- Email: [email protected]
relevant exposure
Available online at www.sciencedirect.com
Toxicology Letters 176 (2008) 68–76
Exposure-triggered reproductive toxicity testing under the REACH legislation: A proposal to define significant/relevant exposure Ulrike Bernauer ∗ , Gerhard Heinemeyer, Barbara Heinrich-Hirsch, Beate Ulbrich, Ursula Gundert-Remy Federal Institute for Risk Assessment, Thielallee 88-92, D-14195 Berlin, Germany Received 27 July 2007; received in revised form 19 October 2007; accepted 19 October 2007 Available online 26 October 2007
Abstract Under the new REACH legislation, toxicological testing is required in relation to annual tonnages produced or imported. Requirements for toxicological information increase when production volume increases. The respective information requirements are laid down in the REACH Annexes VII–X. Concerning human toxicology, certain toxicological tests may be waived under specific conditions. Aside from waiving criteria such as technical feasibility, exposure plays a decisive role in the waiving process with the consequence that toxicological testing will not be required in case of “no relevant exposure”, “limited exposure”, “no exposure” or “no significant exposure” (as expressed in the documents). However, up to now criteria are lacking which precisely define these terms. Attempts have been made to establish cut-off criteria between “non-relevant” and “relevant” (detrimental) exposure based on external exposure concentrations and the threshold of toxicological concern (TTC) principle. In this paper we make a proposal and describe a strategy how to define the currently insufficiently described terms “relevant/significant” exposure. We propose to define relevant/significant exposure based on an endpoint-specific TTC approach, starting from a comparison of the tentative external exposure to the specific TTC. This can be followed by a refinement of exposure estimates and may culminate in the experimental determination of internal and target tissue exposure. This strategy enables a well-founded assessment of what “no relevant exposure” is and safeguards an appropriate level of protection of the general population. The feasibility of the approach is demonstrated for reproductive toxicity endpoints. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: REACH; Exposure-based waiving; Threshold of toxicological concern (TTC); Reproductive toxicity
1. Definition/description of the problem In December 2006 the European Parliament and the European Council have passed the new legislation for the registration, evaluation, authorization and restric-
∗ Corresponding author. Tel.: +49 30 8412 3705; fax: +49 30 8412 3851. E-mail address: [email protected] (U. Bernauer).
tion of chemicals (REACH) (European Commission, 2006). This step was preceded by activities that had started in 2001 with the publication of a statement on future chemicals regulation and risk reduction entitled “White Paper: Strategy for a Future Chemicals Policy” (European Commission, 2001). The leading principles of the new strategy are the application of identical standards to new and existing chemicals, the achievement of a higher level of health and environmental protection and the decision to consider a number of hazardous proper-
0378-4274/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2007.10.008
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
ties of chemicals as unacceptable until proven otherwise. The commission proposed to use non-animal test systems primarily and tailor-made testing approaches taking into account exposure and existing test results. The toxicological information requirements in the framework of the registration procedure for substances are laid down in the REACH Annexes VII–X. Exemption from conducting individual toxicity tests (“waiving”) is possible in cases where exposure is to be neglected (REACH Annexes VIII–XI). This consideration is based on the main aim of the legislation to regulate risk. Risk is defined by hazard on the one hand and exposure on the other hand. Hence, according to the regulation it is not required to identify specific inherent toxicological properties, the hazard, if there is no exposure. This is based on the consideration that without exposure there is no risk. However, it is difficult to define what constitutes “no exposure” and the REACH Annexes VIII–XI use different terms to describe the conditions that allow waiving based on exposure considerations (“no relevant exposure”, “limited exposure”, “no exposure”, “no significant exposure” and “unlikely exposure”), but refrain from defining the level of exposure at which exposure is thought to be so small that no risk can result irrespective of the inherent toxicity of the chemical. Until now, a precise definition of the terms “no relevant exposure”, “limited exposure”, “no exposure”, “no significant exposure” and “unlikely exposure” has not been developed. Therefore it has been requested that the terminology for exposure in REACH should be more precisely described (Bunke et al., 2006). For practical implementation of the REACH legislation, the commission has initiated a number of projects, the REACH implementation projects (RIPs). Within several of the RIPs, working groups aim at establishing rules for a scientifically adequate justification of waiving human health-related tests based on exposure considerations (e.g. RIP3.3 Repro, 2007; RIP3.2-1, general framework of exposure scenarios). The aim of this paper is to arrive at a definition of the terms used in conjunction with exposure-based waiving (see above) in the context of toxicological testing and to introduce the endpoint-specific threshold of toxicological concern (TTC) concept as a benchmark to distinguish between relevant and non-relevant exposure. This threshold as defined in this paper is the exposure level, below which specific adverse effects are not expected to occur. The considerations we present here are based on empirical data derived from the EU existing substances program. We will concentrate on the endpoints fertility and developmental toxicity as it has been estimated that about
69
80% of the animals included in testing according to REACH requirements have to be used in reproductive toxicity testing (H¨ofer et al., 2004). Hence, the need for defining scientific acceptable justification of waiving is high for this endpoint. We already introduced the approach of applying the TTC concept in the REACH testing strategy in an earlier publication (Gundert-Remy et al., 2005). We did, however, not explain the concept and the data basis in detail. In the following we will elaborate on both the TTC and on the exposure assessment and its refinement by estimating or measuring internal exposure. 2. The endpoint-specific TTC concept 2.1. The endpoint-specific TTC concept as a benchmark for toxicologically relevant (significant) exposure Exposure-driven testing requires a decision making on whether the level of estimated exposure of a chemical substance will result in a specific toxic effect or not under the condition that there is no toxicological information for the specific endpoint under consideration for this chemical substance. There are two possibilities how to approach the question. The first approach is to infer the effect from the chemical structure (QSAR, quantitative structure–activity relationship) or to apply the family approach where structural similarities are used to predict the effect. With more complex endpoints like toxic effects on reproduction (reproductive toxicity) this approach did not yield satisfactory results. The second approach is known as the concept of the TTC. In the context of exposure-driven testing we would propose to use this concept in the modification of an endpoint-specific TTC as a benchmark to define an exposure level below which specific adverse effects are not expected to occur. The TTC concept was introduced by Frawley (Frawley, 1967). It has been proposed with the intention to assess the toxicity of substances with low exposures, e.g. in food (Kroes et al., 2000, 2004). Combes et al. (2003) suggested integrating a TTC concept into the REACH procedure in order to minimise the need of toxicological testing, however without developing a clear strategy. In the existing proposals to use TTC values for decision making, a generic TTC concept intended to protect against all types of toxicity (endpoints) has been developed and values for a general TTC have been published (Table 1). The concept has been refined by an additional step, grouping the chemicals in different classes
70
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
Table 1 Human exposure threshold of toxicological concern (TTC) values which are used in different regulatory frameworks (taken from Norwegian Scientific Committee for Food Safety, 2006) Type of chemical
TTC (g/person/day)
TTC (g/kg bw/day)
Genotoxic compounds Non-genotoxic compounds Organophosphates Cramer class IIIa Cramer class IIa Cramer class Ia
0.15 1.5 18 90 540 1800
0.0025 0.025 0.3 1.5 9 30
a
In the Cramer classes, chemicals are assigned to one of three structural classes on the basis of a decision tree (Cramer et al., 1978; Munro et al., 1996). Cramer class I comprises substances with simple chemical structures, for which efficient modes of metabolism exist, suggesting a low degree of oral toxicity. Class II comprises substances which are less clearly innocuous than class I substances. Most class II substances have functional groups similar to class I, but functional groups might be more reactive or substances might have more complex structures compared to class I chemicals. Class III comprises substances with structures that do not indicate strongly that they are innocuous or substances with indications of significant toxicity or substances with reactive functional groups.
based on their structure (Cramer et al., 1978; Munro et al., 1996) (Table 1). In the context of exposure-based waiving, we would propose to use an “endpoint-specific TTC” for the toxicological endpoint under consideration. For the establishment of endpoint-specific TTCs existing data on NOAEL (no-observed-adverse effect level) or NOAEC (no-observed-adverse effect concentration) values for specific endpoints can be used to determine their empirical distribution and to define a cut-off value, as demonstrated in this paper for fertility and developmental toxicity. We are, however, aware that there are limitations in databases for certain endpoints such as allergenicity or endocrine disruption (Barlow, 2005). 2.2. Procedure to derive the endpoint-specific TTC for fertility and developmental toxicity As a database we selected chemical substances assessed under the EU existing chemicals program. NOAELs/NOAECs for fertility and developmental toxicity of the 91 substances for which these endpoints had been tested were taken from the finalised and draft risk assessment reports (RARs) as available at the ECB website in June 2007 (http://www.ecb.jrc.it). Hence, the NOAELs/NOAECs are the outcome of assessments which have been peer reviewed by the experts of the EU member states. We analysed the data separately for the endpoints fertility and developmental toxicity and also for the route of exposure, namely the oral and the
Fig. 1. (A) Frequency distribution of oral NOAELs for fertility (dotted) and developmental toxicity (black) of EU existing chemicals and of NOAELs for foetotoxicity of pharmaceuticals (hatched). Scaling of the y-axis is different for industrial chemicals and pharmaceuticals. (B) Frequency distribution of inhalative NOAECs for fertility (dotted) and developmental toxicity (black) of EU existing chemicals.
inhalation route. In total, 58 NOAELs for fertility and 62 NOAELs for developmental toxicity were obtained for the oral route and concerning the inhalation route, 24 NOAECs were obtained for fertility and 24 NOAECs were obtained for developmental toxicity. The empirical distribution of the data is depicted in Fig. 1A and B. NOAELs range from 1.5 mg/kg bw/day to 17,260 mg/kg bw/day (fertility) and 1.1 mg/kg bw/day to 2430 mg/kg bw/day (developmental toxicity). The inhalation studies resulted in NOAEC ranges from 0.1 mg/m3 to 36,900 mg/m3 (fertility) and from 0.05 mg/m3 to 24,500 mg/m3 (developmental toxicity). We did some simulation to explore the distribution which however was difficult because of the limited number of data. Hence, we would propose to use the lowest value represented in the data set as an appropriate cut-off rather than a predefined percentile. In the next step we selected an appropriate assessment/safety factor. The considerations for this were as follows: the standard default factors for the extrapolation from animal to man and for taking into account the intraspecies variability are 10, respectively. Both cannot
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
be reduced in the absence of data. We propose to add an extra factor of 10 because of the uncertainty inherent in the small database (91 substances for which the endpoints fertility/developmental toxicity have been tested) and because the effects we are dealing with are serious health effects. The described procedure is in accordance with the procedure to establish DNEL(s) as described in Annex I 1.4.1 letter (b) of the regulation (EC) No. 1907/2006 of the European Parliament and of the Council where it is stated that among other factors the nature and the severity of the effect shall be taken into account. This is in line with the proposal by other authors that an additional factor of 10 should be applied when in deriving TTC values (Cheeseman et al., 1999; Frawley, 1967; Rulis, 1986). The resulting overall assessment/safety factor for the oral route is 1000. Hence, the proposed TTC values for oral uptake are 1.5 g/kg bw/day for fertility and 1.0 g/kg bw/day for developmental toxicity (see Table 2). For the NOAECs (inhalation), no assessment factor has been applied for interspecies extrapolation taking into account the experimental conditions (6 h exposure/day; 5 days/week in the rat versus 24 h exposure/day; 7 days/week in humans), the minute volume (0.35 m3 /day in rats versus 15.2 m3 /day in humans) and the body weight (0.2 kg for rats versus 75 kg for humans). Hence, the overall assessment/safety factor in deriving TTC values for inhalation exposure is 100 (factor 10 for interindividual variability and factor 10 as additional factor). The resulting TTC values for uptake via inhalation are 1 g/m3 (fertility) and 0.5 g/m3 (developmental toxicity) (see Table 2). In order to further elucidate whether the extra safety factor of 10 is appropriate we used data from pharmaceuticals: we used a database of NOAEL and LOAEL values for embryo- and foetotoxicity obtained after oral
Table 2 Cut-off values of human exposure for reproductive toxicity endpoints derived from the database contained in the EU risk assessment programme of existing chemicals
Fertility Developmental toxicity
TTC oral (g/kg bw/day)a
TTC inhalative (g/m3 )b
1.5 1.0
1.0 0.5
a Derived from the lowest oral values for the respective endpoints represented in the database (existing chemicals EU) and divided by a safety factor of 1000. b Derived from the lowest inhalative values for the respective endpoints represented in the database (existing chemicals EU). Corrected for experimental conditions minute volume and body weight and divided by a safety factor of 100.
71
testing of 507 different pharmaceuticals in the rat. The database was established from prenatal toxicity studies submitted to the German Federal Institute for Drugs and Medical Devices between 1985 and 2000 (Beate Ulbrich, personal communication). NOAEL values ranged from 0.0001 mg/kg bw/day to 9600 mg/kg bw/day. In this much larger database of pharmacologically active substances, the 5th percentile is 0.1 mg/kg bw/day. Applying a default factor of 100 (10 for interspecies extrapolation, 10 for intraspecies variability), the resulting TTC value is 1 g/kg/day. This value, derived in a “conventional” way from a much larger database (compared to industrial chemicals) is identical to the one obtained from industrial chemicals by applying a safety factor of 1000. We can conclude therefore, that uncertainty of the data and severity of effect would very well justify an extra safety factor of 10 in the case of industrial chemicals (see Fig. 1A and B). The resulting oral endpoint-specific TTC values are close to the value of 0.1 mg/person/day which has been proposed in a German project for a general cut-off for consumer exposure (Bunke et al., 2006), however higher than the general TTC of 1.5 g/day (0.02 g/kg) proposed by Kroes et al. (2000, 2004) and lower than Cramer class I and II chemicals (Cramer et al., 1978). 3. Exposure 3.1. Modeled external exposure assessment and its refinement Contact of humans to chemicals is possible at the workplace (e.g. production, packaging, distribution and transport), for the general population (consumer) in daily life through the use of products containing chemicals and indirectly via the environment. Proposals for the exposure assessment under REACH have been worked out in the scoping study of the RIP3.2 project and work is ongoing in the RIP3.2 project where one of the authors (G.H.) is involved. Due to the particular definition of the exposure scenario (ES) under REACH, risk management measures (RMMs) have to be taken into consideration when estimating exposure. Van Engelen et al. (2007) explain the procedure of exposure assessment as follows. Firstly, a tentative scenario must be defined. This includes the operational conditions and the RMMs. Secondly, the operational conditions comprising the “traditional” exposure factors and model parameters, have to be taken into consideration. Thirdly, RMMs are to be considered separately. It must be kept in mind, however, that changing an operational condition, for example opening the windows, may represent a RMM, where the
72
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
RMM is characterized by the change and the operational condition by the actual ventilation rate. The concept of RMMs is described by Bruinen de Bruin et al. (2007). By using the scenario and the assumptions it must be justified why the product or the process is safe. In the process of exposure assessment, the “tentative exposure scenario” is representing the starting point. The assumptions taken to estimate the exposure and the models are very conservative. Also, tentative scenarios are often characterized in a broad way, i.e. they comprise various uses for the products, or a “use and exposure category”, which is a broad scenario per REACH definition. Therefore, at this tier it is intended to overestimate the exposure. In case the resulting assessment shows that exposure is not safe the tentative scenario must be refined. Several approaches are available that by refinement can shift a scenario closer to reality. Whereas for risk characterization further refinement of external exposure is a suitable option if the tier 1 exposure assessment exceeds the TTC, it is not appropriate as a justification of waiving. In this case, a refinement of exposure parameters by considering internal exposure might be the more appropriate next step.
2003), absorption from the gastrointestinal tract (oral route) can be regarded as negligible when log Pow (where P is the octanol/water partition coefficient) is below −2 or above 7 and when the molecular weight is above 1000. Uptake and absorption by the inhalation route may be derived from information on melting point, boiling point and vapour pressure at ambient temperature. Volatile compounds (compounds with a vapour pressure above 20 Pa at 20 ◦ C and a boiling point below 200 ◦ C) will be prone to enter the body by the inhalation route. Particles with sizes above 5 m are unable to reach alveoli whereas pulmonary resorption is possible when particle size is below 2 m. Dermal absorption is assumed to be negligible when log Pow is below −1 or above +4 and molecular weight is above 500 (De Heer et al., 1999). A number of skin absorption models have been described, all of them using physico-chemical properties of the substances (e.g. molecular weight, log Pow ) as the most important determinants (Fiserova-Bergerova et al., 1990; Guy and Potts, 1992; Wilschut et al., 1995).
The relation between dose and toxic effects (doseresponse-relationship) is routinely expressed as the relation between the level of external dose (exposure) and the observed effect. However, internal exposure is the more relevant metric for toxic effects compared to external exposure, especially when the compound is poorly absorbed. Since the extent of absorption is disregarded in models estimating the external dose, one possibility of refining the exposure model would be to estimate internal exposure by including the extent of absorption. Therefore, we propose to conduct a tiered approach for estimating or measuring absorption. A tiered approach to the generation of these data for substances where the external exposure is higher than the endpoint-specific TTC is presented below.
3.2.2. In vitro approach Many in vitro data on physico-chemical properties and ADME (absorption, distribution, metabolism and excretion) processes in combination with physiological parameters can be incorporated into a physiologically based toxicokinetic (PBTK) model. This allows an a priori estimation of the overall plasma and tissue kinetic behaviour of substances under in vivo conditions (NCEA, 2006; Poulin and Theil, 2002). The minimal data required for developing PBTK models for a chemical in any given species are partition coefficients, biochemical constants and physiological parameters, route-specific absorption parameters and in vivo kinetic data of the substance for model evaluation. However, it should be mentioned here, that there are no established in vitro models to determine oral or inhalative absorption. For the in vitro determination of dermal absorption, established models exist which are described in detail in OECD test guideline 428 (OECD, 2004).
3.2.1. In silico approaches based on physico-chemical properties of the substance In the absence of experimental absorption data of the compound, physico-chemical properties and the chemical structure of a substance may offer a rough estimate of the extent of absorption (Bernauer et al., 2005). There are, of course, a number of uncertainties when modelling uptakes from physico-chemical data. However, in accordance with the guidance given in the technical guidance document (TGD) of the EU (European Commission,
3.2.3. In vivo toxicokinetic investigations All relevant steps of ADME can be investigated by in vivo toxicokinetic investigations (OECD test guideline 417). It is very common to use radiolabelled substance with the advantage of high sensitivity and the disadvantage that discrimination between parent compound and metabolites requires the two steps of separation and measurement. Absorption can be investigated by the quantification of substance/radioactivity (and metabolites) in exhaled air, urine and faeces/bile. In addition,
3.2. Assessment of internal exposure
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
plasma and tissue concentrations can be determined at various times after dosing. By analysis of the data, the internal exposure can be estimated and quantified. 4. Example An example taken from the schedule of the registration of new chemical entities is given to illustrate the process of decision making we are proposing. This can also be applied to decision making under the new REACH legislation. According to the existing legislation of new chemical entities, a volume triggered testing scheme exists. Concerning the specific endpoints fertility and developmental toxicity, reproductive toxicity studies (either a one-generation study (OECD test guideline 415) or a two-generation study (OECD test guideline 416) and a developmental toxicity study (OECD test guideline 414)) have to be performed when the production volume exceeds 100 tons per year. In vivo toxicokinetic studies which would allow the determination of absorption can be required at the same production volume. The substance under consideration was a butane derivative with three phenolic substituents and a molecular weight around 600 g/mol. From the physical state (solid at room temperature) and physico-chemical properties (water solubility < 0.025 mg/L, fat solubility 15.2 g/100 g fat, log Pow = 8.5) a poor oral absorption could be anticipated. However, the molecular weight was below the threshold of 1000 and the high lipophilicity indicated a potential for bioaccumulation. The likelihood of low absorption was supported by data from acute oral toxicity testing which resulted in a LD 50 value above 5000 mg/kg. The substance is envisaged to be used as an antioxidant for synthetic resins and as a stabilizer for thermal paper. When the notifying company announced that the annual production level exceeded 100 tons per year, further testing (toxicokinetics, testing for reproductive toxicity) was requested by the legislation. Because of the presumed low absorption of the substance it was agreed by industry and authorities, that the decision to perform a developmental toxicity study according to OECD test guideline 414 should be deferred until the outcome of in vivo toxicokinetic investigations was known. These were designed to estimate the extent of absorption into the systemic circulation by using oral application of 14 C-radiolabelled substance. It could be demonstrated by analysis of urine, faeces and exhaled air that only 0.1% of the applied dose was absorbed. The question arose whether the low absorption of 0.1% would justify the request of reproductive toxicity studies. Under the new chemicals regulation, when justified by available information on the substance (e.g. low acute
73
oral toxicity and low absorption), the developmental toxicity test according to OECD test guideline 414 can be carried out by using a limit dose of 1000 mg/kg bw/day. For the chemical substance in question, administration of the limit dose of 1000 mg/kg bw/day would result in 1 mg/kg bw/day (0.1%) of systemically available substance (internal dose) which is 1000-fold higher than the endpoint-specific TTC for developmental toxicity derived from the EU risk assessments of existing substances. Hence, in this case, despite of the low absorption value, exposure-based waiving for reproductive/developmental toxicity testing cannot be justified on the basis of the results from toxicokinetic investigations. However, we could give the advice to perform the test with the limit dose only for animal welfare reasons. As this example provides a successful proof of principle, we propose to apply such a procedure when making decision about the necessity to test substances for reproduction toxicity under REACH. The procedure starts with the estimation of external exposure which is to be compared with the endpointspecific TTC for developmental toxicity and for fertility. When the external exposure is higher than the TTC, refinement of exposure assessment can be performed (see outcome of RIP3.2) which may also include the determination of amounts absorbed or systemically available (see Fig. 2). 5. Discussion In this paper we present an approach which could help to decide whether exposure-based waiving according to REACH Annexes can be justified on the basis of available compound-specific data and the principle of an endpoint-specific TTC. In the context of human risk assessment, exposure is defined in general as “contact between an agent and a target.” (WHO, 2007). In the recent past, external concentrations and doses served as the basis for the derivation of cut-off-criteria for workplace and consumer exposures. In a German national project the following external concentrations have been regarded as being of no concern for the consumer: 60–100 g/person/day for oral and dermal exposure and 3–10 g/m3 for inhalation exposure (Bunke et al., 2006). For the oral route, the results we arrive at are very similar to those of Bunke et al. (2006) whereas for the inhalation route the values we derived are 10-fold lower. The proposed values are based on the concept of the TTC (Kroes and Kozianowski, 2002; Kroes et al., 2000, 2004). This concept is considered as a “pragmatic risk assessment tool” that is based on the principle of establishing a human threshold value for chemicals, below
74
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
which there is a very low probability of an appreciable risk to human health. It proposes that a minimum value of exposure can be identified for at least most chemicals, in the absence of a complete toxicity data set (Norwegian Scientific Committee for Food Safety, 2006). Concerning external exposure it should be considered, that substances may be poorly absorbed in humans. In this case, internal exposure may be low in spite of high external exposure. Therefore, internal exposure (for systemic effects) should be regarded as the essential measure when discussing relevant and non-relevant exposures, particularly in cases where the extent of absorption is low (NCEA, 2006).
The principle of our approach to separate relevant from non-relevant exposures consists of several steps. In the first step, external exposure is estimated according to the proposed tiered approach (REACH RIP3.2 Scoping Study, Van Engelen et al., 2007) and its extent is compared to an endpoint-specific TTC. In a refinement of exposure assessment, if necessary, internal exposure is determined. In a tiered approach, internal exposure can be estimated in silico and, if necessary, an experimental determination can be performed (see Fig. 2). The necessity of establishing a strategy is based on the fact that up to now no sound and scientifically robust definitions exist for the terms “relevant exposure”
Fig. 2. Proposed decision tree for an assessment whether exposure based-waiving for toxicological testing might be applicable. In this decision tree, the TTC principle as well as internal/target tissue exposure are being considered.
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
or “significant exposure”, used in the REACH regulation. In order to establish exposure levels at which no adverse affects are expected, the concept of the endpoint-specific TTC has been introduced into the discussion. The general TTC principle has been proposed and successfully applied for various kinds of questions and applications, e.g. for cosmetics or food additives (Norwegian Scientific Committee for Food Safety, 2006). The procedure relies on external exposure concentrations. However, it is a well-known fact, that internal exposure represents a more important metric than external exposure, in particular for systemically induced effects. Therefore, where necessary, it might be appropriate to refine exposure assessment by determining internal exposure. It is worthwhile to mention that in the context of reproductive toxicity we are considering continuous or repeated exposure and not single exposure. However, the distinction between single and repeated exposure should carefully be made as in selected cases in animal studies even single dose exposure applied in the sensitive window did lead to developmental effects (van Raaij et al., 2003). The proposed strategy presents a solution with respect to exposure-based waiving. It is a refinement compared to the already existing cut-off criteria for waiving based on external exposure (e.g. Bunke et al., 2006). The Nordic Council of Ministers (2005) came to the conclusion, that application of the TTC principle (at or above 100 tpa (tons per year)) within REACH would be premature and raised some concerns, for example that the TTC concept has not been evaluated for the diverse group of industrial chemicals and for different routes of exposure. We are well aware, that the derivation of the TTC has limitations. However, we did take into consideration those uncertainties by introducing an additional factor of 10 to the standard default factors. The use of an additional safety factor of 10 is further supported by analysis of a database derived from pharmaceuticals, where a higher number of substances (507 substances) had been tested in more detail (see Fig. 1). Kroes et al. (2004) came to the conclusion that the usual 100-fold safety factor would be sufficient to cover teratogenic effects. However, the decision for an additional safety factor considering inter alia the fact that we are dealing with serious health effects (reproductive toxicity) is in line with the caution prescribed by a recent publication of Bokkers and Slob (2007) who came to the conclusion that applying an interspecies default assessment factor of 10 could result in human exposure scenarios that are insufficiently protective. Thus, our proposed strategy holds the opportunity for the protection of the population from the adverse effects of chemicals and at the same time for
75
avoiding unnecessary testing and thus respecting animal protection. Conflict of interest statement None. References Barlow, S., 2005. Threshold of toxicological concern (TTC) a tool for assessing substances of unknown toxicity present at low levels in the diet. In: ILSI Europe (Ed.), ILSI Europe concise monograph series, Brussels. Bernauer, U., Oberemm, A., Madle, S., Gundert-Remy, U., 2005. The use of in vitro data in risk assessment. Basic Clin. Pharmacol. Toxicol. 96, 176–181. Bokkers, B.G.H., Slob, W., 2007. Deriving a data-based interspecies assessment factor using the NOAEL and the benchmark dose approach. Crit. Rev. Toxicol. 37, 355–373. Bruinen de Bruin, Y., Hakkinen, B., Lahaniatis, M., Papameletiou, D., del Pozo, C., Reina, V., Van Engelen, J., Heinemeyer, G., Viso, A.C., Rodriguez, C., Jantunen, M., 2007. Risk management measures for chemicals in consumer products—issues and challenges for documentation, assessment, and communication across the supply chain. J. Expo. Sci. Environ. Epidemiol. 1, 1–12. Bunke, D., Schneider, K., J¨ager, I., 2006. Exposure-based waiving concrete specifications of the waiving-conditions in the context of the registration procedure according to REACH, project report, ¨ Oko-Institut Freiburg. Cheeseman, M.A., Machuga, E.J., Bailey, A.B., 1999. A tiered approach to threshold of regulation. Food Chem. Toxicol. 37, 387–412. Combes, R., Barratt, M., Balls, M., 2003. An overall strategy for the testing of chemicals for human hazard and risk assessment under the EU REACH system. ATLA 31, 7–19. Cramer, G.M., Ford, R.A., Hall, R.L., 1978. Estimation of toxic hazard—a decision tree approach. Food Cosmet. Toxicol. 16, 255–276. De Heer, C., Wilschut, A., Stevenson, H., Hackkert, B.C., 1999. Guidance document on the estimation of dermal absorption according to a tiered approach: an update. TNO report V98 1237, p. 27, Zeist, The Netherlands. European Commission, 2001. White paper, strategy for a future chemicals policy. COM (2001) 88 final, Brussels. European Commission, 2003. Technical Guidance Document in Support of the Commission Directive 93/67/EEC on Risk Assessment for New Notified Substances and the Commission Regulation (EC) 1488/94 on Risk Assessment for Existing Substances and the Commission Regulation (EC) 1488/94 on Risk Assessment for Existing Substances and Directive 98/8/EC of the European Parliament and of the Council Concerning the Placing of Biocidal Products on the Market. Joint Research Centre, Institute for Health and Consumer Protection, European Chemicals Bureau, Ispra, Italy. European Commission, 2006. Regulation (EC) No. 1907/2006 of the European Parliament and of the Council of 18, December 2006. Fiserova-Bergerova, V., Pierce, J.T., Droz, P.O., 1990. Dermal absorption potential of industrial chemicals: criteria for skin notation. Am. J. Ind. Med. 17, 617–635. Frawley, J.P., 1967. Scientific evidence and common sense as a basis for food-packaging regulations. Food Cosmet. Toxicol. 5, 293–308.
76
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
Gundert-Remy, U., Heinemeyer, G., Kunde, M., Platzek, T., 2005. EUChemikalienrecht und Verbraucherschutz, Proceedings zum ersten BfR-Forum Verbraucherschutz am 23. und 24. Juni 2005. Guy, R.H., Potts, R.O., 1992. Structure-permeability relationships in percutaneous penetration. J. Pharm. Sci. 81, 603–604. H¨ofer, T., Gerner, I., Gundert-Remy, U., Liebsch, M., Schulte, A., Spielmann, H., Vogel, R., Wettig, K., 2004. Animal testing and alternative approaches fort he human health assessment under the proposed new European chemicals regulation. Arch. Toxicol. 78, 549–564. Kroes, R., Kozianowski, G., 2002. Threshold of toxicological concern (TTC) in food safety assessment. Toxicol. Lett. 127, 43– 46. Kroes, R., Galli, C., Munro, I., Schilter, B., Tran, L.-A., Walker, R., W¨urtzen, G., 2000. Threshold of toxicological concern for chemical substances present in the diet: a practical tool for assessing the need for toxicity testing. Food Chem. Toxicol. 38, 255–312. Kroes, R., Renwick, A.G., Cheeseman, M., Kleiner, J., Mangelsdorf, I., Piersma, A., Schilter, B., Schlatter, J., van Schothorst, F., Vos, J.G., W¨urtzen, G., 2004. Structure-based thresholds of toxicological concern (TTC): guidance for application to substances present at low levels in the diet. Food Chem. Toxicol. 42, 65–83. 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 Chem. Toxicol. 34, 829–876. NCEA (National Center for Environmental Assessment), 2006. Approaches for the Application of Physiologically Based Pharmacokinetic (PBPK) Models and Supporting Data in Risk Assessment (EPA/600/R-05/043F), prepared by the National Center for Environmental Assessment (NCEA) within EPA’s Office of Research and Development (ORD). Norwegian Scientific Committee for Food Safety, 2006. Present and suggested use of the threshold of toxicological concern (TTC) principle in areas of risk assessment relevant for the Norwegian Scientific Committee for Food Safety (VKM). OECD, 1983. OECD Guideline for the testing of chemicals: 415 “OneGeneration Reproduction Toxicity Study” (adopted May 26, 1983), OECD, Paris.
OECD, 1984. OECD Guideline for the testing of chemicals: 417 “Toxicokinetics” (adopted April 4, 1984), OECD, Paris. OECD, 2001. OECD Guideline for the testing of chemicals: 414 “Prenatal Developmental Toxicity Study” (adopted January 22, 2001), OECD, Paris. OECD, 2001. OECD Guideline for the testing of chemicals: 416 “TwoGeneration Reproduction Toxicity Study” (adopted January 22, 2001), OECD, Paris. OECD, 2004. OECD Guideline for the testing of chemicals: 428 “Skin Absorption: In Vitro Method” (adopted April 13, 2004), OECD, Paris. Poulin, P., Theil, F.P., 2002. Prediction of pharmacokinetics prior to in vivo studies. II. Generic physiologically based pharmacokinetic models of drug disposition. J. Pharm. Sci. 91, 1358–1370. RIP3.2-1, 2007. WP1: development of the concept of exposure scenarios. General framework of exposure scenarios – Scoping Study – a final report. Technical Guidance Document on preparing the Chemical Safety Report under REACH – Scoping Study – Phase 1A. (REACH implementation project 3.2-1A). Service Contract Number 22551-2004-12 F1SC ISP BE. RIP3.3, 2007. EWG Repro report final—agreed edited version. Rulis, A.M., 1986. De minimis and the threshold of regulation. In: Felix, C.W. (Ed.), Food Protection Technology. Chelsea/Lewis, Michigan, pp. 29–37. The Nordic Council of Ministers, 2005. Threshold of toxicological concern (TTC). Literature review and applicability. TemaNord 559, Copenhagen, ISBN 92-893-1196-7. Available from: URL: http://www.norden.org/pub/miljo/sk/TN2005559.pdf. Van Engelen, J., Heinemeyer, G., Rodriguez, C., 2007. Consumer exposure scenarios: development, challenges and possible solutions. J. Environm. Expos. Anal. 17, S26–S33. van Raaij, M.T.M., Janssen, P.A.H., Piersma, A.H., 2003. The relevance of developmental toxicity endpoints for acute limit setting. RIVM rapport 601900004. WHO/IPCS, 2007. Draft guidance document on characterizing and communicating uncertainty in exposure assessment, in press. Wilschut, A., ten Berge, W.F., Robinson, P.J., McKone, T.E., 1995. Estimating skin permeation: The validation of five mathematical skin permeation models. Chemosphere 30, 1275–1296.
Toxicology Letters 176 (2008) 68–76
Exposure-triggered reproductive toxicity testing under the REACH legislation: A proposal to define significant/relevant exposure Ulrike Bernauer ∗ , Gerhard Heinemeyer, Barbara Heinrich-Hirsch, Beate Ulbrich, Ursula Gundert-Remy Federal Institute for Risk Assessment, Thielallee 88-92, D-14195 Berlin, Germany Received 27 July 2007; received in revised form 19 October 2007; accepted 19 October 2007 Available online 26 October 2007
Abstract Under the new REACH legislation, toxicological testing is required in relation to annual tonnages produced or imported. Requirements for toxicological information increase when production volume increases. The respective information requirements are laid down in the REACH Annexes VII–X. Concerning human toxicology, certain toxicological tests may be waived under specific conditions. Aside from waiving criteria such as technical feasibility, exposure plays a decisive role in the waiving process with the consequence that toxicological testing will not be required in case of “no relevant exposure”, “limited exposure”, “no exposure” or “no significant exposure” (as expressed in the documents). However, up to now criteria are lacking which precisely define these terms. Attempts have been made to establish cut-off criteria between “non-relevant” and “relevant” (detrimental) exposure based on external exposure concentrations and the threshold of toxicological concern (TTC) principle. In this paper we make a proposal and describe a strategy how to define the currently insufficiently described terms “relevant/significant” exposure. We propose to define relevant/significant exposure based on an endpoint-specific TTC approach, starting from a comparison of the tentative external exposure to the specific TTC. This can be followed by a refinement of exposure estimates and may culminate in the experimental determination of internal and target tissue exposure. This strategy enables a well-founded assessment of what “no relevant exposure” is and safeguards an appropriate level of protection of the general population. The feasibility of the approach is demonstrated for reproductive toxicity endpoints. © 2007 Elsevier Ireland Ltd. All rights reserved. Keywords: REACH; Exposure-based waiving; Threshold of toxicological concern (TTC); Reproductive toxicity
1. Definition/description of the problem In December 2006 the European Parliament and the European Council have passed the new legislation for the registration, evaluation, authorization and restric-
∗ Corresponding author. Tel.: +49 30 8412 3705; fax: +49 30 8412 3851. E-mail address: [email protected] (U. Bernauer).
tion of chemicals (REACH) (European Commission, 2006). This step was preceded by activities that had started in 2001 with the publication of a statement on future chemicals regulation and risk reduction entitled “White Paper: Strategy for a Future Chemicals Policy” (European Commission, 2001). The leading principles of the new strategy are the application of identical standards to new and existing chemicals, the achievement of a higher level of health and environmental protection and the decision to consider a number of hazardous proper-
0378-4274/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2007.10.008
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
ties of chemicals as unacceptable until proven otherwise. The commission proposed to use non-animal test systems primarily and tailor-made testing approaches taking into account exposure and existing test results. The toxicological information requirements in the framework of the registration procedure for substances are laid down in the REACH Annexes VII–X. Exemption from conducting individual toxicity tests (“waiving”) is possible in cases where exposure is to be neglected (REACH Annexes VIII–XI). This consideration is based on the main aim of the legislation to regulate risk. Risk is defined by hazard on the one hand and exposure on the other hand. Hence, according to the regulation it is not required to identify specific inherent toxicological properties, the hazard, if there is no exposure. This is based on the consideration that without exposure there is no risk. However, it is difficult to define what constitutes “no exposure” and the REACH Annexes VIII–XI use different terms to describe the conditions that allow waiving based on exposure considerations (“no relevant exposure”, “limited exposure”, “no exposure”, “no significant exposure” and “unlikely exposure”), but refrain from defining the level of exposure at which exposure is thought to be so small that no risk can result irrespective of the inherent toxicity of the chemical. Until now, a precise definition of the terms “no relevant exposure”, “limited exposure”, “no exposure”, “no significant exposure” and “unlikely exposure” has not been developed. Therefore it has been requested that the terminology for exposure in REACH should be more precisely described (Bunke et al., 2006). For practical implementation of the REACH legislation, the commission has initiated a number of projects, the REACH implementation projects (RIPs). Within several of the RIPs, working groups aim at establishing rules for a scientifically adequate justification of waiving human health-related tests based on exposure considerations (e.g. RIP3.3 Repro, 2007; RIP3.2-1, general framework of exposure scenarios). The aim of this paper is to arrive at a definition of the terms used in conjunction with exposure-based waiving (see above) in the context of toxicological testing and to introduce the endpoint-specific threshold of toxicological concern (TTC) concept as a benchmark to distinguish between relevant and non-relevant exposure. This threshold as defined in this paper is the exposure level, below which specific adverse effects are not expected to occur. The considerations we present here are based on empirical data derived from the EU existing substances program. We will concentrate on the endpoints fertility and developmental toxicity as it has been estimated that about
69
80% of the animals included in testing according to REACH requirements have to be used in reproductive toxicity testing (H¨ofer et al., 2004). Hence, the need for defining scientific acceptable justification of waiving is high for this endpoint. We already introduced the approach of applying the TTC concept in the REACH testing strategy in an earlier publication (Gundert-Remy et al., 2005). We did, however, not explain the concept and the data basis in detail. In the following we will elaborate on both the TTC and on the exposure assessment and its refinement by estimating or measuring internal exposure. 2. The endpoint-specific TTC concept 2.1. The endpoint-specific TTC concept as a benchmark for toxicologically relevant (significant) exposure Exposure-driven testing requires a decision making on whether the level of estimated exposure of a chemical substance will result in a specific toxic effect or not under the condition that there is no toxicological information for the specific endpoint under consideration for this chemical substance. There are two possibilities how to approach the question. The first approach is to infer the effect from the chemical structure (QSAR, quantitative structure–activity relationship) or to apply the family approach where structural similarities are used to predict the effect. With more complex endpoints like toxic effects on reproduction (reproductive toxicity) this approach did not yield satisfactory results. The second approach is known as the concept of the TTC. In the context of exposure-driven testing we would propose to use this concept in the modification of an endpoint-specific TTC as a benchmark to define an exposure level below which specific adverse effects are not expected to occur. The TTC concept was introduced by Frawley (Frawley, 1967). It has been proposed with the intention to assess the toxicity of substances with low exposures, e.g. in food (Kroes et al., 2000, 2004). Combes et al. (2003) suggested integrating a TTC concept into the REACH procedure in order to minimise the need of toxicological testing, however without developing a clear strategy. In the existing proposals to use TTC values for decision making, a generic TTC concept intended to protect against all types of toxicity (endpoints) has been developed and values for a general TTC have been published (Table 1). The concept has been refined by an additional step, grouping the chemicals in different classes
70
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
Table 1 Human exposure threshold of toxicological concern (TTC) values which are used in different regulatory frameworks (taken from Norwegian Scientific Committee for Food Safety, 2006) Type of chemical
TTC (g/person/day)
TTC (g/kg bw/day)
Genotoxic compounds Non-genotoxic compounds Organophosphates Cramer class IIIa Cramer class IIa Cramer class Ia
0.15 1.5 18 90 540 1800
0.0025 0.025 0.3 1.5 9 30
a
In the Cramer classes, chemicals are assigned to one of three structural classes on the basis of a decision tree (Cramer et al., 1978; Munro et al., 1996). Cramer class I comprises substances with simple chemical structures, for which efficient modes of metabolism exist, suggesting a low degree of oral toxicity. Class II comprises substances which are less clearly innocuous than class I substances. Most class II substances have functional groups similar to class I, but functional groups might be more reactive or substances might have more complex structures compared to class I chemicals. Class III comprises substances with structures that do not indicate strongly that they are innocuous or substances with indications of significant toxicity or substances with reactive functional groups.
based on their structure (Cramer et al., 1978; Munro et al., 1996) (Table 1). In the context of exposure-based waiving, we would propose to use an “endpoint-specific TTC” for the toxicological endpoint under consideration. For the establishment of endpoint-specific TTCs existing data on NOAEL (no-observed-adverse effect level) or NOAEC (no-observed-adverse effect concentration) values for specific endpoints can be used to determine their empirical distribution and to define a cut-off value, as demonstrated in this paper for fertility and developmental toxicity. We are, however, aware that there are limitations in databases for certain endpoints such as allergenicity or endocrine disruption (Barlow, 2005). 2.2. Procedure to derive the endpoint-specific TTC for fertility and developmental toxicity As a database we selected chemical substances assessed under the EU existing chemicals program. NOAELs/NOAECs for fertility and developmental toxicity of the 91 substances for which these endpoints had been tested were taken from the finalised and draft risk assessment reports (RARs) as available at the ECB website in June 2007 (http://www.ecb.jrc.it). Hence, the NOAELs/NOAECs are the outcome of assessments which have been peer reviewed by the experts of the EU member states. We analysed the data separately for the endpoints fertility and developmental toxicity and also for the route of exposure, namely the oral and the
Fig. 1. (A) Frequency distribution of oral NOAELs for fertility (dotted) and developmental toxicity (black) of EU existing chemicals and of NOAELs for foetotoxicity of pharmaceuticals (hatched). Scaling of the y-axis is different for industrial chemicals and pharmaceuticals. (B) Frequency distribution of inhalative NOAECs for fertility (dotted) and developmental toxicity (black) of EU existing chemicals.
inhalation route. In total, 58 NOAELs for fertility and 62 NOAELs for developmental toxicity were obtained for the oral route and concerning the inhalation route, 24 NOAECs were obtained for fertility and 24 NOAECs were obtained for developmental toxicity. The empirical distribution of the data is depicted in Fig. 1A and B. NOAELs range from 1.5 mg/kg bw/day to 17,260 mg/kg bw/day (fertility) and 1.1 mg/kg bw/day to 2430 mg/kg bw/day (developmental toxicity). The inhalation studies resulted in NOAEC ranges from 0.1 mg/m3 to 36,900 mg/m3 (fertility) and from 0.05 mg/m3 to 24,500 mg/m3 (developmental toxicity). We did some simulation to explore the distribution which however was difficult because of the limited number of data. Hence, we would propose to use the lowest value represented in the data set as an appropriate cut-off rather than a predefined percentile. In the next step we selected an appropriate assessment/safety factor. The considerations for this were as follows: the standard default factors for the extrapolation from animal to man and for taking into account the intraspecies variability are 10, respectively. Both cannot
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
be reduced in the absence of data. We propose to add an extra factor of 10 because of the uncertainty inherent in the small database (91 substances for which the endpoints fertility/developmental toxicity have been tested) and because the effects we are dealing with are serious health effects. The described procedure is in accordance with the procedure to establish DNEL(s) as described in Annex I 1.4.1 letter (b) of the regulation (EC) No. 1907/2006 of the European Parliament and of the Council where it is stated that among other factors the nature and the severity of the effect shall be taken into account. This is in line with the proposal by other authors that an additional factor of 10 should be applied when in deriving TTC values (Cheeseman et al., 1999; Frawley, 1967; Rulis, 1986). The resulting overall assessment/safety factor for the oral route is 1000. Hence, the proposed TTC values for oral uptake are 1.5 g/kg bw/day for fertility and 1.0 g/kg bw/day for developmental toxicity (see Table 2). For the NOAECs (inhalation), no assessment factor has been applied for interspecies extrapolation taking into account the experimental conditions (6 h exposure/day; 5 days/week in the rat versus 24 h exposure/day; 7 days/week in humans), the minute volume (0.35 m3 /day in rats versus 15.2 m3 /day in humans) and the body weight (0.2 kg for rats versus 75 kg for humans). Hence, the overall assessment/safety factor in deriving TTC values for inhalation exposure is 100 (factor 10 for interindividual variability and factor 10 as additional factor). The resulting TTC values for uptake via inhalation are 1 g/m3 (fertility) and 0.5 g/m3 (developmental toxicity) (see Table 2). In order to further elucidate whether the extra safety factor of 10 is appropriate we used data from pharmaceuticals: we used a database of NOAEL and LOAEL values for embryo- and foetotoxicity obtained after oral
Table 2 Cut-off values of human exposure for reproductive toxicity endpoints derived from the database contained in the EU risk assessment programme of existing chemicals
Fertility Developmental toxicity
TTC oral (g/kg bw/day)a
TTC inhalative (g/m3 )b
1.5 1.0
1.0 0.5
a Derived from the lowest oral values for the respective endpoints represented in the database (existing chemicals EU) and divided by a safety factor of 1000. b Derived from the lowest inhalative values for the respective endpoints represented in the database (existing chemicals EU). Corrected for experimental conditions minute volume and body weight and divided by a safety factor of 100.
71
testing of 507 different pharmaceuticals in the rat. The database was established from prenatal toxicity studies submitted to the German Federal Institute for Drugs and Medical Devices between 1985 and 2000 (Beate Ulbrich, personal communication). NOAEL values ranged from 0.0001 mg/kg bw/day to 9600 mg/kg bw/day. In this much larger database of pharmacologically active substances, the 5th percentile is 0.1 mg/kg bw/day. Applying a default factor of 100 (10 for interspecies extrapolation, 10 for intraspecies variability), the resulting TTC value is 1 g/kg/day. This value, derived in a “conventional” way from a much larger database (compared to industrial chemicals) is identical to the one obtained from industrial chemicals by applying a safety factor of 1000. We can conclude therefore, that uncertainty of the data and severity of effect would very well justify an extra safety factor of 10 in the case of industrial chemicals (see Fig. 1A and B). The resulting oral endpoint-specific TTC values are close to the value of 0.1 mg/person/day which has been proposed in a German project for a general cut-off for consumer exposure (Bunke et al., 2006), however higher than the general TTC of 1.5 g/day (0.02 g/kg) proposed by Kroes et al. (2000, 2004) and lower than Cramer class I and II chemicals (Cramer et al., 1978). 3. Exposure 3.1. Modeled external exposure assessment and its refinement Contact of humans to chemicals is possible at the workplace (e.g. production, packaging, distribution and transport), for the general population (consumer) in daily life through the use of products containing chemicals and indirectly via the environment. Proposals for the exposure assessment under REACH have been worked out in the scoping study of the RIP3.2 project and work is ongoing in the RIP3.2 project where one of the authors (G.H.) is involved. Due to the particular definition of the exposure scenario (ES) under REACH, risk management measures (RMMs) have to be taken into consideration when estimating exposure. Van Engelen et al. (2007) explain the procedure of exposure assessment as follows. Firstly, a tentative scenario must be defined. This includes the operational conditions and the RMMs. Secondly, the operational conditions comprising the “traditional” exposure factors and model parameters, have to be taken into consideration. Thirdly, RMMs are to be considered separately. It must be kept in mind, however, that changing an operational condition, for example opening the windows, may represent a RMM, where the
72
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
RMM is characterized by the change and the operational condition by the actual ventilation rate. The concept of RMMs is described by Bruinen de Bruin et al. (2007). By using the scenario and the assumptions it must be justified why the product or the process is safe. In the process of exposure assessment, the “tentative exposure scenario” is representing the starting point. The assumptions taken to estimate the exposure and the models are very conservative. Also, tentative scenarios are often characterized in a broad way, i.e. they comprise various uses for the products, or a “use and exposure category”, which is a broad scenario per REACH definition. Therefore, at this tier it is intended to overestimate the exposure. In case the resulting assessment shows that exposure is not safe the tentative scenario must be refined. Several approaches are available that by refinement can shift a scenario closer to reality. Whereas for risk characterization further refinement of external exposure is a suitable option if the tier 1 exposure assessment exceeds the TTC, it is not appropriate as a justification of waiving. In this case, a refinement of exposure parameters by considering internal exposure might be the more appropriate next step.
2003), absorption from the gastrointestinal tract (oral route) can be regarded as negligible when log Pow (where P is the octanol/water partition coefficient) is below −2 or above 7 and when the molecular weight is above 1000. Uptake and absorption by the inhalation route may be derived from information on melting point, boiling point and vapour pressure at ambient temperature. Volatile compounds (compounds with a vapour pressure above 20 Pa at 20 ◦ C and a boiling point below 200 ◦ C) will be prone to enter the body by the inhalation route. Particles with sizes above 5 m are unable to reach alveoli whereas pulmonary resorption is possible when particle size is below 2 m. Dermal absorption is assumed to be negligible when log Pow is below −1 or above +4 and molecular weight is above 500 (De Heer et al., 1999). A number of skin absorption models have been described, all of them using physico-chemical properties of the substances (e.g. molecular weight, log Pow ) as the most important determinants (Fiserova-Bergerova et al., 1990; Guy and Potts, 1992; Wilschut et al., 1995).
The relation between dose and toxic effects (doseresponse-relationship) is routinely expressed as the relation between the level of external dose (exposure) and the observed effect. However, internal exposure is the more relevant metric for toxic effects compared to external exposure, especially when the compound is poorly absorbed. Since the extent of absorption is disregarded in models estimating the external dose, one possibility of refining the exposure model would be to estimate internal exposure by including the extent of absorption. Therefore, we propose to conduct a tiered approach for estimating or measuring absorption. A tiered approach to the generation of these data for substances where the external exposure is higher than the endpoint-specific TTC is presented below.
3.2.2. In vitro approach Many in vitro data on physico-chemical properties and ADME (absorption, distribution, metabolism and excretion) processes in combination with physiological parameters can be incorporated into a physiologically based toxicokinetic (PBTK) model. This allows an a priori estimation of the overall plasma and tissue kinetic behaviour of substances under in vivo conditions (NCEA, 2006; Poulin and Theil, 2002). The minimal data required for developing PBTK models for a chemical in any given species are partition coefficients, biochemical constants and physiological parameters, route-specific absorption parameters and in vivo kinetic data of the substance for model evaluation. However, it should be mentioned here, that there are no established in vitro models to determine oral or inhalative absorption. For the in vitro determination of dermal absorption, established models exist which are described in detail in OECD test guideline 428 (OECD, 2004).
3.2.1. In silico approaches based on physico-chemical properties of the substance In the absence of experimental absorption data of the compound, physico-chemical properties and the chemical structure of a substance may offer a rough estimate of the extent of absorption (Bernauer et al., 2005). There are, of course, a number of uncertainties when modelling uptakes from physico-chemical data. However, in accordance with the guidance given in the technical guidance document (TGD) of the EU (European Commission,
3.2.3. In vivo toxicokinetic investigations All relevant steps of ADME can be investigated by in vivo toxicokinetic investigations (OECD test guideline 417). It is very common to use radiolabelled substance with the advantage of high sensitivity and the disadvantage that discrimination between parent compound and metabolites requires the two steps of separation and measurement. Absorption can be investigated by the quantification of substance/radioactivity (and metabolites) in exhaled air, urine and faeces/bile. In addition,
3.2. Assessment of internal exposure
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
plasma and tissue concentrations can be determined at various times after dosing. By analysis of the data, the internal exposure can be estimated and quantified. 4. Example An example taken from the schedule of the registration of new chemical entities is given to illustrate the process of decision making we are proposing. This can also be applied to decision making under the new REACH legislation. According to the existing legislation of new chemical entities, a volume triggered testing scheme exists. Concerning the specific endpoints fertility and developmental toxicity, reproductive toxicity studies (either a one-generation study (OECD test guideline 415) or a two-generation study (OECD test guideline 416) and a developmental toxicity study (OECD test guideline 414)) have to be performed when the production volume exceeds 100 tons per year. In vivo toxicokinetic studies which would allow the determination of absorption can be required at the same production volume. The substance under consideration was a butane derivative with three phenolic substituents and a molecular weight around 600 g/mol. From the physical state (solid at room temperature) and physico-chemical properties (water solubility < 0.025 mg/L, fat solubility 15.2 g/100 g fat, log Pow = 8.5) a poor oral absorption could be anticipated. However, the molecular weight was below the threshold of 1000 and the high lipophilicity indicated a potential for bioaccumulation. The likelihood of low absorption was supported by data from acute oral toxicity testing which resulted in a LD 50 value above 5000 mg/kg. The substance is envisaged to be used as an antioxidant for synthetic resins and as a stabilizer for thermal paper. When the notifying company announced that the annual production level exceeded 100 tons per year, further testing (toxicokinetics, testing for reproductive toxicity) was requested by the legislation. Because of the presumed low absorption of the substance it was agreed by industry and authorities, that the decision to perform a developmental toxicity study according to OECD test guideline 414 should be deferred until the outcome of in vivo toxicokinetic investigations was known. These were designed to estimate the extent of absorption into the systemic circulation by using oral application of 14 C-radiolabelled substance. It could be demonstrated by analysis of urine, faeces and exhaled air that only 0.1% of the applied dose was absorbed. The question arose whether the low absorption of 0.1% would justify the request of reproductive toxicity studies. Under the new chemicals regulation, when justified by available information on the substance (e.g. low acute
73
oral toxicity and low absorption), the developmental toxicity test according to OECD test guideline 414 can be carried out by using a limit dose of 1000 mg/kg bw/day. For the chemical substance in question, administration of the limit dose of 1000 mg/kg bw/day would result in 1 mg/kg bw/day (0.1%) of systemically available substance (internal dose) which is 1000-fold higher than the endpoint-specific TTC for developmental toxicity derived from the EU risk assessments of existing substances. Hence, in this case, despite of the low absorption value, exposure-based waiving for reproductive/developmental toxicity testing cannot be justified on the basis of the results from toxicokinetic investigations. However, we could give the advice to perform the test with the limit dose only for animal welfare reasons. As this example provides a successful proof of principle, we propose to apply such a procedure when making decision about the necessity to test substances for reproduction toxicity under REACH. The procedure starts with the estimation of external exposure which is to be compared with the endpointspecific TTC for developmental toxicity and for fertility. When the external exposure is higher than the TTC, refinement of exposure assessment can be performed (see outcome of RIP3.2) which may also include the determination of amounts absorbed or systemically available (see Fig. 2). 5. Discussion In this paper we present an approach which could help to decide whether exposure-based waiving according to REACH Annexes can be justified on the basis of available compound-specific data and the principle of an endpoint-specific TTC. In the context of human risk assessment, exposure is defined in general as “contact between an agent and a target.” (WHO, 2007). In the recent past, external concentrations and doses served as the basis for the derivation of cut-off-criteria for workplace and consumer exposures. In a German national project the following external concentrations have been regarded as being of no concern for the consumer: 60–100 g/person/day for oral and dermal exposure and 3–10 g/m3 for inhalation exposure (Bunke et al., 2006). For the oral route, the results we arrive at are very similar to those of Bunke et al. (2006) whereas for the inhalation route the values we derived are 10-fold lower. The proposed values are based on the concept of the TTC (Kroes and Kozianowski, 2002; Kroes et al., 2000, 2004). This concept is considered as a “pragmatic risk assessment tool” that is based on the principle of establishing a human threshold value for chemicals, below
74
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
which there is a very low probability of an appreciable risk to human health. It proposes that a minimum value of exposure can be identified for at least most chemicals, in the absence of a complete toxicity data set (Norwegian Scientific Committee for Food Safety, 2006). Concerning external exposure it should be considered, that substances may be poorly absorbed in humans. In this case, internal exposure may be low in spite of high external exposure. Therefore, internal exposure (for systemic effects) should be regarded as the essential measure when discussing relevant and non-relevant exposures, particularly in cases where the extent of absorption is low (NCEA, 2006).
The principle of our approach to separate relevant from non-relevant exposures consists of several steps. In the first step, external exposure is estimated according to the proposed tiered approach (REACH RIP3.2 Scoping Study, Van Engelen et al., 2007) and its extent is compared to an endpoint-specific TTC. In a refinement of exposure assessment, if necessary, internal exposure is determined. In a tiered approach, internal exposure can be estimated in silico and, if necessary, an experimental determination can be performed (see Fig. 2). The necessity of establishing a strategy is based on the fact that up to now no sound and scientifically robust definitions exist for the terms “relevant exposure”
Fig. 2. Proposed decision tree for an assessment whether exposure based-waiving for toxicological testing might be applicable. In this decision tree, the TTC principle as well as internal/target tissue exposure are being considered.
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
or “significant exposure”, used in the REACH regulation. In order to establish exposure levels at which no adverse affects are expected, the concept of the endpoint-specific TTC has been introduced into the discussion. The general TTC principle has been proposed and successfully applied for various kinds of questions and applications, e.g. for cosmetics or food additives (Norwegian Scientific Committee for Food Safety, 2006). The procedure relies on external exposure concentrations. However, it is a well-known fact, that internal exposure represents a more important metric than external exposure, in particular for systemically induced effects. Therefore, where necessary, it might be appropriate to refine exposure assessment by determining internal exposure. It is worthwhile to mention that in the context of reproductive toxicity we are considering continuous or repeated exposure and not single exposure. However, the distinction between single and repeated exposure should carefully be made as in selected cases in animal studies even single dose exposure applied in the sensitive window did lead to developmental effects (van Raaij et al., 2003). The proposed strategy presents a solution with respect to exposure-based waiving. It is a refinement compared to the already existing cut-off criteria for waiving based on external exposure (e.g. Bunke et al., 2006). The Nordic Council of Ministers (2005) came to the conclusion, that application of the TTC principle (at or above 100 tpa (tons per year)) within REACH would be premature and raised some concerns, for example that the TTC concept has not been evaluated for the diverse group of industrial chemicals and for different routes of exposure. We are well aware, that the derivation of the TTC has limitations. However, we did take into consideration those uncertainties by introducing an additional factor of 10 to the standard default factors. The use of an additional safety factor of 10 is further supported by analysis of a database derived from pharmaceuticals, where a higher number of substances (507 substances) had been tested in more detail (see Fig. 1). Kroes et al. (2004) came to the conclusion that the usual 100-fold safety factor would be sufficient to cover teratogenic effects. However, the decision for an additional safety factor considering inter alia the fact that we are dealing with serious health effects (reproductive toxicity) is in line with the caution prescribed by a recent publication of Bokkers and Slob (2007) who came to the conclusion that applying an interspecies default assessment factor of 10 could result in human exposure scenarios that are insufficiently protective. Thus, our proposed strategy holds the opportunity for the protection of the population from the adverse effects of chemicals and at the same time for
75
avoiding unnecessary testing and thus respecting animal protection. Conflict of interest statement None. References Barlow, S., 2005. Threshold of toxicological concern (TTC) a tool for assessing substances of unknown toxicity present at low levels in the diet. In: ILSI Europe (Ed.), ILSI Europe concise monograph series, Brussels. Bernauer, U., Oberemm, A., Madle, S., Gundert-Remy, U., 2005. The use of in vitro data in risk assessment. Basic Clin. Pharmacol. Toxicol. 96, 176–181. Bokkers, B.G.H., Slob, W., 2007. Deriving a data-based interspecies assessment factor using the NOAEL and the benchmark dose approach. Crit. Rev. Toxicol. 37, 355–373. Bruinen de Bruin, Y., Hakkinen, B., Lahaniatis, M., Papameletiou, D., del Pozo, C., Reina, V., Van Engelen, J., Heinemeyer, G., Viso, A.C., Rodriguez, C., Jantunen, M., 2007. Risk management measures for chemicals in consumer products—issues and challenges for documentation, assessment, and communication across the supply chain. J. Expo. Sci. Environ. Epidemiol. 1, 1–12. Bunke, D., Schneider, K., J¨ager, I., 2006. Exposure-based waiving concrete specifications of the waiving-conditions in the context of the registration procedure according to REACH, project report, ¨ Oko-Institut Freiburg. Cheeseman, M.A., Machuga, E.J., Bailey, A.B., 1999. A tiered approach to threshold of regulation. Food Chem. Toxicol. 37, 387–412. Combes, R., Barratt, M., Balls, M., 2003. An overall strategy for the testing of chemicals for human hazard and risk assessment under the EU REACH system. ATLA 31, 7–19. Cramer, G.M., Ford, R.A., Hall, R.L., 1978. Estimation of toxic hazard—a decision tree approach. Food Cosmet. Toxicol. 16, 255–276. De Heer, C., Wilschut, A., Stevenson, H., Hackkert, B.C., 1999. Guidance document on the estimation of dermal absorption according to a tiered approach: an update. TNO report V98 1237, p. 27, Zeist, The Netherlands. European Commission, 2001. White paper, strategy for a future chemicals policy. COM (2001) 88 final, Brussels. European Commission, 2003. Technical Guidance Document in Support of the Commission Directive 93/67/EEC on Risk Assessment for New Notified Substances and the Commission Regulation (EC) 1488/94 on Risk Assessment for Existing Substances and the Commission Regulation (EC) 1488/94 on Risk Assessment for Existing Substances and Directive 98/8/EC of the European Parliament and of the Council Concerning the Placing of Biocidal Products on the Market. Joint Research Centre, Institute for Health and Consumer Protection, European Chemicals Bureau, Ispra, Italy. European Commission, 2006. Regulation (EC) No. 1907/2006 of the European Parliament and of the Council of 18, December 2006. Fiserova-Bergerova, V., Pierce, J.T., Droz, P.O., 1990. Dermal absorption potential of industrial chemicals: criteria for skin notation. Am. J. Ind. Med. 17, 617–635. Frawley, J.P., 1967. Scientific evidence and common sense as a basis for food-packaging regulations. Food Cosmet. Toxicol. 5, 293–308.
76
U. Bernauer et al. / Toxicology Letters 176 (2008) 68–76
Gundert-Remy, U., Heinemeyer, G., Kunde, M., Platzek, T., 2005. EUChemikalienrecht und Verbraucherschutz, Proceedings zum ersten BfR-Forum Verbraucherschutz am 23. und 24. Juni 2005. Guy, R.H., Potts, R.O., 1992. Structure-permeability relationships in percutaneous penetration. J. Pharm. Sci. 81, 603–604. H¨ofer, T., Gerner, I., Gundert-Remy, U., Liebsch, M., Schulte, A., Spielmann, H., Vogel, R., Wettig, K., 2004. Animal testing and alternative approaches fort he human health assessment under the proposed new European chemicals regulation. Arch. Toxicol. 78, 549–564. Kroes, R., Kozianowski, G., 2002. Threshold of toxicological concern (TTC) in food safety assessment. Toxicol. Lett. 127, 43– 46. Kroes, R., Galli, C., Munro, I., Schilter, B., Tran, L.-A., Walker, R., W¨urtzen, G., 2000. Threshold of toxicological concern for chemical substances present in the diet: a practical tool for assessing the need for toxicity testing. Food Chem. Toxicol. 38, 255–312. Kroes, R., Renwick, A.G., Cheeseman, M., Kleiner, J., Mangelsdorf, I., Piersma, A., Schilter, B., Schlatter, J., van Schothorst, F., Vos, J.G., W¨urtzen, G., 2004. Structure-based thresholds of toxicological concern (TTC): guidance for application to substances present at low levels in the diet. Food Chem. Toxicol. 42, 65–83. 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 Chem. Toxicol. 34, 829–876. NCEA (National Center for Environmental Assessment), 2006. Approaches for the Application of Physiologically Based Pharmacokinetic (PBPK) Models and Supporting Data in Risk Assessment (EPA/600/R-05/043F), prepared by the National Center for Environmental Assessment (NCEA) within EPA’s Office of Research and Development (ORD). Norwegian Scientific Committee for Food Safety, 2006. Present and suggested use of the threshold of toxicological concern (TTC) principle in areas of risk assessment relevant for the Norwegian Scientific Committee for Food Safety (VKM). OECD, 1983. OECD Guideline for the testing of chemicals: 415 “OneGeneration Reproduction Toxicity Study” (adopted May 26, 1983), OECD, Paris.
OECD, 1984. OECD Guideline for the testing of chemicals: 417 “Toxicokinetics” (adopted April 4, 1984), OECD, Paris. OECD, 2001. OECD Guideline for the testing of chemicals: 414 “Prenatal Developmental Toxicity Study” (adopted January 22, 2001), OECD, Paris. OECD, 2001. OECD Guideline for the testing of chemicals: 416 “TwoGeneration Reproduction Toxicity Study” (adopted January 22, 2001), OECD, Paris. OECD, 2004. OECD Guideline for the testing of chemicals: 428 “Skin Absorption: In Vitro Method” (adopted April 13, 2004), OECD, Paris. Poulin, P., Theil, F.P., 2002. Prediction of pharmacokinetics prior to in vivo studies. II. Generic physiologically based pharmacokinetic models of drug disposition. J. Pharm. Sci. 91, 1358–1370. RIP3.2-1, 2007. WP1: development of the concept of exposure scenarios. General framework of exposure scenarios – Scoping Study – a final report. Technical Guidance Document on preparing the Chemical Safety Report under REACH – Scoping Study – Phase 1A. (REACH implementation project 3.2-1A). Service Contract Number 22551-2004-12 F1SC ISP BE. RIP3.3, 2007. EWG Repro report final—agreed edited version. Rulis, A.M., 1986. De minimis and the threshold of regulation. In: Felix, C.W. (Ed.), Food Protection Technology. Chelsea/Lewis, Michigan, pp. 29–37. The Nordic Council of Ministers, 2005. Threshold of toxicological concern (TTC). Literature review and applicability. TemaNord 559, Copenhagen, ISBN 92-893-1196-7. Available from: URL: http://www.norden.org/pub/miljo/sk/TN2005559.pdf. Van Engelen, J., Heinemeyer, G., Rodriguez, C., 2007. Consumer exposure scenarios: development, challenges and possible solutions. J. Environm. Expos. Anal. 17, S26–S33. van Raaij, M.T.M., Janssen, P.A.H., Piersma, A.H., 2003. The relevance of developmental toxicity endpoints for acute limit setting. RIVM rapport 601900004. WHO/IPCS, 2007. Draft guidance document on characterizing and communicating uncertainty in exposure assessment, in press. Wilschut, A., ten Berge, W.F., Robinson, P.J., McKone, T.E., 1995. Estimating skin permeation: The validation of five mathematical skin permeation models. Chemosphere 30, 1275–1296.