Scientific methodologies for the
assessment of combined effects of
chemicals – a survey and literature
review
Use of novel and alternative
methods in the assessment
of effects from combined
exposure to multiple
chemicals
Stephanie Bopp, Elisabet Berggren, Aude
Kienzler, Sander van der Linden, Andrew
Worth
2015
EUR 27471 EN
Scientific methodologies for the
assessment of combined effects of
chemicals – a survey and literature
review
This publication is a Technical report by the Joint Research Centre, the European Commission’s in-house science
service. It aims to provide evidence-based scientific support to the European policy-making process. The scientific
output expressed does not imply a policy position of the European Commission. Neither the European
Commission nor any person acting on behalf of the Commission is responsible for the use which might be made
of this publication.
JRC Science Hub
https://ec.europa.eu/jrc
JRC97522
EUR 27471 EN
ISBN 978-92-79-51925-3 (PDF)
ISSN 1831-9424 (online)
doi:10.2788/093511 (online)
© European Union, 2015
Reproduction is authorised provided the source is acknowledged.
All images © European Union 2015
How to cite: Bopp S, Berggren E, Kienzler A, van der Linden S, Worth A (2015) Scientific methodologies for the
combined effects of chemicals – a survey and literature review; EUR 27471 EN; doi:10.2788/093511.
Table of contents
ABSTRACT .........................................................................................................................4
1.
Introduction ........................................................................................................5
2.
Introduction to main concepts and terminology in the assessment of
mixtures / effects from combined exposure .......................................................6
2.1.
Similar vs dissimilar mode of action (MoA) and grouping of
chemicals ..............................................................................................7
2.2.
Interactions ...........................................................................................9
3.
New scientific tools for hazard assessment and how they could be used for
assessing mixtures/effects from combined exposure .......................................11
3.1.
Adverse Outcome Pathways (AOPs) .................................................11
3.2.
In vitro methods .................................................................................12
3.3.
Omics .................................................................................................13
3.4.
Quantitative Structure-Activity Relationships (QSARs) ...................14
3.5.
Read-across ........................................................................................16
3.6.
Toxicokinetic and toxicodynamic modelling .....................................18
3.7.
Dynamic Energy Budget (DEB) models ............................................19
3.8.
Threshold of Toxicological Concern (TTC) ......................................20
3.9.
Integrated Approach to Testing and Assessment (IATA) ..................21
4.
Status of current mixture risk assessment based on a dedicated expert
survey ...............................................................................................................22
4.1.
Information on respondents ...............................................................22
4.2.
General experience with mixture toxicity/risk assessment ................23
4.3.
Experience with the whole-mixture and component-based
approaches ..........................................................................................25
4.4.
Expert opinions on mixture toxicity assessment ................................27
4.5.
Use of novel tools in mixture toxicity assessment .............................30
4.5.1.
Use of
in vitro assays in mixture toxicity assessment ........................30
4.5.2.
Use of omics approaches in mixture toxicity assessment ..................32
4.5.3.
Use of (Q)SAR models in mixture toxicity assessment .....................34
4.5.4.
Use of read-across approaches in mixture toxicity assessment .........37
4.5.5.
Use of PBTK modelling in mixture toxicity assessment ...................38
4.5.6.
Use of the TTC approach in mixture toxicity assessment .................40
4.5.7.
Use of AOPs in mixture toxicity assessment .....................................41
4.5.8.
Use of DEB models in mixture toxicity assessment ..........................42
4.5.9.
Use of IATA frameworks in mixture toxicity assessment .................43
4.5.10. Summary on the use of novel tools in the assessment of mixtures ....44
4.6.
Frameworks for the risk assessment of combined exposure to
multiple chemicals .............................................................................44
4.7.
General/additional remarks by experts in the survey .........................45
4.8.
Conclusions from the expert survey...................................................45
5.
Conclusions ......................................................................................................47
2
REFERENCES ..................................................................................................................50
LIST OF ABBREVIATIONS AND DEFINITIONS ........................................................56
LIST OF FIGURES ...........................................................................................................58
3
Abstract
Exposure of humans and wildlife to chemicals via food, consumer products, the
environment etc. can imply exposure to an infinite number of different combinations of
chemicals in mixtures. It is practically impossible to test all these possible mixtures
experimentally and it is therefore needed to find smart strategies to assess the potential
hazards using new tools that rely less on
in vivo testing and incorporate instead
alternative experimental and computational tools. In this report the current state of the
art for the application of these alternative tools for assessing the hazard of chemical
mixtures is briefly reviewed. The focus is hereby on the adverse outcome pathway (AOP)
concept,
in vitro methods, omics techniques,
in silico approaches such as quantitative
structure activity relationships (QSARs) and read-across, toxicokinetic and dynamic
energy budget (DEB) modelling, and on integrated approaches to testing and
assessment (IATA).
Furthermore, an expert survey was performed to collect up to date information and
experience on the current use of different approaches for assessing human and
environmental health risks from exposure to chemical mixtures, with a view to informing
the development of a consistent assessment approach. An online survey was performed
among experts in the field of combined exposure assessment in the period of January to
March 2014, addressing both, human health and environmental risk assessment. Fifty-
eight experts from 21 countries, different stakeholder groups and sectors of legislation
participated in the survey. The main sectors where most experience is already gained in
assessing mixtures are in the area of plant protection products and chemicals under
REACH. These were also rated highest regarding the priority for performing mixture
assessments. Experts have experience with the whole mixture as well as the component-
based approaches applying them to both, intentional and unintentional mixtures. Mostly
concentration addition (CA) based methods are used for predicting mixture effects.
Regarding the use of novel and alternative tools in the risk assessment of mixtures,
expert opinions are split between those applying them (often more in a research
context) and those that generally think these tools are valuable but their use is currently
limited because of lack of guidance, lack of data, or lack of expertise. A general need for
clear guidance for combined exposure assessments was identified.
Overall, a high potential in applying novel tools and scientific methodologies for the
assessment of chemical mixtures can be identified. They allow deriving meaningful
information on individual mixture components or whole mixtures, enabling a better
understanding of the underlying mechanisms of mixture effects. Their main strengths lie
in their integrated use and smart combination to put different aspects regarding the
hazard from combined exposure to multiple chemicals into context. In order to benefit
from these tools in the hazard assessment of mixtures, more guidance on their use is
needed to facilitate a more widespread application.
4
1. Introduction
Humans and the environment are continuously exposed to a multitude of substances via
different routes of exposure. Some of these chemical mixtures are intentional and thus
have a known composition, e.g. personal care products, food additives and pesticide
formulations. However, in many cases, mixtures are unintentional and (largely) of
unknown composition, e.g. the combination of substances in surface water, drinking
water and air.
While many pieces of EU legislation are in place to protect humans and the environment
against adverse effects of chemicals including mixtures, in many cases it remains
unclear how to consider the combined exposure to chemical mixtures in the risk
assessment. Current regulatory requirements do not generally address the exposure to a
single substance by multiple pathways and routes of exposure, following its possible
different uses (i.e. the so-called "aggregate exposure"). Exposure to multiple
components from one or different exposure pathways (i.e. combined exposure) might
also pose a health problem even if the individual components are present at levels below
their respective NOAELs (No Observed Adverse Effects Levels), as these levels are
derived from single compound testing. However the different existing legislations do
often not take into account such risks or do not provide clear guidance on how to
perform risk assessment for aggregate and combined exposures. A detailed review of
regulatory requirements and related guidance can be found in Kienzler et al. (2014).
In order to reflect the actual exposure scenarios, there is a need to develop a consistent,
cross-sectorial approach to deal with the combined exposure to multiple chemicals. In
order to develop such a harmonised approach it is important to consider the current
scientific state of the art in this area. The objective of this report is to give an overview
of the current practices, tools and scientific developments in assessing risks from
combined exposure to chemical mixtures. For this purpose, the report provides some
general background information (Section 2), a brief overview of new scientific tools in
relevant areas of (eco)toxicology based on current literature (Section 3) and presents
the results of a recent survey among experts in the field of mixture risk assessment
(Section 4).
5
2. Introduction to main concepts and terminology in the
assessment of mixtures / effects from combined exposure
The hazard of chemical mixtures can be assessed as a whole (whole-mixture approach),
or based on the individual components of the mixture (components-based approach).
Whole mixture effects can be assessed by testing the mixture itself, but can also be
based on data generated with a mixture of similar composition (i.e. close in composition
regarding components and proportions). If adverse effects are found in relevant toxicity
studies, a quantitative assessment can be carried out directly from these data. This
approach allows consideration of any unidentified materials in the mixtures and any
interactions among mixture components, but it does not identify the chemicals
responsible for the mixture effects or interactions, and does not provide any information
on the toxicity of individual mixture components. Moreover, this approach is restricted to
mixtures that do not significantly change in their composition, and is therefore not
recommended as a general approach (SCHER, SCCS, SCENIHR, 2012).
Another approach, which is generally used when the components of the mixture are
known, is to mathematically predict the combined action of the components. The choice
of the mathematical approach to use depends mainly on considerations whether the
mixture components act by the same mode of action (MoA) or whether they are acting
independently (Groten et al., 2001). Its optimal use is therefore dependent on the
knowledge of the composition of the mixture and the corresponding MoA of the
individual components, or on the information regarding their association with groups of
chemicals demonstrating similar or identical MoA (assessment groups). Such information
may be based on chemical structures and structure-activity relationships (either
qualitative or quantitative), molecular modelling, structural alerts or on toxicological
responses or effects (SCHER, SCCS, SCENIHR, 2012).
Three basic types of action for combination of chemicals are usually addressed by: (i)
dose or
concentration addition (CA), applied to substances with a similar MoA; (ii)
independent action (IA) or response addition, applied to substances with a dissimilar
MoA; and (iii) interactions between substances in the mixture. The term interaction
includes all forms of joint action that deviate from the above additivity concepts. Hence,
the combined effect of two or more substances is either greater (synergistic,
potentiating, supra-additive) or less (antagonistic, inhibitive, sub-additive, infra-additive)
than that predicted on the basis of dose addition or response addition. Both, CA and IA,
are based on the assumption that substances do not influence each other's toxicity by
interacting at the biological target site, and they have been suggested as default
approaches in regulatory risk assessment of chemical mixtures (SCHER, SCCS,
SCENIHR, 2012), although chemical mixtures are rarely composed of either only
similarly or of only dissimilarly acting substance. SCHER, SCCS, SCENIHR (2012)
recommend applying CA if no MoA information is available, as it is regarded as more
protective.
Another way of addressing risks form combined exposure is to apply an additional
mixture assessment factor (MAF), which could be easily implemented in single substance
RA. Detailed information on the ongoing discussion can be found in Backhaus (2015). A
generally applicable MAF is hard to find due to the huge variety in mixture scenarios and
the need to be protective but not overly conservative. Therefore, Backhaus (2015)
investigated further possibilities to develop a protective MAF concept based on
addressing the most important uncertainties that are supposed to be covered by a
suitable MAF. These uncertainties are incomplete knowledge of the mixture composition
(compounds and concentrations), incomplete knowledge on hazard of mixture
components, possible synergistic interactions, and uncertainties related to the sole use
of CA. An exercise was performed to consider four different exposure scenarios with
mixtures of 15-42 components. It was observed that single substance RA and risk
management and mitigation significantly lowered the risk of the mixture, however was
6
insufficient to ensure protection against mixture effects. The Maximum Cumulative Ratio
(Price and Han, 2011) resulted as an adequate approximation for a MAF, which ranged
from 2 to 17 in the investigated examples, highlighting the need for considering specific
exposure scenarios. A scheme is suggested to determine the value of MAF depending on
the information available on the mixture (e.g. known number of components of the
mixture, information on individual risk quotients of the components, information on
interactions). Applying a MAF in the RA of single substances is however difficult since
appropriate risk management and risk mitigation measures might need to be developed
for scenarios in which many actors contribute to an overall risk with chemical emissions
that individually have a risk quotient below 1. The conclusion of the report is that the
risk quotient of an individual chemical should not only be viewed as a measure of risk in
itself but also as a measure of the contribution of the compound to the overall risk and a
combined exposure scenarios, overcoming the concept that chemicals with a risk
quotient below 1 are automatically safe even in complex exposure scenarios.
For further information on the underlying concepts please refer to e.g. Kortenkamp et al.
(2009), SCHER, SCCS, SCENIHR (2012) or Kienzler et al. (2014). Here, recent literature
regarding the approaches for mixtures of similarly and dissimilarly acting compounds are
further addressed, as well as regarding the considerations of interactions in the risk
assessment of chemical mixtures.
2.1. Similar vs dissimilar mode of action (MoA) and grouping of
chemicals
As mentioned above, usually mixtures of components with similar MoA are addressed
based on the concept of concentration addition (CA) and compounds with different MoA
are addressed based on the concept of independent action (IA). For deciding on the right
concept, the distinction of (dis)similar MoA and the related grouping of chemicals in a
mixture are crucial. Typically two main approaches are used for deciding if mixture
components act in a similar or dissimilar way and to perform related groupings of
mixture components: (1) investigating whether components follow a common MoA, or
(2) whether they elicit common phenomenological effects or affect the same target
organs.
EFSA's PPR Panel (2014) developed a methodology for forming cumulative assessment
groups (CAGs) for pesticides in the context of setting maximum residue levels for
pesticides in food. The proposed methodology follows a phenomenological approach
based on organ or system toxicity. Thus all pesticides causing a specific effect are
included in one CAG even if the underlying MoA is unknown. Due to the low exposure
levels of residues, interactions are not expected to occur and the PPR Panel based the
approach on concentration addition. EFSA's PPR Panel (2013) further discussed the
assessment of pesticides with dissimilar MoA, however restricting its considerations to
substances with dissimilar MoA but common adverse effects on the same organ or
system. The PPR Panel concluded from the reviewed literature that no case showed more
conservative predictions of combined toxicity using IA based approaches where at the
same time the predictions were also more accurate than based on CA. The use of IA for
predicting combination effects requires demonstration that modes of action of individual
substances in a mixture are strictly independent, a condition that can rarely be met in
practice. In addition, CA can be applied with existing data and has less data
requirements than IA (Kortenkamp et al., 2009). The PPR Panel therefore recommended
using cumulative risk assessment methods derived from CA also for the assessment of
mixtures of pesticides with dissimilar modes of action, provided they produce a common
adverse outcome.
Junghans et al. (2006) investigated the suitability of the two concepts of CA and IA to
predict the combined effects of realistic environmental mixtures. The exercise was
performed on a realistic exposure scenario for agricultural field run-off water considering
7
25 pesticides. Effects on the reproduction of the freshwater algae
Scenedesmus
vacuolatus were well predicted by CA, in accordance with the finding that toxicity was
dominated by a group of similarly acting photosystem II inhibitors, even if the mixture
included also pesticides with diverse and partly unknown MoA. Predictions based on IA
slightly underestimated mixture toxicity, however, the difference in predictions based on
CA and IA was rather small (factor of 1.3). The authors concluded that CA provides a
precautionary but not overprotective approach for combined effect predictions of
pesticide mixtures under realistic exposure scenarios, irrespective of the similarity or
dissimilarity of their mechanisms of action.
Apart from the combined effects of pesticides, the combination effect of endocrine
disruptors is relatively well researched. Kortenkamp (2007) reviewed literature on
combination effects of endocrine disruptors (EDs). Examples in the literature clearly
demonstrate that combinations of EDs with similar effects are able to act together at
doses that when used alone do not lead to observable effects. The experimental
evidence is in line with the assumption of dose addition. For EDs it seems best to follow
a phenomenological approach to produce workable grouping criteria. There are
arguments against using a toxicity equivalency factor (TEF) approach or other CA based
approaches for EDs since it would ignore potential interactions, however, there is
overwhelming evidence showing that groups of estrogenic, anti-androgenic, and thyroid-
disrupting chemicals act together in an additive way. For the time being it is proposed to
group EDs according to their ability to provoke similar effects rather than according to
similar mechanism of action. Given that the TEF concept unrealistically assumes parallel
dose-response curves it should not be used. Dose addition should be preferred for
calculating quantitative additivity expectations. Further research might trigger
adaptations to such a temporary approach. Comparatively little is known about
combined effects of EDs belonging to different classes, how these might interact and
produce combined effects. Also combinations with chemicals which are not producing the
same effects under analysis but that can modulate the effects of other chemicals should
be investigated further since such effects cannot be predicted by CA. It should be
explored whether the direction of modulation could be anticipated qualitatively e.g. by
analysing interaction at the level of metabolism and transport. Further research should
particularly focus on combinations of EDs that belong to different categories.
Apart from most of the recommendations and current practices of focusing on combined
effects on chemicals with similar MoA, recently also the relevance of combinations of
dissimilarly acting compounds was highlighted. Based on the relatively well studied
adverse effects of mixtures of pharmaceuticals, Hadrup (2014) suggested that chemicals
with dissimilar mechanisms of action could be of bigger concern in the toxicological risk
assessment of chemical mixtures than chemicals with a similar mechanism of action.
Examples such as e.g. in cancer and HIV treatment, show that pharmacological
combination therapy targeting different mechanisms of action is more effective than a
strategy where only one mechanism is targeted. Another argument is that also in many
diseases several organ systems concomitantly contribute to the pathophysiology,
implying that a grouping based on common target organs may be inadequate. In further
considerations, it should be however considered that in pharmacology usually higher
doses are applied, whereas at lower concentrations some specific effects might not
occur.
Goodson et al. (2015) reviewed actions on key pathways and mechanisms related to
carcinogenesis for 85 chemicals. The analysis suggested that the cumulative effects of
individual (non-carcinogenic) chemicals acting on different pathways and a variety of
related systems, organs, tissues and cells, could in combination produce carcinogenic
synergies. Additional basic research on carcinogenesis and low-dose effects of chemical
mixtures is needed. However, the concept of assessing combined effects strictly based
on grouping chemicals according to their MoA/AOP, might need to be revisited in order
not to underestimate cancer-related risks. Risk assessment for combined exposure
8
should consider synergies of chemicals acting via dissimilar processes, acting on
different targets and tissues, and consider synergies between certain pathways.
Overall, evidence in the literature supports the application of concentration addition as a
first, protective approach. It is therefore also the default approach to start from in
several international recommendations and frameworks, independent of components'
similar or dissimilar mode of action. However, once a detailed risk assessment for a
mixture is performed, chemical grouping should be considered and based on common
target organs and/or common mode of action (MoA). The choice of the approach
depends strongly on the context of the risk assessment as well as on the information on
which to base the grouping of components. Irrespective of the starting point for
grouping, it is recommended to use all available information on the mixture and its
components: physico-chemical properties, structural alerts, (Q)SAR and read-across
information, evidence from omics,
in vitro (high throughput screening or other) or
in
vivo experimental data, depending on availability.
2.2. Interactions
Toxicological interactions modulate toxicokinetic and/or toxicodynamic mechanisms of
individual chemicals. Toxicokinetic interactions could be e.g. induction of metabolising
enzymes, alterations in uptake mechanisms, all processes linked to influence of
individual chemicals on ADME of others. Toxicodynamic interactions can be based on e.g.
modulation of homeostasis or repair mechanisms.
Boobis et al. (2011) performed a literature review, identifying 90 studies demonstrating
synergisms in mammalian test systems performed at low doses (i.e. close to the point of
departure, POD) for individual chemicals. Only in 6 of the 90 studies useful quantitative
information on the magnitude of synergy was reported. In those six studies the
difference between observed synergisms and predictions by CA did not deviate by more
than a factor of 4.
Cedergreen (2014) performed a systematic literature review for binary mixtures within
three groups of environmentally relevant chemicals (pesticides, metals, antifouling
agents). Synergy was defined as a minimum two-fold deviation from CA predictions.
Synergy was found in 7%, 3% and 26 % of the pesticide, metal and antifoulant
mixtures, respectively. The extent of synergy was rarely more than factor of 10. Based
on some more in depth analysis Cedergreen concluded that true synergistic interactions
between chemicals are rare and often occur at high concentrations. Using standard
models as CA is regarded as the most important step in the RA of chemical mixtures.
Kamo and Yokomizo (2015) performed a modelling exercise addressing three theoretical
scenarios of non-interacting chemicals, directly and indirectly interacting chemicals. The
results showed that combined effects obey CA only when the MoA of the components of
the mixture are exactly the same. However, nonlinear effects vanished when the
chemical concentrations were low, suggesting that the current management procedure
of assuming CA is rarely inappropriate because environmental concentrations of
chemicals are generally low.
Approaches to address interactions and unravel the mechanisms are shown e.g. in
sections 3.6 and 3.7. In guidance by ECHA (2014) for biocidal products and EFSA's PPR
Panel (2013) for pesticides, a deviation between CA predictions and measured mixture
toxicities by a factor of 5 or more is regarded a synergistic/antagonistic interaction which
has to be considered further. More generic approaches to address interaction in mixture
hazard assessments look at the use of classical CA based methods and adding an
additional safety factor to account for possible (unidentified) interaction effects. This
might be an option for specific cases and compound classes as discussed e.g. in
Backhaus et al. (2013), where an interaction factor of 2 for biocidal mixtures is proposed
9
based on the evidence that in the majority of cases synergistic effects will not lead to
higher deviations from CA prediction.
Overall, evidence in the literature indicates that the interactions at lower concentration
levels such as environmental concentrations are rare and if observed, leading to
deviations from CA predictions that are relatively small. However, more knowledge could
be gained from additional case studies covering different sectors to further underpin this.
10
3. New scientific tools for hazard assessment and how they
could be used for assessing mixtures/effects from
combined exposure
Exposure of humans and wildlife to chemicals via food, consumer products, the
environment etc. can imply exposure to an infinite number of different combinations of
chemicals in mixtures. It is practically impossible to test all these possible mixtures
experimentally, especially
in vivo. Therefore, smart strategies are needed to assess the
potential hazards using new tools that rely less on
in vivo testing and incorporate instead
alternative experimental and computational tools. These tools are ideally simpler, faster,
and more robust in providing the necessary toxicological information of defined and /or
undefined mixtures.
In the following the applicability, benefits and limitations of the main current methods
and concepts are discussed in the context of hazard assessment of mixtures based on
recent literature.
3.1. Adverse Outcome Pathways (AOPs)
The Adverse Outcome Pathway (AOP) methodology is an approach which provides a
framework to collect, organise and evaluate relevant information on chemical, biological
and toxicological effects of chemicals. This approach supports the use of a mode (and/or
mechanism) of action basis for understanding adverse effects of chemicals (OECD,
2013). The approach is based on the concept that toxicity results from the chemical first
reaching and then interacting with an initial target or targets in the organism. Thus, an
AOP is a sequence of events, starting from the molecular initiating event (MIE; at
macromolecular level), via intermediate key events (KEs, at cellular or organ level) to
the
in vivo outcome of interest (adverse outcome, AO; at organism and population
level).
As described in the sections below, the prediction and assessment of mixture effects
often considers mechanistic information in order to determine whether mixture
components follow a similar or dissimilar mode of toxic action and hence should be
assessed together or not. This is most often used in order to group mixture components
and to decide whether to apply CA or IA based approaches for mixture effect predictions
(Borgert et al, 2004). Ankley et al. (2010) illustrated how effects caused by mixtures of
chemicals that act via the same molecular initiating event or affect pathways that
converge at common intermediate steps, can be aggregated for risk characterization.
Thus AOPs provide a valuable framework for grouping mixture components based on the
Mode of Action (MoA) into cumulative assessment groups (CAGs). Examples in the
literature sometimes show a grouping of chemicals based on similar target organs, but
often follow the AOP concept even if not presented as such. When chemicals are grouped
based on their MoA, one needs to keep in mind that depending on the dose ranges,
chemicals might produce different effects by different mechanisms (Borgert et al, 2004),
thus following different AOPs. Thus the mechanistic considerations need to take due
account of the relevant exposure concentrations.
In the Solutions project presented in Altenburger et al. (2015), water quality monitoring
is performed by a combination of chemical and bioanalytical tools for targeted and non-
targeted screening of components in environmental mixtures. The bioanalytical tools
should capture MIEs and KEs in order to inform about the toxic pressure
in situ. Effects
at various biological levels should be assessed (molecular, cellular, organism and
population level), which is in line with the AOP concept.
Using AOPs and toxicokinetic modelling, results from
in vitro testing can be put into
context and used to support mixture risk assessment with a reduced number of animal
testing (section 3.2).
11
3.2. In vitro methods
In vitro models usually consist of cell lines that are cultured in the laboratory under
standard conditions. The main advantage of
in vitro models is that they allow the study
of biological responses under such controlled conditions, where
in vivo models might be
influenced by non-chemical stressors that can make the assessment of chemical induced
effect(s) even more complex and challenging. In addition, most used cell lines can be
cultured relatively easily, they can be used in a high throughput setting, which makes it
possible to test for different effects and different combinations of compounds in parallel.
As such,
in vitro tools provide a pivotal role in toxicity pathway testing, as e.g. put
forward by the NRC report on Toxicity Testing in the 21st Century: A Vision and a
Strategy (NRC, 2007).
Most
in vitro tools currently applied in risk assessment consist of cell lines that are
designed to respond to specific effects, so called mechanistic assays. Generally, they
respond to the activation of receptors and/or specific pathways, as a result of e.g.
receptor activation or triggering of cellular repair mechanisms. As such, they can be
used to elucidate the mechanism(s) of action of a compound or combination of
compounds. By considering the effects in a broader context, e.g. as specific steps in an
AOP,
in vitro tests can provide important information on MoA/AOP, e.g. for subsequent
grouping of chemicals or for prioritizing compounds for risk assessment (Caldwell et al,
2014).
The application of
in vitro tools in the assessment of chemicals mixtures can be divided
into two approaches: top-down and bottom up. In the top-down approach,
in vitro assays are used to assess the overall amount of toxicity, effect, receptor activation etc.
triggered by a complex mixture. This effect-based environmental monitoring is
increasingly being applied to assess environmental mixtures, in part because of their
ability to give an overall response to the mixture present, and in part because the
compounds causing the effects were – and to a large extent still are – largely unknown
(Tang et al., 2013, 2014). Because of this,
in vitro tools are also widely used in
identifying previously unknown or unconsidered effects, but are also used to identify the
compound(s) responsible by combining sophisticated chemical analysis with
in vitro measurements, in a process frequently called Effect Directed Analysis (EDA) (Burgess et
al., 2013, Beyer et al., 2014).
While approaches like EDA start top-down, more bottom approaches are also utilized, in
which many chemicals are screened for activity in a wide range of
in vitro assays (Beyer
et al. (2014). Several recent initiatives have been launched to profile the effects of a
compound using a wide range of
in vitro assays, like the ToxCast program from the US
EPA, or the Tox21 by the National Institute of Health (NIH). These assays include assays
focusing on specific pathways or effects, e.g. mitochondrial toxicity (Attene-Ramos et al.,
2013a), cell viability and nuclear receptor assays including hormone receptors and
metabolic pathways (Huang et al., 2011) and various other endpoints (Tice et al., 2013).
While many of these initiatives initially focused on environmental chemicals, these
approaches are promising for all chemical risk assessment of many compounds,
including pharmaceutical, personal care products and food ingredients (Rovida et al.,
2015). Regardless of the approach, linking the total mixture toxicity measured
in vitro tools to compound concentrations is dependent on the mathematical model that is used
to describe the overall predicted effects based on individual concentration. Similarly,
in
vitro tools can be used to assess the validity of specific models to predict mixture
effects.
A major hurdle in the acceptance of results obtained by
in vitro tools is the question how
to translate
in vitro findings to adverse
in vivo effects (which is the actual protection
goal). As
in vitro studies generally cannot take into account some of the complexity of
the whole organisms, including uptake, metabolism and feedback mechanisms,
in vitro
12
to
in vivo extrapolation (IVIVE) is currently an important research topic (see e.g.
Adeleye et al., 2015). While
in vitro assays can provide important information regarding
the mode of action (the toxicodynamics), better predictions can be obtained by
extrapolating the
in vitro exposure conditions to the
in vivo situation using toxicokinetic
models (see section 3.6). Single
in vitro assays cannot cover all the parameters
necessary to make the translation to
in vivo. Dedicated
in vitro assays can help in
identifying and quantifying the parameters needed to validate the uptake, metabolism
and excretion models needed for the
in vitro to
in vivo translation.
While it is difficult to predict the actual
in vivo effects based solely on
in vitro concentrations,
in vitro tools might be used for regulatory screening purposes by relying
on a threshold (both human and environmental) below which relevant effects are not
expected to occur
in vivo. Different approaches have been put forward, mainly focusing
on endocrine pathways (Brand et al., 2013, Jarosova et al., 2014, Escher et al., 2015).
Based on their ability to predict
in vivo or chemical analysis results (Browne et al., 2015,
Sonneveld et al., 2006), several
in vitro tools are accepted by regulatory bodies for the
assessment of single compounds and mixtures. E.g. the US EPA is currently considering
the acceptance of ToxCast ER model data as alternative for the
in vivo Uterotrophic
Assay (Browne et al, 2015). Some regulation do already allow for the use of
in vitro assay measurements for mixtures. For example, EU regulations 589/2014 and 709/2014
specifically allow the use of
in vitro methods to give an indication of the total TEQ level
in food and feed respectively, expressing the results as Bioanalytical Equivalents (BEQ),
rather than the calculated TEQ based on individual congener concentrations analysed
chemically.
3.3. Omics
Omics techniques allow a global analysis of gene transcripts (transcriptomics, also called
gene expression profiling), of proteins (proteomics) and of small molecule metabolites
(metabolomics) including their relative abundance (see e.g. Schirmer et al 2011;
Villeneuve and Garcia-Reyero 2011). Omics are suitable to study effects at low doses
which are more relevant for environmental mixture exposure due their high sensitivity.
However, the effects observed at omics level need to be interpreted with care since the
molecular responses do not necessarily lead to an adverse outcome at the physiological
level (Borgert 2007; Beyer et al 2014). Furthermore, mechanistic information on the
mode of toxic action and affected pathways can be derived, which makes the tools
valuable in the context of mixture toxicity as well as single substance toxicity
investigations.
Altenburger et al (2012) reviewed literature on the application of omics techniques in
investigations of chemical mixtures. Among the 41 studies found (published 2002-2011),
most were transcriptomics studies. Many studies investigate the mode of action of single
substances and try to predict responses upon exposure to chemical mixtures. Omics
techniques can help identifying toxicological mechanisms of individual compounds by a
non-biased discovery driven approach (Beyer et al 2014). They can facilitate the
identification of key molecular events and complex sequential events caused by
stressors. They can support building a more complete overview of stress-response
profiles (e.g. toxicity pathways), both for single stressors and mixtures; identify key
MoAs; to mechanistically understand the potential for interactions between mixture
components; the selection of robust biomarkers for mixture prediction models in ERA.
In the reviewed literature, Altenburger et al (2012) found no clear relationship between
the exposure doses, the number of chemicals in a mixture and the number of related
affected gene responses. In some cases responses of specific pathways upon exposure
to individual compounds were replaced by more general stress response upon exposure
to mixtures. However, by delivering more mechanistic information also on individual
components, omics results can help in generating hypotheses on possible interactions
13
between mixture components (El-Masri, 2007). This can feed into the development of
mechanistic models used to simulate results that can be tested by model-designed
experiments. Dardenne et al. (2008a) used multi-endpoint bacterial gene profiling in
combination with cluster and principal component analysis in order to explore to what
extent compounds can be grouped according to their toxicological mechanism of action.
Several clusters of MoA could be identified and results be improved by combining
different clustering techniques. Projection of two environmental samples in the principal
component analysis (PCA) plane allowed identifying the mixed mode of action of these
samples, which can be useful for deriving first information on samples of unknown
composition.
Several studies try to also quantitatively interpret the data, using the concepts of
concentration addition and independent action as is usually done for apical endpoints.
The major limitation for applying these concepts to omics studies is the usually limited
number of tested concentrations (Altenburger et al., 2012). Another limitation that
hampers the application of a classical toxic unit approach, is the difficulty of deriving
effect concentrations (ECx) since the maximum induction levels for different genes vary
with different chemicals (depending e.g. on the cytotoxicity of a specific compound)
(Dardenne et al., 2008b). Dardenne et al. (2008b) investigated the effects of individual
substances and binary mixtures on 14 stress gene promoters. Mixture responses were
fitted applying both, CA and IA models. In many cases both models were able to predict
the mixture response from the individual compound responses. Differences between CA
and IA predictions were rather small. Deviations from CA and IA occurred, sometimes
with deviations being in opposite directions (i.e. synergistic or antagonistic) at high and
low dosage level. The choice of the best fitting model could not be made objectively
based on similar or dissimilar mode of action.
In summary, to date the major benefit of applying omics in the context of mixtures is to
use them for unravelling MoAs of the individual components in order to group them and
facilitate appropriate predictions of mixture toxicity.
3.4. Quantitative Structure-Activity Relationships (QSARs)
Quantitative Structure-Activity Relationship (QSAR) models can be used to obtain
information on the properties and activities of substances based on their chemical
structure alone, and can thus be used to fill data gaps in the safety assessment of
chemicals. Predictive approaches, such as QSARs, are essential for estimating mixture
toxicity as the number of possible mixtures is extremely large (Kim and Kim, 2015).
There are three main ways in which QSARs can be applied for the assessment of
mixtures: (1) for predicting (missing) information on individual compounds (physico-
chemical properties, toxicological effects) (2) for predicting directly or stepwise the
combined effects and interactions of chemicals in a mixture (3) for assessing whether
chemicals will act in a similar or dissimilar way to perform their grouping.
Altenburger et al. (2003) outlined how QSARs could support mixture toxicity evaluations.
All organic compounds (also those with specific MoA), will exert baseline toxicity
resulting from non-specific effects, related to partitioning into membranes and
adsorption to macromolecules. Thus all organic chemicals will contribute in an additive
way to baseline effects of a mixture also at very low concentrations. Thus a mixture will
be at least as toxic as corresponds to the sum of the fractional baseline toxic
concentrations of the components. This holds true however only in the absence of
antagonism that could result from metabolic detoxification of some components induced
by other components of the mixture. In a mixture with differently acting chemicals,
specific effects might not be triggered due to very low concentrations of the individual
components. However, their contributions to baseline toxicity will remain and might add
up to overall significant effects of the mixture. QSARs can therefore be used to predict
14
the baseline toxicity of a mixture and when comparing to measured toxicity of a mixture
help identifying if specific effects were occurring in addition.
Altenburger et al. (2003) also found that it is sometimes assumed that if one QSAR can
be applied to predict the toxicity of all mixture components, the compounds will by
default follow CA. However, this is not always the case since compounds may contribute
to different extent to narcotic and specific toxicity in a QSAR. Therefore, this assumption
is not generally valid.
When comparing experimentally measured mixture toxicities with those predicted by
QSARs, deviations observed were bigger for binary mixtures than for multiple mixtures
(Altenburger et al., 2003). This might be due to greater experimental variability for more
complex mixtures as well as due to an increasing degree of compensation of deviations
with increasing number of components in the mixture.
Direct prediction of mixture toxicity by QSARs is rather rare. It is only possible if the
detailed mixture composition is known. In some cases (Altenburger et al 2003),
predicted mixture toxicity is higher than experimentally determined, mainly due to
inadequate predictions for some individual components. If predictions are based on
observations at higher concentrations, this can lead to overestimations of toxicity, since
at low concentrations the specific effects of some compounds may not be triggered yet
(e.g. in the case of pesticides) and hence they act as nonspecific toxicants. If the
mixtures are composed of compounds that are predicted adequately individually, also
the joint toxicity is mostly predicted with a sufficient degree of accuracy.
QSARs can be used to discriminate classes of toxicants, i.e. to assess mixture
components for similar or dissimilar mode of action. QSAR-based tools to look for
functional similarities comprise molecular indices, topological indices and atom pairs,
physicochemical and quantum chemically derived stereoelectronic descriptors, together
with subsequent discriminant analysis. A sequential analysis applying the different
approaches can enhance the predictive capacity when dealing with several different
classes of chemicals (Altenburger et al., 2003). For example, Mwense et al. (2006) use a
set of molecular descriptors to determine the overall structural similarity and
dissimilarity within a mixture based on all the pairwise similarities and dissimilarities
between the constituent molecules. Then the degree of similarity vs dissimilarity is used
to weight the relative contributions of concentration addition and independent action in a
mathematical model based on both approaches.
QSARs can be used for modelling exposure concentrations and have been proposed for
calculating internal exposure concentrations by modelling internal distribution and
metabolism (Altenburger et al., 2003). This is especially important when chemicals are
reactive or interacting with protein macromolecules. For example, Verhaar et al (1997)
illustrated an approach to assess the toxicity of complex multi-component mixtures,
where QSARs can provide input parameters for PBPK models and a lumping analysis to
reduce mixture complexity to a limited number of compound categories.
Kim and Kim (2015) reviewed recently developed computational methods based on
QSARs for predicting mixture toxicity in environmental risk assessment (ERA). They
searched for related peer-reviewed articles published 2011-2013 in the fields of
toxicology, environmental science and ecology and engineering. They identified empirical
QSARs developed mainly based on partition coefficients. In the case of single substance
QSARs these are usually based on Kow, in the case of mixtures it is proposed to base
them on Kmd (i.e. the C18 EmporeDisk/water partition coefficient of the mixture). Kmd was
found to be promising for predicting the mixture toxicity of some chemicals (halogenated
benzenes, phenols, petroleum, PCBs, organochlorines, herbicides). However, these
QSARs can only assess EC50s of non-interacting mixtures ignoring synergistic effects.
Tang et al. (2013) proposed an approach for deriving effect based water quality trigger
15
values. Trigger values are derived in two steps, firstly estimating the individual
chemicals EC50 for non-specific MoA/baseline toxicity by QSARs and secondly calculating
mixture toxicity by CA. This represents a strategic approach to quantitatively derive
reference concentrations of mixture toxicity of different pollutants regulated by water
quality guidelines. Also non-empirical QSARs based on quantum chemistry and molecular
docking processes were identified. Quantum-based QSARs based on atomic charges
were developed that allow the prediction of mixtures at non-equitoxic concentrations of
the components. Two-step models are available that first assign chemicals via structural
similarity to relevant MoA and then calculate the toxicity for similar chemicals based on
CA and in a second step for the dissimilar groups based on IA.
In summary, QSARs can provide valuable input to assessing the toxicity of mixtures.
Some general challenges and limitations for application of QSARs to predict mixture
toxicities remain:
1. The principal difficulties in dealing with mixtures, limit the quantitative application of
QSARs in environmental field research (Altenburger et al., 2003). The restrictions
mostly relate to the characterization of proper QSAR input parameters, since so far
the impact of molecular properties on the mode of interaction in mixtures is
essentially unknown.
2. QSARs to date mainly use CA for mixture toxicity prediction; interactions (especially
synergisms) need to be further addressed. The lack of quality data (on molecular and
biological mechanisms) to increase understanding of synergism is essentially the
most critical challenge in modelling it. Molecular docking based QSAR models have
the potential to contribute to the prediction of synergisms. The main influencing
factors for synergisms are bioavailability, internal transportation, metabolization,
binding at the target site and excretion. Thus molecular docking theory seems most
promising to address this (Kim and Kim, 2013).
3. Most current QSAR developments focus on binary mixtures, QSARs enabling the
assessment of multi-component mixtures need to be further developed. Only two out
of eleven reviewed QSAR models were able to address multi-component mixtures
(Kim and Kim (2015).
4. It is important to acknowledge that the combined effect can be rather different when
considering predictions based on EC50 values instead of considering low dose effects
at concentrations below the NOEC (Kim and Kim, 2015). This is supported by the
observation that at low concentrations, the specific effects of these compounds may
not be triggered yet and hence they act as nonspecific toxicants (Altenburger et al.
2003). Most QSAR models for mixtures predict an EC50 for the mixture, which is
probably not relevant for environmental exposures. QSAR models that are able to
predict multi-point estimates around threshold effect levels should be developed
(Kim and Kim, 2015).
5. Most current QSARs for mixtures focus on acute rather than chronic toxicity. QSARs
based on molecular docking might help to improve chronic predictions, but at the
moment are only able to predict toxicity for binary mixtures (Kim and Kim, 2015).
Nevertheless, descriptive QSARs can be very useful for deriving basic understanding of
relevant interactions and molecular mechanisms. They can help in designing and
interpreting studies to link biological effects with chemical analysis (Altenburger et al
2003).
3.5. Read-across
In the following some general principles of read-across and specific issues to consider
when applying read-across to mixtures will be briefly discussed. For more detailed
general considerations please refer to the OECD Guidance on Grouping of Chemicals
(OECD, 2014) or to the ECHA Read-Across Assessment Framework (ECHA, 2015).
16
Read-across is a technique that allows predicting endpoint or test information for a
chemical (target chemical) based on the information available on the same endpoint or
test for one or more similar chemicals (source chemical(s)) (OECD, 2014). Two main
approaches for read-across are usually distinguished, i.e. the analogue and the category
approach. The analogue approach is usually used to read-across between a small
number of similar chemicals, in a simplest case from a single source chemical to a target
chemical. The category approach is usually applied to read-across between/within whole
groups of similar chemicals, mostly used for structurally similar groups of chemicals and
chemical families (ECHA, 2015).
Read-across can be applied for the prediction of various properties in order to fill data
gaps on e.g. physico-chemical properties, environmental fate, human health effects and
ecotoxicity. Read-across can be performed in a qualitative or quantitative way.
Qualitative predictions usually address the absence or presence of a certain property or
activity. Quantitative read-across instead predicts a value for a certain property or
endpoint, e.g. a dose-response relationship and effect concentrations (such as NO(A)EL,
LO(A)EL) (OECD, 2014).
The basis for read-across from one or more chemicals to another chemical or a group of
chemicals is the similarity in structure, properties and activities of the involved
chemicals. Structural similarity provides a convenient means of identifying likely
analogues. Similarity may be based on common functional groups, common chemical
class, or common precursor or breakdown products (i.e. similar metabolic or degradation
pathway). In addition to forming groups based on structural similarity, groups can be
further developed based on biological information. The adverse outcome pathway (AOP)
framework (see section 3.1) can help to group chemicals according to the molecular
initiating events (MIEs) or key events (KE) that they trigger. For this purpose it is not
needed that the whole sequence of events of an AOP is known, however, a reasonable
link needs to be made with the adverse outcome that is to be predicted by read-across
(OECD, 2014).
Read-across can be of value in the assessment of mixtures mainly in two ways:
• Read-across for untested constituents of a mixture in a component based
approach.
• Read-across for similar mixtures in a whole mixture approach.
For the first case, approaches as described above can be followed for the individual
mixture components. In mixtures of structurally diverse compounds, read-across for
several constituents might be an option. In cases of mixtures of substances of one
chemical class, a category approach might be followed to read-across among different
components of the class for which less information is available. An example is presented
for phthalates by Health Canada/Environment Canada (2015).
An example for the second case is the application of read-across to complex substances
such as MCS (Multi-Constituent Substances) and UVCBs (substances of Unknown or
Variable composition, Complex reaction products or Biological material). However, read-
across is limited to substances with sufficient knowledge about the composition (identity
and properties of constituents) and understanding of key structures that are determining
the mixture's behaviour. Category members are often grouped based on how these are
manufactured, defined and used, which can provide boundaries for the constituents
chemical space (OECD, 2014).
The OECD QSAR toolbox is a software application that allows identifying and filling of
data gaps for chemical hazard assessment. It comprises (eco)toxicological experimental
data and prediction tools which can be used for grouping of chemicals and data gap
filling (for details see OECD, 2009 and webpage with related tutorials). The OECD
17
toolbox allows also the assessment and prediction of mixtures if the chemical
components are known. One can enter the individual components, gather the available
experimental data on certain endpoints for the individual components and the toolbox
can provide a prediction of the endpoint for the mixture giving the choice of similar or
different MoA consideration. Thus this tool allows performing data gap filling by read-
across and mixture endpoint prediction in one workflow.
3.6. Toxicokinetic and toxicodynamic modelling
Toxicokinetics (TK) describe the concentration and time-dependent fate of a substance
within an organism whereas toxicodynamics (TD) describe the subsequent interaction
with biological targets and how this may lead to adverse health effects (Bessems et al.,
2015). In the context of human health and environmental risk assessment usually the
terminology of toxicokinetics and toxicodynamics is used, however, sometimes the terms
pharmacokinetics (PK) and pharmacodynamics (PD) are used synonymously depending
on the origin of the models and underlying data.
TK/TD considerations can support the assessment of chemical mixtures in several ways
with the main areas of application being:
• Determination of internal exposure concentrations, e.g. enabling a relation
between body concentrations and
in vitro experiments (i.e. IVIVE,
in vitro to
in
vivo extrapolations), of relevance for single chemicals as well as for chemical
mixtures.
• Considering the simultaneous or sequential exposure to different mixture
components, assessing the probability that those reach the same target.
• Predicting interactions among mixture components on TK and TD level.
The classically applied methods based on CA and IA lack a mechanistic basis and are
thus of limited utility for high-to-low dose or animal-to-human extrapolations. PBPK
models allow the extrapolation between doses, routes and species. PBPK/PD models can
also help in detecting shifts in the mechanism of action at varying doses and in
predicting interactions and their respective thresholds (El-Masri, 2007).
Most of the identified relevant studies available in the literature, investigate the utility of
physiologically based (PB) PK/PD models to assess and predict interactions of chemicals
in a mixture, i.e. looking at deviations from strict additivity. Tan et al. (2011) reviewed
PBPK/PD modelling efforts to investigate the chemical interactions at the PK and PD
levels. Most interactions studied to date focus on PK interactions. PK interactions mean
that mixture components influence each other's target tissue dose. This can occur if one
chemical in a mixture affects the absorption, distribution, metabolism or excretion
(ADME) of other components of the mixture, e.g. by inducing or inhibiting metabolising
enzymes, competing for transporters etc. PD interactions mean that mixture components
influence each other's target tissue response based on one unit of target tissue dose,
e.g. if one chemical impairs repair or homeostasis mechanisms. Examples can be found
in Tan et al. (2011); however, most examples address higher (occupational) exposure
where interactions are more likely to occur than at lower (environmental) exposure
levels.
PBPK/PD modelling is often hampered by the limited availability of input parameters.
Verhaar et al. (1997) proposed an integrated approach using PBPK/PD modelling with
QSAR analysis and lumping analysis to predict mixture toxicity. QSARs were used to
derive needed input parameters for unknown chemicals, e.g. partition coefficients,
metabolic rate constants, and pharmacodynamics parameters. Since with a higher
number of components in a mixture this becomes rather cumbersome, they proposed to
combine it with a lumping analysis to build categories to reduce the number of
components of the mixture to be addressed.
18
PBPK modelling can also support cumulative risk assessment including both, exposure to
multiple chemicals and non-chemical stressors. Wason et al. (2012) investigated the
cumulative risk for children exposed to multiple pesticides and non-chemical stressors,
such as dietary factors. The study developed a general framework for such approaches,
is however, not further discussed here since the focus is on chemical stressors.
In summary, the integration of TK in the assessment of mixture hazards is of value in
order to generate a better mechanistic understanding and currently mostly used to
predict and interpret interactions between mixture components.
3.7. Dynamic Energy Budget (DEB) models
The approach of dynamic energy budgets (DEBs) is applied in the ecotoxicology area. It
considers toxicity as a process over time that depends firstly on the build-up of the
internal concentration, and secondly on a hazard model that describes the adverse effect
as a chance process, e.g. using a killing rate (Løkke et al., 2013). Compared to the
classical approaches based on CA and IA that interpret effects on different endpoints
separately, DEB models have the advantage that by their mechanistic basis they allow
extrapolating experimental results to e.g. other time points, time-varying exposures,
other mixtures, other organisms, and other (non-chemical) stressors, such as e.g. food
limitation (Baas et al., 2010).
DEB models mostly use few parameters which have a clear biological meaning (e.g. such
as killing or elimination rates). In order to develop DEB models, the results of a time
series of toxicity endpoints (e.g. data on survival) are needed, which might add extra
costs; however, this allows some increased understanding of the underlying mechanisms
and better predictions of mixture effects.
The first step in the DEB model addresses toxicokinetics, to determine the internal
concentration looking at uptake and elimination in a one compartment model. Then a
model description of the processes in the organism follows looking at assimilation,
maintenance, growth, reproduction, death. Resulting effects on the organism can lead to
changes in toxicokinetics for other mixture components and are considered in a feedback
loop. The advantage of this modelling framework is that it can be applied to a large
variety of different species, in contrast to e.g. PBPK/PD models which are very species
specific (Baas et al., 2010). For each component in the mixture, three toxicity
parameters are needed: the no effect concentration, the killing rate and the elimination
rate. In addition one extra parameter to correct for the mortality in unexposed controls
is needed. Among different examples of applying the DEB model to binary mixtures, it
was also proven to work for more complex mixtures. Baas et al. (2009) addressed
effects of mixtures of 80 different compounds in surface water on Daphnia, where 92%
of the cases were correctly predicted. Using this approach the chemicals driving the
effects could also be identified.
Most examples in the literature are dedicated to predicting mortality and similar
endpoints. Studies by Jager et al. (2010, 2014) showed however that DEB modelling can
be also applied for sub-lethal endpoints elicited by mixtures.
In summary, for the time being DEB models are not yet regularly used in the
assessment of mixtures. They are however a promising tool, since they look at effects in
a more integrated and mechanistic way. However, more work has to be done to specify
what type of information is needed to identify the various mechanisms of action, and to
quantify the importance of a correct choice for the population effects (Jager et al.,
2014).
19
3.8. Threshold of Toxicological Concern (TTC)
The Threshold of Toxicological Concern (TTC) is a methodology that can be applied to
assess potential human health concerns for chemicals based on their chemical
characteristics and estimated exposure. It is applicable to substances for which the
chemical structure is known but for which there are few or no relevant toxicity data. The
classification of chemicals according to their chemical structure is an essential
component of the current TTC approach (Cramer classes, Cramer et al., 1978, Munro et
al., 1996). It was first developed for substances in food contact materials. EFSA's
Scientific Committee (2012) investigated the applicability of the TTC approach for its
own work (i.e. food-related risk assessments) and recommended a revision and
refinement of the Cramer Classes and underlying database. Nevertheless, the Cramer
classification scheme is considered conservative and protective of human health (EFSA's
Scientific Committee, 2012). EFSA's Scientific Committee has identified several cases
where the TTC approach should not be applied, one of them being "mixtures of
substances containing unknown chemical structures".
In their opinion on the "Toxicity and Assessment of Chemical Mixtures" (SCHER, SCCS,
SCENIHR, 2012), the Scientific Committees discussed the application of the TTC
approach in the assessment of mixtures. They recommend the TTC approach could be
used at a screening level for comparing first estimates of combined exposure to the TTC.
For representative substances in assessment groups where data on limit values are
missing, QSAR predictions, read-across, or the TTC approach could be used to fill the
data gaps. A TTC-like approach can be used to eliminate combinations that are of no
concern, if conservative exposure concentrations are used.
Exemplary case studies exist in the literature where the TTC approach was used to
assess human health risk from chemical mixtures in surface water. Price et al. (2009)
explored the use of the TTC approach for the evaluation of the chronic non-carcinogenic
effects of hypothetical and actual examples of chemical mixtures and it proved to
provide conservative estimates of mixture toxicity. They therefore propose to use the
TTC in screening assessments of mixtures where compound specific data for components
of a mixture are missing. Along these lines, also Boobis et al. (2011) showed the
application of the TTC approach in a Tier 0 risk assessment with the intention to
prioritize the need for further evaluation of chemical mixtures. Data were based on
surface water monitoring data in order to create a hypothetical mixture of 10 compounds
from different classes (fragrances, pesticides, surfactants, personal care products,
solvents, petrochemicals). They applied some worst case assumptions (i.e. direct
consumption of surface water without treatment and investigating lifelong chronic
exposure at maximum detected levels). The TTC using ToxTree and the concentration
addition based Hazard Index were applied. In the specific case, no risk was identified.
The case study confirmed the utility of using the TTC approach at Tier 0 as suitable tool
for mixture assessments. Terry et al. (2015) showed the application of the TTC approach
for facilitating the risk assessment of parent substances and their environmental
metabolites using a pesticide as an example. The TTC approach was used for some
metabolites with low predicted concentrations. Utilizing information on mode of action,
relative potency, hazard characterisation, read across, predicted exposure and TTC
provided a robust database minimizing animal use for the assessment.
In summary, the TTC approach provides a useful tool to be used at lower tiers in the
assessment of mixtures, providing limit values for mixture components with missing
information. It is currently limited to its application in the area of human health.
However, an international activity is currently ongoing to develop the ecoTTC approach
for environmental hazard assessment, which is based mainly on aquatic toxicity data
(Belanger et al. 2015).
20
3.9. Integrated Approach to Testing and Assessment (IATA)
With the more regular use of new techniques like e.g.
in vitro testing, omics approaches,
computational methods, there is a need to develop strategies for evaluating the data
generated and interpret them in a joint approach. There is also a need to strategically
direct testing efforts in order to save resources. Efforts for developing such strategies
are ongoing in various fora and usually include a framework to integrate test and non-
test information in a weight of evidence approach. Under the OECD, these approaches
are called Integrated Approaches to Testing and Assessment (IATA). They "integrate
existing knowledge based on classes of chemicals with the results of biochemical and
cellular assays, computational predictive methods, exposure studies, and other sources
of information to identify requirements for targeted testing or develop assessment
conclusions. In some cases, the application of IATA could lead to the refinement,
reduction, and/or replacement of selected conventional tests (e.g., animal toxicity
tests)." (NAFTA, 2012).
The development of IATA is nowadays strongly linked to the development and
availability of AOPs. The AOPs offer the biological framework to build around the testing
and assessment strategy (Tollefsen et al., 2014). A testing and data interpretation
strategy can be developed by addressing MIEs and KEs in an AOP.
Related testing strategies can follow a battery approach (all tests performed and results
collected), sequential or tiered approaches (results are collected in a given sequence and
further testing is stopped when sufficient information is available), or result-driven
further testing approach (depending on results next most valuable testing is decided).
The integration of results from the different information sources can occur at different
levels, i.e. form raw data to summary/category level. Different approaches
(deterministic, decision trees, scoring approaches etc.) can be applied (Tollefsen et al.,
2014).
IATA provide another framework to collect information on individual mixture components
as well as on whole mixtures, allowing a more structured (and if AOP based more
mechanistically relevant) way of data generation and interpretation.
21
4. Status of current mixture risk assessment based on a
dedicated expert survey
In order to gain an overview of the current practices and experiences with assessing
effects and risks from combined exposure, a survey among experts in authorities,
academia and industry was performed. The online survey was published in the EU survey
platform (https://ec.europa.eu/eusurvey/) on 23/01/2015 and closed on 23/03/2015.
The link to the survey was sent to experts in the field that were identified in the
following ways: (1) from scientific literature, (2) experts involved in developing the
WHO/ICPS framework for combined exposure to chemical mixtures, (3) experts involved
in developing the Opinion of the SCHER/SCHENIHR/SCCS on "Toxicity and Assessment
of Chemical Mixtures", (4) participants of the recent EFSA Scientific Colloquium N°21 on
"Harmonisation of human and ecological risk assessment of combined exposure to
multiple chemicals", (5) representatives of the ongoing OECD project on combined
exposure to multiple chemicals, (6) members of the EURL ECVAM Stakeholder Forum
(ESTAF) and the PARERE expert network dedicated to the Preliminary Assessment of
Regulatory Relevance of alternative test methods proposed for validation.
The questionnaire contained the following sections:
• General information on the respondent
• Experience with different approaches for the assessment of mixture toxicity
• Expert opinions regarding the use of specific approaches for mixture toxicity
assessment
• Views on the use of novel tools for the assessment of mixture toxicity
• Comments on existing frameworks for the risk assessment of combined exposure
to multiple chemicals
• Possibility to provide references or files regarding relevant projects, publications,
case studies.
4.1. Information on respondents
Fifty-eight valid responses were received and evaluated. Responses were received from
48 experts from 16 different EU countries and 10 experts from 5 non-EU countries
(Figure 1). Survey participants represented experts from academia, authorities, and
industry in nearly equal parts (Figure 2). Participants had experience in the following
sectors: chemicals (multiconstituent and UVCB substances under REACH), plant
protection products, biocides, pharmaceuticals, cosmetics, food or feed additives, food or
feed contaminants, surface water, drinking water, waste streams, soil, air, medical
devices, alloys, botanicals, cigarette smoke, landfill leachate, solid waste from industrial
combustion processes, tobacco toxicants, jewellery, toys, food contact materials, and
sewage sludge.
22
Country of respondent
2 3
10
Austria
Belgium
2
3
Denmark
France
4
Germany
Greece
1
Italy
Latvia
10
Netherlands
Portugal
5
Romania
Slovenia
3
Spain
Sweden
2
7
1
2 2
United Kingdom
Other
1
Figure 1 Country of Respondent. Survey participants were from 16 different EU
countries and 5 non-EU countries (indicated as "Other": Canada, Japan, Norway, Serbia,
Switzerland, US).
Type of Affiliation
2
Academia
16
Authority
20
Consultancy
Industry
Other
3
17
Figure 2 Survey respondents' affiliation
4.2. General experience with mixture toxicity/risk assessment
The majority of the respondents (48) had performed risk assessment(s) for chemical
mixtures in the area of human health risk assessment (HRA) or environmental risk
assessment (ERA) (Figure 3). Those that replied yes (i.e. had already performed mixture
risk assessment) indicated that they had mainly assessed chemical mixtures in the
context of the authorisation of commercial products and research & development (Figure
4).
23
Did you ever need to perform a RA of chemical mixtures for HRA or
ERA?
Yes
1
3
4
No, because the safety factors in single
2
compound assessment provide sufficient
protection against mixtures.
No, because there is a lack of legal
requirements.
No, due to the large knowledge gap on
the toxicity of the individual compounds
48
No, because of other reasons.
Figure 3 Replies to the question "Did you ever need to perform a risk assessment of
chemical mixtures for HRA or ERA?"
For which purpose(s) did you assess the overall toxicity of chemical
mixtures?
0
10
20
30
Authorisation of commercial products (i.e.
26
prospective RA)
Research and development
25
General exposure monitoring
15
RA of contaminated sites
12
Priority setting of risk reduction measures
6
Control of emission permits
4
Control of remediation works and their success
6
Other
2
Figure 4 Replies to the question "For which purpose(s) did you assess the overall
toxicity of chemical mixtures?" Other purposes indicated were 1) development of CEFIC
MIAT conceptual framework and CLP workplace monitoring.
24
4.3. Experience with the whole-mixture and component-based
approaches
Those participants having experience in performing mixture risk assessments (i.e. that
answered "yes" to the question in Figure 3) were asked about their experience in
applying whole mixture and component-based approaches for assessing intentional and
unintentional mixtures (Figure 5). They further specified for which type of commercial
product (intentional mixture) or sample (unintentional mixture) they applied such
approaches (Figure 6) and which kind of component-based approaches they are mostly
using (Figure 7).
Have you ever applied a mixture assessement using whole
mixutre/component based approaches for un/intentional mixtures?
60
11
11
11
12
50
40
12
17
19
28
30
20
35
30
27
10
19
0
whole mixture approach / whole mixture approach /
component based
component based
intentional mixtures
unintentional mixtures
approach / intentional
approach / unintentional
mixtures
mixtures
no reply
No
Yes
Figure 5 Experience in performing mixture risk assessments using whole mixture or
component-based approaches applied to intentional or unintentional mixtures.
25
Intentional mixtures
Unintentional mixtures
0
10
20
0
5
10
13
5
Chemicals
Sediment
3
12
5
Soil
8
14
PPPs
16
4
Food contaminants
9
6
4
Cosmetics
Surface Water
9
3
2
Air
5
5
Biocides
5
1
Biota
3
Food / feed
2
1
Drinking Water
5
additives
1
1
Ground Water
3
1
Pharmaceuticals
1
2
Industrial Waste Water
1
1
3
Municipal Waste Water
Others
2
3
0
Feed contaminants
0
0
Marine Water
1
whole-mixture approach
6
Other
component-based approach
6
Figure 6 Replies to the question "For which type of mixture(s) have you applied a whole
mixture or component-based approach?" Chemicals were further specified in the survey
as "multiconstituent or UVCB substances under REACH". Other types of mixtures
mentioned were medical devices, alloys, botanicals, food ware materials, jewellery, toys,
cigarette smoke/tobacco toxicants, human tissue extracts, landfill leachates, solid waste
from industrial combustion processes, sewage sludge.
Respondents were asked which kind of tests they used for the whole-mixture and
component based approaches for assessing intentional and unintentional mixtures. A
wide range of tests reflecting all major tests usually used in single substance testing was
mentioned.
Experts were further asked about their experiences with different component-based
approaches. The approaches most frequently used by the participating experts are the
direct application of the Concentration Addition (CA) equation, the Hazard Index (HI) as
well as Toxic Equivalency Factor (TEF) (Figure 7). Concentration Addition based
approaches seem much more wide spread than independent action (IA) based
predictions.
26
How often have you used the following component-based approaches?
60
50
34
40
37
34
46
43
53
50
50
47
50
54
30
20
9
12
15
10
10
15
6
6
12
5
6
6
6
6
5
4
2
0
1
2
3
2
5
2
Direct
TUS
TEF
RPF
PODI
HI
HI(Int)
MCR
Other CA
Direct
Other
Application
based
application approaches
of CA
approaches of the IA
formula
formula
Never+no reply
Occassionally
Frequently
Figure 7 Replies to question "How often have you used the following component-based
approaches?" (CA=concentration addition, TUS=Toxic Unit Summation, TEF=Toxic
Equivalency Factor, RPF=Relative Potency Factor, PODI=Point of Departure Index,
HI=Hazard Index, HI(int)=Hazard Index including interactions, MCR=Maximum
Cumulative Ratio, IA=Independent Action)
Experts could also provide information on additional component-based approaches they
are using. For example, ADIadjusted, AOELadjusted, ARfDadjusted (based on common
target organ/common toxicity; use of biotic ligand models (BLM); use of TTC and Margin
of Safety (MoS) were mentioned.
Experts were then asked about their experience with other approaches not fitting into
the CA or IA category. Some examples mentioned were that mixtures were assessed and
toxicity related to the main two constituents; effect-directed fractionation (e.g. toxicity
identification evaluation TIE) for identifying most potent components in a mixture;
approaches based on TKTD modelling, to assess dynamic effects of mixtures over time;
history of safe use via well-established dietary intakes. These additional approaches
were mainly used for research and development purposes.
4.4. Expert opinions on mixture toxicity assessment
Experts were asked about their opinions on mixture toxicity assessment. Experts gave
their view on priority mixtures (Figure 8). Chemicals under REACH and plant protection
products were mentioned most often. Experts were further asked whether they see a
need to address interactions (e.g. synergisms/antagonisms) in the mixture risk
assessment. Multiple options with possible reasons to take combination effects into
account or not were given in the survey. Interestingly, several experts selected yes and
no options together. The overall picture can be found in Figure 9. Among the experts
that selected the "other" option, the main reasoning provided in the free text option was
that interactions are considered rare, especially at low/environmental concentrations.
However, most experts agreed that interactions should be taken into account on a case-
by-case basis and one expert proposed to look at interactions especially in the case of
active substances like plant protection products, biocides, pharmaceuticals.
27
Chemicals
10
11
3
3
1
PPPs
11
10
3
1 2
Pharmaceuticals
6
6
5
2 1
Biocides
4
9
3
2 1
Food or feed contaminants
6
8
2
3
academia
Surface Water
5
5
3
1 1
authority
Drinking Water
7
5
1 1
industry
Waste streams
3
3
4
consultancy
Food or feed additives
2
3
1
3
Soil
3
5
1
other
Air
3
3
2
Cosmetics
1
3
1
Other
1 1
0
5
10
15
20
25
30
Figure 8 Replies to the question "Which type of mixture(s) or samples would you
identify as highest priority for risk assessment that needs to take mixture effects into
account?" divided by stakeholder group. Chemicals were further specified in the survey
as "multiconstituent or UVCB substances under REACH". Other mixtures of importance
mentioned were those present in human tissues and container systems.
Based on your experience or knowledge about the toxicity of mixtures, do you
consider it important to include interactions (e.g. potential synergism or
antagonism) in the assessment?
0
10
20
30
Yes, but only if there is strong evidence from data
3
9
14
2 2
available, read-across etc.
Yes, case-specific interaction information should be
9
7
8
2
incorporated in the assessment.
No, it is sufficiently protective to base mixture
3
2 1 1
assessments on Concentration Addition models.
academia
No, interactions are rare and do not need to be addressed
3
3
specifically.
authority
No, but a default conservative safety factor should be
industry
applied which does not rely on case-specific interaction
2 1
information.
consultancy
Other
1 1
7
other
Figure 9 Replies to the question "Based on your experience or knowledge about the toxicity
of mixtures, do you consider it important to include interactions (e.g. potential synergism or
antagonism) in the assessment?" divided by stakeholder group. Among the experts that
selected the "other" option, the main reasoning provided in the free text field was that
interactions are considered rare, especially at low/environmental concentrations. However,
most experts agreed that interactions should be taken into account on a case-by-case basis
and one expert proposed to look at interactions especially in the case of active substances
like plant protection products, biocides, pharmaceuticals.
28
Experts were asked about approaches that they consider particularly valuable or have
abandoned based on their experience as well as information on software tools and
databases relevant to assessing mixture effects (Figure 10).
0
20
40
60
Are there any approaches or methodologies for
assessing the risk of chemical mixtures which you
4
53
1
have used or evaluated in the past but which you
have abandoned due to negative experience?
Are there any approaches or methodologies for
assessing the risk of chemical mixtures which you
consider particularly valuable for specific samples,
29
27
2
endpoints, or other purposes and which you
would recommend for a more extensive use in EU
Member States?
Are there any software tools or databases or
online resources that you find particularly useful
15
40
3
in carrying out mixture assessments?
Yes
No
no reply
Figure 10 Replies to the questions on approaches the experts have abandoned due to
negative experience, specific approaches that the experts would recommend for further
considerations, and on use of relevant software tools and databases.
The approach being mentioned most often as being abandoned was mainly Independent
Action with the general reason that it is too data demanding. Specific approaches that
were recommended for more extensive use were the MCR approach (Price and Han,
2011), mixture toxic pressure approach (e.g. De Zwart and Posthuma, 2006), whole
mixture testing at human relevant levels, use of TTC,
in vitro methods, TKTD and DEB
modelling. Software tools and databases mentioned of use for carrying out mixture
assessments were CREME, CARES, MIXTOX, DEBtox, ToxTree, DEREK, Acropolis tool
MCRA, US EPA BMDS, ToxCalcMix, Metals Classification Tool MeClas.
29
4.5. Use of novel tools in mixture toxicity assessment
Experts were asked about the use of novel tools in the assessment of mixtures. An
overview of the general responses regarding their use can be found in Figure 11.
Do you apply novel tools in the RA of mixtures?
60
2
2
3
3
5
3
3
3
50
28
27
40
30
no reply
33
No
46
41
43
47
30
Yes, for HRA and ERA
7
7
Yes, for ERA
5
1
20
2
3
Yes, for HRA
6
11
1
1
10
19
1
4
18
2
2
15
6
7
9
7
9
2
0
Figure 11 Replies to the question "Do you apply
in vitro tools/omics approaches/
(quantitative) structure activity relationships ((Q)SARs)/ read-across/ physiologically
based pharmacokinetic (PBPK) modelling/ the toxicological threshold of concern (TTC)
concept/ Adverse Outcome pathways (AOPs)/ dynamic energy budget (DEB) models for
human health risk assessment (HRA), environmental risk assessment (ERA) or both?"
Experts were then asked further on their reasons for using specific tools or not using
them. Results are presented in the following sections 4.5.1-4.5.9.
4.5.1.
Use of in vitro assays in mixture toxicity assessment
In vitro tools are increasingly applied, both in human and environmental risk
assessment. Not only because they are powerful tools to investigate TK processes, but
also because they can provide key information on the mechanism of action. Especially
with complex mixtures like environmental samples, they might provide insight in toxicity
beyond the traditional chemical analysis.
28 experts replied that they use
in vitro assays in the assessment of mixtures. The tests
mentioned were tests from the OECD endocrine disruptor testing framework, cell-based
transactivation assays, irritation/sensitisation assays, epiocular assay, dermal
absorption, Ames test, micronucleus assay, genotoxicity, mutagenicity, comet assay,
ROS production, cytotoxicity, microtox, algae assays, zebrafish embryo tests, bacterial
reporters, rodent and human cell lines and human tissue cultures.
The reasons for using
in vitro assays are presented in Figure 12. Among the other
reasons, one expert replied that the use of
in vitro tools combined with PBPK modelling
is
the way forward for chemical risk assessment. Other 28 experts replied that they are
not using
in vitro assays and were asked for the reason(s) (Figure 13). The main reasons
30
for not using
in vitro methods are the lack of legal drivers and lack of guidance and
validation.
What are the main reasons you are using these in vitro assays?
0
5
10
15
The results are quickly available
15
The mechanism of action tested has a clear in
15
vivo relevance
The method is cost efficient
11
The method allows for whole mixture testing
10
The results are easy to interpret
8
In vivo method should be the last resort
8
The assay(s) can provide key TK information
2
Other reason(s)
4
Figure 12 Replies to the question "What are the main reasons you are using these
in
vitro assays?" Other reasons mentioned were for research purpose, Ames test because it
is a data requirement in some regions, and combination of
in vitro testing with PBPK
modelling is seen as the way forward for chemical risk assessment.
31
What are the main reason(s) you are not using in vitro assays?
0
5
10
The method lacks ecological relevance
7
There are no legal drivers for this method
7
The method is not validated
6
The method lacks (in vivo) predictability (in vitro to in
4
vivo extrapolation)
The method lacks the proper metabolisation steps
4
There is no generally accepted protocol (e.g. OECD) for
4
this test
The method is too costly
3
The results from the test are difficult to interpret
2
The method is too complicated
1
The samples are generally too complex to test reliably
1
There are no suitable reference matierals available
1
There is a lack of proper QA/QC for this method
1
The method is too labour intensive
0
The method is not reliable enough
0
The method takes too much time to produce results
0
Other reason(s)
11
Figure 13 Replies to the question "What are the main reasons you are not using
in vitro assays? The main "other reasons" given were lack of EU guidance and validation, and in
the case of cosmetics that
in vitro tests are less relevant for the final products as clinical
tests are performed with cosmetic formulations.
4.5.2.
Use of omics approaches in mixture toxicity assessment
Omics technologies are increasingly applied to gain insight in the mechanism of action of
compounds and mixtures, at the transcription level (transcriptomics), the protein level
(proteomics) or even the whole metabolome (metabolomics). Genomic tools can
potentially be used also to investigate differences in responses between individuals and
species.
10 experts had experience with using omics tools in mixture risk assessment. Most
experience is available for transcriptomics, and to a lesser extent with proteomics and
metabolomics (Figure 14). The main reasons for using omics technologies are presented
in Figure 15. The main advantages mentioned were the possibility to study overall
effects and ability to gain mechanistic information as well as the sensitivity of the
methods. 46 experts replied they were not using omics technologies for mixture risk
32
assessment. The main reasons for not using omics are presented in Figure 16, with one
of the main problems identified being the lack of clear guidance and protocols.
Which type of omics technologies are you using?
5
Transcirptomics
Proteomics
10
Metabolomics
4
Figure 14 Replies to the question "Which type of omics technologies are you using?"
What are the main reasons you are using omics based approaches?
0
2
4
6
The approach allows an overall analysis of effects
6
The method is very sensitive (detects changes
6
before toxicologically relevant)
The approach allows for whole mixture testing
5
The approach has a clear in vivo relevance
2
The method is very cost efficient
1
Other reason(s)
2
Figure 15 Replies to the question "What are the main reasons you are using omics
based approaches?" The other reasons given were the possibility to study mechanisms of
action.
33
What are the main reason(s) you are not using omics based assays?
0
5
10
15
There are no legal drivers for this method
13
There is no generally accepted protocol (e.g. OECD)
13
for this method
The method is not validated
10
The data analysis is too complex
9
The method lacks (in vivo) predictability
9
The method is too complicated
7
The method lacks ecological relevance
6
There are no suitable reference materials available
3
The method is too costly
3
The method is too labour intensive
2
The samples are generally too complex to test
2
reliably
The method is not robust enough
1
There is a lack of proper QA/QC for this method
1
The method takes too much time to result
0
Other reason(s)
12
Figure 16 Replies to the question "What are the main reasons you are not using omics
based assays?" The main "other reason" provided was the unavailability of the relevant
practical tools in the facilities of the responding experts, as well as the lack of clear
guidance.
4.5.3.
Use of (Q)SAR models in mixture toxicity assessment
(Quantitative) structure activity relationship ((Q)SAR) models are mathematical models
that have been developed to predict a number of physico-chemical and (eco)toxicological
properties of chemicals without performing
in vitro or
in vivo tests. The models use
information on the structure of the chemical, and combine this with a database with
toxicological information on related/comparable chemicals.
25 of the participating experts are using (Q)SAR in the mixture risk assessment. They
consider (Q)SARs very useful because they represent an alternative to
in vivo studies to
some extent and they can serve for prioritising experimental testing, and results are
quickly available (Figure 17). The purposes for which the experts apply (Q)SAR methods
are shown in Figure 18 with the main purpose being hazard identification. The endpoints
for which (Q)SARs are mainly used are endocrine activity in cell based transactivation
assays, mortality, reprotoxicity, acute (eco)toxicity, mutagenicity, genotoxicity, cancer
34
alerts, skin sensitisation, toxicokinetics, secondary poisoning, logKow, logKoc, BCF,
biodegradation, fate and behaviour. Experts were also asked to provide information on
the QSAR tools they are using (Figure 19).
30 experts replied that they are not using QSARs for mixture assessments. The main
reasons for not using (Q)SARs are presented in Figure 20.
What are the main reasons you are using (Q)SARs?
0
5
10
15
The results are quickly available
14
The method can replace in vivo testing (to
13
some extent)
The method is cost efficient
12
The mechanism of action considered has a
6
clear in vivo relevance
Other reason(s)
4
Figure 17 Replies to the question "What are the main reasons you are using (Q)SARs?"
"Other reasons" mentioned were the application to metabolites and impurities where no
data are available, and the possibility to identify substances of concern and drive further
experimental testing. Also the use for chemical fate aspects was mentioned.
For what purpose(s) are you applying (Q)SARs?
0
5
10
15
20
Hazard identification
21
Priority setting
9
Mode of Action identification
7
Exposure assessment
5
Other
5
Figure 18 Replies to the question "For what purpose are you applying (Q)SARs? "Other
reasons" mentioned were data gap filling, prediction of physic-chemical properties,
grouping of metabolites and impurities, and fate predictions.
35
Which (Q)SAR models do you apply?
0
5
10
15
OECD Toolbox
14
Derek (Nexus)
11
ECOSAR
8
Toxtree
7
TOPKAT
5
CAESAR
5
Danish EPA QSAR database
5
Lazar
2
OASIS TIMES
2
MultiCASE
1
DEMETRA
1
OncoLogic
0
HazardExpert
0
Meteor (Nexus)
0
Metabolexpert
0
CHEMPROP
0
Other(s)
7
Figure 19 Replies to the question "Which (Q)SAR models do you apply?" Other models
mentioned were DEBtox, VEGA platform models, ACD Labs Percepta, Episuite, SarPy.
36
What are the main reasons you are not using (Q)SARs?
0
5
10
The method is not properly validated
9
No good models are available for the inteded application
8
Too much expert judgement is needed to run the
7
method and interpret results
The method lacks (in vivo) predictability
5
There are no legal drivers for this method
5
Appropriate software is not available to the user
4
Competing models show conflicting results
2
There is a lack of information regarding the identity of
2
the compounds
The method lacks ecological relevance
0
There is a lack of proper QA/Q for this method
0
The method is too costly
0
Other reason(s)
11
Figure 20 Replies to the question "What are the main reasons you are not using
(Q)SARs?". The main "other reasons" given were that databases are not validated, lack
of guidance, (Q)SARs are applied to single substances but not to mixtures because only
qualitative QSAR information is used, (Q)SARs are not relevant because of available
testing information (e.g. for PPPs), too many uncertainties associated with (Q)SARs, lack
of knowledge/training.
4.5.4.
Use of read-across approaches in mixture toxicity
assessment
Read-across is a technique for predicting endpoint information for one substance by
using data on the same endpoint from (an)other substance(s). This can be performed
with a limited set of substances (analogue approach) or within a large group of
substances (category approach).
28 experts replied that they are using read-across in the assessment of mixtures. The
main reason given for using read-across was to replace
in vivo testing to some extent
(Figure 21). Regarding the endpoints for which read-across is applied, experts were
mostly answering that they use read-across for any kind of endpoint where
needed/information is suitable and where read-across is accepted by ECHA for REACH
assessments. Endpoints mentioned include the whole spectrum of short and long-term
animal study endpoints, as well as ecotoxicity. 27 experts replied they are not using
read-across approaches, mainly because of lack of expert knowledge (Figure 22).
37
What are the main reasons you are using a read-across approach?
0
5
10
15
The method can replace in vivo testing (to some extent)
16
The results are quickly available
11
The method allows for whole mixture consideration
9
The method is very cost efficient
9
The mechanism of action tested has a clear in vivo
5
relevance
Other reason(s)
6
Figure 21 Replies to the question "What are the main reasons you are using a read-
across approach?" Main "other reasons" were that read-across is a useful tool for the use
for metabolites and impurities, where toxicity information is limited, it is encouraged
under REACH, can be used to support
in vitro test results and refine experimental
design.
What are the main reasons you are not using a read-across approach?
0
5
10
The method requires too much expert
9
information/expert knowledge
The approach is non-formalised
8
The results from the read-across approach are difficult
7
to interpret
The method lacks (in vivo) predictability
5
There are no legal drivers for this method
5
The method is too complicated
2
The method lacks ecological relevance
0
Other reason(s)
11
Figure 22 Replies to the question "What are the main reasons you are not using a read-
across approach?" Main "other reasons" given were a lack of guidance and validation.
4.5.5.
Use of PBTK modelling in mixture toxicity assessment
Physiologically based (PB) models are used for modelling toxicokinetic (TK) processes
(PBTK) or TK and toxicodynamic (TD) processes (PBTKTD). The PBTK models are
especially useful to assess hazard, as they provide a quantitative means to address TK
processes. PBTKTD models link the TK and TD dimension and therefore are generally
more complex.
38
12 experts replied that they are using PBTK modelling in the assessment of mixtures.
The main reasons for using PBTK models were the replacement of
in vivo testing to a
certain extent, since PBTK models allow relating
in vitro experiments to
in vivo internal
exposures (Figure 23). The main purpose for which PBTK modelling is used in mixture
assessment is to correlate
in vitro concentrations to
in vivo scenarios/to link internal and
external dose, to identify TK interactions, to understand metabolite generation, species
comparison, route of exposure, assess effects of intermittent exposure often mentioned
in the context of PPPs.
What are the main reasons you are using a PBTK model?
0
2
4
6
The method can replace in vivo testing (to some
6
extent)
The method has a clear in vivo relevance
3
The method is cost efficient
2
The results are quickly available
1
Other reason(s)
3
Figure 23 Replies to the question "What are the main reasons you are using a PBTK
model?" The main "other reasons" were to study mechanisms, study mechanisms by
which mixtures affect life-history traits, assess intermittent exposures
What are the main reasons you are not using PBTK models?
0
5
10
15
20
Data on many important parameters are lacking
16
The method is too complicated
11
The method is not validated
10
There are no legal drivers for this method
9
The method takes too much time to produce results
7
The method is not robust enough
3
The method lacks (in vivo) predictability
3
There is a lack of proper QA/QC for this method
1
Other reason(s)
19
Figure 24 Replies to the question "What are the main reasons you are not using a PBTK
model?" The main "other reasons" given were lack of knowledge/expertise/training, lack
of guidance and tackling related uncertainties.
39
41 experts are not using PBTK modelling for mixture assessments. The main reason for
not using PBTK modelling is a lack of knowledge/training as well as the lack of many
input parameters (Figure 24).
4.5.6.
Use of the TTC approach in mixture toxicity assessment
The approach of the threshold of toxicological concern (TTC) is based on historical
toxicological data, that show empirically that there is a threshold below which toxicity
does not occur (for non-cancer effects) or likelihood of tumour incidence is negligible
(cancer effects). There are thresholds derived for different classes of compounds
(Cramer classes).
22 experts replied that they apply the TTC approach in the assessment of mixtures. The
main reason is that it can allow the waiving of additional testing (Figure 25). 33 experts
are not using the TTC approach with the main reason being that it was developed for
human health assessments and not for environmental assessments (Figure 26).
What are the main reasons you are using the TTC approach?
0
5
10
15
The method can waive additional testing
14
The results are quickly available
13
The method is very cost efficient
10
Other reason(s)
5
Figure 25 Replies to the question "What are the main reasons you are using the TTC
approach?" Main "other reasons" were the availability for chemicals without chronic data,
the use to identify and prioritise testing needs, avoiding testing, and application to
ingredients of biological origin and residual impurities.
40
What are the main reasons you are not using the TTC approach?
0
5
10
15
The method was developed for human health only,
11
not for environmental assessments
Difficulties in assigning compound to the right class
7
There are no legal drivers for this method
7
The method lacks (in vivo) predictability
5
The method is not applicable to the compounds under
3
consideration, e.g. hormones
There is a lack of information on the identity of the
3
compounds
Other reason(s)
13
Figure 26 Replies to the question "What are the main reasons you are not using the
TTC approach?" Main "other reasons" were the need of updating the TTC approach, the
lack of experience, and dealing with higher substance concentrations.
4.5.7.
Use of AOPs in mixture toxicity assessment
An Adverse Outcome Pathway (AOP) is an analytical construct, describing a sequential
chain of causally linked events at different levels of biological organisation that lead to
an adverse health or ecotoxicological effect. AOPs might provide insight into the
relevance of combinational effects when assessing the toxicity of mixtures.
12 experts replied that they are using AOPs in the assessment of mixtures. The main
reasons for applying AOPs were that the method shows insight into potential interactions
and that the considered mechanisms have a clear
in vivo relevance (Figure 27). 43
experts are not applying AOPs with the main reason being the limited availability of
relevant, sufficiently developed AOPs (Figure 28).
What are the main reasons you are using an AOP approach?
0
2
4
6
8
10
The method shows insight in potential
9
interactions
The mechanism of action considered has a
9
clear in vivo relevance
The approach has clear ecological relevance
2
Other reason(s)
1
Figure 27 Replies to the question "What are the main reasons you are using an AOP
approach?" The "other reason" mentioned was to explore its potential for risk
assessment.
41
What are the main reasons you are not using an AOP approach?
0
5
10
15
20
Relevant AOPs are not available or not sufficiently
20
developed
The approach is non-formalised
10
There are no legal drivers for this method
8
The results are difficult to interpret
5
Other reason(s)
18
Figure 28 Replies to the question "What are the main reasons you are not using an AOP
approach?" The "other reasons" mentioned are mainly the lack of expertise/experience,
and difficulties to implement this new concept in risk assessment.
4.5.8.
Use of DEB models in mixture toxicity assessment
Dynamic Energy Budget (DEB) models aim to identify (simple) quantitative rules for the
processes of uptake of substrate by organisms and the use for maintenance, growth,
maturation and reproduction. By linking a TK model to the DEB, effects of chemicals can
be incorporated as well.
8 experts replied they are using DEB models in the assessment of mixtures with the
main reason being the inclusion of multiple stressors, the ecological relevance, and the
in vivo relevance (Figure 29). 47 experts were not using DEB models with the main
reason being on the one hand that it is a concept usually applied in ecotoxicology rather
than for human health assessments and on the other hand the lack of expertise and the
lack of validation (Figure 30).
What are the main reasons you are using DEB models?
0
1
2
3
4
The mechanism of action under consideration has a clear in vivo
4
relevance
The method has a clear ecological relevance
4
The method allow the inclusion of multiple stressors
4
The method is very cost efficient
3
The results are quickly available
1
Other reason(s)
1
Figure 29 Replies to the question "What are the main reasons you are using DEB
models?" The "other reason" was stating that DEB models are the most promising
approach to interpret and predict effects of stressors on growth and reproduction and to
learn from experimental testing.
42
What are the main reasons you are not using DEB models?
0
5
10
15
20
There are no legal drivers for this method
9
Much of the required information for the model is
7
lacking
The method is too complex
6
The method is not robust enough
5
The method lacks (in vivo) predictability
1
The method lacks ecological relevance
1
Other reason(s)
24
Figure 30 Replies to the question "What are the main reasons you are not using DEB
models?" Main "other reasons" were a general lack of expertise/experience, for experts
in human RA its focus on environmental RA, and a lack of validation.
4.5.9.
Use of IATA frameworks in mixture toxicity assessment
IATA provide a framework to integrate existing knowledge based on classes of chemicals
with the results of biochemical and cellular assays, computational predictive methods,
exposure studies, and other sources of information to identify requirements for targeted
testing or develop assessment conclusions.
8 experts had experience in applying IATA frameworks to the assessment of mixtures
(Figure 31). The IATA mentioned are skin and eye irritation, integrated testing strategies
(ITS) described in the ECHA guidance and some tailored customised approaches.
Do you use any IATA framework for the assessment of
mixtures?
3
8
Yes
No
no reply
47
Figure 31 Replies to the question "Do you use any IATA framework for the assessment
of mixtures?"
43
4.5.10. Summary on the use of novel tools in the assessment of
mixtures
Many experts are already using several of the new tools. Some are still more frequently
used as research activity, but start being applied also in regulatory context. Many tools
are considered as promising but not yet ready for regulatory purposes. Often mentioned
reasons for not applying novel tools are the lack of legal drivers, lack of standardisation /
validation, lack of guidance, and a lack of expertise.
4.6. Frameworks for the risk assessment of combined exposure
to multiple chemicals
Several international frameworks for addressing combined exposure to chemical
mixtures were developed in recent years. Experts were asked in the survey about their
experience with the most widespread three frameworks, i.e. the WHO/IPCS framework
(WHO/IPCS 2009, Meek et al 2011), Proposal by the three non-food scientific
committees of the European Commission (SCHER, SCENIR, SCCS, 2012), and the
proposal by CEFIC MIAT (Price et al., 2012). 35 experts were familiar with at least one of
the three frameworks. Most experience was shown for the WHO/IPCS framework (Figure
32).
Which framework(s) do you apply?
WHO/IPCS 2009, Meek et al 2011
9
Proposal by the three non-food
Scientific Committees of the European
20
Commission from 2012
8
Proposal by CEFIC Mixtures Industry Ad-
Hoc team (MIAT) 2012
Figure 32 Replies to question "Which [international] framework(s) [for addressing
combined exposure] do you apply?"
Experts were then asked to provide feedback on their experience with the above-
mentioned frameworks. The experiences were generally positive. The main limitation
mentioned especially for the WHO/IPCS and the SCHER/SCENIHR/SCCS framework is
that they provide a more conceptual framework and less practical guidance.
The experts commented positively on the WHO/IPCS, rating it an easy and transparent
approach. Critics were that it is rather general and lacks criteria when refinement should
be stopped. The data available usually allow only to perform Tier 1 and 2 assessments
and not to go to higher tiers. However, it was appreciated that Tier 0 and Tier1 can
usually be performed with the data available.
44
The SCHER, SCENIHR, SCCS framework is considered useful for organising data and
deciding how to perform the assessment, but it is more conceptual and provides limited
guidance on refined assessments.
The CEFIC MIAT framework was judged as useful since it comprises practical tools. Most
input received on this framework was from experts involved in its development.
Experts were then asked about other frameworks they are applying for assessing
combined exposure. 15 experts replied they had experience with other frameworks, 40
replied to have no experience with other frameworks (Figure 33). The most often
mentioned example was the cumulative assessment for plant protection products as
developed by EFSA (EFSA PPR Panel, 2013). Furthermore, experts mentioned the US
EPA, transitional guidance on mixture toxicity assessment of biocidal products (2014);
EFSA guidance on Birds and Mammals (2009); Backhaus and Faust 2012 (Environ. Sci.
Technol., 46 (5)); UK Combined toxicity assessment for plant protection products (non-
dietary and dietary risk assessment); Methodology and the information to be specified
for the assessment of pesticides used in tank-mixtures subject to prior assessment in
accordance with the French Ministerial Order of 7 April 2010; Metals Classification tool
(MeClas)/Transformation Dissolution Protocol.
Are you familiar with or do you apply any other framework to
assess the toxicity of mixtures?
3
15
Yes
No
no reply
40
Figure 33 Replies to the question "Are you familiar with or do you apply any other
framework to assess the toxicity of mixtures?"
4.7. General/additional remarks by experts in the survey
In the end of the questionnaire, experts had the possibility to add some general remarks
or provide additional information in the form of references or uploading files. Experts
were providing many useful references.
4.8. Conclusions from the expert survey
The expert survey was a success with 58 experts participating and providing extensive
information in the free text fields. The results from the survey allow to derive a clear
picture on the current status of assessing mixtures.
45
The main sectors where most experience is already gained in assessing mixtures are in
the area of plant protection products and chemicals. These were also rated highest
regarding the priority for performing mixture assessments. However, mixture
assessments are also performed in many other areas.
Experts have experience with the whole mixture as well as the component-based
approaches applying them to both, intentional and unintentional mixtures.
Mostly concentration addition (CA) based methods are used for predicting mixture
effects. In contrast, several experts would not recommend the further use of
independent action (IA) based approaches, mainly because of the higher need for input
data for IA and considering the small differences in predictions by IA compared to CA.
Experts were asked about their opinion on considering interactions. Most experts stated
that interactions should be considered if there is specific evidence for interactions and on
a case-by-case basis.
Regarding the use of novel tools in the risk assessment of mixtures (such as
in vitro methods, omics, (Q)SARs, read-across, PBTK modelling, TTC approaches, AOPs, DEB
models, IATA), expert opinions are split between those applying them (often more in a
research context) and those that generally think these tools are valuable but their use is
currently limited because of lack of guidance, lack of data, or lack of expertise.
Experts had experience with assessing mixtures, both in the context of human health
and environmental risk assessment. Apart from some tools that were developed for a
specific application in HRA (e.g. TTC approach) or ERA (e.g. DEB modelling), there
seems to be no clear difference in the opinions/experiences provided.
A general need for clear and harmonised guidance for combined exposure assessments
can be identified from the survey.
46
5. Conclusions
Humans and the environment are continuously exposed to a multitude of substances via
different routes of exposure. However, the risk assessment of chemicals for regulatory
purposes does not generally take into account the “real life” exposure to multiple
substances, but mainly relies on the assessment of individual substances. A previous
review on regulatory requirements for the assessment of mixtures shows that combined
exposure is nowadays taken up in several pieces of legislation, however a harmonised
consistent approach on performing mixture assessments across different regulatory
sectors is still lacking (Kienzler et al., 2014).
Assessment in different sectors
Our expert survey showed that mixture risk assessment is taken up in various fields. The
main sectors where most experience is already gained in assessing mixtures are in the
area of plant protection products (PPPs) and chemicals falling under REACH, which is
however also linked to the fields of work of the respondents. PPPs and chemicals under
REACH were also rated highest regarding the priority for performing mixture
assessments by the experts. The survey respondents seemed to rate active substances
(such as biocides, plant protection products and pharmaceuticals) overall of higher
concern, also regarding a potential presence of interactions (i.e. including the risk of
synergistic effects). However, mixture assessments are also performed in many other
areas. Looking at the available experience derived through the survey and at case
studies in the literature (Kienzler et al., unpublished report), many examples can be
found where environmental exposure or occupational exposure to mixtures were
retrospectively assessed, based on monitored exposure data. Examples in the
prospective risk assessment are rarer. One major achievement in this area is
represented by the cumulative risk assessment and respective cumulative assessment
groups as developed by EFSA's PPR Panel (EFSA PPR, 2014). Further development of
consistent approaches for prospective risk assessment is still needed.
Component-based assessment approaches: concentration addition vs independent action
Currently, many experts have experience with the whole mixture as well as the
component-based approaches applying them to both, intentional and unintentional
mixtures, without a clear trend in their combination. Regarding component-based
approaches, survey respondents mostly use concentration addition (CA) based methods
for predicting mixture effects. Several experts would even not recommend the further
use of independent action (IA) based approaches, mainly because of the higher need for
input data for IA and considering the small differences in predictions by IA compared to
CA. Overall, evidence in the literature supports the application of concentration addition
as a first, protective approach. It is therefore also the default approach to start from in
several international recommendations and frameworks, independent of components'
similar or dissimilar mode of action. However, once a detailed risk assessment for a
mixture is performed, relevant groupings will be based on common target organs and/or
MoA (e.g. based on AOPs). The choice of the approach used depends strongly on the
context of the risk assessment as well as on the information on which to base the
grouping of components. Irrespective of the starting point for grouping, it is
recommended to use all available information on the mixture and its components:
physico-chemical properties, structural alerts, (Q)SAR and read-across information,
evidence from omics,
in vitro (high throughput screening or other) or
in vivo experimental data, depending on availability.
Considering interactions in mixture assessments
Current evidence in the literature suggests that interactions (synergistic or antagonistic
effects) at lower concentration levels such as environmental concentrations are rare and
47
if observed, leading to deviations from CA predictions that are relatively small. In the
survey, experts were asked about their opinion on considering interactions. Most experts
stated that interactions should be considered if there is specific evidence for interactions
and on a case-by-case basis. Experts commented that the highest potential for
interactions should be assumed for active substances (such as PPPs, biocides,
pharmaceuticals). Some experts answered that they consider interactions rare and
hazard sufficiently covered using CA based approaches. Only very few experts agreed to
introduce a default safety factor to cover potential interactions, which is sometimes
proposed in the literature (for overview see Backhaus, 2015). However, more knowledge
could be gained from additional case studies covering different sectors to further
underpin this.
To address and predict interactions, toxicokinetic and toxicodynamic modelling are
valuable tools. PBTK and DEB modelling can support gaining further mechanistic
understanding and looking at effects of individual chemicals in a mixture in an integrated
approach. Toxicokinetic information to feed into these models, can be gained e.g. from
in vitro studies. Also read-across information from similar mixtures can be used to
identify mixtures where interactions could play a role and should be further investigated.
Application of "novel tools" in the assessment of mixtures
In this report the current state of the art of the application of alternative tools for
assessing the hazard of chemical mixtures was briefly reviewed. The focus is hereby on
the adverse outcome pathway (AOP) concept,
in vitro methods, omics techniques,
in
silico approaches such as quantitative structure activity relationships (QSARs) and read-
across, toxicokinetic and dynamic energy budget (DEB) modelling, and on integrated
approaches to testing and assessment (IATA). A brief summary of the main possibilities
for the application of each of these tools in the context of mixtures is given below:
• AOPs are already used for the grouping of chemicals according to their MoA, which is
an important step in the assessment of mixtures. Apart from this, they can help in
putting results from
in vitro tests and computational modelling into context, e.g. in
developing AOP based IATA, that allow for the integration of different data types.
•
In vitro methods can support the assessment of mixtures in many ways, mainly in
"whole-mixture testing" applied e.g. in effect-based environmental monitoring or in
deriving relevant information on individual mixture components. Performing high-
throughput screening of many chemicals in many different assays enables the
characterisation of chemicals regarding their MoA which can help e.g. in the grouping
of chemicals. AOPs can help in integrating data from diverse
in vitro tests that
address different steps in a chain of biological events. If
in vitro test results are
interpreted in connection with toxicokinetic information, even quantitative
information on potency can be derived.
• The main potential of
omics techniques (transcriptomics, proteomics, metabolomics)
regarding the assessment of mixtures lies in the investigation of affected pathways
for unravelling MoAs and investigating possible interactions, which can again support
the grouping of chemicals.
• QSAR models can be used to obtain information on the properties and activities of
substances from chemical structure alone, and can thus be used to fill data gaps in
the safety assessment of chemicals. There are three main ways in which QSARs can
be applied for the assessment of mixtures: (1) for predicting (missing) information
on individual compounds (physico-chemical properties, toxicological effects) (2) for
predicting directly or stepwise the combined effects and interactions of chemicals in a
mixture (3) for assessing whether chemicals will act in a similar or dissimilar way.
• Read-across can be of value in the assessment of mixtures mainly in two ways, i.e.
to predict missing information for untested constituents of a mixture in a component
based approach, or to read-across for similar mixtures in a whole mixture approach.
48
• Toxicokinetics (TK) and toxicodynamics (TD) considerations can support the
assessment of chemical mixtures in several ways with the main areas of application
being (1) determination of internal exposure concentrations, e.g. enabling a relation
between body concentrations and
in vitro experiments (i.e. IVIVE,
in vitro to
in vivo extrapolations), of relevance for single chemicals as well as for chemical mixtures,
(2) considering the simultaneous or sequential exposure to different mixture
components, assessing the probability that those reach the same target, and (3)
predicting interactions among mixture components on TK and TD level.
• The approach of dynamic energy budgets (DEBs) is applied in the ecotoxicology area.
For the time being DEB models are not yet regularly used in the assessment of
mixtures. They are however a promising tool, since they look at effects in a more
integrated and mechanistic way, potentially integrating chemical and non-chemical
stressors.
• The TTC approach is recommended in the literature for use at a screening level
mixture assessment, for comparing first estimates of combined exposure to the TTC.
It can serve eliminating combinations that are of low concern and to prioritise
mixtures for further assessment. Currently, the TTC approach is limited to application
in human health risk assessment; however, a corresponding ecoTTC approach is
under development.
• In the context of the risk assessment from combined exposure, IATA provide another
framework to collect information on individual mixture components as well as on
whole mixtures, allowing a more structured (and if AOP based more mechanistically
relevant) way of data generation and interpretation.
In the survey, expert opinions regarding these methodologies and tools were split
between those applying them (often more in a research context) and those that
generally think these tools are valuable but see their use as currently limited because of
a lack of guidance, lack of data, or lack of expertise.
Overall, a high potential in applying novel tools and scientific methodologies for the
assessment of chemical mixtures can be identified. They allow deriving meaningful
information on individual mixture components or whole mixtures, enabling a better
understanding of the underlying mechanisms of mixture effects. Their main strengths lie
in their integrated use and smart combination to put different aspects regarding the
hazard from combined exposure to multiple chemicals into context. In order to benefit
from these tools in the hazard assessment of mixtures, more guidance on their use is
needed to facilitate a more widespread application.
49
References
Adeleyea Y, Andersen M, Clewell R, Davies M, Dent M, Edwards S, Fowler P, Malcomber
S, Nicol B, Scott A, Scott S, Sun B, Westmoreland C, White A, Zhang Q,
Carmichael PL, 2015. Implementing Toxicity Testing in the 21st Century (TT21C):
Making safety decisions using toxicity pathways, and progress in a prototype risk
assessment. Toxicology, 332, 102–111.
Altenburger R, Ait-Aissa S, Antcak P, Backhaus T, Barceló D, Seiler TB, Brion F, Busch
W, et al., 2015. Future water quality monitoring — Adapting tools to deal with
mixtures of pollutants in water resource management. Science of the Total
Environment, 512–513, 540–551.
Altenburger R, Nendza M, Schüürmann G, 2003. Mixture toxicity and its modeling by
quantitative structure-activity relationships. Environmental Toxicology and Chemistry,
22(8), 1900-1915.
Altenburger R, Scholz S, Schmitt-Jansen M, Busch W, Escher BI, 2012. Mixture toxicity
revisited from a toxicogenomic perspective. Environmental Science and Technology,
46, 2508-2522.
Ankley GT, Bennett RS, Erickson RJ, Hoff DJ, Hornung MW, Johnson RD, Mount DR,
Nichols JW, Russom CL, Schmieder PK, Serrano JA, Tietge JE, Villeneuve DL, 2010.
Adverse outcome pathways: a conceptual framework to support ecotoxicology
research and risk assessment. Environmental Toxicology and Chemistry, 29 (3), 730–
741.
Attene-Ramos MS, Huang R, Sakamuru S, Witt KL, Beeson GC, Shou L, Schnellmann RG,
Beeson CC, Tice RR, Austin CP, Xia M, 2013. Systematic study of mitochondrial
toxicity of environmental chemicals using quantitative high throughput screening.
Chemical Research in Toxicology, 26 (9), 1323–1332
Baas J, Jager T, Kooijman B, 2010. A review of DEB theory in assessing toxic effects of
mixtures. Science of the Total Environment, 408, 3740–3745.
Baas J, Willems J, Kraak M, Vandenbrouck T, Jager T, Kooijman SALM, 2009. Prediction
of daphnid survival after in situ exposure to complex mixtures. Environmental Science
and Technology, 43, 6064–6069.
Backhaus T, 2015. An additional assessment factor (MAF) – A suitable approach for
improving the regulatory risk assessment of chemical mixtures? Report for Swedish
Chemicals
Agency
(KEMI)
5/15.
http://kemi.se/Documents/Publikationer/Trycksaker/Rapporter/Rapport-5-15.pdf
Backhaus T, Altenburger R, Faust M, Frein D, Frische T, Johansson P, Kehrer A, Porsbring
T. 2013. Proposal for environmental mixture risk assessment in the context of the
biocidal product authorization in the EU. Environmental Sciences Europe, 25 (1), 4.
Belanger SE, Sanderson H, Embry MR, Coady K, DeZwart D, Farr BA, Gutsell S, Halder
M, Sternberg R, Wilson P, 2015. It is time to develop ecological thresholds of
toxicological concern to assist environmental hazard assessment. Environmental
Toxicology and Chemistry, accepted article.
Bessems J, Coecke S, Gouliarmou V, Whelan M, Worth A (2015) EURL ECVAM strategy
for achieving 3Rs impact in the assessment of toxicokinetics and systemic toxicity.
EUR27315
EN,
50
http://publications.jrc.ec.europa.eu/repository/bitstream/JRC96418/eurl%20ecvam%
20toxicokinetics%20strategy.pdf
Beyer J, Petersen K, Song Y, Ruus A, Grung M, Bakke T, Tollefsen KE, 2014.
Environmental risk assessment of combined effects in aquatic ecotoxicology: A
discussion paper. Marine Environmental Research, 96, 81-91.
Boobis A, Budinsky R, Collie S, Crofton K, Embry M, Felter, S, Hertzberg R, Kopp D, et
al., 2011. Critical analysis of literature on low-dose synergy for use in screening
chemical mixtures for risk assessment. Critical Reviews in Toxicology, 41 (5), 369–
383.
Boobis A, Budinsky R, Crofton K, Embry M, Felter S, Mihlan G, Mumtaz M, Price P,
Solomon K, Zaleski R, 2011. Annex B - Example case study B: Tier 0 - Substances
potentially detectable in surface water. In: Meek ME, Boobis AR, Crofton KM,
Heinemeyer G, Van Raaij M, Vickers C, 2011. Risk assessment of combined exposure
to multiple chemicals: A WHO/IPCS framework. Regulatory Toxicology and
Pharmacology, 60, S1–S14.
Borgert CJ, 2007. Predicting interactions from mechanistic information: Can omic data
validate theories? Toxicology and Applied Pharmacology, 223 (2), 114-120.
Borgert CJ, Quill TF, McCarty LS, Mason AM, 2004. Can mode of action predict mixture
toxicity for risk assessment? Toxicology and Applied Pharmacology, 201 (2), 85-96.
Burgess RM, Ho KT, Brack W, Lamoree M, 2013. Effects-directed analysis (EDA) and
toxicity identification evaluation (TIE): Complementary but different approaches
for diagnosing causes of environmental toxicity. Environmental Toxicology and
Chemistry, 32 (9), 1935–1945.
Brand W, de Jongh CM, van der Linden SC, Mennes W, Puijker LM, van Leeuwen CJ, van
Wezel AP, Schriks M, Hering MB, 2013. Trigger values for investigation of
hormonal activity in drinking water and its sources using CALUX bioassays.
Environment International, 55, 109–118.
Browne P, Judson RS, Casey WM, Kleinstreuer NC, Thomas RS, 2015. Screening
chemicals for estrogen receptor bioactivity using a computational model.
Environmental Science and Technology, 49 (14), 8804–8814.
Caldwell DJ, Mastrocco F, Margiotta-Casaluci L, Brooks BW, 2014. An integrated
approach for prioritizing pharmaceuticals found in the environment for risk
assessment, monitoring and advanced research. Chemosphere, 115, 4–12.
Cedergreen N, 2014. Quantifying Synergy: a systematic review of mixture toxicity
studies within Environmental Toxicology. PLoS ONE 9(5), e96580.
Cramer GM, Ford RA, Hall RL, 1978. Estimation of toxic hazard - a decision tree
approach. Food and Cosmetic Toxicology, 16, 255-276.
Dardenne F, Van Dongen S, Nobels I, Smolders R, De Coen W, Blust R, 2008a. Mode of
action clustering of chemicals and environmental samples on the bases of bacterial
stress gene inductions. Toxicological Sciences, 101 (2), 206-214.
Dardenne F, Nobels I, De Coen W, Blust R, 2008b. Mixture toxicity and gene inductions:
Can we predict the outcome? Environmental Toxicology and Chemistry, 27 (3), 509-
518.
51
ECHA (European Chemicals Agency). 2014. Transitional Guidance on mixture toxicity
assessment for biocidal products for the environment.
ECHA, 2015. Read-Across Assessment Framework (RAAF). ECHA-15-R-07-EN.
http://echa.europa.eu/documents/10162/13628/raaf_en.pdf
EFSA (European Food Safety Authority), 2013. International framework dealing with
human risk assessment of combined exposure to multiple chemicals. EFSA Journal
2013; 11 (7):3313. [69 pp.] doi:10.2903/j.efsa.2013.3313.
EFSA Panel on Plant Protection Products and their Residues (PPR), 2014. Scientific
Opinion on the identification of pesticides to be included in cumulative assessment
groups on the basis of their toxicological profile (2014 update). EFSA Journal 2013, 11
(7):3293, 131 pp. doi:10.2903/j.efsa.2013.3293
EFSA Panel on Plant Protection Products and their Residues (PPR), 2013. Scientific
Opinion on relevance of dissimilar mode of action and its appropriate application for
cumulative risk assessment of pesticides residues in food. EFSA Journal 2013;11
(12): 3472, 40 pp. doi:10.2903/j.efsa.2013.3472
EFSA Scientific Committee, 2012. Scientific Opinion on exploring options for providing
advice about possible human health risks based on the concept of Threshold of
Toxicological Concern (TTC). EFSA Journal 2012, 10 (7), 2750 [103 pp.]
doi:10.2903/j.efsa.2012.2750.
EFSA Panel on Plant Protection Products and their Residues (PPR), 2013. Guidance on
tiered risk assessment for plant protection products for aquatic organisms in edge-of-
field
surface
waters.
EFSA
Journal
2013,
11
(7):
3290,
268
pp.
doi:10.2903/j.efsa.2013.3290.
El-Masri H, 2007. Experimental and mathematical modelling methods for the
investigation of toxicological interactions. Toxicology and Applied Pharmacology, 223,
148-154.
Escher BI, Neale PA, Leusch FDL, 2015. Effect-based trigger values for in vitro
bioassays: Reading across from existing water quality guideline values. Water
Research, 81, 137–148.
Goodson WH, Lowe L, Carpenter DO, Gilbertson M, Manaf AA, Lopez de Cerain
Salsamendi A, Lasfar A, et al. 2015. Assessing the carcinogenic potential of low-dose
exposures to chemical mixtures in the environment: the challenge ahead.
Carcinogenesis, 36 (Suppl 1), S254-S296.
Groten, J.P., Feron, V.J., Sühnel, J., 2001. Toxicology of simple and complex mixtures.
Trends in Pharmacological Sciences, 22, 316-322.
Hadrup N, 2014. Evidence from pharmacology and pathophysiology suggests that
chemicals with dissimilar mechanisms of action could be of bigger concern in the
toxicological risk assessment of chemical mixtures than chemicals with a similar
mechanism of action. Regulatory Toxicology and Pharmacology, 69 (3), 281-283.
Health Canada / Environment Canada, 2015. Proposed approach for cumulative risk
assessment of certain phthalates under the chemicals management plan, August
2015. http://www.ec.gc.ca/ese-ees/default.asp?lang=En&n=723C9007-1
Huang RL, Xia MH, Cho MH, Sakamuru S, Shinn P, Houck KA, Dix DJ, Judson RS, et al.,
2011. Chemical genomics profiling of environmental chemical modulation of
52
human nuclear receptors. Environmental Health Perspectives, 119 (8), 1142–
1148.
Jager T, Gudmundsdottir EM, Cedergreen N, 2014. Dynamic modeling of sublethal
mixture toxicity in the nematode
Caenorhabditis elegans. Environmental Science and
Technology, 48, 7026−7033.
Jager T, Vandenbrouck T, Baas J, De Coen WM, Kooijman SALM, 2010. A biology-based
approach for mixture toxicity of multiple endpoints over the life cycle. Ecotoxicology,
19, 351−361.
Jarošová B, Bláha L, Giesy JP, Hilscherová K, 2014, What level of estrogenic activity
determined by in vitro assays in municipal waste waters can be considered as
safe? Environment International, 64, 98–109.
Kamo and Yokomizo, 2015. Explanation of non-additive effects in mixtures of similar
mode of action chemicals. Toxicology, 335, 20–26.
Kienzler A, Berggren A, Bessem J, Bopp S, van der Linden S, Worth A (2014)
Assessment of mixtures - Review of regulatory requirements and guidance (EUR
26675 EN)
Kim J, Kim S, 2015. State of the art in the application of QSAR techniques for predicting
mixture toxicity in environmental risk assessment. SAR and QSAR in Environmental
Research, 26 (1), 41-59.
Kortenkamp, A., Backhaus, T., Faust, M., 2009. State of the art report on mixture
toxicity. Final report. 070307/2007/485103/ETU/D.1, 391 p.
Løkke H, Ragas AMJ, Holmstrup M, 2013. Tools and perspectives for assessing chemical
mixtures and multiple stressors. Toxicology, 313, 73– 82.
Meek ME, Boobis AR, Crofton KM, Heinemeyer G, van Raaij M, Vickers C (2011) Risk
assessment of combined exposure to multiple chemicals: A WHO/IPCS framework.
Regulatory Toxicology and Pharmacology, 60 (2), S1–S14.
Munro IC, Ford RA, Kennepohl E, Sprenger JG, 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.
Mwense M, Wang XZ, Buontempo FV, Horan N, Young A, Osborn D, 2006. QSAR
approach for mixture toxicity prediction using independent latent descriptors and
fuzzy membership functions. SAR and QSAR in Environmental Research, 17 (1), 53-
73.
NAFTA (North American Free Trade Agreement), 2012. Technical Working Group on
Pesticides - (Quantitative) Structure Activity Relationship ((Q)SAR) Guidance
Document.
186
pp.
Available
at:
[http://www.epa.gov/oppfead1/international/naftatwg/guidance/qsar-guidance.pdf].
National Research Council, 2007. Toxicity testing in the 21st century: a vision and a
strategy.
OECD, 2009. Guidance document for using the OECD (Q)SAR application toolbox to
develop chemical categories according to the OECD guidance on grouping of
chemicals. Series on Testing and Assessment, No 102. ENV/JM/MONO(2009)5, 24-
Feb-2009.
53
OECD, 2013. Guidance document on developing and assessing adverse outcome
pathways. Series on Testing and Assessment. No. 184, ENV/JM/MONO(2013)6. 17-
Apr-2013.
OECD, 2014. Guidance on grouping of chemicals, second edition. Series on Testing and
Assessment, No. 194, ENV/JM/MONO(2014)4, 14-Apr-2014.
Price PS and Han X, 2011. Maximum cumulative ratio (MCR) as a tool for assessing the
value of performing a cumulative risk assessment. International Journal of
Environmental Research and Public Health, 8 (6), 2212-2225.
Price PS, Hollnagel HM, Zabik JM, 2009. Characterizing the noncancer toxicity of
mixtures using concepts from the TTC and quantitative models of uncertainty in
mixture toxicity. Risk Analysis, 29 (11), 1534-1548.
SCHER, SCCS, SCENIHR (2012) Opinion on the toxicity and assessment of chemical
mixtures.
Schirmer K, Fischer BB, Madueira DJ, Pillai S (2010) Transcriptomics in ecotoxicology.
Analytical and Bioanalytical Chemistry, 397 (3), 917-923.
Sonneveld E, Riteco JAC, Jansen HJ, Pieterse B, Brouwer A, Schoonen WG, van der Burg
B, 2006. Comparison of In Vitro and In Vivo Screening Models for Androgenic and
Estrogenic Activities. Toxicological Sciences, 89 (1), 173-187.
Tan YM, Clewell H, Campbell J, Andersen M, 2011. Evaluating pharmacokinetic and
pharmacodynamic interactions with computational models in supporting cumulative
risk assessment. International Journal of Environmental Research and Public Health, 8
(5), 1613-1630.
Tang JYM, McCarty S, Glenn E, Warne MSJ, Escher BI, 2013. Mixture effects of organic
micropollutants present in water: Towards the development of effect-based water
quality trigger values for baseline toxicity. Water Research, 47 (10), 3300-3314.
Tang JYM, Busetti F, Charrois JWA, Escher BI, 2014. Which chemicals drive biological
effects in wastewater and recycled water. Water Research, 60, 289–299.
Tice RR, Austin CP, Kavlock RJ, Bucher JR, 2013. Improving the human hazard
characterization of chemicals: a Tox21 update. Environmental Health Perspectives,
121 (7), 756–765.
Rovida C, Asakura S, Daneshian M, Hofman-Huether H, Leist M, Meunier L, Reif D, Rossi
A, Schmutz M, Valentin J-P, Zurlo J and Hartung T, 2015, Toxicity testing in the
21st Century Beyond Environmental Chemicals. Altex, 32 (3),171-181.
Terry C, Rasoulpour RJ, Knowles S, Billington R, 2015. Utilizing relative potency factors
(RPF) and threshold of toxicological concern (TTC) concepts to assess hazard and
human risk assessment profiles of environmental metabolites: A case study.
Regulatory Toxicology and Pharmacology, 71 (2), 301-317.
Toleffsen KE, Scholz S, Cronin MT, Edwards SW, de Knecht J, Crofton K, Garcia-Reyero
N, Hartung T, Worth A, Patlewicz G, 2014. Applying Adverse Outcome Pathways
(AOPs) to support Integrated Approaches to Testing and Assessment (IATA).
Regulatory Toxicology and Pharmacology, 70, 629–640
Verhaar HJM, Morroni JR, Reardon KF, Hays SM, Gaver DP, Carpenter RL, Yang RSH,
1997. A proposed approach to study the toxicology of complex mixtures of petroleum
54
products: the integrated use of QSAR, lumping analysis and PBPK/PD modeling.
Environmental Health Perspectives, 105 (Suppl 1), 179-195.
Villeneuve and Garcia-Reyero (2011) Predictive ecotoxicology in the 21st Century.
Environmental Toxicology and Chemistry, 30 (1), 1-8.
Von Stackelberg K, Guzy E, Chu T, Henn BC, 2015. Exposure to mixtures of metals and
neurodevelopmental outcomes: a multidisciplinary review using an adverse outcome
pathway framework. Risk Analysis, 35 (6), 971-1016.
Wason SC, Smith TJ, Perry MJ, Levy JI, 2012. Using physiologically-based
pharmacokinetic models to incorporate chemical and non-chemical stressors into
cumulative risk assessment: a case study of pesticide exposures. International
Journal of Environmental Research and Public Health, 9, 1971-1983.
WHO/IPCS (2009) Assessment of combined exposures to multiple chemicals: report of a
WHO/IPCS international workshop on aggregate/cumulative risk assessment.
55
List of abbreviations and definitions
AOP
Adverse Outcome Pathway
CA
Concentration Addition
CEFIC MIAT
European Chemical Industry Council Mixtures Industry Ad-hoc Team
CLP
Classification, labelling and packaging
DEB
Dynamic Energy Budget modelling
ECHA
European Chemicals Agency
EFSA
European Food Safety Authority
ERA
Environmental Risk Assessment
ESTAF
EURL ECVAM Stakeholder Forum
EURL ECVAM
European Union Reference Laboratory for alternatives to animal
testing
HI
Hazard Index
HIint
Hazard Index considering Interactions
HRA
Human Health Risk Assessment
IA
Independent Action
IATA
Integrated approaches to testing and assessment
MCR
Maximum Cumulative Ratio
MoA
Mode of Action
OECD
Organisation for Economic Co-operation and Development
PARERE
EURL ECVAM's Network for Preliminary Assessment of Regulatory
Relevance
PBTK
Physiologically Based Toxicokinetic modelling
PODI
Point of Departure Index
PPP
Plant Protection Product
QA/QC
Quality Assurance/Quality Control
QSAR
Quantitative Structure Activity Relationship
RA
Risk Assessment
REACH
Registration, Evaluation, Authorisation and Restriction of Chemicals
RPF
Relative Potency Factor
56
SCCS
Scientific Committee on Consumer Safety
SCENHIR
Scientific Committee on Emerging and Newly Identified Health Risks
SCHER
Scientific Committee on Health and Environmental Risks
TEF
Toxic Equivalence Factor
TKTD
Toxicokinetic/Toxicodynamic modelling
TTC
Threshold of Toxicological Concern
TUS
Toxic Unit Summation
WHO/IPCS
World Health Organisation/ International Programme on Chemical
Safety
57
List of figures
Figure 1 Country of Respondent. Survey participants were from 16 different EU
countries and 5 non-EU countries (indicated as "Other": Canada, Japan, Norway, Serbia,
Switzerland, US). ................................................................................................ 23
Figure 2 Survey respondents' affiliation ................................................................ 23
Figure 3 Replies to the question "Did you ever need to perform a risk assessment of
chemical mixtures for HRA or ERA?" ...................................................................... 24
Figure 4 Replies to the question "For which purpose(s) did you assess the overall
toxicity of chemical mixtures?" Other purposes indicated were 1) development of CEFIC
MIAT conceptual framework and CLP workplace monitoring. ..................................... 24
Figure 5 Experience in performing mixture risk assessments using whole mixture or
component-based approaches applied to intentional or unintentional mixtures. ........... 25
Figure 6 Replies to the question "For which type of mixture(s) have you applied a whole
mixture or component-based approach?" Chemicals were further specified in the survey
as "multiconstituent or UVCB substances under REACH". Other types of mixtures
mentioned were medical devices, alloys, botanicals, food ware materials, jewellery, toys,
cigarette smoke/tobacco toxicants, human tissue extracts, landfill leachates, solid waste
from industrial combustion processes, sewage sludge. ............................................. 26
Figure 7 Replies to question "How often have you used the following component-based
approaches?" (CA=concentration addition, TUS=Toxic Unit Summation, TEF=Toxic
Equivalency Factor, RPF=Relative Potency Factor, PODI=Point of Departure Index,
HI=Hazard Index, HI(int)=Hazard Index including interactions, MCR=Maximum
Cumulative Ratio, IA=Independent Action) ............................................................. 27
Figure 8 Replies to the question "Which type of mixture(s) or samples would you
identify as highest priority for risk assessment that needs to take mixture effects into
account?" divided by stakeholder group. Chemicals were further specified in the survey
as "multiconstituent or UVCB substances under REACH". Other mixtures of importance
mentioned were those present in human tissues and container systems. .................... 28
Figure 9 Replies to the question "Based on your experience or knowledge about the
toxicity of mixtures, do you consider it important to include interactions (e.g. potential
synergism or antagonism) in the assessment?" divided by stakeholder group. Among the
experts that selected the "other" option, the main reasoning provided in the free text
field was that interactions are considered rare, especially at low/environmental
concentrations. However, most experts agreed that interactions should be taken into
account on a case-by-case basis and one expert proposed to look at interactions
especially in the case of active substances like plant protection products, biocides,
pharmaceuticals. ................................................................................................. 28
Figure 10 Replies to the questions on approaches the experts have abandoned due to
negative experience, specific approaches that the experts would recommend for further
considerations, and on use of relevant software tools and databases. ........................ 29
Figure 11 Replies to the question "Do you apply
in vitro tools/omics approaches/
(quantitative) structure activity relationships ((Q)SARs)/ read-across/ physiologically
based pharmacokinetic (PBPK) modelling/ the toxicological threshold of concern (TTC)
concept/ Adverse Outcome pathways (AOPs)/ dynamic energy budget (DEB) models for
human health risk assessment (HRA), environmental risk assessment (ERA) or both?" 30
Figure 12 Replies to the question "What are the main reasons you are using these
in
vitro assays?" Other reasons mentioned were for research purpose, Ames test because it
is a data requirement in some regions, and combination of
in vitro testing with PBPK
modelling is seen as the way forward for chemical risk assessment. .......................... 31
58
Figure 13 Replies to the question "What are the main reasons you are not using
in vitro assays? The main "other reasons" given were lack of EU guidance and validation, and in
the case of cosmetics that
in vitro tests are less relevant for the final products as clinical
tests are performed with cosmetic formulations. ..................................................... 32
Figure 14 Replies to the question "Which type of omics technologies are you using?" 33
Figure 15 Replies to the question "What are the main reasons you are using omics
based approaches?" The other reasons given were the possibility to study mechanisms of
action. ............................................................................................................... 33
Figure 16 Replies to the question "What are the main reasons you are not using omics
based assays?" The main "other reason" provided was the unavailability of the relevant
practical tools in the facilities of the responding experts, as well as the lack of clear
guidance. ........................................................................................................... 34
Figure 17 Replies to the question "What are the main reasons you are using (Q)SARs?"
"Other reasons" mentioned were the application to metabolites and impurities where no
data are available, and the possibility to identify substances of concern and drive further
experimental testing. Also the use for chemical fate aspects was mentioned. .............. 35
Figure 18 Replies to the question "For what purpose are you applying (Q)SARs? "Other
reasons" mentioned were data gap filling, prediction of physic-chemical properties,
grouping of metabolites and impurities, and fate predictions. .................................... 35
Figure 19 Replies to the question "Which (Q)SAR models do you apply?" Other models
mentioned were DEBtox, VEGA platform models, ACD Labs Percepta, Episuite, SarPy. . 36
Figure 20 Replies to the question "What are the main reasons you are not using
(Q)SARs?". The main "other reasons" given were that databases are not validated, lack
of guidance, (Q)SARs are applied to single substances but not to mixtures because only
qualitative QSAR information is used, (Q)SARs are not relevant because of available
testing information (e.g. for PPPs), too many uncertainties associated with (Q)SARs, lack
of knowledge/training. ......................................................................................... 37
Figure 21 Replies to the question "What are the main reasons you are using a read-
across approach?" Main "other reasons" were that read-across is a useful tool for the use
for metabolites and impurities, where toxicity information is limited, it is encouraged
under REACH, can be used to support
in vitro test results and refine experimental
design. .............................................................................................................. 38
Figure 22 Replies to the question "What are the main reasons you are not using a read-
across approach?" Main "other reasons" given were a lack of guidance and validation. . 38
Figure 23 Replies to the question "What are the main reasons you are using a PBTK
model?" The main "other reasons" were to study mechanisms, study mechanisms by
which mixtures affect life-history traits, assess intermittent exposures ...................... 39
Figure 24 Replies to the question "What are the main reasons you are not using a PBTK
model?" The main "other reasons" given were lack of knowledge/expertise/training, lack
of guidance and tackling related uncertainties. ........................................................ 39
Figure 25 Replies to the question "What are the main reasons you are using the TTC
approach?" Main "other reasons" were the availability for chemicals without chronic data,
the use to identify and prioritise testing needs, avoiding testing, and application to
ingredients of biological origin and residual impurities. ............................................. 40
Figure 26 Replies to the question "What are the main reasons you are not using the
TTC approach?" Main "other reasons" were the need of updating the TTC approach, the
lack of experience, and dealing with higher substance concentrations. ....................... 41
Figure 27 Replies to the question "What are the main reasons you are using an AOP
approach?" The "other reason" mentioned was to explore its potential for risk
assessment. ....................................................................................................... 41
59
Figure 28 Replies to the question "What are the main reasons you are not using an AOP
approach?" The "other reasons" mentioned are mainly the lack of expertise/experience,
and difficulties to implement this new concept in risk assessment. ............................ 42
Figure 29 Replies to the question "What are the main reasons you are using DEB
models?" The "other reason" was stating that DEB models are the most promising
approach to interpret and predict effects of stressors on growth and reproduction and to
learn from experimental testing. ........................................................................... 42
Figure 30 Replies to the question "What are the main reasons you are not using DEB
models?" Main "other reasons" were a general lack of expertise/experience, for experts
in human RA its focus on environmental RA, and a lack of validation. ........................ 43
Figure 31 Replies to the question "Do you use any IATA framework for the assessment
of mixtures?" ...................................................................................................... 43
Figure 32 Replies to question "Which [international] framework(s) [for addressing
combined exposure] do you apply?" ...................................................................... 44
Figure 33 Replies to the question "Are you familiar with or do you apply any other
framework to assess the toxicity of mixtures?" ....................................................... 45
60
Europe Direct is a service to help you find answers to your questions about the European Union
Free phone number (*): 00 800 6 7 8 9 10 11
(*) Certain mobile telephone operators do not allow access to 00 800 numbers or these calls may be billed.
A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server http://europa.eu
How to obtain EU publications
Our publications are available from EU Bookshop (http://publications.europa.eu/howto/index_en.htm),
where you can place an order with the sales agent of your choice.
The Publications Office has a worldwide network of sales agents.
You can obtain their contact details by sending a fax to (352) 29 29-42758.
L
B
-NA-27471-EN-N
JRC Mission
As the Commission’s
in-house science service,
the Joint Research Centre’s
mission is to provide EU
policies with independent,
evidence-based scientific
and technical support
throughout the whole
policy cycle.
Working in close
cooperation with policy
Directorates-General,
the JRC addresses key
societal challenges while
stimulating innovation
through developing
new methods, tools
and standards, and sharing
its know-how with
the Member States,
the scientific community
and international partners.
Serving society
Stimulating innovation
Supporting legislation
doi:10.2788/093511
ISBN 978-92-79-51925-3
2