Application of the threshold of toxicological concern (TTC) to the safety evaluation of cosmetic ingredients

Available online at www.sciencedirect.com

Food and Chemical Toxicology 45 (2007) 2533–2562 www.elsevier.com/locate/foodchemtox

Application of the threshold of toxicological concern (TTC) to the safety evaluation of cosmetic ingredients q,qq R. Kroes a, A.G. Renwick b,*, V. Feron c, C.L. Galli d, M. Gibney e, H. Greim f, R.H. Guy g, J.C. Lhuguenot h, J.J.M. van de Sandt i b

a Institute for Risk Assessment Sciences, Utrecht University, c/o Seminariehof 38, NL- 3768 EE Soest, The Netherlands School of Medicine, University of Southampton, Biomedical Sciences Building, Bassett Crescent East, Southampton SO16 7PX, UK c Business Unit Toxicology and Applied Pharmacology, TNO Quality of Life, P.O. Box 360, NL 3700 AJ Zeist, The Netherlands d Laboratory of Toxicology, University of Milan, Via Balzaretti 9, Milan 20133, Italy e UCD Institute of Food and Health, University College Dublin, Belfield, Dublin 4, Ireland f Technical University of Munich, Hohenbachernstrasse 15-17, D-85354 Freising-Weihenstephan, Germany g University of Bath, Department of Pharmacy and Pharmacology, Claverton Down, Bath BA2 7AY, UK h ENSBANA, Universite´ de Bourgogne, 1 Esplanade Erasme, F-21000 Dijon, France i TNO Quality of Life, Utrechtseweg 48, 3704 HE Zeist, The Netherlands

Received 1 November 2006; accepted 15 June 2007

Abstract The threshold of toxicological concern (TTC) has been used for the safety assessment of packaging migrants and flavouring agents that occur in food. The approach compares the estimated oral intake with a TTC value derived from chronic oral toxicity data for structurally-related compounds. Application of the TTC approach to cosmetic ingredients and impurities requires consideration of whether route-dependent differences in first-pass metabolism could affect the applicability of TTC values derived from oral data to the topical route. The physicochemical characteristics of the chemical and the pattern of cosmetic use would affect the long-term average internal dose that is compared with the relevant TTC value. Analysis has shown that the oral TTC values are valid for topical exposures and that the relationship between the external topical dose and the internal dose can be taken into account by conservative default adjustment factors. The TTC approach relates to systemic effects, and use of the proposed procedure would not provide an assessment of any local effects at the site of application. Overall the TTC approach provides a useful additional tool for the safety evaluation of cosmetic ingredients and impurities of known chemical structure in the absence of chemical-specific toxicology data.  2007 Elsevier Ltd. All rights reserved.

Abbreviations: AUC, area under the plasma concentration–time curve; BHA, butylated hydroxyl anisole; BHT, butylated hydroxyl toluene; Cmax, maximum observed concentration; Csat, saturation concentration in water; EFSA, European Food Safety Authority; Jmax, maximum flux; log Kp, permeability coefficient; logP, log of the octanol:water partition coefficient; MW, molecular weight; NOAEL, no observed adverse effect level; OP, organophosphate; SCF, Scientific Committee on Food; TTC, threshold of toxicological concern. q This paper is the output of an expert group organised by Colipa (The European Cosmetic Toiletry and Perfumery Association; Comite´ de Liaison de la Parfumerie), Avenue Herrman Debroux 15A, B-1160 Auderghem, Brussels, Belgium; observers who attended one or more meetings were W. Aulmann, Henkel KGaA, 40191 Du¨sseldorf, Germany, M. Bouvier d’Yvoire, European Commission, Joint Research Centre, Institute for Health and Consumer Protection, European Centre for the Validation of Alternative Methods, via Enrico Fermi 1, 21020 Ispra (VA), Italy, G. Nohynec, L’Oreal Recherche, Centre C. Zviak, 90 rue du General Roguet, Clichy Cedex F – 92583, France, T. Peetso, European Commission, Health and Consumer Directorate, 1, rue de Gene`ve, 1140 Brussels, Belgium and P. Wagstaffe, European Commission, Management of Scientific Committees, 200 rue de la Loi, 1049 Brussels, Belgium. qq This paper is one of the last of the major scientific publications of the late Professor Robert Kroes who died in December 2006. The participants at the meetings and his co-authors will remember him as an enthusiastic, stimulating and knowledgeable chairman, a renowned toxicologist and pathologist, and a greatly missed colleague and friend. * Corresponding author. Tel.: +44 01229 588894. E-mail address: [email protected] (A.G. Renwick). 0278-6915/$ - see front matter  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.fct.2007.06.021

2534

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Keywords: Risk assessment; Threshold of toxicological concern (TTC); Cosmetic ingredients; Trans-dermal absorption

1. Introduction The concept that ‘‘safe levels of exposure’’ for humans can be identified for individual chemicals with known toxicological profiles is well accepted in risk assessment. The Threshold of Toxicological Concern (TTC) concept is a pragmatic approach to establish such an exposure level for chemicals for which toxicity data are not available, and is based on chemical structure and the toxicity data of structurally related chemicals. In previous publications (Munro et al., 1996, 1999; Barlow et al., 1999; Kroes et al., 2000, 2004, 2005; Renwick, 2004; ILSI, 2004), the TTC approach was described for chemicals in food in relation to general toxicity endpoints as well as for specific endpoints including carcinogenicity, teratogenicity, reproductive toxicity and immunotoxicity. In addition, consideration was given to structural alerts for high potency carcinogens, endocrine disrupting chemicals, food allergens and to the potential for metabolism and accumulation. A decision tree for applying the TTC procedure, which incorporates a tiered approach, was proposed as a preliminary step in the safety evaluation of food chemicals (Kroes et al., 2004). There is ongoing public concern that humans are exposed to a large number of chemicals of diverse structures, and consequently a need for the safety evaluation of tens of thousands of chemicals. At the same time there is a strong pressure to reduce our reliance on animal experimentation and to rely increasingly on in vitro and in silico data (see the recent Report of the Royal Commission on Environmental Pollution in the UK; Royal Commission, 2003). Without a scientifically-based predictive approach, an unnecessarily large number of chemicals would have to be investigated at high costs with a considerable use of experimental animals. The safety evaluation of chemicals should not be based solely on hazard identification but requires consideration of the two factors that contribute to risk characterization, i.e. hazard characterization and the potential human exposure. The TTC approach provides such a predictive tool for chemicals that have not been subject to toxicity testing. The application of the TTC approach provides a method to assess the potential risk to human health related to exposure to a chemical using • already available data (including chemical structure plus any in vitro, in vivo and/or in silico information), • information on potential exposure, and • the predicted in vivo toxicity based on (chronic) toxicity data of compounds that have similar chemical structures. The use of the TTC approach to date, its potential future applications and its incorporation into the Risk Assessment paradigm was reviewed recently (Kroes et al.,

2005). The TTC as applied to chemicals in food is defined as the highest nominal oral dose which is without appreciable risk to human health after daily lifetime exposure. Toxicological safety testing is not deemed necessary or warranted if the average daily dietary intake is at or below the level of the TTC. The TTC concept may also provide an appropriate tool to evaluate/prioritise the need for toxicity testing and assessment of substances that are present in cosmetic end products and/or their ingredients, including impurities or degradation products. Such a use of the TTC approach will significantly contribute to a reduction in the use of animals for safety tests. This paper reports the deliberations of an expert group that addressed the use of the TTC in the safety evaluation of cosmetic ingredients and provides guidance for the application of the TTC approach for these cosmetic ingredients. 2. History of the TTC approach The generic exposure threshold concept was the scientific basis of the US Food and Drug Administration’s Threshold of Regulation for indirect food additives (Frawley, 1967; Rulis, 1986; Federal Register, 1993; see also Food and Drug Administration, 1983, 1993 and 1995) and the concept of a TTC evolved from the review (Munro, 1990) of the Threshold of Regulation as applied by the Food and Drug Administration in the USA for food contact chemicals. The concept was further developed (Munro et al., 1996, 1999), based on an extensive analysis of available chronic oral toxicity data of substances, which were divided into three chemical classes of toxic potential on the basis of their structure using the decision tree of Cramer et al. (1978). The TTC approach can be applied to low concentrations in food of chemicals that lack toxicity data provided that their structure is known and that there exist valid estimates of intake. The approach was adopted by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) to evaluate flavouring substances (JECFA, 1993, 1995 and 1999; Munro et al., 1999), and since 1997 a procedure that incorporates the different TTCs for the three Cramer et al. (1978) structural classes has been used for the safety evaluation of over 1250 flavouring substances (Renwick, 2004). Subsequent work by an ILSI Europe Expert Group (Kroes et al., 2000 and 2004) extended the approach adopted by the JECFA (JECFA, 1993, 1995, 1999 and 2003) and incorporated the analyses of Cheeseman et al. (1999) into an additional TTC value for compounds with certain structural alerts for possible activity as genotoxic carcinogens. Compounds containing an N-nitroso- or azoxy-group or that were aflatoxin-like compounds were

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

excluded from this additional TTC value because of their high potency. In the TTC decision tree developed by Kroes et al. (2004), proteins, heavy metals and polyhalogenated dibenzo-p-dioxins and related compounds were excluded because the databases used to derive the TTC values did not include toxicity data on proteins, heavy metals such as cadmium, lead and mercury, and because the uncertainty factors used to derive the TTC values would not allow for extreme species differences in elimination seen with polyhalogenated-dibenzo-p-dioxins, polyhalogenated-dibenzofurans and polyhalogenated-biphenyls. In addition, a well established risk assessment method exists for the evaluation of dioxin-like compounds and, therefore, the TTC approach is not necessary for this type of chemical. The Kroes et al. (2004) decision tree provides a systematic structured approach for the consistent application of the TTC approach to chemicals that are present in food at low concentrations excluding those structures discussed above. A review of prior knowledge and use should always be performed preceding application of the decision tree. The decision tree approach starts with the identification and evaluation of possible structural alerts for genotoxicity and high potency carcinogenicity. This step excludes substances with structural alerts for high potency genotoxicity (aflatoxin-like compounds, N-nitroso-compounds, azoxycompounds) from consideration and applies a generic highly conservative threshold for all other structural alerts of 0.15 lg/person/day (0.0025 lg/kg bw/day) (Fig. 1) [see Kroes et al. (2004) for the derivation of this value]. The decision tree considers substances lacking structural alerts for genotoxicity in a sequence of steps related to their structure and the estimates of intake. Analysis of neurotoxicity data showed that the cumulative distributions of no

2535

observed adverse effect levels (NOAELs) for organophosphates was lower than that for other neurotoxins. A TTC of 18 lg/person/day was proposed for organic phosphates. The thresholds of toxicological concern for Cramer et al. (1978) structural classes were established by Munro et al. (1996, 1999) based on an analysis of data from chronic toxicity studies on 137, 28 and 448 compounds in classes I, II and III, respectively. Class I are compounds that have simple chemical structures and efficient modes of metabolism which would suggest a very low order of oral toxicity, class II have structural features which are less innocuous and may contain active functional groups, while class III have structural features that permit no strong initial presumption of safety or may even suggest significant toxicity. For more details on the assignment of substances to structural classes I, II and III, see Cramer et al. (1978) and Kroes et al. (2004). The cumulative distributions of NOAELs for the compounds in each Cramer et al. (1978) structural class were plotted and a log-normal distribution was fitted (Munro et al., 1996). The 5th percentile values for the classes I, II and III NOAEL distributions were calculated to be 3.0, 0.91 and 0.15 mg/kg body weight/day. These values provide a 95% probability that the NOAEL from a chronic animal bioassay on an unstudied compound from the class would be below the relevant 5th percentile value. The 5th percentile NOAEL values were converted to corresponding human intakes by dividing by the usual 100-fold uncertainty factor and then multiplied by 60 to scale to the adult human body weight. The no observed adverse effect level (NOAEL) from a chronic toxicity study in rodents is usually divided by an uncertainty factor of 100 (WHO, 1987) to derive an acceptable daily intake. These analyses gave thresholds of toxicological concern of 1800,

Fig. 1. Consideration of potential genotoxicity using the TTC approach developed by Kroes et al. (2004). Estimated intake refers to chronic exposure since the TTC values are derived from a database where chronic oral exposure was evaluated. If application of the TTC approach suggests that compoundspecific data are required the nature of the additional data will depend on the predicted duration of exposure and nature and size of the population exposed.

2536

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Fig. 2. Consideration of potential non-genotoxic effects using the TTC approach developed by Kroes et al. (2004). Estimated intake refers to chronic exposure since the TTC values are derived from a database were chronic oral exposure as evaluated. The required compound specific data depend on the predicted duration and extend/magnitude of the population exposed.

540 and 90 lg per person per day for structural classes I, II and III, respectively (30, 9 and 1.5 lg/kg bw/day, see Fig. 2). 3. Applications of the TTC approach to non-food chemicals The TTC approach is more broadly applicable than just to chemicals in food, and has potential value in the assessment of risks in other exposure scenarios. A paper dealing with potential future applications of the TTC approach and its incorporation in the Risk Assessment paradigm (see Fig. 3) has been published recently (Kroes et al., 2005). The TTC concept has been accepted by various international and/or national regulatory agencies for consideration of the safety of substances in the human diet and human pharmaceutical preparations, and has been proposed for other chemicals that produce minimal human systemic exposure (Bridges, 2003a). Regulatory TTC values have been defined for migrant substances from packaging material in food (Food and Drug Administration, 1993), flavouring substances in food (JECFA, 2003; Renwick, 2004) or genotoxic impurities in pharmaceutical preparations (EMEA, 2003, 2004; Mu¨ller et al., 2006). The approach was also used by the former EC Scientific Committee on Food and is now used by the European Food Safety Authority to evaluate flavouring substances (EFSA, 2004). The TTC approach has also been endorsed

by the WHO International Program on Chemical Safety for the risk assessment of chemicals (IPCS, 1998) and by the EU Scientific Committee on Toxicology, Ecotoxicology and the Environment (Bridges, 2003b). Recently, the approach has also been suggested for application to aquatic environmental exposure (de Wolf et al., 2005). Application of the TTC approach could be extended to other categories of chemical use, such as constituents of cosmetics and consumer products, as well as (trace) contaminants or impurities present in regulated compounds/ substances, such as food additives, pesticides and cosmetics. A recent paper has suggested that the TTC approach is applicable to constituents of consumer products (Blackburn et al., 2005), and also to occupational exposure in drug manufacturing (Dolan et al., 2005). In addition, since the approach is based on safety evaluations relating to daily intake throughout life, the approach could be used to define exposure levels that would be of negligible concern for contaminants and naturally occurring plant constituents. The TTC approach could also be used to indicate analytical data needs, based on the levels of an impurity in a product that would result in an intake that was not a health concern (as for example for indirect food additives, as it is used in the USA). To extend the TTC method to non-oral exposures, ideally appropriate databases would be compiled for such exposure routes. In the absence of suitable databases, it could be appropriate to use existing databases for oral

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

2537

Fig. 3. Application of the TTC principle in the Risk Assessment Paradigm.

exposure, but route-to-route extrapolation would have to be considered. 4. Application of the TTC to cosmetic ingredients Human exposure to cosmetic products and their ingredients occurs primarily via the topical route, although oral and inhalation exposures also occur. Oral exposures can be evaluated by the usual TTC approach, as outlined for the food area, whereas inhalation exposures would require different considerations and are beyond the scope of this report. If the TTC approach were to be applied to cosmetics, comparison of the TTC values derived from oral toxicity studies with the systemic exposure to chemicals present in cosmetics would need to take into account the different route of human exposure. Topical application and oral ingestion can result in different proportions of the applied dose entering the body as the parent compound (the bioavailability). A difference in bioavailability may arise from (i) more extensive metabolism in the intestine and liver, compared with the skin, prior to reaching the general circulation, or (ii) slower and incomplete transfer across the skin compared with the intestinal wall, due to the physicochemical properties of the compound. Moreover, the slower absorption after topical application results in a different shape to the plasma concentration–time curve even if the same total fraction of the dose is absorbed (Fig. 4).

Fig. 4. Typical patterns of systemic (internal) exposure after oral and topical treatments. (The graphs have been modeled using an elimination half-life of 3.5 h with an absorption half-life of 12 min for the oral dose and 14 h for topical treatment).

In addition, application of the TTC values to cosmetics has to consider various aspects of potential consumer exposure to the cosmetic end product(s). Exposure would need to consider products that contain the relevant ingredient (and/ or impurities) at the highest concentrations, and also the sum from daily cosmetics use when the ingredient/impurity is present in more than a single cosmetic end product. In addition, the duration and frequency of human exposure to the respective cosmetic end product (e.g. rinse-off or leave-on cosmetics) has to be considered. Finally, some cosmetic products do not result in daily, but intermittent,

2538

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

consumer exposure. For example, direct or oxidative hair dyes are typically used at intervals of 3–8 weeks. Therefore, these modifying exposure parameters should be taken into account by applying appropriate, conservative default correction values to the exposure estimation as described below. These different aspects are considered in Sections 4.2–4.7. The TTC approach as described by Kroes et al. (2004) relates only to potential systemic toxicity of substances arising from dermal exposure. The Expert Group recognised that while the approach potentially can also be applied to local effects, the databases on local effects and on substances producing local effects, such as sensitisation or irritation, are currently too limited to be used as a basis for the derivation of valid TTC values for local effects. Regarding the potential systemic toxicity arising from dermal exposure the Expert Group agreed that substances such as proteins, heavy metals and substances with specific structural alerts of concern, which were excluded in the decision tree developed by Kroes et al. (2004), should also be excluded if the TTC approach is used as a first step in the safety evaluation of cosmetic ingredients. The Expert Group examined and discussed a number of issues related to the chemical nature and effects of ingredients and their exposure when used as cosmetics: (i) Similarity between cosmetic ingredients and the chemicals on which the Cramer classes for chemicals in food were based. (ii) Differences in metabolism between the dermal and oral routes of application. (iii) Default adjustment factors for percutaneous penetration to assess the systemic exposure for topically applied cosmetics. (iv) Default adjustment factors for percutaneous penetration to assess the systemic exposure for rinse off products. (v) Default adjustment factors for intermittent use of cosmetic products resulting in intermittent human exposure. (vi) Total (aggregate) exposure to the cosmetic ingredient. (vii) Simultaneous exposure to different cosmetic ingredients. These considerations are discussed below in more detail. 4.1. Similarity between cosmetic ingredients and chemical classes in the Munro et al. (1996) database from which the TTC values for chemicals in food were derived In order to address this question the chemical classes of fragrance ingredients (Bickers et al., 2003) and the substances listed in the first EC update inventory (EC, 2000) were used as being representative of cosmetic ingredients and products, and these databases were compared to the chemicals from which the TTC values for chemicals in food were derived.

4.1.1. Fragrance ingredients (based on Bickers et al., 2003) A total of 2155 substances are listed in Table 1 of Bickers et al. (2003). Many of these are used as flavouring substances and a number of these have been among the 1614 flavors that have been evaluated by the JECFA over the past 8 years using a scheme (Munro et al., 1996) in which the TTC approach is applied (Renwick, 2004). Some functional groups listed in Bickers et al. (2003) have not been evaluated by the JECFA as flavours, examples include aromatic amines, oximes and nitriles. Thus much of the ground work necessary for the evaluation of the fragrances using a TTC approach has already been performed by the JECFA because many fragrances can be assigned to one of the existing groups of chemicals evaluated by the JECFA. Application of the TTC approach will require consideration of the estimated systemic exposure from topical application, and whether first-pass metabolism and detoxication differ in extent between percutaneous penetration and oral exposure (see below). 4.1.2. Substances listed in the First Update EU Inventory (EC, 2000) In addition to the fragrances discussed above the Inventory contains dyes and colourants, food ingredients, low molecular weight organic compounds with a variety of uses, inorganic salts of various metals, normal organic constituents of the human body, pharmaceutical-type compounds, plant and animal extracts, polymeric compounds and surfactants, emollients, humectants and emulsifying agents. These groups are discussed below regarding the applicability of the TTC approach. 4.1.2.1. Dyes and colourants. Most of these substances are not approved for food or medicinal use. A wide variety of aromatic amino compounds is included in this group. Because similar compounds are in the Munro et al. (1996) database, this group could be assessed using the TTC approach. There exists a significant database on some compounds enabling a comparison with representative members of the group where relevant. 4.1.2.2. Food-type components. Food components such as acetylated glycerides, aspartame, BHA, BHT, caffeine, cinnamaldehyde, erythritol, glyceryl esters and sucrose esters have approved food uses. Comparisons with SCF and EFSA approvals may cover the cosmetic use of some chemicals in the Inventory, and for other compounds the TTC approach could be applied. 4.1.2.3. Low molecular weight organic compounds with a variety of uses. These compounds would be well suited to safety evaluation using the TTC approach as described in the paper of Kroes et al. (2004). Examples include amyl salicylate, the numerous CI compounds (e.g. CI 10006, etc.) and dibenzothiophene.

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

4.1.2.4. Inorganic salts of various metals. These include a mixture of compounds that occur naturally at quite high levels in the human body and in foods (such as sodium, calcium and zinc salts), plus other potentially more toxic compounds (such as manganese, nickel and strontium compounds). Inorganic compounds are not well represented in the databases used to derive the TTC values that are presented in Munro et al. (1996) and Kroes et al. (2004). It seems prudent to exclude heavy metals and heavy metal salts from the TTC application as has been suggested previously (Kroes et al., 2004). 4.1.2.5. Normal organic constituents of the human body. These can be assumed to be generally of no safety concern, unless they have intrinsic potent biological activity, such as hormonal activity. Examples of topical compounds with no safety concern include amino acids, adenosine triphosphate, cholesterol, creatinine, ergosterol, glucosamine, pyridoxine, tocopherol esters, urea and uric acid. The TTC approach would not be necessary for such compounds. 4.1.2.6. Pharmaceutical-type compounds. Some of these compounds have been evaluated as constituents of medicinal products. For such compounds there are extensive data that would support a compound-specific risk assessment, such that a TTC-based approach would not be needed Examples in the Inventory include acetaminophen (paracetamol), chlorhexidine, various parabens, ketoconazole and lithium compounds. However, the TTC approach could be applied when exposures are in the low range and if pharmacological activity can be excluded. In general however, chemical structures that may have or are suspected to have pharmacological properties should be excluded for application of the TTC, unless the exposure is below that producing any effect. 4.1.2.7. Plant and animal extracts. Despite the presence of animal extracts in the Inventory, in reality only plant extracts are used in cosmetics and these are the focus in this category. The Inventory contains numerous examples of hydrolysates, resins and oleoresins, such as the various citrus, coriander, echinacea, geranium, helianthus, prunus and wheat germ preparations. The TTC approach could not be applied to these complex mixtures because the starting point is the known chemical structure. The JECFA is currently developing a procedure for the evaluation of plant materials used as flavors. 4.1.2.8. Polymeric compounds. The polymeric compounds would not be absorbed intact to any significant extent. High molecular weight polymers are unlikely to be absorbed through the skin and the gastrointestinal tract. The guidance from the European Food Safety Agency (EFSA, 2005) indicates that only the polymeric fraction below 1000 Da has to be regarded as toxicologically relevant to the assessment of polymeric substances that might be swallowed. In case of oral intake (e.g. lipsticks), this

2539

fraction can be accounted for as indicated below.1 For topical applications the Default adjustment factors as described in Chapter 4.7 should be used for the fraction below 1000 Da. For polymeric compounds application of the TTC approach could be applied to the safety evaluation of any residual monomers and oligomers contaminants following percutaneous absorption. 4.1.2.9. Surfactant, emollients, humectants and emulsifying agents. There are many compounds in this group, which are generally macro-components of cosmetics. Although the TTC approach could be applied to these chemical structures, in practice the exposures would be likely to exceed any relevant TTC value in the Kroes et al. (2004) procedure, and chemical-specific data would be needed to reach a safety conclusion. Application of the TTC approach could be of practical value in the safety assessment of low level impurities in such macro-components. In conclusion there is considerable potential for the TTC approach using the expanded decision tree of Kroes et al. (2004) for many of the groups as discussed above. Comparisons of the chemical structures for cosmetic ingredients with chemicals in the database used to develop the TTC values in the decision tree of Kroes et al. (2004) is not the only issue that needs consideration; of equal importance is the possible impact of the route of administration. The validity of the TTC values for the topical route could be influenced by route-specific differences in biotransformation between topical and oral administration (Section 4.2), while the potential systemic exposure of the compound under evaluation will be influenced by the fact that the skin, unlike the gastrointestinal tract, is a significant barrier to the absorption of many chemicals (Section 4.3). 4.2. Differences in metabolism between the dermal and per oral routes of application There are major differences in the rates and extents of transfer across the skin compared with across the gastrointestinal tract. When the TTC approach is used for cosmetic ingredients it is suggested that exposure is modified according to molecular characteristics, such as lipid solubility and molecular weight, which predict the rate and extent of transfer across the skin and thus the extent of systemic exposure. Any physicochemical influence on the extent of transfer into the general circulation is not considered in this analysis (see Section 4.3) and this section concentrates on the influence of route of administration on the extent of presystemic metabolism. Presystemic metabolism is that which occurs as the chemical is transferred between the site of administration 1

Example: If the fraction below 1000 Da of a high-molecular weight substance is determined as F = 2% (i.e. 98% are above 1000 Da), and the amount of substance (S) applied as a lipstick is 2 mg/day, then the amount of absorbable substance (Sab) is given by correcting the applied amount of substance S by the factor F: Sab = F · S = 0.02 · 2 mg = 0.04 mg = 40 lg.

2540

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

and the systemic circulation. In the context of pharmacokinetics this process is usually called ‘‘first-pass’’ metabolism because it occurs during the first passage of an oral dose of a the drug through the liver; subsequent post-absorption passages through the liver are responsible for the removal of the drug from the body, i.e. the clearance of the compound from the circulation. In this paper, the term ‘‘pre-systemic’’ is used to describe metabolism within the gastrointestinal tract and liver during absorption of an oral dose and also within the skin as the chemical is absorbed following topical application. Route-dependent differences in presystemic metabolism could be important because (i) a route-dependent difference in the extent of metabolism could result in a different balance of parent compound and metabolites reaching the systemic circulation after dermal and oral administration, (ii) a difference in pathways of metabolism in different tissues could result in metabolites being formed by one route and not by the other and (iii) a difference in the rates of absorption could result in saturation of metabolism, which might influence the pattern of systemic exposure, for example saturation of hepatic metabolism might occur after an oral bolus dose but not after dermal exposure. The database used by Munro et al. (1996) to derive the TTC values that were incorporated into the decision tree (Kroes et al., 2004) largely comprised low molecular weight organic compounds. There have been a number of reviews of the metabolism of foreign chemicals in the skin (Merk et al., 2004; Swanson, 2004) and in the gastrointestinal tract and liver (Ding and Kaminsky, 2003; Doherty and Charman, 2002). This paper does not attempt to review the different pathways of metabolism, but explores the basic principles in the context of applying the TTC values to cosmetic ingredients. The text concentrates on the influence of route-dependent differences in metabolism on systemic exposure to the chemical and its metabolites, because systemic effects would be relevant to both dermal and oral administration. 4.2.1. Route-dependent differences in the extent of presystemic metabolism of foreign compounds Foreign compound metabolism is complex with numerous pathways, each of which may involve a number of enzymes and isoenzymes that may show differential expression in the skin, gastrointestinal tract and liver. The consequences of metabolism on the potential for toxicity are also complex, because while many enzyme-catalyzed metabolic reactions result in detoxication of the substrate, the same metabolic reaction may be responsible for the activation of a different substrate into a toxic form. In general, extensive presystemic metabolism is far more likely following oral administration than following topical application (Merk et al., 2004) because

(i) there are higher levels of enzyme activity in the gut wall and especially in the liver compared with the skin, (ii) the chemical has to pass through a greater mass of metabolizing tissue after oral administration, (iii) there would be a greater uptake from extracellular fluid by hepatocytes than by cells in the dermis or epidermis, because hepatocytes have a brush border which is related to their function as a metabolic barrier between the gut lumen and the general circulation, (iv) the blood flow in the gastrointestinal tract carries the chemical to the main site of presystemic metabolism in the hepatocytes, while the blood flow in the dermis removes the chemical from possible sites of presystemic metabolism deeper in the dermis and in subcutaneous tissue. In theory, differences in the metabolism of chemicals after oral and topical exposure could result in one of two possible scenarios. (a) Increased systemic exposure to the parent compound following topical treatment associated with lower amounts of any presystemic metabolites. Many of the pathways of foreign compound metabolism have been detected in the skin, although the levels of enzyme activity per mg of tissue are generally much lower than in the gastrointestinal tract and liver (Merk et al., 2004). In consequence, it can be predicted that as a general rule there will be greater presystemic metabolism of that fraction of an oral dose that passes from the gut lumen into the general circulation than of the absorbed fraction of a topical dose. This means that a greater proportion of the total systemic exposure will be to the parent compound after topical application, associated with lower levels of the presystemic metabolites. (b) Decreased systemic exposure to the parent compound following topical treatment associated with higher amounts of presystemic metabolite(s). This would be an extremely unlikely scenario, but could theoretically arise if the chemical were metabolized by an isoenzyme that was expressed to a greater extent in the skin than in the gut or liver. For example, Nacetyl-transferase (NAT1) is expressed at very high levels in the skin (Kawakubo et al., 1990; Reilly et al., 2000), but the activity in human epidermis skin is lower than that in human liver (Nohynek et al., 2005). The patterns of systemic exposure would be the different from those described under (a) above, with proportionately less parent compound and more presystemic metabolite reaching the systemic circulation after topical application. The activities of enzymes involved in foreign compound metabolism are higher in the liver than in the skin, so that this is only a theoretical possibility and will not be considered further.

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

The toxicological consequences of a route-dependent difference in presystemic metabolism would depend on the toxicological activities of the parent compound and the presystemic metabolite. If the rates and extents of presystemic metabolism in the gut and liver following oral dosage are higher than those in the skin following topical exposure, then the oral database of Munro et al. (1996) could either 1. under-predict toxicity after topical application if the parent compound were the active form and first-pass metabolism of the oral dose resulted in detoxication or 2. over-predict toxicity after topical application if the presystemic metabolite were the active form. The consequences for this in relation to the structuralclass database used by Munro et al. (1996) and Kroes et al. (2004) are discussed below. This analysis relates to that fraction of the dose that is transferred into the systemic circulation; non-metabolic influences on the fractional absorption are considered in Section 4.3. 4.2.2. Influence of the extent and rate of absorption on the potential for metabolism 1. The extent of absorption: The extent of absorption is dependent on the rate of diffusion of the chemical through the tissue, the surface area available for absorption and the time over which the chemical is in contact with the site of absorption. For both oral and topical exposures the greatest rates of absorption are seen with slightly lipid soluble compounds, and there is poor or incomplete absorption for highly polar or highly lipid soluble compounds. Highly polar compounds in the Munro et al. (1996) database that show incomplete absorption from the gastrointestinal tract would not undergo significant absorption across the skin. Highly lipid soluble compounds could also show incomplete absorption by both routes, with a topical dose being sequestered in the stratum corneum and lost as this layer is replaced, and with an oral dose passing through the gut without forming a micellar solution (e.g. liquid paraffin). In general there is much lower absorption after topical application and a greater potential for ‘‘physical’’ loss of the chemical prior to absorption due to removal of part of the ‘‘dose’’ by washing, loss on clothing etc. Such nonmetabolic, physicochemical considerations are taken into account in the determination of systemic exposure by the use of adjustment factors to the TTC values for each structural class (see Section 4.3). 2. The rate of absorption: The rate of absorption would have a profound influence on the shape of the plasma concentration–time curve (even if there is a similar extent of absorption topically and orally) (Fig. 4). Oral absorption of lipid soluble compounds occurs largely across the wall of the small intestine, is efficient and results in the rapid appearance of the peak concentration (Cmax), with the rate of increase dependent on the absorption rate constant. In

2541

contrast, the profile following topical application is a very gradual increase, with a rate that reflects the elimination rate constant (flip-flop kinetics). Therefore there would be a much lower maximum concentration (Cmax) after topical administration, even if the total area under the plasma concentration–time curve (AUC) to infinity were to be the same. This difference could have an impact on the risk assessment if the toxicity of concern were due to the peak concentration rather than the AUC. Often toxicity arises after exposure is at or above a minimum plasma concentration for some minimum duration; topical exposure usually gives low plasma levels which may not exceed the minimum concentration necessary to elicit a toxicological effect. The toxicity data used to derive the TTC values was for sub-chronic or chronic systemic effects after oral administration, and it would be reasonable to assume that these would depend on AUC rather than Cmax. If Cmax were important for toxicity, then the oral database would greatly over-predict the risk following topical administration. This has been shown for the risk assessment of human topical exposure to pesticides on the basis of animal oral toxicity data, which resulted in an overestimation of human risk (Ross et al., 2000). If the adverse effects were related to the AUC of the parent compound, the oral database would only under-predict the potential effect following topical administration if there was significant first-pass metabolism in the gut or liver to inactive metabolites, such that the AUC of the active parent compound were smaller after oral than after topical exposure, after physicochemical differences have been allowed for (Sections 4.3 and 5). 4.2.3. Route-dependent differences in metabolism in relation to the TTC values derived by Munro et al. (1996) and Kroes et al. (2004) Based on the discussion above, the only plausible scenario in which the oral database of Munro et al. (1996) and Kroes et al. (2004) may under-predict the potential toxicity following topical application is where the toxicity is due to the parent compound (or a systemic metabolite that would be formed as a proportion of the systemic exposure to the parent compound) and intestinal/hepatic presystemic metabolism results in detoxication – the fractional transfer, due to physicochemical characteristics, is considered in Section 4.3. The greater the extent of presystemic metabolism the lower would be the systemic exposure to the parent compound. As an example a twofold route-dependent difference would arise if the compound underwent 50% presystemic metabolism in the gut/liver and zero presystemic metabolism via the topical route (100% bioavailability), while a fivefold difference would require that there was 80% presystemic metabolism via the oral route (20% bioavailability) and zero presystemic metabolism via the topical route (100% bioavailability). High presystemic (first-pass) metabolism in the liver indicates that the compound is a good substrate for hepatic

2542

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

enzymes. This metabolic characteristic would result in a high metabolic removal from the general circulation (high clearance) and a short half-life. In consequence, compounds which undergo high first-pass metabolism have short half-lives. In contrast, compounds with long halflives are not good substrates for metabolism and therefore would show negligible first-pass metabolism. Thus insights into the potential for presystemic metabolism can be derived from data on the route and rate of elimination from the general circulation. The following analysis considers the different TTC values in the decision tree of Kroes et al. (2004) (Fig. 2). The TTC values were derived from analyses (Munro et al., 1996; Kroes et al., 2004) of the distributions of NOAEL values for compounds sharing the same chemical characteristics. The TTC values were based on the 5th percentile of each distribution (in mg/kg body weight per day) divided by the usual 100-fold uncertainty factor and multiplied by 60 to allow for an average body weight of 60 kg. (i) Organophosphate (OP) TTC (18 lg per person/day). This TTC value (based on a 5th percentile NOAEL of 0.03 mg/kg bw/day) is not directly relevant because such compounds would not be used as cosmetic ingredients. This TTC value was included in the Kroes et al. (2004) decision tree to allow for the possibility that a novel (non-regulated) OP might be identified in human food. This TTC would only be relevant to cosmetics if an OP were present as a minor contaminant in one of the ingredients. Although the TTC value of OPs could be applied, any such contaminant detected would probably be a regulated OP with an established chemical-specific database and the chemical-specific database should be used to assess any risk if potential human exposure would surpass the OP related TTC. (ii) Structural class III TTC (90 lg per person/day). The structural class III TTC value was derived from the database of Munro et al. (1996), which included OPs and polyhalogenated compounds as well as low molecular weight compounds of potential relevance to cosmetics. There were 448 chemicals in the database and the calculated 5th percentile NOAEL (0.15 mg/kg bw /day) approximates to that of the 22nd most potent compound (Appendix Table 1). This TTC value would only be insufficiently conservative for topical exposures if 1. the chemicals used to derive the 5th percentile value were all relevant to cosmetic ingredients and 2. any presystemic metabolism after oral exposure resulted in detoxification and 3. those chemicals with NOAEL values close to the 5th percentile NOAEL had very low oral bioavailabilities, such that the oral data would underestimate the systemic exposure to the parent compound and resulting toxicity via the topical route with higher bioavailability.

In order to apply the oral NOAEL values to topical exposure, each NOAEL value would have to be corrected to allow for the absolute oral bioavailability. For example, a compound with 10% oral bioavailability due to 90% firstpass metabolism and an oral NOAEL of 1 mg/kg bw/day would have a ‘‘bioavailability-corrected NOAEL’’ of 0.1 mg/kg bw /day, i.e. the oral NOAEL would have been 0.1 mg/kg bw /day if 100% of the oral dose had reached the general circulation as the toxic parent compound. Under these circumstances the distribution of ‘‘bioavailability-corrected NOAEL’’ values may differ from that used by Munro et al. (1996) to determine the 5th percentile NOAEL. The class III TTC value would be inappropriate for topical exposures if correction for oral bioavailability resulted in more than 5% of chemicals having ‘‘bioavailability-corrected NOAEL’’ values below 0.15 mg/kg bw /day. As discussed above the decision tree of Kroes et al. (2004) considers OPs separately against an OP TTC value of 18 lg per person/day. In consequence, OP compounds in the class III database of Munro et al. (1996) should have been deleted in the Kroes et al. (2004) decision tree, but the original class III threshold of Munro et al. (1996) was retained. Therefore OPs should be deleted from data used to assess the validity of the TTC value for class III compounds when the Kroes et al. (2004) decision tree is applied to cosmetics. Deletion of potent OPs reduces the list of class III compounds to a total of 428 and therefore the 5th percentile will approximate to the NOAEL of the 21st most potent of the remaining compounds (Table 1), i.e. about 0.3 mg/kg bw/ day. Thus, deleting OPs from consideration doubles the TTC value for class III. This would allow all compounds in the class III database to undergo 50% presystemic detoxication without affecting the validity of the TTC value for non-oral routes of administration. In effect this difference is equivalent to an additional safety factor of 2 if the TTC values in the Kroes et al. (2004) paper are applied. Analysis of the extent of any presystemic inactivation following oral dosing for the 61 compounds with NOAELs of 1.0 mg/kg bw /day or less is given in Appendix Table 1. Calculation of numerical ‘‘bioavailability-corrected NOAEL’’ values was not possible due to the lack of quantitative oral bioavailability data. However, sufficient metabolic data were available to allow qualitative analysis of the potential for presystemic detoxication. None of the 61 compounds in Appendix Table 1 has been shown to undergo major presystemic detoxication after oral dosage. Adequate data to allow conclusions about the potential for first-pass metabolism were not available on 14 compounds. A total of 20 compounds were predicted to undergo negligible first-pass metabolism based on structural characteristics; 12 were polyhalogenated compounds and 8 were polar unmetabolized compounds or showed steric hindrance of possible sites of metabolism. In addition, 11 compounds would undergo presystemic metabolism to toxic products such that oral toxicity would exceed topical toxicity.

Table 1 Skin absorption data of 15 cosmetic ingredients tested under in-use conditions MW

Diethanolamine

105.1

p-Phenylene-diamine

108.2

Mexoryl SX

562

p-Methoxycinnamic acid 2-ethylhexyl ester

290.4

UVASORB K2A

Pigment Red 57

Log P

Vehicle/ formulation

Dose of active applied (mg/cm2)

Dose of formulation applied (mg/cm2)

Exposure time (min)

In vivo/ in vitro

Species Relative absorption Reference (% of dose)

0.0042 0.0077 0.1093 0.1089 1.31 g/subject 0.44

13 13 17.5 17.5 70 g/subject 20

5 5 30 30 30 30

In vitro

Kraeling et al. (2004)

In vivo In vitro

Human 1.0 at 24 h 2.0 at 24 h 1.4 at 24 h 2.4 at 24 h Human 0.43 U+F 2.4 at 24 h

Sunscreen emulsion 0.04 0.04

2 2

240 240

In vivo In vitro

Human 0.014 at 24 h 0.16 at 24 h

Benech-Kieffer et al. (2003)

6.0

NC-O/W NC-W/O O/W W/O

0.04 0.04 0.04 0.04

8 8 8 8

1440 1440 1440 1440

In vitro

Pig

Jimenez et al. (2004)

765

4.7

2.5% emulsion 5% emulsion 10% emulsion

0.25 0.53 0.82

8.8 8.8 8.8

1440 1440 1440

In vitro

Human 0.91 at 2h 0.52 at 24 h 0.33 at 24 h

SCCNFP/0814/ 04

430.4

3.6

Commercial formulation

1.00

200

30

In vitro

Pig

0.01 at 24 h

SCCNFP/0795/ 04

Ethyl lauroyl arginate HCl 421

1.4

Propylene glycol/ water 30/70

0.10

4.8

1440

In vitro

Pig

5.4 at 24 h

SCCNFP/0837/ 04

Imezine BD

492.5

2.0

Hair dye formulation

0.09

20

30

In vitro

Human 0.9 at 24 h

SCCNFP/0875/ 05

Acid Yellow 23

534.4

10.2

Hair dye formulation

5

1000

30

In vitro

Pig

SCCNFP/0786/ 04

4-Methyl-benzylidene camphor

254.4

5.9

O/W formulation

0.25

5

360

In vivo

Human 0.75–1.9 at 24 h

SCCNFP/0779/ 04

Tetrabromophenol Blue

985.6

1.67

100

30

In vitro

Pig

0.77 at 72h

SCCNFP/0794/ 04

Benzethonium chloride

448.1

0.002

2

1440

In vitro

Human 4.72 at 24 h

Caffeine

194.2

3

960

In vitro

Human 1.44 at 16h

SCCNFP/0762/ 03 Potard et al. (1999)

Catechol

110.1

0.88 Hair dye formulation

20

30

In vitro

Human 1.9 at 24 h

Jung et al. (2003)

Acid Yellow 3

375.3 + 477.0

0.7

1200

30

In vitro

Pig

SCCNFP/0789/ 04

1.43 Diluted shampoo A Diluted shampoo B Hair dye C Hair dye D 0.25 Oxidative hair dye formulation 3.9

10.33 Oxidative formulation 4.00 Unspecified cream

0.07 Anti-cellulite cream 0.09 0.12

Formulation at pH 6 3.0

1.25 at 24 h 0.5 at 24 h 3.4 at 24 h 5.4 at 24 h

0.26 at 24 h

0.37 at 24 h

Hueber-Becker et al. (2004)

2543

Abbreviations: MW, molecular weight; Log P, log of the octanol:water partition coefficient; NC-O/W, nanocapulses in oil-in-water; NC-W/O, nanocapulses in water-in-oil; O/W, oil-in-water; W/O, water-in-oil; U+F, urine and faeces collected over 24 h. SCCNFP references refer to reports of the Scientific Committee on Cosmetic Products and Non-food Products intended for Consumers, with the first number giving the reference number and the second number the year of publication http://ec.europa.eu/health/ph_risk/committees/04_sccp/04_sccp_en.htm.

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Cosmetic ingredient

2544

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

It is clear from this detailed analysis that the majority of class III compounds do not undergo presystemic detoxication after oral dosage; indeed many would show higher toxicity after oral dosage because hepatic first-pass metabolism results in the generation of a toxic metabolite. The data given in Appendix Table 1 indicate that if a 5th percentile NOAEL could be calculated for topical exposure to class III compounds it would be considerable higher than the value of 0.15 mg/kg bw/day calculated by Munro et al. (1996) based on oral data. (iii) Structural class II TTC (540 lg per person/day). The structural class II TTC value was derived from the database of Munro et al. (1996) based on oral data for 28 chemicals for which the calculated 5th percentile NOAEL was 0.91 mg/kg bw /day. Analysis of the potential for presystemic detoxication was undertaken for the seven compounds with NOAEL values of 10 mg/kg bw /day or less (Appendix Table 2). Overall, there was no evidence that the oral NOAEL values would under-predict the toxicity profiles of these compounds following topical administration. (iv) Structural class I TTC (1800 lg per person/day). The structural class I TTC value was derived from the database of Munro et al. (1996) based on oral data for 137 chemicals for which the calculated 5th percentile NOAEL was 3.0 mg/kg bw /day. Analysis of the potential for presystemic detoxication was undertaken for the 24 compounds with NOAEL values of 30 mg/kg bw/day or less (Appendix Table 3). It was more difficult to predict the consequences of metabolism on the NOAEL values for these less toxic compounds. Only three of the compounds would undergo first-pass metabolism to active and toxic products. Overall, there was no evidence that the oral NOAEL values would under-predict the toxicity profiles of these compounds following topical administration. 4.2.4. Conclusions regarding route-dependent differences in metabolism The 5th percentile oral NOAEL values used to derive the TTC values for class III compounds would overestimate the potential toxicity of the same compounds following topical exposure, even if 100% of the topical dose entered the general circulation as the parent compound. Although fewer compounds were used to derive the 5th percentile oral NOAEL values for class II and class I, the available data indicated that these values and the resulting TTC values would also be relevant to topical exposures. 4.3. Default adjustment factors for percutaneous absorption to assess the systemic exposure for topically applied cosmetics Exposure estimation is of critical importance in application of the TTC approach. As discussed above, the TTC

values used in the decision tree of Kroes et al. (2004) are valid for cosmetic ingredients but exposure estimations have to consider how physicochemical properties influence the fractional absorption. Because most chemicals do not readily cross the skin permeability barrier, the use of the TTC concept in the safety evaluation of cosmetic ingredients and cosmetic products requires an estimate of the absorption across the skin. At present, the rather arbitrary value of 10% absorption of the amounts contacting the skin is used as default value for risk assessment of plant protection products (EC, 2002, 2004) and industrial chemicals (ECB, 2003). The EU Draft Guidelines on skin absorption of pesticides have suggested default skin absorption values of 10% or 100% on the basis of certain physical/chemical values of the compound in question (SCP, 2002). It is important to emphasize that these are conservative cut-off values, rather than quantitative predictions. A cut-off value of 10% absorption is not applied to cosmetic ingredients, and guidelines on percutaneous absorption have suggested that a default absorption value of 100% should be applied in the absence of experimental data (SCCP, 2006). While there is merit in the use of one or two conservative default factors, it is important to appreciate that absorption will also depend upon the manner in which the chemical is presented to the skin (i.e., the ‘formulation’ in which it is applied), the area of skin exposed and the duration of skin contact/exposure. These issues are discussed further below. 4.3.1. Experimental absorption data Data from the open literature were investigated to obtain insights into the extent of actual absorption of cosmetic ingredients under in-use conditions; the results for 15 compounds (hair dyes, preservatives, and sunscreens), with MW ranging from 105 to 986 and estimated log P values from 10.2 to 10.3 were identified for detailed examination (Table 1). The estimated absorption was less than 8% for each compound investigated (Table 1). For in vitro studies absorption was defined as the systemically available amount plus that present within the skin, since the amounts found in the epidermis (except for the stratum corneum) and dermis act as a sink and are considered to be absorbed and are added to the amount found in the receptor fluid. The amounts that are retained by the stratum corneum (when collected at the end of the experiment) are not considered to be absorbed, because the stratum corneum is lost by abrasion so that these amounts do not contribute to the systemic dose. Based on this analysis of published data, the assumption of 100% absorption is not scientifically supportable. It is proposed that (conservative) default absorption factor(s) are developed to correct the applied dose to provide an internal exposure suitable for comparison with the relevant TTC value. The proposed approach is based on defining the maximum absorbability of a chemical based on its physicochemical properties and then using adjustment factors related to the specific exposure scenario involved.

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

4.3.2. Theoretical principles regarding percutaneous absorption The maximum flux (Jmax), at which a chemical can cross the skin, is theoretically achieved when it is maintained in a saturated solution (or in neat chemical form) on the surface. The relevant equation that applies in these circumstances is Fick’s first law: J max

D ffi  K skin=vehicle  C sat vehicle h

where D is the chemical’s diffusivity across the skin (typically that through the stratum corneum, the skin’s least permeable and outermost layer), h is the diffusion pathlength, Kskin/vehicle is the compound’s partition coefficient between the skin and the vehicle contacting the surface, and C sat vehicle is its saturation solubility in the vehicle. Given that Kskin/vehicle may be defined as follows: K skin=vehicle ¼

C sat skin C sat vehicle

ð2Þ

sat where C sat skin and C vehicle are the concentrations in skin and vehicle respectively. Eq. (1) can then be reduced to a simpler form

J max ¼

D  C sat skin h

ð3Þ

which shows that the maximum flux achievable across the barrier is independent of the formulation, providing that the formulation is saturated. That is, Jmax should be constant as long as the chemical is at its maximum thermodynamic activity in the vehicle (i.e., that it is saturated), and provided that the excipients in the formulation do not change the skin’s barrier properties (e.g., exhibit permeation-enhancing, or retarding, characteristics). Hence, if the value of Jmax can be predicted from first principles, it should then be possible, with knowledge of the degree of saturation of the chemical in a particular formulation, to calculate the maximum amount absorbed across the skin following a specific scenario. This objective can be achieved using an algorithm derived by Potts and Guy (1992) from an extensive database of the permeability coefficients of approximately 100 chemicals across human skin in vitro following their application in water. The permeability coefficient of a chemical (Kp) from an aqueous vehicle is defined as Kp ¼

D  K skin=water h

ð4Þ

Multiple regression analysis of the experimental values of Kp against various physicochemical variables led to the derivation of an equation (Potts and Guy, 1992), which has been shown to have reasonable predictive power: log K p ¼ 2:7 þ 0:71  log P  0:0061  MW

In the above form, the units of permeability coefficient are cm/h. For very lipophilic chemicals, it is necessary to correct the value of Kp calculated from Eq. (5) to take into account the contribution of the living skin layers (viable epidermis and dermis) to the permeation process (Cleek and Bunge, 1993): K corr ¼ p

ð1Þ

ð5Þ

where P is the octanol–water partition coefficient of the chemical and MW is its molecular weight.

2545

Kp 1þ

pffiffiffiffiffiffiffi K p  MW 2:6

ð6Þ

Combining Eqs. (1), (4) and (6) yields: J max ¼ K corr  C sat water p

ð7Þ

The permeability coefficient can be calculated from Eqs. (5) and (6) and readily available physicochemical parameters (MW and log P), for which a very large database of values exists, or which can be calculated with many different approaches available on the internet (http://www.daylight.com/release/index.html; http://www.syrres.com/esc/ kowdemo.htm). Equally, there is a considerable number of aqueous solubilities tabulated and/or accessible via the web (e.g http://146.107.217.178/lab/alogps/). Using Eqs. (5) and (7), maximum fluxes were determined for 62 chemicals in the EDETOX database (http:// www.ncl.ac.uk/edetox/theedetoxdatabase.html) for which in vitro permeation data across human skin have been reported in the literature. The respective octanol-water partition coefficients and aqueous solubilities were found either in the EDETOX database (EDETOX, 2004), or from one or more websites (http://www.daylight.com/ release/index.html; http://www.syrres.com/esc/kowdemo. htm http://146.107.217.178/lab/alogps/). The results are presented in Table 2. 4.3.3. Practical implications for percutaneous absorption The compounds included in this analysis ranged in molecular weight from 60 to more than 450 Da; their values of log P were between 2.2 and +6.1, i.e., the partition coefficients spanned more than eight orders of magnitude. The calculated values of Jmax similarly covered a wide range from 30 pg/cm2/h for methotrexate to more than 1 mg/cm2/h for nicotine. The form of the Potts and Guy (1992) equation (Eq. (5)) implies that skin permeability will become more difficult with increasing molecular weight, if all other factors remain equal. Macroscopically, the database of 62 chemicals investigated conforms to this behavior (see Fig. 5). However, while this graph is visually persuasive, it must be emphasized that the y-axis is logarithmic and that, for any particular value of molecular weight, a range of Jmax values is apparent. This reflects the fact, of course, that each subset of MW values includes compounds which may differ considerably in their lipophilicity (as shown on the graph) and that this parameter can be the ultimate determinant of the maximum flux. The latter point is amplified in Fig. 6. While the Potts and Guy (1992) correlation (Eq. (5)) suggests that

2546

Table 2 Calculated maximum fluxes for 62 chemicals in the EDETOX database MW

Log P

log Kp

Kp (cm/h)

Csat (mg/cm3)

Jmax (mg/cm2/h)

Methotrexate Benzo[a]pyrene Aldosterone Griseofulvin Acyclovir Chlorpyrifos T2 Toxin Cortisone Hydrocortisone Lindane Estradiol Methylene-bis-(2-chloroaniline) Dinitrochlorobenzene Progesterone Testosterone Parathion Parathion methyl Diazinon Butachlor Triclosan Fluorouracil Mannitol Propranolol Nitro-1,4-benzenediamine Cinnamyl anthranilate Methylenedianiline Propoxur Caffeine Pentachlorophenol Cinnamic acid Coumarin Acetylsalicylic acid DEET Nicotinate benzyl Nicotinic acid Nitrobenzene Methyl-4-hydroxybenzoate Urea Lidocaine Cinnamyl alcohol Phenylphenol Salicylic acid Benzoic acid Dimethylnitrosamine

454.45 252.32 360.44 352.77 225.21 350.59 466.57 360.46 362.47 290.83 272.37 267.00 202.55 314.45 288.40 291.26 263.21 304.35 311.86 289.55 130.08 182.17 257.34 153.14 253.30 198.27 209.25 194.20 266.34 148.16 146.15 180.16 191.28 213.24 123.11 123.11 152.14 60.10 234.34 134.18 170.21 138.12 122.10 74.08

1.85 6.13 1.08 2.18 1.56 4.96 2.27 1.47 1.61 3.72 4.01 3.91 2.17 3.87 3.32 3.83 2.86 3.81 4.50 4.76 0.89 2.20 3.48 0.53 4.74 1.59 1.52 0.07 5.12 2.13 1.39 1.19 2.18 2.40 0.36 1.85 1.96 2.11 2.44 1.95 3.09 2.26 1.87 0.57

6.786 0.113 4.132 3.304 5.181 1.317 3.934 3.855 3.768 1.833 1.514 1.553 2.395 1.870 2.102 1.757 2.275 1.851 1.407 1.087 4.125 5.373 1.799 3.258 0.880 2.781 2.897 3.934 0.689 2.091 2.605 2.954 2.319 2.297 3.195 2.137 2.236 4.565 2.397 2.134 1.544 1.938 2.117 3.557

1.63E07 1.30E+00 7.37E05 4.96E04 6.57E06 4.82E02 1.16E04 1.39E04 1.70E04 1.47E02 3.06E02 2.80E02 4.02E03 1.35E02 7.90E03 1.75E02 5.30E03 1.41E02 3.91E02 8.19E02 7.48E05 4.22E06 1.59E02 5.52E04 1.32E01 1.66E03 1.27E03 1.16E04 2.04E01 8.09E03 2.48E03 1.11E03 4.79E03 5.04E03 6.37E04 7.28E03 5.80E03 2.72E05 4.00E03 7.34E03 2.85E02 1.15E02 7.63E03 2.77E04

0.17 1.14E06 0.05 8.66E03 1.62 7.50E04 0.33 0.28 0.39 7.31E03 3.60E03 4.62E03 0.037 1.17E02 2.46E02 1.49E02 5.50E02 5.29E02 2.01E02 1.00E02 11.07 220 6.17E02 2.14 1.10E02 0.99 1.86 20.8 1.40E02 0.49 1.88 4.42 1.62 2.04 17.8 1.95 2.53 550 4.07 2.36 0.69 2.09 3.44 98.5

2.78E08 1.48E06 3.68E06 4.29E06 1.06E05 3.61E05 3.83E05 3.90E05 6.64E05 1.07E04 1.10E04 1.29E04 1.49E04 1.58E04 1.94E04 2.60E04 2.92E04 7.44E04 7.86E04 8.19E04 8.28E04 9.29E04 9.79E04 1.18E03 1.45E03 1.64E03 2.35E03 2.42E03 2.86E03 3.97E03 4.67E03 4.91E03 7.76E03 1.03E02 1.13E02 1.42E02 1.47E02 1.50E02 1.63E02 1.73E02 1.97E02 2.41E02 2.62E02 2.73E02

Jmax (lg/cm2/h) 0.00003 0.00148 0.00368 0.00429 0.0107 0.0361 0.0383 0.0390 0.0664 0.107 0.110 0.129 0.149 0.158 0.194 0.260 0.292 0.744 0.786 0.819 0.828 0.929 0.979 1.18 1.45 1.64 2.35 2.42 2.86 3.97 4.67 4.91 7.76 10.3 11.3 14.2 14.7 15.0 16.3 17.3 19. 7 24.1 26.2 27.3

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Chemical

2547

MW, molecular weight; Log P, log of the octanol:water partition coefficient; log Kp, permeability coefficient; Csat, saturation concentration in water; Jmax, maximum flux.

Benzene Nicotinate hexyl Phenoxyethanol Nicotinate methyl Butoxyethanol Safrole Chloroform Benzyl alcohol Nicotinate ethyl Methoxypropan-2-ol Ethoxyethanol Nicotinate butyl Dimethylamine Phenol Catechol Dimethylethylamine Nicotine

78.12 207.27 138.17 137.14 118.18 162.19 119.38 108.13 151.17 90.12 90.12 129.22 45.10 94.11 110.11 73.14 162.23

2.13 3.51 1.16 0.83 0.83 3.45 1.97 1.10 1.32 0.49 0.32 2.27 0.38 1.46 0.88 0.70 1.17

1.664 1.472 2.719 2.947 2.832 1.240 2.030 2.579 2.685 3.598 3.477 1.877 3.245 2.237 2.747 2.649 2.859

2.17E02 3.37E02 1.91E03 1.13E03 1.47E03 5.75E02 9.34E03 2.64E03 2.06E03 2.52E04 3.33E04 1.33E02 5.68E04 5.78E03 1.79E03 2.24E03 1.38E03

1.79 1.43 26.9 47.6 44.9 1.24 8.07 43.1 56 470 530 18.35 520 94.1 460 460 1000

3.88E02 4.82E02 5.13E02 5.37E02 6.61E02 7.13E02 7.53E02 1.14E01 1.16E01 1.19E01 1.76E01 2.44E01 2.95E01 5.44E01 8.23E01 1.03E+00 1.38E+00

38.8 48.2 51.3 53.7 66.1 71.3 75.3 114 116 119 176 244 295 544 823 1030 1382

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Fig. 5. Predicted maximum flux of chemicals in the EDETOX database (EDETOX, 2004), that have been studied in vitro using human skin, as a function of molecular weight.

Fig. 6. Predicted maximum flux of chemicals in the EDETOX database (EDETOX, 2004), that have been studied in vitro using human skin, as a function of octanol–water partition coefficient (P). ‘‘Contours’’ of molecular weight are illustrated to visualize the concomitant impact of penetrant size on absorption.

permeability coefficients will continue to increase with increasing lipophilicity, the Cleek and Bunge (1993) correction (Eq. (6)) takes into account the fact that these progressively hydrophobic compounds have ever-diminishing water solubilities. As a result, when Jmax is calculated, the resulting value becomes a reflection of the fact that such lipophilic molecules cannot ‘escape’ from the stratum corneum and diffuse further through the viable skin. Thus, the global pattern observed (albeit with considerable, apparent scatter caused by the different MWs of the compounds considered – as shown on the graph) is that Jmax for very water-soluble compounds is small (low log P); the maxi-

2548

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

mum flux then increases with increasing lipophilicity towards the highest values at around log P  2–2.5. Can the modeling described above be used to estimate absorption for a specific exposure to be used in the context of the TTC approach? From the maximum flux (Jmax in lg/ cm2/h), the maximum potential absorption (Qabs lg) of a chemical following topical exposure is simply calculated from the following expression: Qabs ¼ J max  A  texp  DS

ð8Þ

2

where A (cm ) is the area of skin contact with the formulation, texp (h) is the duration of exposure, and DS is the degree of saturation of the chemical in the vehicle. For a saturated solution, a suspension, or for a neat chemical, DS = 1; for more typical scenarios, in which the chemical is far from being saturated in the formulation, DS is less than 1. Eq. (8) reflects a number of key features for the determination of the absorbability of a chemical. First of all, the maximum flux calculation draws upon the physicochemical parameters that define the molecule’s ability to cross the skin barrier – as discussed above, this property (assuming a vehicle that does not change the membrane) is independent of the formulation. However, the nature of the formulation is important under non-saturated conditions, and this is reflected in the degree of saturation (DS). DS is not a concentration and is defined as DS ¼

C vehicle C sat vehicle

ð9Þ

that is, DS is the ratio of the actual concentration of the chemical in the vehicle to its saturation solubility in that vehicle. It follows that a compound may be present at the same concentration in two formulations but that its DS in the different vehicles may be very different. Therefore, the amount (or the percentage of the applied dose) of a particular chemical that is absorbed may differ widely from formulation to formulation. In consequence, it is totally inappropriate to assign a single value to the potential for a chemical to be absorbed under use conditions, without taking into consideration the nature of the vehicle and the level at which the chemical is present therein. Unfortunately, the DS of an active in a formulation is usually not known. Nevertheless, the calculation of Jmax allows chemicals to be classified in terms of their inherent maximum absorbability across the skin, as a function of their physicochemical properties as indicated in Table 3. An additional question is whether there is a MW beyond which percutaneous absorption can generally be considered negligible. It is clear that absorption decreases with increase in MW, and it is reasonable to assume that larger molecules are less well-absorbed than small ones. It is possible to calculate the maximum percutaneous flux for a chemical of relatively high MW as a function of it lipophilicity and water solubility. Predicted fluxes for a hypothetical set of compounds with MW = 500 Da, and with

Table 3 Classification of chemicals (on the basis of their physicochemical properties) in terms of their potential to be absorbed across the skin Jmax (lg/cm2/h)

MW (Da)

log P

Category

Jmax = 0

Non-reactive chemicals > 1000 Da > 300  200–300  150–250 60–200 <150

Any

Negligible

< 1 or > 5 > 2.0, 2.5  1.0–2.0  0.5–3.5 0.5–2.0

Low Medium low Medium high High High

Jmax < 0.1 0.1 100

Jmax, maximum flux; MW, molecular weight; log P, log of the octanol:water partition coefficient.

Fig. 7. Predicted maximum flux of chemicals with MW 500 and varying water solubilities (Cs) and log P values. Two example chemicals (having MW close to 500) from the EDETOX database (EDETOX, 2004) are included in the figure: T2 (log P = 2.27, Cs = 0.33 mg/mL) and methotrexate (log P = 1.85, Cs = 0.17 mg/mL).

varying log P and saturation solubilities in water, are depicted in Fig. 7; two examples from the EDETOX database table (methotrexate and T2 toxin) are included for comparison. It is clear that fluxes above 0.1 lg/cm2/h are predicted only for a combination of physicochemical characteristics unlikely to be found in the same molecule – that is, high MW, high lipophilicity and high water solubility. As an example T2 toxin is quite lipophilic (log P = 2.27) and has a reasonable water solubility (0.33 mg/mL), yet its predicted maximum flux is only 0.04 lg/cm2/h; that is, comfortably in the ‘‘low’’ category of skin absorption potential (Table 3). It can be safely concluded, therefore, that the percentage dose absorbed of cosmetic ingredients with MW > 500 will be negligible and certainly less than 10%. Therefore as discussed before, the polymeric fraction above 1000 Da and other non-reactive chemicals above 1000 Da should be considered not to be bioavailable via the oral, as well as the dermal routes (see paragraph 4; EFSA, 2005) However one should always bear in mind that the potential risk associated with the weak permeation of a highly toxic compound may be similar to that of a well-absorbed but weakly toxic substance. The experimental data for a broad range of 15 cosmetic ingredients (Table 1) indicated that not more than 8% of

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

2549

Table 4 Proposed default adjustment factors for the % dose absorbed of cosmetic ingredients across the skin Jmax (lg/cm2/h)

Default% dose absorbed per 24 h

Non-reactive chemicals with MW > 1000 Jmax < 0.1 0.1 10

Negligible 10 40 80

Jmax, maximum flux.

the applied dose was absorbed percutaneously in 24 h. When experimental data on absorption are available these data should be used as the exposure assessment for comparison with the TTC value. If however such data are not available, it is proposed that conservative default adjustment factors be assigned to the categories of permeant identified in Table 3. Specifically the suggested dose absorbed default adjustment factors are given in Table 4. In summary, it is proposed that the most appropriate method to estimate the systemic exposure over 24 h to a cosmetic ingredient, following a single application, should be based on calculation of Jmax according to Eqs. (5)–(7) and assignment of the default % absorption per day according to Table 4. It should be emphasized that this approach involves several worst-case assumptions: (i) that the compound is applied at its saturation concentration in the formulation (while cosmetic ingredients will generally be present at much lower levels so that the flux will be less than Jmax); (ii) that no depletion of the chemical within the formulation occurs during the period of exposure (texp) (whereas in fact cosmetic ingredients are usually applied in finite amounts which do not sustain a constant concentration gradient (driving force) for transfer across the skin) (see Fig. 8); (iii) that the formulation does not affect the characteristics of the skin barrier; (iv) that, by using the maximal flux over the entire exposure time, the lower flux during diffusional lag time is ignored. 4.4. Additional default adjustment factors for rinse-off cosmetic products Rinse-off cosmetics are products that remain in contact with human skin only for a limited time (<1 h) and are subsequently rinsed off. They include shampoos, shower gels, hair conditioners, hair dyes and hair bleaching agents. Current EU Guidelines of cosmetic safety evaluation propose a default retention factor of 0.01 or 0.1 (1% or 10%) for rinse-off products (SCCNFP, 2003). This factor will be relevant to the safety evaluation of ingredients or

Fig. 8. The theoretical calculations (solid line) assume that the chemical’s flux across the skin is constant during the entire period of observation (in this case, 24 h). These conditions are approached only occasionally in ‘real-life’ (dashed line) – for example, when a transdermal patch containing an excess amount of drug is administered for a time insufficient to cause significant depletion in the reservoir; however, in this case, the maximum flux is reached only after a certain lag-time necessary for the steady-state, linear concentration gradient to be developed across the skin. In-use applications of cosmetics, on the other hand, typically employ a finite ‘dose’ of the active, such that its concentration at the skin surface depletes appreciably, and resulting in a flux versus time profile (dotted line) which increases to a maximum before tailing off as more and more of the compound is absorbed. It follows that, in the case that the % dose absorbed of a cosmetic ingredient in 24 h is 80%, there would remain only 20% of the applied substance at the skin surface one day post-application of the product.

their impurities present in cosmetic rinse-off products (shampoos, shower gels and hair dyes) using the TTC approach. 4.5. Default adjustment factors for intermittent use of cosmetic products resulting in intermittent human exposure The TTC values shown in Figs. 1 and 2 were derived from chronic animal studies and therefore allow for exposure every day throughout life. Intermittently used cosmetics are products that are used in intervals of > 1 week and include products, such as self-tanning agents, removers of body hair, hair dyes, permanent waiving, hair straightening and bleaching agents. Some cosmetic products, such as oxidative or direct hair dyes, produce a consumer exposure at intervals of 2–3 (e.g. direct hair dyes) to 6–8 (e.g. oxidative hair dyes) weeks, respectively. Although the potential exposure per event may be the same as for daily-used or intermittently-used products, the time-averaged (e.g. mean annual) consumer exposure from intermittent use of cosmetic products will obviously be proportionately lower than that from a daily used product. Consequently, it is proposed to take into account intermittent use (time interval >7days) of relevant cosmetic end products by the use of adjustment factors. It is well recognised from animal experiments that there are 3–10-fold differences in NOAEL values between acute and sub-chronic studies and between sub-chronic and chronic studies (Dourson and Stara, 1983; Dourson et al., 1996 and Vermeire et al., 1999). Additional uncertainty

2550

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

factors of 3–10-fold have been used in the risk assessment of environmental contaminants for which there are subchronic but no chronic toxicity data (IPCS, 1994, 1999). Therefore similar adjustment factors could be applied to cosmetic ingredients where the pattern of exposure is intermittent rather than daily, and when exposure on the day of use is compared with TTC values derived from chronic daily treatment. It is therefore proposed by the expert group that the estimated intake be decreased by default adjustment factors of 3-fold for ingredients used only once per week and 10-fold for ingredients used less frequently. However, before such a correction factor is applied it is important that a literature search is undertaken on the ingredient and structural analogues to compare the acute toxic potential and its potency following single and multiple dosages. Values lower or higher than the default correction factor should be used if the available data support a different value.

mates to a cosmetic ingredient from different product formulations will be far more complex than is the case for food and a probabilistic approach will almost certainly be needed. It is not the purpose of the present paper to set out optimal methods for the estimation of total (aggregate) exposure to cosmetic ingredients. When a safety assessor considers applying the TTC concept within the safety evaluation of a cosmetic ingredient or its impurities, the risk assessment should take into account the maximal concentration of the ingredient in the final cosmetic product as well as the type and use conditions of the cosmetic end product. Some of the well established principles that have been developed for the measurement of total consumer exposure to food chemical intake could be adapted to exposure estimates for cosmetic ingredients.

4.6. Aggregate (total) exposure to the cosmetic ingredient

Traditionally, hazard identification and characterisation, exposure assessment, and risk characterisation have focused on single chemicals. Given the reality of concurrent or sequential exposure to multiple agents, insight into the health consequences of mixtures or combined exposures is needed to be able to judge whether current approaches to risk assessment offer adequate protection to humans. Whether a mixture or combined exposure constitutes a safety concern mainly depends on two factors, one, the mode of combined action (similar or dissimilar joint action, or synergistic or antagonistic interaction) and two, the margin of exposure (the margin between the actual exposure level and the lowest-observed-toxic-effect level) of the individual chemicals in the mixture or combined exposure (Cassee et al., 1998, 1999). One of the major lessons learned from two decades of toxicological research on chemical mixtures (Ito et al., 1995a,b; Feron and Groten, 2002; Groten et al., 2004a), is that, in general, exposure to mixtures of chemicals at exposure levels that are non-toxic for each individual chemical in the mixture does not result in a health concern. However, there are exceptions to this rule; for instance, when a mixture consists of chemicals with a similar mode of action (and the same target organ) dose addition is to be expected, and this should be taken into account in the safety evaluation of such a mixture (Jonker et al., 1996; Groten et al., 2004a). Although synergistic interactions can never be fully excluded on theoretical grounds, they are likely to occur only when there is a biologically active dose level of at least one of the chemicals involved. Therefore, synergism is not relevant to the exposure levels at which the TTC approach would result in a conclusion that there were no safety concerns. Given the very low levels of combined dermal exposure to similarly acting compounds from topical use of cosmetics and the low levels of fragrances used in cosmetics (with

A key component of the TTC approach is an accurate assessment of exposure of chemicals which may be present in multiple sources. The same chemical in different brands within a beauty care category would need to be considered alongside the exposure to the same chemical from other cosmetic categories. As with any exposure assessment, the need for estimates of aggregate exposure should be considered. The challenges and methods for conducting these analyses would be no different for TTC than for risk assessments based on chemical-specific data. Current cosmetic exposure parameters (SCCNFP/ 06901/03) in the EU are based on industry’s surveys conducted 15 years ago. More recent estimates of exposure in EU consumers have been presented by industry to the SCCNFP (COLIPA, 2005, unpublished data). A key principle is that exposure estimates should always begin with a simple and basic method (Kroes et al., 2002) and should increase in complexity only in so far as is necessary to establish a reasonable basis for determining if exposure is more than or less than the relevant TTC value. Many of the methods for collecting food intake data, such as the food frequency questionnaire, could readily be applied to cosmetic ingredients. There are however some unique challenges to the estimation of exposure to cosmetic ingredients. In the case of food, it is assumed that almost all ingredients are fully absorbed in the small intestine, unless specifically intended not to be so absorbed (mineral oils, sucrose polyesters and polyols for example). Thus the total exposure of a target food chemical does not usually take account what are normally minor differences in absorption from different food matrices. However, in the case of cosmetic ingredients, the rate of dermal transfer is influenced greatly by product formulation (see Section 4.3). Establishing total exposure esti-

4.7. Simultaneous dermal exposure to different cosmetic ingredients

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

the exception of perfumes), combined exposure to cosmetic ingredients may be of health concern in exceptional cases only. If such an exceptional case emerges, its potential health risk can be assessed using methods for risk characterisation of combined exposures that have been developed and are readily available (see reviews by Jonker et al., 2004; Feron et al., 2004; Groten et al., 2004b). 5. Procedure to apply the TTC approach to the safety evaluation of cosmetic ingredients The analyses given above show that it is scientifically justified to use the TTC approach, and the database underlying the TTC values established for food chemicals, for the safety evaluation of cosmetic ingredients. The TTC values as described earlier (Munro et al., 1996 and Kroes et al., 2004) are appropriate for the safety evaluation of systemic exposures resulting from the use of cosmetic ingredients and products. Proteins, heavy metals and substances with specific structural alerts of concern, which were excluded in the decision tree developed by Kroes et al. (2004), should also be excluded if the TTC approach is used for cosmetic ingredients. In addition (new) chemicals that may have or are suspected to have pharmacological properties should be excluded for application of the TTC. The TTC values for systemic exposure to cosmetic ingredients and the use-related adjustment factors are to be regarded as provisional and could be subject to refinement when new data are developed on hazard or exposure (i.e. use pattern of cosmetic products, percutaneous absorption). Although theoretically the TTC approach could also be applied for topical (local) effects, TTC values for local topical effects have not been developed, and at present the databases on substances producing local (topical) effects are too limited to be used as a basis for the derivation of valid TTC values. In application of the TTC approach, appropriate exposure assessment is of prime importance. The Expert Group suggests following the methodology as described in the SCCNFP ‘‘Notes of Guidance for the Testing of Cosmetic Ingredients and their Safety Evaluation’’ (SCCNFP, 2003). Depending on the use of the ingredients or products (e.g. cosmetics producing human oral exposure, cosmetics used under occlusion, cosmetics used without occlusion, rinseoff products) and the chemical characteristics of the compound under evaluation, default adjustment factors are suggested for percutaneous absorption and intermittent exposure. It is proposed that these default adjustment factors are incorporated into the exposure assessment models as described by the SCCNFP. The results of the exposure assessment should then be used in the TTC decision tree shown in Figs. 1 and 2. In this preliminary risk assessment (see Fig. 3) all available data regarding the substance evaluated should be taken into consideration.

2551

The following steps are suggested for application of the TTC approach to cosmetic ingredients and impurities: 1. Define product type, its intended use and related skin surface area involved. 2. Define concentration of ingredient in the product. 3. Estimate external exposure per day (SCCNFP, 2003; EPA, 1997). 4. Estimate skin absorption of the ingredient based on its physical and chemical characteristics (Section 4.3). 5. If a rinse off product apply retention factor (Section 4.4). 6. Establish use pattern: e.g. daily or intermittent use, if the latter is the case apply the default factor related to the use interval (Section 4.5). 7. Calculate adjusted internal exposure per person per day (this will be the long-term average internal dosage for a 60 kg person). 8. Where relevant, calculate total (aggregate) exposure when several cosmetic products contain this target ingredient (Section 4.6). 9. Use this average aggregate internal dosage in the TTC decision tree (Figs. 1 and 2) (Note – the resulting assessment will relate to systemic but not to local effects). The decision tree comprises a series of steps, each one framed as a question, to which the answer, either ‘Yes’ or ‘No’, will carry the assessor through to the next step. The questions relate to whether the ingredient is suitable for assessment via the TTC approach, the presence of absence of structural alerts for genotoxicity, and, depending on its structure, how the level of exposure relates to the relevant human exposure threshold. For any ingredient taken through the decision tree process, one of two conclusions will be drawn: either, the substance is predicted not to be a health concern, or, further risk assessment is necessary using compoundspecific toxicity data. Conflict of interest statement The paper was developed during meetings that were supported financially by Colipa, Avenue Herrman Debroux 15A, Brussels, Belgium. Members of the Expert Group were reimbursed for their expenses, and received a per diem payment for time involved on the project. In addition, R.H.G. acts as a consultant to L’Oreal, and TNO undertake contract research for cosmetic companies although J.J.M. vdS was not individually involved in these. Appendix See Appendix Tables 1–3.

2552

Appendix Table 1 Metabolic considerations for Class III compounds in the Munro database with NOAEL values of 1 mg/kg bw/day or less (minus OPs) Munro no.

Name

NOAEL

Potential for presystemic metabolism orally

Suitability of oral NOAELa

168 448

Dieldrin Zeranol

0.005 0.02

+++ +?

Avermectin B1

0.03

327

Patulin

0.04

423

0.04

422

Trenbolone hydroxide, 17-alphaTrenbolone acetate

0.044

385

Sodium fluoroacetate

0.05

106 235 399

Chlordane Hexachlorobenzene Terbutryn

0.055 0.08 0.1

292 194

Mirex Diquat

0.17 0.19

Acrylamide

0.2

173

Dihydroavermectin-Bla 22,23

0.2

293

Molinate

0.2

232 233

Heptachlor Heptachlor epoxide

0.25 0.25

Aldicarb

0.3

Organochlorine – resistant to metabolism – negligible first-pass metabolism predicted Rapidly absorbed from the intestine; undergoes metabolism via oxidation and conjugation with glucuronic acid; short half-life of parent compound suggests the possibility of extensive first-pass metabolism1 Few toxicokinetic data available; the structural analogue ivermectin has a long elimination half-life in various animal species and humans indicating slow metabolism and therefore limited first-pass metabolism, however, P-glycoprotein in the intestine inhibits the absorption of ivermectin and abermectin and limits systemic toxicity; secretion of the parent compound from the general circulation into the gut lumen is a major route of elimination2 Rapidly absorbed across the gut wall; reacts with glutathione in vitro and is retained in red blood cells; toxicity in the intestine and liver is likely to be greater after oral dosage3 Few toxicokinetic data available; predicted to be conjugated with glucuronic acid so that some first-pass metabolism is possible after oral dosage Undergoes rapid and complete hydrolysis after intravenous dosage so that parent compound would not be detected after either oral or topical dosage; subsequent oxidation in the liver yields a range of metabolites including reactive compounds, which may be linked to the toxicity; oral toxicity predicted to be similar to or exceed dermal toxicity4 Parent compound detected in blood after oral dosage; evidence of some conversion to inorganic fluoride, but renal excretion is probably the main route of elimination; has a very short elimination half-life; a route difference in toxicity is unlikely5 Organochlorine-resistant to metabolism – negligible first-pass metabolism predicted Organochlorine – resistant to metabolism – negligible first-pass metabolism predicted Undergoes numerous pathways of metabolism including S-demethylation, conversion of the methylthio group to a hydroxy group, N-deethylation, oxidation and conjugation with glucuronic acid; the rate of elimination and extent of first-pass metabolism have not been defined6 Organochlorine – resistant to metabolism – negligible first-pass metabolism predicted Parent compound detected in human tissues after an overdose; toxicity is produced in the intestine and in the liver via redox cycling and oxidative damage; highly polar compound unlikely to undergo any first-pass metabolism; oral toxicity predicted to exceed dermal toxicity7 Rapidly and extensively absorbed; metabolized to glycidamide, the reactive metabolite; oral bioavailability is 60–90%, but there is greater bioactivation and toxicity after oral dosage; conjugated with glutathione to a similar extent after oral and dermal application; oral toxicity predicted to be similar to or exceed dermal toxicity8 Few toxicokinetic data available; the structural analogue ivermectin has a long elimination half-life in various animal species and humans indicating slow metabolism and therefore limited first-pass metabolism, however P-glycprotein in the intestine inhibits the absorption of ivermectin and abermectin and limits systemic toxicity; secretion of the parent compound from the general circulation into the gut lumen is a major route of elimination2 Oxidized in the liver to ring hydroxyl and sulphoxide metabolites; the sulphoxide is considered to be an active metabolite causing testicular toxicity and reacts with glutathione to form a conjugate; oral toxicity predicted to be similar to or exceed dermal toxicity9 Organochlorine – resistant to metabolism – negligible first-pass metabolism predicted Organochlorine – epoxide group could undergo conjugation with glutathione via a transferase; the similarity of the NOAEL to that for heptachlor suggests similar systemic bioavailability and therefore negligible firstpass metabolism is predicted Undergoes hepatic first-pass metabolism to the toxic sulphoxide metabolite; oral toxicity predicted to be similar to or exceed dermal toxicity10

62

35

+++ ++? +++

+++?

+++ +++ ?

+++ +++

+++

+?

+++

+++ +++

+++

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

30

+?

0.3

241

0.3

320

Hexahydro-1,3,5-trinitro-1,3,5triazine Oxyfluorfen

0.3

436

Tridiphane

0.33

401 344 379

Tetrachlorobenzene 1,2,4,5Photodieldrin Rotenone

0.34 0.35 0.38

174

Dihydroavermectin-Bl b, 22,23-

0.4

185

Dinitrobenzene, m-

0.4

228

Glufosinate-ammonium

0.4

137 138

Cyhalothrin Cypermethrin

0.5 0.5

317

Oxadiazon

0.5

383

Simazine

0.52

375 36 101 247

Reserpine Aldicarb sulfone Carbon tetrachloride Hydroxypropyl methanethiolsulfonate,-2Tralomethrin Assure Albendazole

0.58 0.6 0.71 0.75 0.75 0.9 1

Amitraz

1

50 78 100

Amphetamine sulfate, dlBiphenthrin Carbofuran

1 1 1

102

Carbosulfan

1

123

Chlorpromazine

1

421 59 33

48

Few toxicokinetic data available; predicted to undergo slow metabolism as the 2-chloro group would block +++ conjugation of the OH group and the 4-chloro would block oxidation Produces neurotoxicity and local gastrointestinal effects after oral dosage; extent of first-pass metabolism ? undefined?11 Causes experimental porphyria via an action on protoporphyrinogen oxidase activity in liver and kidney; oral +++ toxicity predicted to be similar to or exceed dermal toxicity12 Metabolized by microsomal epoxide hydrolase and glutathione conjugation which are induced on chronic +++ treatment; epoxide hydrolase is a low capacity pathway and the low Km value for tridiphane indicates limited first-pass metabolism13 Organochlorine – resistant to metabolism – negligible first-pass metabolism predicted +++ A metabolite and degradation product of dieldrin; chlorination would minimize first-pass metabolism +++ CNS effects produced by parent compound, not metabolites; metabolized by CYP3A4 and therefore extensive + first-pass metabolism is possible14 Few toxicokinetic data available; the structural analogue ivermectin has a long elimination half-life in various +? animal species and humans indicating slow metabolism and therefore limited first-pass metabolism, however P-glycoprotein in the intestine inhibits the absorption of ivermectin and abermectin and limits systemic toxicity; secretion of the parent compound from the general circulation into the gut lumen is a major route of elimination2 Neurotoxicity depends on the maintenance of high concentrations for many hours; molecule has no sites for +++ extensive first-pass metabolism; hepatic nitro-reduction would yield a potentially toxic metabolite15 Long half-life and eliminated largely unchanged by renal excretion which indicate a low potential for first-pass +++ metabolism; similar toxicity by both oral and dermal routes16 Eliminated by metabolism, but the extent of first-pass metabolism is not known17 ? Toxicity is due to parent compound as hydrolysis result in detoxication; there is negligible absorption after +++? topical application to humans; the extent of first-pass metabolism after oral dosage is not known18 Causes experimental porphyria via an action on protoporphyrinogen oxidase activity in liver and is a +++ peroxisome inducer; oral toxicity predicted to be similar to or exceed dermal toxicity19 Undergoes cytochrome P450 mediated N-monodealkylation and isopropylhydroxylation; effects on female +++ reproduction mediated by parent compound and metabolites; any first-pass metabolism would have little influence on toxicity20 Few toxicokinetic data available ? Sulphones are metabolically stable and therefore negligible first-pass metabolism is predicted +++ Hepatotoxicity linked to local metabolism (classic example); oral toxicity is predicted to exceed dermal toxicity +++ No toxicokinetic data are available; highly polar compound therefore low hepatic first-pass metabolism would +++? be expected, but the fate in the gut lumen is more difficult to predict Neurotoxicity arises from parent compound and its metabolite deltamethrin; few data available ? No toxicokinetic data identified ? Undergoes hepatic metabolism to the sulphoxide (active metabolite) and slowly to the sulphone (inactive +++ metabolite); very extensive first-pass metabolism; toxicity probably related to the sulphoxide metabolite; oral toxicity predicted to be similar to or exceed dermal toxicity21 Rapidly metabolized to an active metabolite and other products; short half-life; oral toxicity is predicted to be +++? similar to or to exceed dermal toxicity22 Metabolized by aromatic oxidation and deamination; high oral bioavailability23 +++ No relevant toxicokinetic data identified ? Carbofuran toxicity correlates with parent compound; it has a short half-life and is rapidly oxidized to +? hydroxyl metabolites; first-pass detoxication after oral dosage is possible24 Carbosulfan shows slightly lower acute toxicity after oral compared with intravenous administration +++ indicating possible first-pass detoxication (but the data were for peak effects); effects correlate better with the plasma levels of its metabolite carbofuran, indicating that any first-pass metabolism would represent a bioactivation step25 Oral bioavailability in humans is 20–40%26 ++ (continued on next page)

2553

Dichlorophenol, 2,4-

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

162

Name

NOAEL

Potential for presystemic metabolism orally

Suitability of oral NOAELa

151 163

Dichloro-2-propanol, 1,3Dichlorophenoxyacetic acid, 2,4-

1 1

+++ +++

197 212 224 231 234 236 239 240 311

Diuron Express Fluvalinate Harmony Hexabromobenzene Hexachlorobutadiene Hexachloroethane Hexachlorophene Ochratoxin A

1 1 1 1 1 1 1 1 1

314

Olaquindox

1

325 415 427

Paraquat Thiobencarb Trichloroacetonitrile

1 1 1

442

Trisulfuron

1

Metabolized in liver the active chemical species; oral toxicity is predicted to exceed dermal toxicity27 Produces primarily hepatic and muscle tissue damage; alters intestinal and hepatic xenobiotic metabolizing enzymes; metabolized by conjugation with glycine and glucuronic acid; oral toxicity predicted to be similar to or exceed dermal toxicity28 No relevant toxicokinetic data identified No toxicokinetic data identified No relevant toxicokinetic data identified No relevant toxicokinetic data identified Organochlorine – resistant to metabolism - negligible first-pass metabolism Organochlorine-low hepatic metabolism-negligible first-pass metabolism Organochlorine – resistant to metabolism - negligible first-pass metabolism Organochlorine – resistant to metabolism - negligible first-pass metabolism Eliminated largely by oxidation, but with a very long half-life indicating very slow metabolism and therefore negligible first-pass metabolism29 Few toxicokinetic data available; short half-life indicating rapid metabolism and therefore possible first-pass metabolism30 Highly polar compound – no oral first-pass metabolism Induces toxic neuropathies, probably via the parent compound; no toxicokinetic data identified31 Metabolized to a limited extent to cyanide and excreted in the urine as thiocyanate; parent compound probably the active species; high toxicity after oral dosage, but influence of route on metabolism on toxicity is undefined32 No relevant toxicokinetic data identified

? ? ? ? +++ +++ +++ +++ +++ +? +++ ? ?

?

Munro Number–reference number given in the paper of Munro et al., 1996. References 1 Bories, G. and Suarez, A.F., 1989. Profiling of free and conjugated [3H]zeranol metabolites in pig plasma. Journal of Chromatography 489, 191–197; Heitzman, R.J., 1983. The absorption, distribution and excretion of anabolic agents. Journal of Animal Science 57, 233–238; Migdalof, B.H., Dugger, H.A., Heider, J.G., Coombs, R.A. and Terry, M.K., 1983. Biotransformation of zeranol: disposition and metabolism in the female rat, rabbit, dog, monkey and man. Xenobiotica 13, 209–221; Baldwin, R.S., Williams, R.D. and Terry, M.K., 1983. Zeranol: a review of the metabolism, toxicology, and analytical methods for detection of tissue residues. Regulatory Toxicology and Pharmacology 3, 9–25. 2 Gokbulut, C., Boyacioglu, M. and Karademir, U., 2005. Plasma pharmacokinetics and faecal excretion of ivermectin (Eqvalan paste) and doramectin (Dectomax, 1%) following oral administration in donkeys. Research in Veterinary Sciences 79, 233–238; Perez, R., Godoy, C., Palma, C., Cabezas, I., Munoz, L., Rubilar, L., Arboix, M. and Alvinerie, M., 2003. Plasma profiles of ivermectin in horses following oral or intramuscular administration. Journal of Veterinary Medicine A Physiology Pathology and Clinical Medicine 50, 297–302; Guzzo, C.A., Furtek, C.I., Porras, A.G., Chen, C., Tipping, R., Clineschmidt, C.M., Sciberras, D.G., Hsieh, J.Y. and Lasseter, K.C., 2002. Safety, tolerability, and pharmacokinetics of escalating high doses of ivermectin in healthy adult subjects. Journal of Clinical Pharmacology 42, 1122–1133; Laffont, C.M., Toutain, P.L., Alvinerie, M. and Bousquet-Melou, A., 2002. Intestinal secretion is a major route for parent ivermectin elimination in the rat. Drug Metabolism and Disposition 30, 626–630; Lankas, G.R., Cartwright, M.E. and Umbenhauer, D., 1997. P-glycoprotein deficiency in a subpopulation of CF-1 mice enhances avermectin-induced neurotoxicity. Toxicology and Applied Pharmacology 143, 357–365. 3 Rychlik, M., Kircher, F., Schusdziarra, V. and Lippl, F., 2004. Absorption of the mycotoxin patulin from the rat stomach. Food and Chemical Toxicology 42, 729–735; Mahfoud, R., Maresca, M., Garmy, N. and Fantini, J., 2002. The mycotoxin patulin alters the barrier function of the intestinal epithelium: mechanism of action of the toxin and protective effects of glutathione. Toxicology and Applied Pharmacology 181, 209–218; Dailey, R.E., Blaschka, A.M. and Brouwer, E.A., 1977. Absorption, distribution, and excretion of [14C]patulin by rats. Journal of Toxicology and Environmental Health 3, 479–489. 4 Evrard, P. and Maghuin-Rogister, G., 1988. In vitro metabolism of trenbolone: study of the formation of covalently bound residues. Food Additives and Contaminants 5, 59–65; Rico, A.G., 1983. Metabolism of endogenous and exogenous anabolic agents in cattle. Journal of Animal Science 57, 226–232; Pottier, J., Cousty, C., Heitzman, R.J. and Reynolds, I.P., 1981. Differences in the biotransformation of a 17 beta-hydroxylated steroid, trenbolone acetate, in rat and cow. Xenobiotica 11, 489–500. 5 Gooneratne, S.R., Eason, C.T., Dickson, C.J., Fitzgerald, H. and Wright, G., 1995. Persistence of sodium monofluoroacetate in rabbits and risk to non-target species. Human and Experimental Toxicology 14, 212–216; Eason, C.T., Gooneratne, R., Fitzgerald, H., Wright, G. and Frampton, C., 1994. Persistence of sodium monofluoroacetate in livestock animals and risk to humans. Human and Experimental Toxicology 13, 119–122. 6 Larsen, G.L., Bakke, J.E. and Feil, V.J., 1978. Metabolism of [14C] terbutryn (2-(t-butylamino)-4-(ethylamino)-6-(methylthio-s-triazine)) by rats and goat. Biomedical Mass Spectrometry 5, 382–390.

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Munro no.

2554

Appendix Table 1 (continued)

2555

(continued on next page)

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

7 Hantson, P., Wallemacq, P. and Mahieu, P., 2000. A case of fatal diquat poisoning: toxicokinetic data and autopsy findings. Journal of Toxicology and Clinical Toxicology 38, 149–152; Kurisaki, E. and Sato, H., 1979. Tissue distribution of paraquat and diquat after oral administration in rats. Forensic Science International 14, 165–170; Gupta, S., Husser, R.C., Geske, R.S., Welty, S.E. and Smith, C.V., 2000. Sex differences in diquat-induced hepatic necrosis and DNA fragmentation in Fischer 344 rats. Toxicological Sciences 54, 203–211; Madhu, C., Gregus, Z. and Klaassen, C.D., 1992. Marked interanimal differences in susceptibility of Sprague-Dawley rats to diquat-induced oxidative stress in the liver: correlation with hepatic uptake of diquat. Journal of Pharmacology and Experimental Therapeutics 263, 1003–1008; Spalding, D.J., Mitchell, J.R., Jaeschke, H. and Smith, C.V., 1989. Diquat hepatotoxicity in the Fischer-344 rat: the role of covalent binding to tissue proteins and lipids. Toxicology and Applied Pharmacology 101, 319–327; Kurisaki, E. and Sato, H., 1979. Tissue distribution of paraquat and diquat after oral administration in rats. Forensic Science International 14, 1654170. 8 Doerge, D.R., Young, J.F., McDaniel, L.P., Twaddle, N.C. and Churchwell, M.I., 2005. Toxicokinetics of acrylamide and glycidamide in B6C3F1 mice. Toxicology and Applied Pharmacology 202, 258–267; Sumner, S.C., Williams, C.C., Snyder, R.W., Krol, W.L., Asgharian, B. and Fennell, T.R., 2003. Acrylamide: a comparison of metabolism and hemoglobin adducts in rodents following dermal, intraperitoneal, oral, or inhalation exposure. Toxicological Sciences 75, 260–270. 9 Jewell, W.T. and Miller, M.G., 1999. Comparison of human and rat metabolism of molinate in liver microsomes and slices. Drug Metabolism and Disposition 27, 842–847; Jewell, W.T., Hess, R.A. and Miller, M.G., 1998. Testicular toxicity of molinate in the rat: metabolic activation via sulfoxidation. Toxicology and Applied Pharmacology 149, 159–166. 10 Montesissa, C., Huveneers, M.B., Hoogenboom, L.A., Amorena, M., De Liguoro, M. and Lucisano, A., 1994. The oxidative metabolism of aldicarb in pigs: in vivo-in vitro comparison. Drug Metabolism and Drug Interactions 11, 127–138; Pelekis, M. and Krishnan, K., 1997. Determination of the rate of aldicarb sulphoxidation in rat liver, kidney and lung microsomes. Xenobiotica 27, 1113–1120. 11 Etnier, E.L., 1989. Water quality criteria for hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX). Regulatory Toxicology and Pharmacology 9, 147–157. 12 Krijt, J., Stranska, P., Maruna, P., Vokurka, M. and Sanitrak, J., 1997. Herbicide-induced experimental variegate porphyria in mice: tissue porphyrinogen accumulation and response to porphyrogenic drugs. Canadian Journal of Physiology and Pharmacology 75, 1181–1187. 13 Magdalou, J. and Hammock, B.D., 1987. Metabolism of tridiphane (2-(3,5-dichlorophenyl)-2(2,2,2-trichloroethyl)oxirane) by hepatic epoxide hydrolases and glutathione S-transferases in mouse. Toxicology and Applied Pharmacology 91, 439–449. 14 Caboni, P., Sherer, T.B., Zhang, N., Taylor, G., Na, H.M., Greenamyre, J.T. and Casida, J.E., 2004. Rotenone, deguelin, their metabolites, and the rat model of Parkinson’s disease. Chemical Research in Toxicology 17, 1540–1548. 15 Xu, J., Nolan, C.C., Lister, T., Purcell, W.M. and Ray, D.E., 1999. Pharmacokinetic factors and concentration-time threshold in m-dinitrobenzene-induced neurotoxicity. Toxicology and Applied Pharmacology 161, 267–273; Hu, H.L., Bennett, N., Lamb, J.H., Ghersi-Egea, J.F., Schlosshauer, B. and Ray, D.E., 1997.Capacity of rat brain to metabolize m-dinitrobenzene: an in vitro study. Neurotoxicology 18, 363–370. 16 Ebert, E., Leist, K.H. and Mayer, D., 1990. Summary of safety evaluation toxicity studies of glufosinate ammonium. Food and Chemical Toxicology 28, 339–349; Hirose, Y., Kobayashi, M., Koyama, K., Kohda, Y., Tanaka, T., Honda, H., Hori, Y., Yoshida, K. and Kikuchi, M., 1999. A toxicokinetic analysis in a patient with acute glufosinate poisoning. Human and Experimental Toxicology 18, 305–308. 17 Chester, G., Sabapathy, N.N. and Woollen, B.H., 1992. Exposure and health assessment during application of lambda-cyhalothrin for malaria vector control in Pakistan. Bulletin of the World Health Organization 70, 615–619; Angerer, J. and Ritter, A., 1997. Determination of metabolites of pyrethroids in human urine using solid-phase extraction and gas chromatography-mass spectrometry. Journal of Chromatography B Biomedical Science Applications 695, 217–226. 18 Grajeda-Cota, P., Ramirez-Mares, M.V. and Gonzalez de Mejia, E., 2004. Vitamin C protects against in vitro cytotoxicity of cypermethrin in rat hepatocytes. Toxicology In Vitro 18:13–19; Luty, S., Latuszynska, J., Halliop, J., Tochman, A., Obuchowska, D., Przylepa, E. and Korczak, E., 1998. Toxicity of dermally applied alpha-cypermethrin in rats. Annals of Agricultural and Environmental Medicine 5:109–116; Cantalamessa, F. 1993. Acute toxicity of two pyrethroids, permethrin, and cypermethrin in neonatal and adult rats. Archives of Toxicology 67, 510–513; Woollen, B.H., Marsh, J.R., Laird, W.J. and Lesser, J.E., 1992. The metabolism of cypermethrin in man: differences in urinary metabolite profiles following oral and dermal administration. Xenobiotica 22, 983–991; Eadsforth, C.V., Bragt, P.C. and van Sittert, N.J., 1988. Human dose-excretion studies with pyrethroid insecticides cypermethrin and alphacypermethrin: relevance for biological monitoring. Xenobiotica 18, 603–614. 19 Richert, L., Price, S., Chesne, C., Maita, K. and Carmichael, N., 1996. Comparison of the induction of hepatic peroxisome proliferation by the herbicide oxadiazon in vivo in rats, mice, and dogs and in vitro in rat and human hepatocytes. Toxicology and Applied Pharmacology 141, 35–43; Krijt, J., van Holsteijn, I., Hassing, I., Vokurka, M. and Blaauboer, B.J., 1993. Effect of diphenyl ether herbicides and oxadiazon on porphyrin biosynthesis in mouse liver, rat primary hepatocyte culture and HepG2 cells. Archives of Toxicology 67, 255–261. 20 Hanioka, N., Jinno, H., Tanaka-Kagawa, T., Nishimura, T. and Ando, M., 1999. In vitro metabolism of simazine, atrazine and propazine by hepatic cytochrome 450 enzymes of rat, mouse and guinea pig, and oestrogenic activity of chlorotriazines and their main metabolites. Xenobiotica 29, 1213–1226; Hanioka, N., Jinno, H., Tanaka-Kagawa, T., Nishimura, T. and Ando, M., 1999. In vitro metabolism of chlorotriazines: characterization of simazine, atrazine, and propazine metabolism using liver microsomes from rats treated with various cytochrome P450 inducers. Toxicology and Applied Pharmacology 156, 195–205; Eldridge, J.C., Fleenor-Heyser, D.G., Extrom, P.C., Wetzel, L.T., Breckenridge, C.B., Gillis, J.H., Luempert, L.G. and Stevens, J.T., 1994. Short-term effects of chlorotriazines on estrus in female Sprague-Dawley and Fischer 344 rats. Journal of Toxicology and Environmental Health 43, 155–167. 21 Velik, J., Baliharova, V., Skalova, L., Szotakova, B., Wsol, V. and Lamka, J., 2005. Liver microsomal biotransformation of albendazole in deer, cattle, sheep and pig and some related wild breeds. Journal of Veterinary Pharmacology and Therapeutics 28, 377–384; Merino, G., Molina, A.J., Garcia, J.L., Pulido, M.M., Prieto, J.G. and Alvarez, A.I., 2003. Effect of clotrimazole on microsomal metabolism and pharmacokinetics of albendazole. Journal of Pharmacy and Pharmacology 55, 757–764; Mirfazaelian, A., Rouini, M.R. and Dadashzadeh, S., 2002. Dose dependent pharmacokinetics of albendazole in human. Biopharmaceutics and Drug Disposition 23, 379–383; Nagy, J., Schipper, H.G., Koopmans, R.P., Butter, J.J., Van Boxtel, C.J. and Kager, P.A., 2002. Effect of grapefruit juice or cimetidine coadministration on albendazole bioavailability. American Journal of Tropical Medicine Hygiene 66, 260–263; Dominguez, L., Fagiolino, P., Gordon, S. and Manta, E., 1995.

2556 R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Bioavailability comparison between albendazole and albendazole sulphoxide in rats and man. Farmacologie 50, 697–702; Rolin, S., Souhaili-el Amri, H., Batt, A.M., Levy, M., Bagrel, D. and Siest, G., 1898. Study of the in vitro bioactivation of albendazole in human liver microsomes and hepatoma cell lines. Cellular and Biological Toxicology 5, 1–14. 22 Pass, M.A. and Mogg, T.D., 1995. Pharmacokinetics and metabolism of amitraz in ponies and sheep. Journal of Veterinary Pharmacology and Therapeutics 18, 210–215; Knowles, C.O. and Benezet, H.J., 1981. Excretion balance, metabolic fate and tissue residues following treatment of rats with amitraz and N’-(2,4-dimethylphenyl)-N-methylformamidine. Journal of Environmental Sciences and Health B 16, 547–555. 23 de la Torre, R., Farre, M., Navarro, M., Pacifici, R., Zuccaro, P. and Pichini, S., 2004. Clinical pharmacokinetics of amfetamine and related substances: monitoring in conventional and nonconventional matrices. Clinical Pharmacokinetics 43, 157–185; Green, C.E., LeValley and S.E., Tyson, C.A., 1986. Comparison of amphetamine metabolism using isolated hepatocytes from five species including human. Journal of Pharmacology and Experimental Therapeutics 237, 931–936. 24 Ferguson, P.W., Dey, M.S., Jewell, S.A. and Krieger, R.I., 1984. Carbofuran metabolism and toxicity in the rat. Fundamental and Applied Toxicology 4, 14–21; Spittler, T.D. and Marafioti, R.A., 1983. Determination of carbofuran and its metabolites. Journal of Chromatography 255, 191–198. 25 Renzi, B.E. and Krieger, R.I., 1986. Sublethal acute toxicity of carbosulfan [2,3-dihydro-2,2-dimethyl-7-benzofuranyl(di-n- butylaminosulfenyl)(methyl)carbamate] in the rat after intravenous and oral exposures. Fundamental and Applied Toxicology 6, 7–15. 26 Koytchev, R., Alken, R.G., Kirkov, V., Neshev, G., Vagaday, M. and Kunter, U., 1994[Absolute bioavailability of chlorpromazine, promazine and promethazine] [Article in German] Arzneimittelforschung 44, 121–125. 27 Hammond, A.H., Garle, M.J., Sooriakumaran, P. and Fry, J.R., 2002. Modulation of hepatocyte thiol content by medium composition: implications for toxicity studies. Toxicology In Vitro 16, 259– 265; Hammond, A.H., Garle, M.J. and Fry, J.R., 1999. The nature of halogen substitution determines the mode of cytotoxicity of halopropanols. Toxicology and Applied Pharmacology 155, 287–291; Katoh, T., Haratake, J., Nakano, S., Kikuchi, M., Yoshikawa, M. and Arashidani, K., 1998. Dose-dependent effects of dichloropropanol on liver histology and lipid peroxidation in rats. Industrial Health 36, 318–323. 28 Paulino, C.A., Guerra, J.L., Oliveira, G.H. and Palermo-Neto, J., 1996. Acute, subchronic and chronic 2,4-dichlorophenoxyacetic acid (2,4-D) intoxication in rats. Veterinary and Human Toxicology 38, 348–352; Hietanen, E., Linnainmaa, K. and Vainio, H., 1983. Effects of phenoxyherbicides and glyphosate on the hepatic and intestinal biotransformation activities in the rat. Acta Pharmacologie et Toxicologie (Copenhagen) 53, 103–112; Knopp, D. and Schiller, F., 1992. Oral and dermal application of 2,4-dichlorophenoxyacetic acid sodium and dimethylamine salts to male rats: investigations on absorption and excretion as well as induction of hepatic mixed-function oxidase activities. Archives of Toxicology 66, 170–174. 29 Zepnik, H., Volkel, W. and Dekant, W., 2003. Toxicokinetics of the mycotoxin ochratoxin A in F 344 rats after oral administration. Toxicology and Applied Pharmacology 192, 36–44; Li, S., Marquardt, R.R., Frohlich, A.A., Vitti, T.G. and Crow, G., 1997. Pharmacokinetics of ochratoxin A and its metabolites in rats. Toxicology and Applied Pharmacology 145, 82–90; Hagelberg, S., Hult, K. and Fuchs, R., 1989. Toxicokinetics of ochratoxin A in several species and its plasma-binding properties. Journal of Applied Toxicology 9, 91–96. 30 Li, T., Qiao, G.L., Hu, G.Z., Meng, F.D., Qiu, Y.S., Zhang, X.Y., Guo, W.X., Yie, H.L., Li, S.F. and Li, S.Y., 1995. Comparative plasma and tissue pharmacokinetics and drug residue profiles of different chemotherapeutants in fowls and rabbits. Journal of Veterinary Pharmacology and Therapeutics 18, 260–273. 31 Pentyala, S.N., Chetty, C.S., Korlinara, G. and Pentyala, S., 1993. Permeability changes in the blood-brain barrier of neonate and adult rats after thiobencarb exposure. Veterinary and Human Toxicology 35, 509–511. 32 Lin, E.L., Daniel, F.B., Herren-Freund, S.L. and Pereira, M.A., 1986. Haloacetonitriles: metabolism, genotoxicity, and tumor-initiating activity. Environmental Health Perspectives 69, 67–71; Christ, S.A., Read, E.J., Stober, J.A. and Smith, M.K., 1996. Developmental effects of trichloroacetonitrile administered in corn oil to pregnant Long-Evans rats. Journal of Toxicology and Environmental Health 47, 233–247. a Suitability of oral NOAEL to be used for topical exposure +++ = negligible first-pass metabolism predicted or hepatotoxicity predicted to be higher after oral dosage; ++ limited first-pass metabolism predicted; +high first-pass metabolism predicted; ? = no data or conclusion uncertain.

Appendix Table 2 Metabolic considerations for Class II compounds in the Munro database with NOAEL values = or <10 mg/kg bw/day Name

27 2

Pyridine Allyl alcohol Propargyl alcohol Thujone Acrylic acid

5.3

26 28 1

16 6

Etretinate Caffeine

NOAEL

Potential for presystemic metabolism orally

Suitability of oral NOAELa

1 4.8

Undergoes N-oxidation and N-methylation; both reactions are of low capacity and negligible first pass metabolism is predicted1 Undergoes metabolic activation in the liver to a toxic metabolite; oral toxicity is predicted to exceed dermal toxicity2

+++ +++

5

Undergoes metabolic activation in the liver to a toxic metabolite; oral toxicity is predicted to exceed dermal toxicity3

+++

5

Undergoes cytochrome P450 mediated oxidation to inactive metabolites; the extent of first-pass metabolism and other toxicokinetic parameters have not been defined4 Undergoes rapid metabolism to carbon dioxide after oral dosage; produces forestomach oedema indicating that the parent compound may be the active form; the metabolic fate is similar after topical and oral administration; it is rapidly oxidized by most tissues, but the extent of firstpass metabolism and other toxicokinetic parameters have not been defined5 Metabolized to etretin and isoetretin; oral bioavailability is about 15–50%; topical toxicity has been shown to be less than oral toxicity6 Oxidized by CYP1A2 in liver; negligible first-pass metabolism after oral dosage7

?

6 10.1

++?

++ +++

Munro Number–reference number given in the paper of Munro et al., 1996. References 1 Damani, L.A. and Crooks, P.A., 1982. Oxidative metabolism of heterocyclic ring systems in The Metabolic Basis of Detoxication Edited by W. B. Jakoby. 2nd ed. pp. 69–89. Academic Press, New York. 2 Karas, M. and Chakrabarti, S.K., 2001. Influence of caffeine on allyl alcohol-induced hepatotoxicity in rats. I. In vivo study. Journal of Environmental Pathology, Toxicology and Oncology 20, 141– 154; Rikans, L.E. and Moore, D.R., 1987. Effect of age and sex on allyl alcohol hepatotoxicity in rats: role of liver alcohol and aldehyde dehydrogenase activities. Journal of Pharmacology and Experimental Therapeutics 243, 20–26; Penttila, K.E., Makinen, J. and Lindros, K.O., 1987. Allyl alcohol liver injury: suppression by ethanol and relation to transient glutathione depletion. Pharmacology and Toxicology 60, 340–344; Jaeschke, H., Kleinwaechter, C. and Wendel, A., 1987. The role of acrolein in allyl alcohol-induced lipid peroxidation and liver cell damage in mice. Biochemical Pharmacology 36, 51–57. 3 Dix, K.J., Coleman, D.P., Fossett, J.E., Gaudette, N.F., Stanley, A.P., Thomas, B.F. and Jeffcoat, A.R., 2001. Disposition of propargyl alcohol in rat and mouse after intravenous, oral, dermal and inhalation exposure. Xenobiotica 31, 357–375; Moridani, M.Y., Khan, S., Chan, T., Teng, S., Beard, K. and O’Brien, P.J., 2001. Cytochrome P450 2E1 metabolically activates propargyl alcohol: propiolaldehyde-induced hepatocyte cytotoxicity. Chemical and Biological Interactions 130–132, 931–942. 4 Hold, K.M., Sirisoma, N.S. and Casida, J.E., 2001. Detoxification of alpha- and beta-Thujones (the active ingredients of absinthe): site specificity and species differences in cytochrome P450 oxidation in vitro and in vivo. Chemical Research in Toxicology 14, 589–595; Hold, K.M., Sirisoma, N.S., Ikeda, T., Narahashi, T. and Casida, J.E., 2000. Alpha-thujone (the active component of absinthe): gamma-aminobutyric acid type A receptor modulation and metabolic detoxification. Proceedings of the National Academies of Science USA 97, 3826–3831. 5 deBethizy, J.D., Udinsky, J.R., Scribner, H.E. and Frederick, C.B., 1987. The disposition and metabolism of acrylic acid and ethyl acrylate in male Sprague-Dawley rats. Fundamental and Applied Toxicology 8, 549–561; Black, K.A., Beskitt, J.L., Finch, L., Tallant, M.J., Udinsky, J.R. and Frantz, S.W., 1995. Disposition and metabolism of acrylic acid in C3H mice and Fischer 344 rats after oral or cutaneous administration. Journal of Toxicology and Environmental Health 45, 291–311; Black, K.A. and Finch, L., 1995. Acrylic acid oxidation and tissue-to-blood partition coefficients in rat tissues. Toxicology Letters 78, 73–78; Winter, S.M. and Sipes, I.G., 1993. The disposition of acrylic acid in the male Sprague-Dawley rat following oral or topical administration. Food and Chemical Toxicology 31, 615–621. 6 Willhite, C.C., Sharma, R.P., Allen, P.V. and Berry, D.L., 1990. Percutaneous retinoid absorption and embryotoxicity. Journal of Investigational Dermatology 95, 523–529; Thongnopnua, P., Massarella, J.W. and Zimmerman, C.L., 1989. The pharmacokinetics of etretinate and its metabolites in the dog. Drug Metabolism and Disposition 17, 473–480; Thongnopnua, P. and Zimmerman, C.L., 1988. Biovailability and pulmonary first-pass removal of etretinate in rats. Research Communications in Chemical Pathology and Pharmacology 61, 269–272. 7 Walton, K., Dorne, J.L. and Renwick, A.G., 2001. Uncertainty factors for chemical risk assessment: interspecies differences in the in vivo pharmacokinetics and metabolism of human CYP1A2 substrates. Food and Chemical Toxicology 39, 667–680; Aramaki, S., Suzuki, E., Ishidaka, O., Momose, A. and Umemura, K., 1991. Pharmacokinetics of caffeine and its metabolites in horses after intravenous, intramuscular or oral administration. Chemical and Pharmaceutical Bulletin (Tokyo) 39, 2999–3002; Blanchard, J. and Sawers, S.J., 1983. The absolute bioavailability of caffeine in man. European Journal of Clinical Pharmacology 24, 93–98. a Suitability of oral NOAEL to be used for topical exposure +++ = negligible first-pass metabolism predicted or hepatotoxicity predicted to be higher after oral dosage; ++ limited first-pass metabolism predicted; +high first-pass metabolism predicted; ? = no data or conclusion uncertain.

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Munro no.

2557

Munro no.

Name

NOAEL

2558

Appendix Table 3 Metabolic considerations for Class I compounds in the Munro database with NOAEL values = or <30 mg/kg bw/day Potential for presystemic metabolism orally

Suitability of oral NOAELa ?

85

Isopropyl alcohol

0.018

132 39

0.5 0.6

105 116

Triethylene glycol Dimethylphenol, 2,6Dimethylphenol, 3,4Oleylamine Riboflavin

3 4

82

Isoamyl salicylate

4.7

7 47

Ascorbic acid Ethyl acrylate

5.5 8.4

95

Methyl methacrylate

8.4

44

Dodecyl gallate

10

80 97

Ionone Methyl-1phenylpentan-2-ol, 4Phenyl-1-propanol, 2. Retinol Styrene

10 10

Presystemic oxidation to the corresponding carboxylic acid, via the aldehyde, is possible; role of metabolism in the unexpected toxic potential is not known No relevant toxicokinetic data have been identified No relevant toxicokinetic data have been identified; the OH group is a site of potential first-pass metabolism, but the 2- and 6-methyl groups would slow the rate of conjugation and enhance oral bioavailability No relevant toxicokinetic data identified; the OH group is a site of potential first-pass metabolism, and the 3- and 4-methyl groups would have little effect on the rate of conjugation Produces apoptosis in vitro, due to the parent compound; the fate in vivo has not been defined1 Nutrient with very high oral bioavailability; absorption is non-linear at high doses; converted to an active intracellular metabolite riboflavin-5 0 -phosphate2 No relevant toxicokinetic data identified; predicted to be hydrolyzed in the gut and liver to the active metabolite salicylic acid; firstpass metabolism is predicted to result in an increase in activity Nutrient with high oral bioavailability; absorption is non-linear at high doses Hydrolyzed very rapidly to acrylic acid which is only in part responsible for the toxicity; conjugated with glutathione indicating reactivity of the parent compound; the rates of hydrolysis and metabolism following topical exposure may be similar to those following oral dosage; covalent binding and toxicity in the forestomach are probably related to the parent compound3 Available data relate largely to nasal deliver and hydrolysis; extensive hydrolysis would be predicted after both oral and topical exposures;4 The toxicity of propyl, octyl and dodecyl esters of gallic acid differ and are probably due to the parent compounds as well as gallic acid; the extent of first-pass metabolism, and possible detoxication have not been defined. No relevant toxicokinetic data have been identified No relevant toxicokinetic data have been identified

10

Conjugated with glucuronic acid but the extent of first-pass metabolism is not known5

?

10 12

High oral bioavailability and greater formation of the active metabolites (retinoic acids) after oral administration6 Produces hepatotoxicity via cytochrome P450-mediated formation of the active metabolite styrene oxide; toxicity is associated with depletion of glutathione; oral toxicity is predicted to exceed dermal toxicity7 Readily hydrolyzed by carboxyl esterase; extent and consequences of intestinal first-pass metabolism have not been defined8

+++ +++

?

25

Predicted to be readily hydrolyzed by carboxyl esterase; extent and consequences of intestinal first-pass metabolism have not been defined Reactive chemical, with most data related to inhalation and effects in the airways; binds extensively to the skin after topical application; data on oral bioavailability not identified; oral toxicity is predicted to exceed dermal toxicity, because of binding in the skin9 Causes peroxisome proliferation in the liver and liver cancer; undergoes almost complete first-pass metabolism to the mono-ethylhexyl analogue, which is probably responsible for the toxic effects; oral toxicity is predicted to exceed dermal toxicity10 Nutrient with high oral bioavailability

25

Nutrient with high oral bioavailability

+++

25

Negligible absorption following dermal administration; the extent of first-pass metabolism and other toxicokinetic parameters have not been defined; oral toxicity is predicted to exceed dermal toxicity11 No relevant toxicokinetic data have been identified; predicted to be metabolized extensively by most tissues

++?

40

115 124 51 48

Ethyl glycol monomethyl ether Ethyl butyrate

14.4

63

Formaldehyde

15

13

Bis(2ethylhexyl)phthalate Glutamate, monosodium Glutamic acid hydrochloride Hydroquinone

18

66 67 73 90

Malonaldehyde, sodium salt

12.5

30

+? ? +++ +++ +++ +?

++ ? ? ?

?

+++?

+++ +++

++?

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

109

1.4

? ++?

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

Munro Number – reference number given in the paper of Munro et al., 1996. References 1 Mizukami, Y., 2002. [Oleylamine (long-chain fatty amine)-induced cell death through MAP kinase pathways in human pancreatic cancer cells] [Article in Japanese] Hokkaido Igaku Zasshi 77, 17–29; Aussel, C., Mahmoudi, A.H., Bernard, G., Breittmayer, J.P. and Bernard, A., 1995. Sphingosine, oleylamine and stearylamine inhibit both CD11a/CD18-dependent and -independent homotypic aggregation: demonstration by cytofluorimetry. Immunology Letters 47, 175–180. 2 Zempleni, J., Galloway, J.R. and McCormick, D.B., 1996. Pharmacokinetics of orally and intravenously administered riboflavin in healthy humans. American Journal of Clinical Nutrition 63, 54–66. 3 Sweeney, L.M., Andersen, M.E. and Gargas, M.L., 2004. Ethyl acrylate risk assessment with a hybrid computational fluid dynamics and physiologically based nasal dosimetry model. Toxicological Sciences 79, 394–403; Linhart, I., Vosmanska, M. and Smejkal, J., 1994. Biotransformation of acrylates. Excretion of mercapturic acids and changes in urinary carboxylic acid profile in rat dosed with ethyl and 1-butyl acrylate. Xenobiotica 24, 1043–1052; McCarthy, T.J., Hayes, E.P., Schwartz, C.S. and Witz, G., 1994. The reactivity of selected acrylate esters toward glutathione and deoxyribonucleosides in vitro: structure-activity relationships. Fundamental and Applied Toxicology 22, 543–548; Frederick, C.B., Potter, D.W., Chang-Mateu, M.I. and Andersen, M.E., 1992. A physiologically based pharmacokinetic and pharmacodynamic model to describe the oral dosing of rats with ethyl acrylate and its implications for risk assessment. Toxicology and Applied Pharmacology 114, 246–260; Ghanayem, B.I., Burka, L.T. and Matthews, H.B., 1987. Ethyl acrylate distribution, macromolecular binding, excretion, and metabolism n male Fisher 344 rats. Fundamental and Applied Toxicology 9, 389–397. 4 Andersen, M.E., Green, T., Frederick, C.B. and Bogdanffy, M.S., 2002. Physiologically based pharmacokinetic (PBPK) models for nasal tissue dosimetry of organic esters: assessing the state-ofknowledge and risk assessment applications with methyl methacrylate and vinyl acetate. Regulatory Toxicology and Pharmacology 36, 234–245; Andersen, M.E., Sarangapani, R., Frederick, C.B. and Kimbell, J.S., 1999. Dosimetric adjustment factors for methyl methacrylate derived from a steady-state analysis of a physiologically based clearance-extraction model. Inhalation Toxicology 11, 899– 926; Andersen, M.E. and Sarangapani, R., 2001. Physiologically based clearance/extraction models for compounds metabolized in the nose: an example with methyl methacrylate. Inhalation Toxicology 13, 397–414. 5 Goenechea, S., Rucker, G., Hoffmann, G., Neugebauer, M. and Langer, M., 1986. [Cleavage of 2-phenylpropan-1-ol and 2-phenylpropan-2-ol-glucuronide, 2 metabolites of isopropylbenzol (Cumol)] [Article in German] Zeitung Rechtsmedizine 97, 83–88. 6 Eckhoff, C., Bailey, J.R., Collins, M.D., Slikker, W. and Nau, H., 1991. Influence of dose and pharmaceutical formulation of vitamin A on plasma levels of retinyl esters and retinol and metabolic generation of retinoic acid compounds and beta-glucuronides in the cynomolgus monkey. Toxicology and Applied Pharmacology 111, 116–127; Tembe, E.A., Honeywell, R. and Buss, N.E., Renwick, A.G., 1996. All-trans-retinoic acid in maternal plasma and teratogenicity in rats and rabbits. Toxicology and Applied Pharmacology 141, 456–472; Buss, N.E., Tembe, E.A., Prendergast, B.D., Renwick, A.G. and George, C.F., 1994. The teratogenic metabolites of vitamin A in women following supplements and liver. Human and Experimental Toxicology 13, 33–43. 7 Carlson, G., 2003. In vitro metabolism of styrene to styrene oxide in liver and lung of Cyp2E1 knockout mice. Journal of Toxicology and Environmental Health A 66, 861–869; Carlson, G.P., 1997. Effects of inducers and inhibitors on the microsomal metabolism of styrene to styrene oxide in mice. Journal of Toxicology and Environmental Health 51, 477–488; Mizutani, T., Irie, Y. and Nakanishi, K., 1994. Styrene-induced hepatotoxicity in mice depleted of glutathione. Research Communications in Molecular Pathology and Pharmacology 86, 361–374. 8 Stott, W.T. and McKenna, M.J., 1985. Hydrolysis of several glycol ether acetates and acrylate esters by nasal mucosal carboxylesterase in vitro. Fundamental and Applied Toxicology 5, 399–404. 9 Robbins, J.D., Norred, W.P., Bathija, A. and Ulsamer, A.G., 1984. Bioavailability in rabbits of formaldehyde from durable-press textiles. Journal of Toxicology and Environmental Health 14, 453– 463. 10 Gray, T.J., Lake, B.G., Beamand, J.A., Foster, J.R. and Gangolli, S.D., 1983. Peroxisomal effects of phthalate esters in primary cultures of rat hepatocytes. Toxicology 28, 167–179; Ljungvall, K., Tienpont, B., David, F., Magnusson, U. and Torneke, K., 2004. Kinetics of orally administered di(2-ethylhexyl) phthalate and its metabolite, mono(2-ethylhexyl) phthalate, in male pigs. Archives of Toxicology 78, 384–389; Huber, W.W., Grasl-Kraupp, B. and Schulte-Hermann, R., 1996. Hepatocarcinogenic potential of di(2-ethylhexyl)phthalate in rodents and its implications on human risk. Critical Reviews in Toxicology 26, 365–481; Pollack, G.M., Shen, D.D. and Dorr, M.B., 1989. Contribution of metabolites to the route- and time-dependent hepatic effects of di-(2-ethylhexyl)phthalate in the rat. Journal of Pharmacology and Experimental Therapeutics 248, 176–181. 11 English, J.C. and Deisinger, P.J., 2005. Metabolism and disposition of hydroquinone in Fischer 344 rats after oral or dermal administration. Food and Chemical Toxicology 43, 483–493. a Suitability of oral NOAEL to be used for topical exposure +++ = negligible first-pass metabolism predicted or hepatotoxicity predicted to be higher after oral dosage; ++ limited first-pass metabolism predicted; + high first-pass metabolism predicted; ? = no data or conclusion uncertain.

2559

2560

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

References Barlow, S.M., Kozianowski, G., Wu¨rtzen, G., Schlatter, J., 1999. Threshold of toxicological concern for chemical substances present in the diet. Food and Chemical Toxicology 39, 893–905. Benech-Kieffer, F., Meuling, W.J.A., Leclerc, C., Roza, L., Leclaire, J., Nohynek, G., 2003. Skin Pharmacology and Applied Skin Physiology 16, 343–355. Bickers, D.R., Calow, P., Greim, H.A., Hanifin, J.M., Rogers, A.E., Saurat, J-H., Sipes, I.G., Smith, R.L., Tagami, H., 2003. The safety assessment of fragrance materials. Regulatory Toxicology and Pharmacology 37, 218–273. Blackburn, K., Stickney, J.A., Carlson-Lynch, H.L., McGinnis, P.M., Chappell, L., Felter, S.P., 2005. Application of the threshold of toxicological concern approach to ingredients in personal and household care products. Regulatory Pharmacology and Toxicology 43, 249–259. Bridges, J., 2003a. The view of risk assessors on the white paper. Strategy for future chemical policy. J. Bridges, President of the CSTEE and Chair of the EU Task Force on the Harmonisation of Risk Assessment. www.eutop.de/chp/Download/BridgesSt.doc. Bridges, J., 2003b. Strategy for a future chemicals policy. The view of the Scientific Committee on Toxicology, Ecotoxicology and the Environment (CSTEE). Professor Jim Bridges, University of Surrey, UK. http://www.eutop.de/chp/Download/BridgesRe.doc. Cassee, F.R., Groten, J.P., van Bladeren, P.J., Feron, V.J., 1998. Toxicological evaluation and risk assessment of chemical mixtures. Critical Reviews in Toxicology 28, 73–101. Cassee, F.R., Su¨hnel, J., Groten, J.P., Feron, V.J., 1999. Toxicology of chemical mixtures, In: Ballantyne, B., Marrs, T., Syversen, T. (Eds.), General and Applied Toxicology, second edition, vol. 1. Macmillan Reference Ltd., pp. 303–319. Cheeseman, M.A., Machuga, E.J., Bailey, A.B., 1999. A tiered approach to threshold of regulation. Food and Chemical Toxicology 37, 387– 412. Cleek, R.L., Bunge, A.L., 1993. A new method for estimating dermal absorption from chemical exposure. 1. General approach. Pharmaceutical Research 4, 497–506. Cramer, G.M., Ford, R.A., Hall, R.L., 1978. Estimation of toxic hazard – a decision tree approach. Food and Chemical Toxicology 16, 255–276. Ding, X., Kaminsky, L.S., 2003. Human extrahepatic cytochromes P450: function in xenobiotic metabolism and tissue-selective chemical toxicity in the respiratory and gastrointestinal tracts. Annual Reviews of Pharmacology and Toxicology 43, 149–173. Doherty, M.M., Charman, W.N., 2002. The mucosa of the small intestine: how clinically relevant as an organ of drug metabolism? Clinical Pharmacokinetics 41, 235–253. Dolan, D.G., Nauman, B.D., Sargent, E.V., Maier, A., Dourson, M., 2005. Application of the threshold of toxicological concern concept to pharmaceutica manufacturing operations. Regulatory Toxicology and Pharmacology 43, 1–9. Dourson, M.L., Stara, J.F., 1983. Regulatory history and experimental support of uncertainty (safety) factors. Regulatory Toxicology and Pharmacology 3, 224–238. Dourson, M., Felter, S.P., Robinson, D., 1996. Evolution of science-based uncertainty factors in noncancer risk assessment. Regulatory Toxicology and Pharmacology 24, 108–120. EC, 2000. The Scientific Committee on Cosmetic Products and Non-food Products intended for consumers (SCCNFP). Opinion concerning The 1st update of the inventory of ingredients employed in cosmetic products, adopted by the SCCNFP during the plenary session of 24 October 2000. EC, 2002. European Commission, Directorate E – food safety, plant health, animal health and welfare. Guidance Document on Dermal Absorption. DG Sanco/222/2000 rev. 6, 27 November 2002. EC, 2004. Guidance document on dermal absorption. Directorate E1 – Plant Health (Sanco/222/2000 rev. 7, 1–15).

ECB (European Chemical Bureau), 2003. Technical guidance document (TGD) on risk assessment in support of Commission Directive 93/67/ EEC on risk assessment for new notified substances. European Commission, Part 1, 1–303. http://ecb.jrc.it/documents/technicalguidancedocument/edition2/tgdpart12ed.pdf. EDETOX, 2004. EDETOX Data Base, University of Newcastle, UK. Available at: http://edetox.ncl.ac.uk/searchinvivo.aspx. EFSA, 2004. European Food Safety Authority. Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food. Opinion on Flavouring Group FGE.03 Acetals of branchedand straight-chain aliphatic saturated primary alcohols and branchedand straight-chain saturated aldehydes, and an orthoester of formic acid, from chemical groups 1 and 2. Opinion expressed on 7 October 2004. Available at: www.efsa.eu.int/science/afc/catindex_en.html. EFSA, 2005.European Food Safety Agency. Note for guidance for petitioners presenting an application for the safety assessment of a substance to be used in food contact materials prior to its authorisation. . EMEA, 2003. The European Medical Control Agency. Committee for Medicinal Products for Human Use. Guideline on the limits of genotoxic impurities. London, 23 June 2003. . EMEA, 2004. European Medicines Evaluation Agency. Committee for Medicinal Products for Human Use (CHMP). Guideline on the limits of genotoxic impurities. CPMP/SWP/5199/02. London, 23 June 2004. . EPA, 1997. US Exposure Factors Handbook. EPA/600/P-95/002Fa. Office of Research and Development, US Environmental Protection Agency. Washington DC. Federal Register, 1993. Food additives; Threshold of regulation for substances used in food-contact articles. Federal Register 58 (195), 52719–52727. Feron, V.J., Groten, J.P., 2002. Toxicological evaluation of chemical mixtures. Food and Chemical Toxicology 40, 825–839. Feron, V.J., van Vliet, P.W., Notten, W.R.F., 2004. Exposure to combinations of substances: a system for assessing health risks. Environmental Toxicology and Pharmacology 18, 215–222. Food and Drug Administration, 1983. Toxicological Principles for the Safety A Assessment of Direct Food Additives and Color Additives Used in Food, Redbook, US Food and Drug Administration, Bureau of Foods, Washington, DC. Food and Drug Administration, 1993. Toxicological Principles for the Safety Assessment of Direct Food Additives and Color Additives Used in Food, Redbook. Food and Drug Administration, 1995. US Food and Drug Administration, Department of Health and Human Services. Food Additives; Threshold of Regulation for Substances Used in Food-Contact Articles 21 CFR Parts 5, 25, 170, 171, and 174. Frawley, J.P., 1967. Scientific evidence and common sense as a basis for food-packaging regulations. Food and Cosmetics Toxicology 5, 293– 308. Groten, J.P., Heijne, W.H.M., Stierum, R.H., Freidig, A.P., Feron, V.J., 2004a. Toxicology of chemical mixtures: a challenging quest along empirical sciences. Environmental Toxicology and Pharmacology 18, 185–192. Groten, J.P., Cassee, F.R., van Bladeren, P.J., de Rosa, C., Feron, V.J. and Su¨hnel, J., 2004b. Misschungen chemischer Stoffe. In: Marquardt H., S. Scha¨fer, (Eds.), Lehrbuch der Toxikologie, Wissenschaftliche Verlagsgesellschaft mbH, Stuttgart, pp. 287–302. Hueber-Becker, F., Nohynek, G.J., Meuling, W.J.A., Benech-Kieffer, F., Toutain, H., 2004. Food and Chemical Toxicology 42, 1227– 1236. ILSI, 2004. Threshold of Toxicological Concern Task Force. International Life Science Institute. http://europe.ilsi.org/file/TTCPoster.pdf. IPCS, 1994. International Programme on Chemical Safety: Assessing human health risks of chemicals: Derivation of Guidance values for health-based exposure limits. Environmental Health Criteria, 170,

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562 73pp, World Health Organisation, International Programme on Chemical safety, Geneva. IPCS, 1998. International Program on Chemical Safety, World Health Organisation, WHO Food Additives Series 35, WHO, Geneva, Switzerland. IPCS, 1999. International Programme on Chemical Safety: Assessing human health risks of chemicals: Principles for the assessment of risk to human health from exposure to chemicals. Environmental Health Criteria 210, World Health Organisation, Geneva. Ito, N., Hasegawa, R., Imaida, K., Kurata, Y., Hagiwara, A., Shirai, T., 1995a. Effect of ingestion of 20 pesticides in combination at acceptable daily intake levels on rat liver carcinogenesis. Food and Chemical Toxicology 33, 159–163. Ito, N., Hagiwara, A., Tamano, S., Hasegawa, R., Imaida, K., Hirose, M., Shirai, T., 1995b. Lack of carcinogenicity of pesticide mixtures administered in the diet at acceptable daily intake (ADI) dose levels in rats. Toxicology Letters (82/83), 513–520. JECFA, 1993. Evaluation of Certain Food Additives and Contaminants. Safety evaluation of flavouring agents. Forty-first Report of the Joint FAO/WHO Expert Committee on Food Additives. WHO Technical Report Series 837. World Health Organization, Geneva. JECFA, 1995. Evaluation of Certain Food Additives and Contaminants. Safety evaluation of flavouring agents. In: Forty-fourth Report of the Joint FAO/WHO Expert Committee on Food Additives, WHO Technical Report Series 859. World Health Organization, Geneva. pp. 2–3. JECFA, 1999. Evaluation of Certain Food Additives and Contaminants. Procedure for the safety evaluation of flavouring agents. In: Fortyninth Report of the Joint FAO/WHO Expert Committee on Food Additives, WHO Technical Report Series 884. World Health Organization, Geneva. pp. 3–6. JECFA, 2003. Summary of Evaluations Performed by the Joint FAO/ WHO Expert Committee on Food Additives (JECFA 1956–2003). In: First through sixty-first meetings. http://jecfa.ilsi.org/. Jimenez, M.M., Pelletier, J., Bobin, M.F., Martini, M.C., 2004. Influence of encapsulation on the in vitro percutaneous absorption of octyl methoxycinnamate. International Journal of Pharmaceutics 272, 45– 55. Jonker, D., Woutersen, R.A., Feron, V.J., 1996. Toxicity of mixtures of nephrotoxicants with similar or dissimilar mode of action. Food and Chemical Toxicology 34, 1075–1082. Jonker, D., Freidig, A.P., Groten, J.P., de Hollander, A.E.M., Stierum, R.H., Woutersen, R.A., Feron, V.J., 2004. Safety evaluation of chemical mixtures and combinations of chemical and non-chemical stressors. Reviews in Environment and Health 19, 83–139. Jung, C.T., Wickett, R.R., Desai, P.B., Bronaugh, R.L., 2003. In vitro and in vivo percutaneous absorption of catechol. Food and Chemical Toxicology 41, 885–895. Kawakubo, Y., Yamazoe, Y., Kato, R., Nishikawa, T., 1990. High capacity of human skin for N-acetylation of arylamines. Skin Pharmacology 3, 180–185. Kraeling, M.E.K., Yourick, J.J., Bronaugh, R.L., 2004. In vitro skin penetration of diethanolamine. Food and Chemical Toxicology 42, 1553–1561. Kroes, R., Galli, C.L., Munro, I., Schilter, B., Tran, L.-A., Walker, R., Wu¨rtzen, G., 2000. Threshold of toxicological concern for chemical substances present in the diet: a practical tool for assessing the need for toxicity testing. Food and Chemical Toxicology 38, 255– 312. Kroes, R., Muller, D., Lambe, J., Lowik, M.R.H., van Klaveren, J., Kleiner, J., Massey, R., Mayer, S., Urieta, I., Verger, P., Visconti, A., 2002. Assessment of intake from the diet. Food and Chemical Toxicology 40, 327–385. Kroes, R., Renwick, A.G., Cheeseman, M., Kleiner, J., Mangelsdorf, I., Piersma, A., Schilter, B., Schlatter, J., van Schothorst, F., Vos, J.G., Wu¨rtzen, G., 2004. Structure-based thresholds of toxicological concern (TTC): guidance for application to substances present at low levels in the diet. Food and Chemical Toxicology 42, 65–83.

2561

Kroes, R., Kleiner, J., Renwick, A., 2005. The threshold of toxicological concern concept in risk assessment. Toxicological Sciences 86, 226– 230. Merk, H.F., Abel, J., Baron, J.M., Krutmann, J., 2004. Molecular pathways in dermatotoxicology. Toxicology and Applied Pharmacology 195, 267–277. Mu¨ller, L., Mauthe, R.J., Riley, C.M., Andino, M.M., De Antonis, D., Beels, C., DeGeorge, J., De Knaep, A.G.M., Ellison, D., Fagerland, J.A., Frank, R., Fritschel, B., Galloway, S., Harpur, E., Humfrey, C.D.N., Jacks, A.S., Jagota, N., Mackinnon, J., Mohan, G., Ness, D.K., O’Donovan, M.R., Smith, M.D., Vudathala, G., Yotti, L., 2006. A rationale for determining, testing, and controlling specific impurities in pharmaceuticals that possess potential for genotoxicity. Regulatory Toxicology and Pharmacology 44, 198–211. Munro, I.C., 1990. Safety assessment procedures for indirect food additives: an overview. Report of a Workshop. Regulatory Toxicology and Pharmacology 12, 2–12. Munro, I.C., Ford, R.A., Kennepohl, E., Sprenger, J.G., 1996. Correlation of structural class with no-observed effect levels: a proposal for establishing a threshold of concern. Food and Chemical Toxicology 34, 829–867. Munro, I.C., Kennepohl, E., Kroes, R., 1999. A procedure for the safety evaluation of flavouring substances. Food and Chemical Toxicology 37, 207–232. Nohynek, G.J., Duche, D., Garrigue, A., Meunier, P.A., Toutain, H., Leclaire, J., 2005. Under the skin: biotransformation of para-aminophenol and para-phenylenediamine in reconstructed human epidermis and human hepatocytes. Toxicology Letters 158, 196–212. Potard, G., Laugela, C., Baillet, A., Schaefer, H., Marty, J.-P., 1999. Quantitative HPLC analysis of sunscreens and caffeine during in vitro percutaneous penetration studies. International Journal of Pharmaceutics 189, 249–260. Potts, R.O., Guy, R.H., 1992. Predicting skin permeability. Pharmaceutical Research 9, 663–669. Reilly, T.P., Lash, L.H., Doll, M.A., Hein, D.W., Woster, P.M., Svensson, C.R., 2000. A role for bioactivation and covalent binding within epidermal keratinocytes in sulfonamide-induced cutaneous drug reactions. Journal of Investigative Dermatology 114, 1164–1173. Renwick, A.G., 2004. Toxicology databases and the concept of thresholds of toxicological concern as used by the JECFA for the safety evaluation of flavouring agents. Toxicology Letters 149, 223–234. Ross, J.H., Dong, M.H., Krieger, R.I., 2000. Conservatism in pesticide exposure assessment. Regulatory Toxicology and Pharmacology 31, 53–58. Royal Commission, 2003. Chemicals in the Environment. Safeguarding the Environment and Human Health. Royal Commission on Environmental Pollution. Twenty Fourth Report. Rulis, A.M., 1986. De minimis and the threshold of regulation. In: Felix, C.W. (Ed.), Food Protection Technology. Lewis Publishers, Chelsea, MI, pp. 29–37. SCCNFP, 2003. The EU Scientific Committee on Cosmetics and NonFood Products. Notes of Guidance for the Testing of Cosmetic Ingredients and their Safety Evaluation. Brussels, 20, October, 2003. SCCP 2006. European Commission. Health and Consumer Protection, Directorate General. Scientific Committee on Consumer Products. The SCCP’s Notes of Guidance for the Testing of Cosmetic Ingredients and Their Safety Evaluation. 6th Revision. Adopted by the SCCP during the 10th plenary meeting of 19 December, 2006. SCP 2002. European Commission. Health and Consumer Protection, Directorate General. Scientific Committee on Plants. Opinion of the Scientific Committee on Plants on Commission Draft Guidance on Dermal Absorption. Opinion adopted by the Scientific Committee on Plants, 24 April, 2002. Swanson, H.I., 2004. Cytochrome P450 expression in human keratinocytes: an aryl hydrocarbon receptor perspective. Chemical Biological Interactions 149, 69–79. Vermeire, T., Stevenson, H., Peiters, M.N., Rennen, M., Slob, W., Hakkert, B.C., 1999. Assessment factors for human health risk

2562

R. Kroes et al. / Food and Chemical Toxicology 45 (2007) 2533–2562

assessment: a discussion paper. Critical Reviews in Toxicology 29, 439–490. WHO, 1987. Principles for the Safety of Food Additives and Contaminants in Food. WHO Environmental Health Criteria. World Health Organization (WHO), International Programme on Chemical Safety (IPCS), in cooperation with the Joint WHO/FAO Expert Committee

on Food Additives (JECFA), Geneva, Switzerland, Environmental Health Criteria No. 70. Wolf de, W., Sieber-Sauer, A., Lecloux, A., Koch, A., Holt, M., Feytel, T., Comber, M., Boeije, G., 2005. Mode of action and aquatic exposure thresholds of no concern. Environmental Toxicity and Chemistry 24, 479–485.