate of the threshold for interactions (i.e., by defining the additive region in which the method is applicable). Another approach that could be used to incorporate information on component interactions in 190 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 the assessment of BTEX mixtures is the BINWOE modification to the HI (ATSDR, 2002). Unlike the PBPK model-based approach, the BINWOE method is qualitative and does not define the exposure region in which its predicted direction of interaction is applicable, and is intended to be used when joint action data are inadequate to support more quantitative methods, such as the PBPK modeling of BTEX. However, BINWOE analysis is relevant in that the binary interaction studies generally support the direction of interaction predicted by the PBPK models for higher levels of exposures. Table 4 shows binary interaction matrices that indicate the plausible direction of interactions when they do occur, as predicted from the interaction studies of metabolic and neurological effects. These matrices show that joint action was less than additive for metabolic effects in all 12 predictions, greater than additive for neurological effects in 2 of 12 predictions, and additive for neurological effects in 2 of 12 predictions. Because less-than-additive metabolic interaction implies greater-than-additive neurotoxicity (because of increased levels of unmetabolized chemicals that can act on neuronal membranes), the overall assessment is that the mixture components are likely to jointly act on the nervous system in a greater-thanadditive manner, which is consistent with the PBPK model predictions for levels of exposure above the interaction threshold. The evaluation of possible hematotoxic and carcinogenic hazards from exposure to BTEX is best approached by evaluating benzene as a single component. PBPK model predictions indicate that (a) toluene, eth- ylbenzene, and xylene are unlikely to influence the hematotoxicity or carcinogenicity of benzene; and (b) benzene, toluene, and xylene are unlikely to affect the carcinogenicity of ethylbenzene, at environmental levels of exposure. However, toluene can inhibit the hematological effects of benzene (Andrews et al., 1977; Gut et al., 1980; Tunek et al., 1981, 1982; Gad-El-Karim et al., 1984; Hsieh et al., 1990; Plappert et al., 1994), and ethylbenzene and xylenes are also expected to reduce the hematotoxic/carcinogenic potential of benzene because of competitive metabolic interactions, but available toxicity data for these pairs of chemicals are indeterminate (Table 4). Similarly, no binary toxicity data exist to support the possible reduction in ethylbenzene carcinogenicity due to competitive metabolic interactions with benzene, toluene, and xylenes. 3.4. Interaction profile for lead, manganese, zinc, and copper The lead, manganese, zinc, and copper mixture was chosen for an interaction profile on the basis of an analysis of the most frequently occurring binary mixtures in completed exposure pathways at hazardous waste sites. These metals are commonly found in soil. The exposure scenario of greatest concern for this mixture is long-term, low-level oral exposure (ATSDR, 2001e). No pertinent health effects data or PBPK models were available for the mixture of lead, manganese, zinc, and copper. In the absence of these data, the potential health hazards of the joint toxic action of this mixture Table 4 Binary interaction matrix for metabolic effects from simultaneous exposure to chemicals of concern On metabolism or toxicity of Benzene Effect of Benzene Toluene Ethylbenzene Xylenes Metabolism < Nervous system ? Hematological and clastogenic ¼ Metabolism < Nervous system ? Hematological and clastogenic ? Metabolism < Nervous system ? Hematological and clastogenic ? Metabolism < Nervous system ? Hematological and clastogenic ? Metabolism < Nervous system ¼ Hematological and clastogenic ¼ Toluene Metabolism < Nervous system ? Hematological and clastogenic< Ethylbenzene Metabolism < Nervous system ? Hematological and clastogenic ? Metabolism < Nervous system ? Hematological and clastogenic ? Xylenes Metabolism < Nervous system ? Hematological and clastogenic ? Metabolism < Nervous system ¼ Hematological and clastogenic ¼ Metabolism < Nervous system ¼ Hematological and clastogenic ? Metabolism < Nervous system ¼ Hematological and clastogenic ? Determine if the interaction of the mixture is additive (¼), greater than additive (>), or less than additive (<). 191 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 might best be assessed by using a component-based approach that estimates end-point-specific HIs. End points of concern for this mixture include the critical effects of the individual components, and toxicity targets in common that might become significant because of additivity or interactions. The critical effects of lead and manganese are neurological; the critical effects of zinc are hematological, which is also a sensitive effect of lead; and the critical effect of copper is hepatic. The qualitative WOE approach is then used to predict the impact of interactions on the end-point-specific HIs. Mechanistic data and interactions data were evaluated to determine how pairs of these chemicals might jointly act in producing neurological, hematological, and hepatic effects. To characterize the overall potential for interactions among lead, manganese, zinc, and copper, BINWOE determinations were derived using the classification scheme described by ATSDR (2002). The BINWOEs are shown in Table 5. Neurological: The predicted direction of joint toxic action for neurological effects, an end point common to lead and manganese, is greater than additive for the effect of manganese on lead on the basis of evidence from rat studies indicating that manganese increases the distribution and/or retention of lead in the brain (Chandra et al., 1981, 1983; Shukla and Chandra, 1987). On the basis of animal studies and mechanistic understanding, the BINWOEs are predicted to be • less than additive for the effects of zinc on lead (Willoughby et al., 1972; Vassilev et al., 1994); • less than additive for the effects of copper on lead (Flora et al., 1982, 1989; Kies and Ip, 1990); • additive (no effect) for the effect of lead on manganese (Chandra et al., 1981); and • indeterminate for the effects of zinc and copper on manganese. This evaluation indicates that the potential health hazard might be less than estimated by the end-pointspecific HI for neurological effects, particularly for waste sites with relatively high hazard quotients for lead, copper, and zinc, and a lower hazard quotient for manganese. However, for mixtures where manganese and lead predominate, the potential health hazard might be greater than estimated by the end-point-specific HI for neurological effects. The indeterminate ratings for two of the BINWOEs (zinc and copper on manganese) are a source of uncertainty in assessments where manganese accounts for a great portion of the apparent neurological hazard. Hematological: The potential health hazard for hematological effects is likely to be lower than indicated by the end-point-specific HI for mixtures where lead, zinc, and copper predominate, because three of the BINWOEs for combinations of these metals were less than additive with moderate to high confidence, and the remaining one was additive. The BINWOE for zinc on lead was less than additive on the basis of studies indicating that excess zinc protects against a number of hematological effects of lead (Cerklewski and Forbes, 1976; Chisolm, 1981; Flora et al., 1982, 1991). Lead is predicted to have no effect (additive) on the hematological toxicity of zinc (Cerklewski and Forbes, 1976; ElGazzar et al., 1978). The BINWOE for copper on lead is predicted to be less than additive on the basis of the protection by supplemental copper against lead-induced hematopoietic effects (Flora et al., 1989) and supported by mechanistic data indicating coexposure to excess copper decreases lead absorption (Flora et al., 1982, 1989; Kies and Ip, 1990). The BINWOE for copper on zinc is less than additive because copper appears to protect against the hematological toxicity of zinc (Smith and Larson, 1946; Magee and Matrone, 1960). The BINWOE for manganese on lead was greater than Table 5 Matrix of BINWOE determinations for toxicity end points of concern for intermediate or chronic simultaneous oral exposure to lead, manganese, zinc, and copper On toxicity of Lead Effect of Lead Manganese Zinc Copper ¼ IIICii neurological ¼ IIB hematological ¼ IIIC hepatic ? Hematological ? Hepatic Manganese >ICii neurological >IIB2ii hematological Zinc <IB neurological <IA hematological ? Neurological Copper <IC neurological <IB hematological ? Neurological <IB hepatic <IIA hematological See Table 1 for definition of classifications and modifiers. Indeterminate ratings (?) were assigned to pairs with no pertinent interactions data or mechanistic information available. BINWOE: binary weight-of-evidence. Direction: ¼ additive; > greater than additive; < less than additive; ? indeterminate. 192 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 additive with low-moderate confidence on the basis of an intravenous study in which coadministration of manganese delayed the recovery of blood ALAD activity from lead inhibition (Chiba and Kikuchi, 1984). The BINWOE for manganese on zinc was indeterminate. Hepatic: The predicted effects of the other mixture components on the hepatic toxicity of copper are less than additive for zinc with high-moderate confidence on the basis of studies showing that excess zinc protected against copper-induced hepatic effects (Ogiso et al., 1974; Reinstein et al., 1984; Jenkins and Hidiroglou, 1989; Pocino et al., 1990). Lead is predicted to have no effect (additive) on the hepatic toxicity of copper (Cerklewski and Forbes, 1977; Flora et al., 1982; Flora et al., 1989). The BINWOE for manganese on copper was indeterminate. Thus, the available data indicate the potential health hazard for hepatic effects might be less than predicted by the hazard quotient for mixtures where zinc and copper predominate. There is uncertainty with regard to the potential effect of manganese due to the lack of pertinent information. Thus, the recommendations for assessing the potential hazard to public health of the joint toxic action of this mixture include the estimation of end-point-specific HIs for neurological effects of lead and manganese and for hematological effects of lead and zinc. The qualitative WOE approach is then used to predict the impact of interactions on the end-point-specific HIs. The hazard quotient for coppers hepatic toxicity (critical effect for oral exposure) is estimated separately and the qualitative WOE is used to predict the impact of interactions on this hazard quotient. The impact of interactions on the endpoint-specific HIs and the copper hazard quotient were discussed above in terms of the WOE approach. 3.5. Interaction profile for arsenic, cadmium, chromium, and lead Arsenic, cadmium, chromium, and lead constitute a quaternary mixture that frequently occurs at hazardous waste sites. This mixture was found in soil at 219 sites out of the 1608 sites for which ATSDR has produced a public health assessment––including waste storage, treatment or disposal, manufacturing and industrial, and government waste sites. The primary route of exposure for this mixture in soil is likely to be oral, and the durations of concern are intermediate and chronic. The profile focuses on inorganic forms of these metals, consistent with the monitoring data, and on chromium (VI), the species of concern for chromium. Because no pertinent health effects data or PBPK models were located for the quaternary mixture, exposure-based assessment of health hazards for this mixture depends on an evaluation of the health effects data for the individual metals and on the joint toxic action and mechanistic data for various combinations of these metals (ATSDR, 2001f). End points of concern for this mixture are neurological, dermal, renal, cardiovascular, hematological, testicular, and carcinogenic effects. The recommendation for assessing the potential hazard to public health of the joint toxic action of lead, arsenic, cadmium, and chromium (VI) is to use the HI and TTDs to estimate endpoint-specific HIs for the effects of concern for the mixture. The BINWOEs are then used to qualitatively predict the impact of interactions on the end-pointspecific HIs. The hazard quotient for arsenics dermal toxicity and the cancer risk estimate for arsenic are estimated separately from the other mixture components, because dermal effects are a unique critical effect (oral exposure to the other components does not affect the skin) and because the other components are not carcinogenic by the oral route (ATSDR, 2001a). The BINWOEs developed for this mixture are presented in Table 6. Neurological: The predicted direction of joint toxic action for neurological effects (an end point common to all four components) is greater than additive for the effect of lead on arsenic (low-moderate confidence) on the basis of effects of combined exposure on reading and spelling in children (Moon et al., 1985); greater than additive for arsenic on lead (moderate confidence) and cadmium on lead (low confidence) on the basis of a study of maladaptive classroom behavior in children (Marlowe et al., 1985). The remaining nine BINWOEs were indeterminate because of a lack of toxicological and mechanistic data. Thus, the potential health hazard might be somewhat greater than estimated by the endpoint-specific HI for neurological effects, particularly for waste sites with relatively high hazard quotients for lead and arsenic, and lower hazard quotients for the other components. Given the indeterminate ratings for the majority of the BINWOEs, confidence in this conclusion would be lower for mixtures where cadmium and chromium (VI) account for a greater portion of the apparent neurological hazard. Renal: On the basis of data from animal studies and on mechanistic information, the BINWOEs for renal toxicity are as follows: • less than additive for lead on arsenic and for arsenic on lead (Fairhall and Miller, 1941); • less than additive for arsenic on chromium (VI) and chromium (VI) on arsenic (Mason and Edwards, 1989); and • less than additive for cadmium on lead (Mahaffey and Fowler, 1977; Mahaffey et al., 1981). The BINWOEs for lead on cadmium and cadmium on arsenic are additive. The potential health hazard for renal effects is likely to be lower than the additive, endpoint-specific HI because five of the BINWOEs were less than additive, two were additive, and five were indeter- 193 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 Table 6 Matrix of BINWOE determinations for toxicity end points of concern for intermediate or chronic simultaneous oral exposure to lead, arsenic, cadmium and chromium (VI) On toxicity of Lead Effect of Lead Arsenic Cadmium Chromium (VI) >IIIB neurological <IIIB hematological <IIIB renal ? Other effects ¼ IIIA cardiovascular ¼ IIC neurological ¼ IIAii renal >IIA testicular ? Neurological ? All effects <IIIB hematological <IIIB2ii testicular ? Other effects <IIIB2ii renal ? Other effects Arsenic >IIB neurological <IIIB hematological <IIIB renal ? Other effects Cadmium >IIIC neurological ¼ IIIA cardiovascular <IIIB hematological <IIA renal >IIA testicular ¼ IIB renal <IIIB hematological ? Other effects Chromium (VI) ? All effects >IIIC dermal <IIIB2ii renal ? Other effects ? All effects ? All effects See Table 1 for definition of classifications and modifiers. Indeterminate ratings (?) were assigned to pairs with no pertinent interactions data or mechanistic information available. BINWOE: binary weight-of-evidence. Direction: ¼ additive; > greater than additive; < less than additive; ? indeterminate. minate. Confidence in the less than additive and additive BINWOEs ranges from low-moderate to high-moderate. Cardiovascular: On the basis of relevant data, the BINWOES for cardiovascular effects of cadmium on lead and lead on cadmium are additive (Perry and Erlanger, 1978; Kopp et al., 1980a,b; Voors et al., 1982) the remaining 10BINWOEs are indeterminate. Thus, the WOE will have no impact on the additive, end-pointspecific HI. For mixtures other than those predominated by lead and cadmium, uncertainty is high. Hematological: The BINWOEs for lead on arsenic and vice versa, and for arsenic on cadmium and vice versa, are less than additive on the basis of apparent protection of each metal on the other metals hematological toxicity (Fairhall and Miller, 1941; Mahaffey and Fowler, 1977; Mahaffey et al., 1981). For similar reasons, the BINWOE for cadmium on lead is less than additive (Mahaffey and Fowler, 1977; Mahaffey et al., 1981). However, the effect of lead on cadmium is predicted to be additive on the basis of apparent additive hematopoietic effects (Thawley et al., 1977). The remaining BINWOEs are indeterminate. Thus, the potential health hazard for hematological effects is likely to be lower than indicated by the end-point-specific HI because five of the BINWOEs were less than additive, one was additive, and six were indeterminate. Confidence in the less than additive and additive BINWOEs is low-moderate. For waste sites where the chromium (VI) predominates, however, uncertainty would be high, be- cause the indeterminate BINWOEs all were for combinations of the other components with chromium (VI). Testicular: The potential health hazard might be higher than the end-point-specific HI for testicular effects for mixtures with relatively high hazard quotients for cadmium and lead because BINWOEs for this pair were greater than additive, with relatively high confidence based on synergistic effects seen in rat studies (Saxena et al., 1989). The BINWOE score for arsenic effects on cadmium testicular toxicity was less than additive, but confidence was low and the impact on the HI will be low. For the other pairs, BINWOEs were indeterminate (six BINWOEs) or not applicable (three BINWOEs for the effect of the other components on arsenic). Dermal: Interactions of the other mixture components on the dermal toxicity of arsenic are indeterminate for lead and cadmium, and greater than additive with low confidence for chromium (VI). Thus, the available data do not indicate a significant impact of interactions on the hazard quotient for the unique critical effect of arsenic, but uncertainty is high because of the lack of pertinent information. Carcinogenic: Data regarding effects of the other mixture components on arsenic carcinogenicity were not available, so BINWOEs are indeterminate and will have no impact on the cancer risk estimate for arsenic. Uncertainty regarding interactions is high due to the lack of pertinent information. 194 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 The predicted direction of interaction for the effects of these mixtures is not consistent across end points, which underscores the uncertainty in extrapolating interactions from one end point to another. It also suggests the possibility that a less sensitive target-organ might have the potential to impact a mixtures health assessment if it is affected synergistically. Concern would be heightened if several chemicals in the mixture affect that target-organ, and if confidence in the interaction (as reflected by the BINWOE scores) is high. 4. Conclusions The interaction profiles described in this paper present predictions that have a wide range of usefulness in the protection of public health. The conclusions from this WOE methodology demonstrate that even when sparse data exist, public health scientists and health assessors can provide guidance to the health risk manager of potential hazards of chemical mixtures. Such methods of evaluating chemical mixtures are valuable to scientists, health assessors, and those charged with health risk management. To the scientist, the interaction profiles provide a hypothesis-generating context for further exploration through laboratory experiments or modeling to refine the ATSDR WOE methodology. For the health assessor, they provide a methodologic approach to evaluating the potential hazards of chemical mixtures at a site. For the health risk manager, they provide a sentinel for more conservative guidelines, especially for high risk populations who are potentially exposed. When interaction profiles are developed by the methods discussed here and are subjected to the peer review process, they condense a large body of data and scientific expertise to conclusions about chemical mixtures that are highly useful in the protection of public health. Acknowledgements The authors thank Joan Colman, Peter McClure, Lisa Ingerman, Gary Diamond, Stephen Bosch, and Marc Odin for contributions in development of the Interaction Profiles. The authors also thank Pam Wigington for her editorial comments. References Abadin, H.G., Hibbs, B.F., Pohl, H.R., 1997. Breast-feeding exposure of infants to cadmium, lead, and mercury: a public health viewpoint. Toxicol. Ind. Health 13 (4), 495–517. Andrews, L.S., Lee, E.W., Wither, C.M., et al., 1977. Effects of toluene on the metabolism, disposition and hemopoietic toxicity of [3 H]benzene. Biochem. Pharmacol. 26, 293– 300. ATSDR, 1990. Toxicological profile for 1,1-dichloroethane. Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 1992. Public health assessment guidance manual. Atlanta: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry. ATSDR, 1995a. Toxicological profile for 1,1,1-trichloroethane. Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 1995b. Toxicological profile for xylenes. Atlanta: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry. ATSDR, 1997a. Toxicological profile for trichloroethylene. Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 1997b. Toxicological profile for tetrachloroethylene. Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 1997c. Toxicological profile for benzene. Atlanta: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry. ATSDR, 1999. Toxicological profile for ethylbenzene. Atlanta: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry. ATSDR, 2000. Toxicological profile for toluene. Atlanta: US Department of Health and Human Services, Agency for Toxic Substances and Disease Registry. ATSDR, 2001a. Interaction profiles for persistent chemicals found in fish. Chlorinated dibenzo-p-dioxins (CDDs), hexachlorobenzene, dichlorodiphenyl dichloroethane (p; p0 DDE), methyl mercury, and polychlorinated biphenyls (PCBs) (draft for public comments). Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 2001b. Interaction profiles for persistent chemicals found in breast milk. Chlorinated dibenzo-p-dioxins (CDDs), hexachlorobenzene, dichlorodiphenyl dichloroethane (p; p0 -DDE), methyl mercury, and polychlorinated biphenyls (PCBs) (draft for public comments). Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 2001c. Interaction profile for 1,1,1-trichloroethane, 1,1-dichloroethane, trichloroethylene, and tetrachloroethylene (draft for public comments). Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 2001d. Interaction profile for benzene, ethylbenzene, toluene, and xylenes (BTEX) (draft for public comments). Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 2001e. Interaction profile for lead, manganese, zinc and copper (draft for public comments). Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 2001f. Interaction profile for arsenic, cadmium, chromium, and lead (draft for public comments). Atlanta: Agency for Toxic Substances and Disease Registry, US Department of Health and Human Services. ATSDR, 2002. Guidance manual for the assessment of joint toxic action of chemical mixtures. Atlanta: Agency for H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 Toxic Substances and Disease Registry, US Department of Health and Human Services. Bannister, R., Davis, D., Zacherewski, T., 1987. Aroclor 1254 as a 2,3,7,8-tetrachlorodibenzo-p-dioxin antagonist: effects on enzyme induction and immunotoxicity. Toxicology 46, 29–42. Birnbaum, L.S., Weber, H., Harris, M.W., 1985. Toxic interaction of specific polychlorinated biphenyls and 2,3,7,8tetrachlorodibenzo-p-dioxin: increased incidence of cleft palate in mice. Toxicol. Appl. Pharmacol. 77, 292–302. Cerklewski, F.L., Forbes, R.M., 1976. Influence of dietary zinc on lead toxicity in the rat. J. Nutr. 106, 689–696. Cerklewski, F.L., Forbes, R.M., 1977. Influence of dietary copper on lead toxicity in the young male rat. J. Nutr. 107, 143–146. Chandra, S.V., Mohd, M., Saxena, D.K., 1981. Behavioral and neurochemical changes in rats simultaneously exposed to manganese and lead. Arch. Toxicol. 49, 495–496. Chandra, S.V., Murthy, R.C., Saxena, D.K., 1983. Effects of pre- and postnatal combined exposure to Pb and Mn on brain development in rats. Ind. Health 21, 273–279. Chiba, M., Kikuchi, M., 1984. The in vivo effects of manganese and zinc on d-aminolevulinic acid dehydratase activity inhibited by lead. Toxicol. Lett. 20, 143–147. Chisolm Jr., J.J., 1981. Dose–effect relationships for lead in young children: evidence in children for interactions among lead, zinc, and iron. In: Lynam, D.R., Piantanida, L.G., Cole, J.F. (Eds.), Environmental Lead: Proceedings of the 2nd International Symposium on Environmental Lead Research. In: Ecotoxicology and Environmental Quality Series. Academic Press, New York, pp. 1–7. Daly, H.B., 1991. Reward reductions found more aversive by rats fed environmentally contaminated salmon. Neurotoxicol. Teratol. 13, 449–453. Davis, D., Safe, S., 1989. Dose–response immunotoxicities of commercial polychlorinated biphenyls (PCBs) and their interaction with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol. Lett. 48, 35–43. Dudek, B., Gralewicz, K., Jakuboski, M., et al., 1990. Neurobehavioral effects of experimental exposure to toluene, xylene and their mixture. Pol. J. Occup. Med. 3 (1), 109–116. El-Gazzar, R.M., Finelli, V.N., Boiano, J., et al., 1978. Influence of dietary zinc on lead toxicity in rats. Toxicol. Lett. 1, 227–234. Fairhall, L.T., Miller, J.W., 1941. A study of the relative toxicity of the molecular components of lead arsenate. Public Health Rep. 56, 1610–1625. Feeley, M.M., Jordan, S.A., Gilman, A.P., 1998. The Health Canada Great Lakes Multigeneration Study: summary and regulatory considerations. Regul. Toxicol. Pharmacol. 27, S90–S98. Flora, S.J.S., Jain, V.K., Behari, J.R., et al., 1982. Protective role of trace metals in lead intoxication. Toxicol. Lett. 13, 51–56. Flora, S.J.S., Coulombe, R.A., Sharma, R.P., et al., 1989. Influence of dietary protein deficiency on lead–copper interaction in rats. Ecotoxicol. Environ. Saf. 18, 75–82. Flora, S.J.S., Kumar, D., Gupta, D., 1991. Interaction of zinc, methionine or their combination with lead at gastrointestinal or post-absorptive level in rats. Pharmacol. Toxicol. 68, 3–7. 195 Frantık, E., Horvath, M., Vodickova, L., 1988. Effects of solvent mixtures on behaviour and seizure characteristics at the utmost additive. Activ. Nerv. Super. (Praha) 30, 260– 264 (in Russian). Frantik, E., Vodickova, L., 1995. Combined effects of binary solvent mixtures. Cent. Eur. J. Occup. Environ. Med. 1 (1), 31–37. Gad-El-Karim, M.M., Harper, B.L., Legator, M.S., 1984. Modifications in the myeloclastogenic effect of benzene in mice with toluene, phenobarbital, 3-methylcholanthrene, Aroclor 1254 and SKF-525A. Mutat. Res. 135, 225–243. Gladen, B.C., Rogan, W.J., 1991. Effects of perinatal polychlorinated biphenyls and dichlorobiphenyl dichloroethene on later development. J. Pediatr. 119, 58–63. Gladen, B.C., Rogan, W.J., Hardy, P., et al., 1988. Development after exposure to polychlorinated biphenyls and dichlorobiphenyl dichloroethene transplacentally and through human milk. J. Pediatr. 113, 991–995. Goldsworthy, T.L., Popp, J.A., 1987. Chlorinated hydrocarbon-induced peroxisomal enzyme activity in relation to species and organ carcinogenicity. Toxicol. Appl. Pharmacol. 88, 225–233. Grandjean, P., Weihe, P., Needham, L.L., et al., 1995a. Relation of a seafood diet to mercury, selenium, arsenic, and polychlorinated biphenyl and other organochlorine concentrations in human milk. Environ. Res. 71, 29–38. Gut, I., Kopecky, J., Nerudova, J., et al., 1980. Metabolic and toxic interactions of benzene and acrylonitrile with organic solvents. In: Gut, I., Cikrt, M., Plaa, G. (Eds.), Industrial and Environmental Xenobiotics: Metabolism and Pharmacogenetics of Organic Chemicals and Metals: Proceedings of an International Conference Held in Prague, Czechoslovakia, 27–30 May 1980. Springer Verlag, Berlin, pp. 255– 262. Haake, J.M., Safe, S., Mayura, K., 1987. Aroclar 1254 as an antagonist of the teratogenicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol. Lett. 38, 299–306. Haddad, S., Tardif, R., Charest-Tardif, G., et al., 1999. Physiological modeling of the toxicokinetic interactions in a quaternary mixture of aromatic hydrocarbons. Toxicol. Appl. Pharmacol. 161, 249–257. Haddad, S., Charest-Tardif, G., Tardif, R., et al., 2000. Validation of a physiological modeling framework for simulating the toxicokinetics of chemicals in mixtures. Toxicol. Appl. Pharmacol. 167, 199–209. Hertzler, D.R., 1990. Neurotoxic behavioral effects of Lake Ontario salmon diets in rats. Neurotoxicol. Teratol. 12, 139–143. Hsieh, G.C., Parker, R.D.R., Sharma, R., et al., 1990. Subclinical effects of groundwater contaminants. III. Effects of repeated oral exposure to combinations of benzene and toluene on immunologic responses in mice. Arch. Toxicol. 64, 320–328. Huisman, M., Koopman-Esseboom, C., Fidler, V., et al., 1995a. Perinatal exposure to polychlorinated biphenyls and dioxins and its effect on neonatal neurological development. Early Hum. Dev. 41, 111–127. Huisman, M., Koopman-Essesboom, C., Lanting, C.I., et al., 1995b. Neurological condition in 18-month-old children perinatally exposed to polychlorinated biphenyls and dioxins. Early Hum. Dev. 43, 165–176. 196 H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 Jenkins, K.J., Hidiroglou, M., 1989. Tolerance of the calf for excess copper in milk replacer. J. Dairy Sci. 72 (1), 150–156. Kies, C., Ip, S.W., 1990. Lead bioavailability to humans from diets containing constant amounts of lead: impact of supplemental copper, zinc and iron. Trace Subst. Environ. Health 24, 177–184. Koopman-Esseboom, C., Morse, D.C., Weisglas-Kuperus, N., et al., 1994a. Effects of dioxins and polychlorinated biphenyls on thyroid hormone status of pregnant women and their infants. Pediatr. Res. 36 (4), 468–473. Koopman-Esseboom, C., Huisman, M., Weisglas-Kuperus, N., et al., 1994b. Dioxin and PCB levels in blood and human milk in relation to living areas in the Netherlands. Chemosphere 29 (911), 2327–2338. Koopman-Esseboom, C., Weisglas-Kuperus, N., de Ridder, M.A.J., et al., 1996. Effects of polychlorinated biphenyl/ dioxin exposure and feeding type on infantsÕ mental and psychomotor development. Pediatrics 97 (5), 700–706. Kopp, S.J., B ar any, M., Erlanger, M., et al., 1980a. The influence of chronic low-level cadmium and/or lead feeding on myocardial contractility related to phosphorylation of cardiac myofibrillar proteins. Toxicol. Appl. Pharmacol. 54, 48–56. Kopp, S.J., Glonek, T., Erlanger, M., et al., 1980b. Altered metabolism and function of rat heart following chronic low level cadmium/lead feeding. J. Mol. Cell. Cardiol. 12, 1407– 1425. Korsak, Z., Sokal, J., Dedyk, A., et al., 1988. Toxic effects of combined exposure to toluene and xylene in animals. I. Acute inhalation study. Pol. J. Occup. Med. 1 (1), 45–50. Korsak, Z., Sokal, J.A., G orny, R., 1992. Toxic effects of combined exposure to toluene and m-xylene in animals. III. Subchronic inhalation study. Pol. J. Occup. Med. Environ. Health. 5 (1), 27–33. Kostyniak, P.J., Stinson, C., Greizerstein, H.B., 1999. Relation of Lake Ontario fish consumption, lifetime lactation, and parity to breast milk polychlorobiphenyl and pesticide concentrations. Environ. Res. A 80, S166–S174. Li, S.M.A., Denomme, M.A., Leece, B., 1989. Hexachlorobenzene: biochemical effects and synergistic toxic interactions with 2,3,7,8-tetrachlorodibenzo-p-dioxin. Toxicol. Environ. Chem. 22, 215–227. Magee, A.C., Matrone, G., 1960. Studies on growth, copper metabolism and iron metabolism of rats fed high levels of zinc. J. Nutr. 72, 233–242. Mahaffey, K.R., Fowler, B.A., 1977. Effects of concurrent administration of lead, cadmium, and arsenic in the rat. Environ. Health Perspect. 19, 165–171. Mahaffey, K.R., Capar, S.G., Gladen, B.C., et al., 1981. Concurrent exposure to lead, cadmium, and arsenic: effects on toxicity and tissue metal concentrations in the rat. J. Lab. Clin. Med. 98, 463–481. Marlowe, M., Cossairt, A., Moon, C., et al., 1985. Main and interaction effects of metallic toxins on classroom behavior. J. Abnorm. Child Psychol. 13 (2), 185–198. Mason, R.W., Edwards, I.R., 1989. Acute toxicity of combinations of sodium dichromate, sodium arsenate and copper sulphate in the rat. Comp. Biochem. Physiol. 93 (1), 121–125. Moon, C., Marlowe, M., Stellern, J., et al., 1985. Main and interaction effects of metallic pollutants on cognitive functioning. J. Learn. Disabil. 18 (4), 217–221. Mumtaz, M.M., Durkin, P.R., 1992. A weight-of-evidence approach for assessing interactions in chemical mixtures. Toxicol. Ind. Health 8, 377–406. Mumtaz, M.M., De Rosa, C.T., Durkin, P.R., 1994. Approaches and challenges in risk assessments of chemical mixtures. In: Yang, R.S.H. (Ed.), Toxicology of Chemical Mixtures: Case Studies, Mechanisms, and Novel Approaches. Academic Press, New York, NY, pp. 565–597. Newsome, W.H., Davies, D., Doucet, J., 1995. PCB and organochlorine pesticides in Canadian human milk––1992. Chemosphere 30 (11), 2143–2153. Ogiso, T., Moriyama, K., Sasaki, S., et al., 1974. Inhibitory effect of high dietary zinc on copper absorption in rats. Chem. Pharm. Bull. 22, 55–60. Patandin, S., Koopman-Esseboom, C., De Ridder, M.A.J., et al., 1998. Effects of environmental exposure to polychlorinated biphenyls and dioxins on birth size and growth in Dutch children. Pediatr. Res. 44 (4), 538–545. Patandin, S., Lanting, C.I., Mulder, P.G.H., et al., 1999a. Effects of environmental exposure to polychlorinated biphenyls and dioxins on cognitive abilities in Dutch children at 42 months of age. J. Pediatr. 134 (1), 33–41. Patandin, S., Dagnelie, P.C., Mulder, P.G.H., 1999b. Dietary exposure to polychlorinated biphenyls and dioxins from infancy until adulthood: a comparison between breastfeeding, toddler, and long-term exposure. Environ. Health Perspect. 107, 45–51. Perry, H.M., Erlanger, M.W., 1978. Pressor effects of chronically feeding cadmium and lead together. Trace Subst. Environ. Health 12, 268–275. Plappert, U., Barthel, E., Seidel, H.J., 1994. Reduction of benzene toxicity by toluene. Environ. Mol. Mutagen. 24, 283–292. Pocino, M., Malave, I., Baute, L., 1990. Zinc administration restores the impaired immune response observed in mice receiving excess copper by oral route. Immunopharmacol. Immunotoxicol. 12 (4), 697–713. Pohl, H.R., Hibbs, B.F., 1996. Breast-feeding exposure of infants to environmental contaminants––a public health risk assessment viewpoint: chlorinated dibenzodioxins and chlorinated dibenzofurans. Toxicol. Ind. Health 12 (5), 593–611. Pohl, H., Hansen, H., Chou, C.-H.S.C., 1997. Public health guidance values for chemical mixtures: current practice and future directions. Regul. Toxicol. Pharmacol. 26, 322– 329. Pohl, H.R., Tylenda, C.A., 2000. Breast-feeding exposure of infants to selected pesticides: a public health viewpoint. Toxicol. Ind. Health 16 (2), 65–77. Purcell, K.J., Cason, G.H., Gargas, M.L., et al., 1990. In vivo metabolic interactions of benzene and toluene. Toxicol. Lett. 52, 141–152. Reinstein, N.H., Lonnerdal, B., Keen, C.L., et al., 1984. Zinc– copper interactions in the pregnant rat: fetal outcome and maternal and fetal zinc, copper and iron. J. Nutr. 114, 1266– 1279. Restum, J.C., Bursian, S.J., Giesy, J.P., et al., 1998. Multigenerational study of the effects of consumption of PCBcontaminated carp from Saginaw Bay, Lake Huron, on mink. 1. Effects on mink reproduction, kit growth and survival, and selected biological parameters. J. Toxicol. Environ. Health A 54, 343–375. H.R. Pohl et al. / Chemosphere 53 (2003) 183–197 Rogan, W.J., 1996. Pollutants in breast milk. Arch. Pediatr. Adol. Med. 150, 981–990. Rogan, W.J., Gladen, B.C., McKinney, J.D., et al., 1986a. Polychlorinated biphenyls (PCBs) and dichlorophenyl dichloroethene (DDE) in human milk: effects of maternal factors and previous lactation. Am. J. Public Health 76, 172–177. Rogan, W.J., Gladen, B.C., McKinney, J.D., et al., 1986b. Neonatal effects of transplacental exposure to PCBs and DDE. J. Pediatr. 109, 335–341. Rogan, W.J., Gladen, B.C., McKinney, J.D., et al., 1987. Polychlorinated biphenyls (PCBs) and dichlorodiphenyl dichloroethene (DDE) in human milk: effects on growth, morbidity, and duration of lactation. Am. J. Public Health 77, 1294–1297. Saxena, D.K., Murthy, R.C., Singh, C., et al., 1989. Zinc protects testicular injury induced by concurrent exposure to cadmium and lead in rats. Res. Commun. Chem. Pathol. Pharmacol. 64, 317–329. Seegal, R.F., 1999. Are PCBs the major neurotoxicant in Great Lakes salmon? Environ. Res. 80, S38–S45. Seegal, R.F., Pappas, B.A., Park, G.A.S., 1998. Neurochemical effects of consumption of Great Lakes salmon by rats. Regul. Toxicol. Pharmacol. 27, S68–S75. Seiji, K., Inoue, O., In, C., et al., 1989. Dose–excretion relationship in tetrachloroethylene-exposed workers and the effect of tetrachloroethylene co-exposure on trichloroethylene metabolism. Am. J. Ind. Med. 16, 675–684. Shukla, G.S., Chandra, S.V., 1987. Concurrent exposure to lead, manganese, and cadmium and their distribution to various brain regions, liver, kidney, and testis of growing rats. Arch. Environ. Contam. Toxicol. 16, 303–310. Smith, S.E., Larson, E.J., 1946. Zinc toxicity in rats: antagonistic effects of copper and liver. J. Biol. Chem. 103, 29–38. Stacey, N.H., 1989. Toxicity of mixtures of trichloroethylene, tetrachloroethylene and 1,1,1-trichloroethylene: similarity of in vitro to in vivo responses. Toxicol. Ind. Health 5 (3), 441–450. Stewart, P., Pagano, J., Sargent, D., et al., 2000. Effects of Great Lakes fish consumption on brain PCB pattern, concentration, and progressive-ratio performance. Environ. Res. 82, 18–32. Tardif, R., Lapare, S., Krishnan, K., et al., 1993a. A descriptive and mechanistic study of the interaction between toluene and xylene in humans. Int. Arch. Occup. Environ. Health 65, S135–S137. Tardif, R., Lapare, S., Krishnan, K., et al., 1993b. Physiologically based modeling of the toxicokinetic interaction 197 between toluene and m-xylene in the rat. Toxicol. Appl. Pharmacol. 120, 266–273. Tardif, R., Lapare, S., Charest-Tardif, G., et al., 1995. Physiologically-based pharmacokinetic modeling of a mixture of toluene and xylene in humans. Risk Anal. 15 (3), 335–342. Tardif, R., Charest-Tarif, G., Brodeur, J., et al., 1997. Physiologically based pharmacokinetic modeling of a ternary mixture of alkyl benzenes in rats and humans. Toxicol. Appl. Pharmacol. 144, 120–134. Thawley, D.G., Willoughby, R.A., McSherry, B.J., et al., 1977. Toxic interactions among Pb, Zn, and Cd with varying levels of dietary Ca and vitamin D: hematological system. Environ. Res. 14, 463–475. Toftgard, R., Nilsen, O.G., 1981. Induction of cytochrome P450 in rat liver after inhalation of aromatic organic solvents. In: Gut, I., Cikrt, M., Plaa, G. (Eds.), Industrial and Environmental Xenobiotics: Metabolism and Pharmacokinetics of Organic Chemicals and Metals: Proceedings of an International Conference Held in Prague, Czechoslovakia, 27–30 May 1980. Springer-Verlag, New York, pp. 307–317. Toftgard, R., Nilsen, O.G., 1982. Effects of xylene and xylene isomers on cytochrome P-450 and in vitro enzymatic activities in rat liver, kidney and lung. Toxicology 23, 197–201. Tunek, A., Olofsson, T., Berlin, M., 1981. Toxic effects of benzene and benzene metabolites on granulopoietic stem cells and bone marrow cellularity in mice. Toxicol. Appl. Pharmacol. 59, 149–156. Tunek, A., Hosgstedt, B., Olofsson, T., 1982. Mechanism of benzene toxicity: effects of benzene and benzene metabolites on bone marrow cellularity, number of granulopoietic stem cells and frequency of micronuclei in mice. Chem. Biol. Interact. 39, 129–138. Vassilev, P.P., Venkova, K., Pencheva, N., et al., 1994. Changes in the contractile responses to carbachol and in the inhibitory effects of verapamil and nitrendipine on isolated smooth muscle preparations form rats subchronically exposed to Pb2þ and Zn2þ . Pharmacol. Toxicol. 75, 129– 135. Voors, A.W., Johnson, W.D., Shuman, M.S., 1982. Additive statistical effects of cadmium and lead on heart-related disease in a North Carolina autopsy series. Arch. Environ. Health 37 (2), 98–102. Willoughby, R.A., MacDonald, E., McSherry, B.J., et al., 1972. Lead and zinc poisoning and the interaction between Pb and Zn poisoning in the foal. Can. J. Comp. Med. 36, 348– 359.