link to page 1 link to page 13
Regulatory Toxicology and Pharmacology 80 (2016) 321e334
Contents lists available at
ScienceDirect
Regulatory Toxicology and Pharmacology
j o u r n a l h o m e p a g e :
w w w . e l s e v i e r . c o m / l o c a t e / y r t p h
Commentary
Regulatory assessment of chemical mixtures: Requirements, current
approaches and future perspectives
Aude Kienzler, Stephanie K. Bopp
*, Sander van der Linden, Elisabet Berggren,
Andrew Worth
European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Systems Toxicology Unit and EURL ECVAM, Via Enrico Fermi
2749, 21027, Ispra, VA, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
This paper reviews regulatory requirements and recent case studies to illustrate how the risk assessment
Received 4 February 2016
(RA) of chemical mixtures is conducted, considering both the effects on human health and on the
Received in revised form
environment. A broad range of chemicals, regulations and RA methodologies are covered, in order to
13 May 2016
identify mixtures of concern, gaps in the regulatory framework, data needs, and further work to be
Accepted 16 May 2016
carried out. Also the current and potential future use of novel tools (Adverse Outcome Pathways, in silico
Available online 20 May 2016
tools, toxicokinetic modelling, etc.) in the RA of combined effects were reviewed.
The assumptions made in the RA, predictive model specifications and the choice of toxic reference
Keywords:
values can greatly influence the assessment outcome, and should therefore be specifically justified. Novel
Mixture
Combined effect
tools could support mixture RA mainly by providing a better understanding of the underlying mecha-
Risk assessment
nisms of combined effects. Nevertheless, their use is currently limited because of a lack of guidance, data,
Regulation
and expertise. More guidance is needed to facilitate their application. As far as the authors are aware, no
prospective RA concerning chemicals related to various regulatory sectors has been performed to date,
even though numerous chemicals are registered under several regulatory frameworks.
© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
1. Introduction
combination are rarely examined.
The hazard and/or risk assessment (RA) requirements for
The number of chemicals and combinations thereof to which
(components of) products on the European market are laid down in
humans and the environment are continuously exposed is poten-
specific EU legislations primarily depending on the intended use of
tially enormous, ever changing in concentration and identity and to
the product. These products, e.g. biocides, pesticides, food or feed
a large extent unknown. This makes it neither realistic nor useful to
additives, pharmaceuticals, can consist of an individual compound
test every possible combination. However, current human risk
or of mixtures of several compounds. As the composition of these
assessment (HRA) and environmental risk assessment (ERA) of
products is generally known, and the relevant compounds are
chemicals mainly focuses on exposure to individual chemicals,
relatively well assessed individually, the RA is performed pro-
mostly considering only a single source.
spectively, based on the properties of the individual constituents.
In 2012, the European Commission published a communication
Where appropriate, tests can also be carried out on the formulated
on the combined effects of chemicals
(EC, 2012), expressing con-
products. However, when several formulated products are used in
cerns about the current limitations of assessing compounds indi-
combination, i.e. for the application of plant protection products
vidually and proposing a path forward to ensure that risks
(PPPs) in the field or for the use of personal care products at home,
associated with chemical mixtures are properly understood and
the combined resulting risk is generally not assessed. Similarly, the
assessed. It states that EU laws set strict limits for the amounts of
prospective RA often considers only one route of exposure, e.g.
particular chemicals allowed in food, water, air and manufactured
linked to occupational exposure to pesticide, and does not consider
products, but that the potential risks of these chemicals in
potential additional sources of exposure such as the intake via food
consumption.
In addition to the regulations that cover the intentional mix-
* Corresponding author.
tures that are present in specific products, several others focus on
E-mail address:
[email address] (S.K. Bopp).
http://dx.doi.org/10.1016/j.yrtph.2016.05.020
0273-2300/© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (
http://creativecommons.org/licenses/by/4.0/).
link to page 2 link to page 13 link to page 13 link to page 13 link to page 14
322
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
Abbreviations
IPCS
International Program on Chemical Safety
MCR
maximum cumulative ratio
ADI
acceptable daily intake
MCS
multi constituent substances
AOP
adverse outcome pathway
MoA
mode of action
ARfD
acute reference dose
MRLs
maximum residue levels
AS
active substance
PBPK
physiologically based pharmacokinetic modelling
BPR
biocidal product regulation
PPP
plant protection products
CA
concentration addition
PPPR
plant protection product regulation
DEB
dynamic energy budget modelling
QSAR
quantitative structure activity relationship
EQS
environmental quality standard
TEFs
toxic equivalency factor
ERA
environmental risk assessment
TTC
threshold of toxicological concern
ILSI-HESI Health and Environmental Sciences Institute
UVCBs
substances of unknown or variable composition,
HI
hazard index
complex reaction products or biological materials
HRA
human risk assessment
WFD
water framework directive
IA
independent action
the exposure to unintentional mixtures. These can be mixtures that
consistent,
comprehensive
and
integrated
approach
across
are unintentionally formed during the production process or mix-
different pieces of legislation. As a step forward, a widely accepted
tures that are found in the environmental matrix after being
framework for the RA of combined exposure to multiple chemicals
emitted (also defined as coincidental mixture,
Table 1). Examples of
was developed in a WHO/IPCS workshop (
Meek et al., 2011). This
regulations that address these types of mixtures are the Water
framework describes a general approach for RA of combined
Framework Directive (WFD), Marine Water Strategy, or Air and Soil
exposure to multiple chemicals that could be adapted to the needs
related regulation.
of specific users. However, its use is often hampered by large data
Because of their varying composition in space and time, due to
gaps on exposure as well as hazard information.
both the environmental fate of chemicals and the constant entry of
This review presents an overview of the current regulatory re-
new pollutants, exposure to coincidental mixtures in the environ-
quirements for chemical mixture assessment, with emphasis on the
ment is never assessed prospectively, and retrospective RAs are
extent to which they address the assessment of intentional and
scarce, even if this is the most common situation. In special cases, if
unintentional mixtures. Recent case studies, specifically focusing
more information on use would be available, some unintentional
on mixture RA beyond the current regulatory requirements are
mixtures might be assessed prospectively, e.g. PPPs tank mixtures
summarized to illustrate what lessons can be learned in terms of
(mixtures of individually assessed formulation that are mixed by
methodology being used and existing regulatory and data gaps. The
the user), or mixtures of chemicals found in the environment after
lessons learned from the case studies are supplemented by the
being emitted at the same place and the same time or sequentially
results from an expert survey. This survey, which was carried out to
(e.g. production plants of specific substances). These cases are
explore the current use of different approaches for assessing hu-
currently not covered under the legislation. However several
man and environmental health risks from combined effects and the
different guidance documents on how to deal with mixtures have
added value of several novel tools that could provide some of the
been published recently, each focusing on a specific group of
missing
information
currently
hampering
the
toxicological
compounds or type of assessment. Examples are the guidance on
assessment of mixtures.
aquatic RA under REACH (
Bunke et al., 2013), the assessment of
mixture effects of biocides (
ECHA, 2014), and for pesticides, how to
2. Mixture terminology and assessment concepts
assess exposure scenarios for RA both using MRL or actual expo-
sures based on monitoring data
(EFSA, 2012).
2.1. Mixture assessment terminology
Although methodologies for assessing the combination effects
of chemicals are being developed and applied by scientists and
While the term mixture might seem a clear term at first sight,
regulators in specific circumstances, so far there is no systematic,
many similar - but different terms - are used in parallel, to indicate
Table 1
Types of mixtures, characterisation and related regulation.
Type of mixture
Definition
Characterisation
Assessment
Example of related regulation
Intentional
Formulated products marketed
Usually of known or
Usually prospective based on the
Plant Protection Products, Biocides,
as such
well-known
properties of the constituents
Pharmaceuticals, Food additives …
composition
supplemented, where appropriate, by
tests carried out on the entire products
Unintentional
Usually from one source;
The composition
If composition unknown, whole-
Water Framework Directive or waste-related
generated by discharge during
can either be
mixture approach.
regulation.
production, transport, use or
known (effluent) or
disposal of goods
unknown
Coincidental
Originating from various
Composition
Usually not required
a) water/soil/air-related regulation;
sources
unknown, varying
b) exposure of workers in the workplace, for
in space and time
which a risk assessment is required for all
hazardous chemicals, including in combination;
c) exposure of humans to multiple chemicals
from food and drinking water.
link to page 14 link to page 14 link to page 3 link to page 12 link to page 12 link to page 13
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
323
the different types of mixtures and pathways of exposure. There-
generally provide reliable estimates of combined effects, they can
fore, to avoid future confusion, there is a clear need to define a
more easily be used with existing toxicity data and are considered
consistent terminology to identify the different types of mixture
to be slightly more conservative than IA models. However, the re-
scenarios. In this work, we adopted the terminology developed in
sults obtained by both models are usually very similar and the
the context of the WHO, OECD, IPCS, ILSI/HESI initiatives (
Meek
difference between the predictions rarely exceed a factor of five
et al., 2011; OECD, 2011; WHO IPCS, 2009). This means that an
(
Backhaus and Faust, 2012; Backhaus et al., 2004; Kortenkamp
exposure to multiple chemicals is defined as a combined exposure,
et al., 2009). Some models are also capable of incorporating in-
should it be by a single route or by multiple routes (which is
teractions (e.g. antagonism or synergism) to some extent. A more
sometimes referenced as “cumulative” exposure). Exposure to a
extensive overview of mixture assessment models is given in
single chemical from multiple sources and by multiple pathways
Kienzler et al. (2014), and the most used (i.e Quotient Ratio, Hazard
and routes is defined as an aggregate exposure
(Fig. 1).
Index, etc …) are presented in
Table S1.
In addition, since the terminology sometimes differs between
Using these models, the overall toxicity of a mixture of known
HRA and ERA, we define the following regarding the routes, path-
composition can to some extent be determined prospectively. For
ways and sources of exposure. The route of exposure refers to the
environmental mixtures, the (eco)toxicity is usually investigated by
way a chemical enters the organism, i.e. via dermal exposure, oral
one of the two following approaches: testing the mixture as a
exposure or inhalation. The pathway of exposure refers to the
whole (using a simple in vivo or in vitro test system) or using the
medium with which the chemicals are taken up, e.g. with drinking
(eco)toxicological data on individual components combined with
water, air, food. The sources of exposure are the places of release of
the chemical-analytical concentration data. The later can then feed
chemicals such as industrial emissions, waste water treatment
into a mathematical model to predict the final (eco)toxicity of the
plant effluents, etc.
mixture. Whole-mixture testing is frequently applied for environ-
mental mixtures, as it allows to assess the (eco)toxicity of mixtures
2.2. Mixture assessment models
of unknown composition; however, the compounds responsible for
the response frequently remain unidentified. The component-
Two main mathematical models exist to assess the combined
based approach is more common, but requires more information
toxicological effect of chemicals, either assuming that individual
regarding identity, concentration and toxicity, including mode of
compounds act via a dissimilar mode of action (independent action,
action (MoA) of the individual compounds.
IA, or Response Addition, assuming the addition of the response) or
by the same mode of action (dose or concentration addition, CA). In
2.3. Mixture assessment tools
CA based models, the total response corresponds to the sum of all
the individual concentrations multiplied with their respective po-
In addition to the mathematical models applied to predict the
tencies. CA models are the most frequently applied, because they
overall toxicity, additional tools are increasingly used to determine
Fig. 1. Aggregate vs combined exposure.
link to page 13 link to page 14 link to page 13 link to page 12 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 4 link to page 4 link to page 4 link to page 4
324
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
when or whether combined RA is needed at all, e.g. by expressing
Those data gaps are numerous, both regarding hazard and exposure
the individual compound contribution compared to the combined
data, for pharmaceuticals
(Backhaus and Karlsson, 2014), pesticides
total toxicity of the known compounds in the mixture. As such,
(
Junghans et al., 2006; Kennedy et al., 2015; Nowell et al., 2014),
these tools do not predict the risk of a mixture per se but rather
cosmetics, etc. and implies to use extrapolations (i.e. acute to
provide a means of investigating data on cumulative exposure to
chronic), which increase the uncertainties of the RA. Models to
human and ecological receptor and identify when cumulative RA is
estimate aggregate exposure of consumers in personal care prod-
most needed. An example of such a tool is the Maximum Cumu-
ucts (PCPs) are being developed (
Delmaar et al., 2015), but a suffi-
lative Ratio (MCR), which is the ratio between the toxicity of the
ciently elaborated data on the frequency of use of those products
mixture (based on CA models) and the toxicity of the most
are still lacking (
Gosens et al., 2013) which hamper refinements of
contributing chemical in the mixture. The MCR approach is
the RA if needed.
currently applied in various contexts, to determine when cumula-
As a result, RA of chemical mixtures requires a lot of assump-
tive assessments are most required and to discriminate between
tions. Their choice can have a large impact on the outcome and
those mixtures requiring further combined RA and those for which
should be carefully documented and justified (
Boon et al., 2015;
a single-substance assessment is sufficient. Therefore, it helps to
Kennedy et al., 2015). This is also the case for single substance as-
decide on the next step of the RA, e.g. undertake a further refine-
sessments, however, of particular importance for the assessment of
ment of mixture RA in case of several main contributors or
mixtures since the uncertainties around single substance assess-
concentrate on only a few components dominating the effects.
ments are adding up when combined risks are assessed.
However, for calculating the MCR, at least a screening level
Moreover, in the case in which different models are combined
mixture RA has to be performed to predict the combined effects to
and used in the same RA (i.e. for dietary and non-dietary exposure),
which the effect of the most contributing compound(s) can be
care must be taken when interpreting the result to recognize
related. Moreover, application of the MCR methodology requires
possible differences in the degree of conservatism between dietary
knowledge of the concentrations of chemicals in a mixture together
and non-dietary exposure models. Furthermore, the assessment of
with health-based reference values for those chemicals. This tool
combined effects for substances of common effects or common
has also been found to be useful for analysing the pattern of
MoA implies that reference values for the specific effect under
chemical-specific contributions to the total exposure levels of
consideration should be used. Toxicity values reported however are
mixtures based on biomonitoring data when Toxic Equivalent
often those driving the single substance risk, i.e. the lowest refer-
Factors (TEFs, see
Table S1) or similar approaches are available (i.e.
ence value which might be for a different effect. Using these
occupational vs. background exposure) (
Han and Price, 2013).
reference values in lower tiers can be a first conservative estimate,
Depending on the scope of the assessment and the available
but might lead to large overestimations of the combined effects.
data, a tiered approach can be followed. By starting from a
screening level with the option for further refinement where
needed, resources can focus on the most important factors
3. Assessment of mixtures under current regulations
contributing to risk. However, the criteria used in developing tiers
need to be balanced, yet sufficiently conservative so that important
As mentioned previously, a general distinction can be made
factors are not inappropriately screened out. In this context, the
between intentional and unintentional mixtures. Intentional mix-
utility of the Threshold of Toxicological Concern (TTC) approach as a
tures are generally well addressed by current regulation through a
Tier 0 assessment tool for chemical co-exposures, especially for a
prospective RA prior to the marketing of the product: the existing
data-poor chemical has been demonstrated (
Meek et al., 2011): a
European regulations dealing with the RA of intentional mixtures
RA based on monitoring data for 10 chemicals found in water was
have been reviewed in detail and their requirements are presented
done using the TTC approach for the substances for which no
in
Table S2. Nevertheless, this assessment is restricted to a partic-
established chronic health standards or health-based guidance
ular use in a given regulatory framework and does not take into
values were available. The resulting HI (see
Table S1 for details) of
account other potential uses related to other regulatory frame-
0.2 suggested that there was no need to further refine the RA.
works, i.e. aggregate exposure, although there might be an overlap
between the different regulatory frameworks. However, if there is
2.4. Methodological issues and hurdles hampering the risk
no reason why chemicals allocated to specific regulatory frame-
assessment of chemical mixtures
works would have non-overlapping risk profiles, then there is also
no reason to expect that mixture RA limited only to chemicals
When having a closer look into RA of chemical mixture case
within one regulation can fully capture the risk that may be present
studies, some methodological issues are recurrent. The data sources
to human consumers (
Evans et al., 2015).
used are various and the data sets more or less complete, this
Overlapping can be illustrated by considering the 428 unique
having a direct impact on the quality of the RA and the related
substances registered as pesticides in the EU
1 DG SANCO database.
uncertainties. Exposure data are usually modelled, from bio-
Of these, 38 are also registered as
biocides,2 55 as industrial
monitoring or published data from surveys on exposure, and
chemicals under REA
CH3 and six are registered within all these
exposure data reliability directly depends on the biomonitoring
three regulatory frameworks. In addition, one substance (Benzoic
practice (
Dewalque et al., 2014; Malaj et al., 2014) and on the
acid) is also registered in a fourth framework as a cosmetic ingre-
quantity of the data. The exposure of persistent and bio-
dient.4 While this illustrates the potential for aggregate exposure
accumulating chemicals is even more challenging as it requires
for individual substances, the combined exposure could be even
taking into account the kinetics of the chemicals and to consider
more relevant when considering different chemicals that share the
the body burden as a starting point for the RA, instead of the daily
intake, as well as the exposure history, as the exposure patterns
1
might change over time.
Source:
DG
SANCO
database,
extraction
on
the
11/05/2015.
Except
Toxicological data are mostly from published databases, but in
microorganisms.
2 Source: ECHA database on the 11/05/2015. Except microorganisms.
case of missing data e.g. the TTC approach, or other methods to fill
3 Source: ECHA database on the 18/09/2014. Except mixtures, reaction products,
data gaps are used. As a matter of fact, data gaps seem to be the
polymers and petroleum-derivatives.
major issue when it comes to deal with RA of chemical mixtures.
4 SCCS, 26/11/2014.
link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
325
same MoA (
Evans et al., 2015).
are specific to particular foodstuffs, and for products and/or pesti-
The assessment of unintentional or coincidental mixtures is
cides for which no specific MRLs are set, a default value of 0.01 mg/
generally not required
(Table S3), although the assessment of
kg applies.
multiple substances from multiple sources is the main issue raised
Reg. 396/2005 states that MRLs should be set in “view of human
by the European Commission when dealing with the assessment of
exposure to combinations of AS and their cumulative and possible
chemical mixtures
(EC, 2012). The requirements regarding mixture
aggregate and synergistic effects” and explicitly addresses the need
assessment of the most important European regulation are re-
for carrying out further work to develop methodology and tech-
ported hereafter.
nical guidelines on pesticides residues allowing to take into account
aggregate, cumulative and synergistic effects. However, established
3.1. Plant protection product (PPPR; Reg 1107/2009), biocidal
procedures for safety assessment of MRLs on the basis of ADI
products regulation (BPR; Reg 528/2012) and maximum residue
(Acceptable Daily Intake) values and food consumption patterns are
levels (MRLs) of pesticides (Regulation 396/2005)
focused on single substance assessments. The regulation also states
that Commission decisions related to MRLs shall take account of the
The regulations on PPPs and biocidal products both focus on the
possible presence of pesticides residues arising from sources other
active substance (AS). AS are assessed by a reporting member state
than current plant protection uses, without specifying which kind
and authorized at the European level, and the preparations made
of other sources, and “their known cumulative and synergistic effects,
with those AS are registered for specific uses at the national level.
when the methods to assess such effects are available”. The meth-
Requirements for PPP AS (Reg 283/2013) and formulations (Reg
odology is currently under development by EFSA see e.g.
(EFSA,
284/2013) can differ and are addressed by different regulations; in
2014a).
particular, chronic testing is usually only required for the AS and
not for the formulation (i.e. mixture). However, the fact that the
3.2. Pharmaceuticals (Dir 2001/83/EC and 2001/82/EC)
PPPR requires more data for the RA of AS than for the formulated
products (
EU, 2013a, 2013b), especially for chronic RA, is often
Regarding pharmaceuticals, both directives on human (Dir
criticized, as formulations are designed to be more effective than
2001/83/EC) and veterinary (Dir 2001/82/EC) pharmaceuticals
the AS itself. Both the PPPR and BPR mention the necessity to take
share some basic features: the RA follows a risk-benefit balance
into account interactions between components and require the
approach, in which the applicant is required to demonstrate that
assessment of cumulative and synergistic effects in the environ-
the potential risks are outweighed by the therapeutic efficacy of the
ment. However, for the PPPs this is restricted to the formulation
product.
itself. It does not apply to the potential combined effect resulting
Both risk to the patient’s health and risk to public health are
from the concomitant use of several formulations, as applied in
considered when dealing with human medicines. Wanted and
practice, or to the combined effects in the environmental matrix
unwanted interactions of substances combined within a medicinal
where they end up. Similarly, the potential aggregate exposure to
product are addressed, as well as potential interactions of the
the same AS coming from other sources is currently not addressed
medicine with other medicinal products, or with alcohol, tobacco,
for PPPs. Conversely, the BPR requires that for biocidal products
and foodstuffs. Studies on pharmacokinetic and pharmacodynamic
that are intended to be used in combination the risks to human
interactions are part of the standard dossier requirements
health, animal health and the environment arising from these
(
Kortenkamp et al., 2009). Thus, the toxicity of the whole product is
combinations shall be assessed. Moreover, when the evaluating
taken into account and potential interactions are deeply assessed
authority considers that there might be some concerns for human
when it comes to human exposure. An ERA is also required, but
health, animal health or the environment because of the cumula-
does not specifically address any aspect of mixture toxicity
(EMEA,
tive effects from the use of biocidal products containing the same or
2006).
different active substances, this concern should be documented
Regarding veterinary medicines, three types of risks are
and included in the conclusions.
considered: risks to the target animal, risks to human health (from
Regarding PPP and human health, the RA is not limited to the
both exposure to residues of the product in foodstuffs and direct
end-user of the PPP but should cover consumer exposure, opera-
exposure, i.e. during administration), and risks to the environment.
tors, workers, residents and bystanders, taking into account, where
Toxicity and ecotoxicity studies and assessments are performed for
relevant, the cumulative exposure to more than one AS. It is how-
the product, its active substances and relevant metabolites. Atten-
ever not specified if this requirement should be met for AS used in
tion is paid to interactions with other medicinal products or feed
combination in the same PPP only, or also when several PPP are
additives with respect to effects in the target animals, but this point
used in combination in the field, in a period that would allow cu-
is less deeply assessed than for human medicinal products
mulative exposure. When estimating the potential and actual
(
Kortenkamp et al., 2009). The ERA of the product is required but
exposure through diet and other sources, the presence of residues
does not take into account the toxicity potentially resulting from
arising from other sources (i.e. use as a biocide or veterinary drug)
the joint occurrence of different residues of veterinary products or
should be taken into account (aggregate exposure), as well as the
of other pollutants.
potential cumulative exposure to more than one AS, where rele-
vant. However it is not specified in detail how to proceed to such an
3.3. Food and feed additives (Reg 1333/2008, 1331/2008 and 429/
assessment, but reference is made that it should be assessed
2008)
“where the scientific methods accepted by the authority to assess
such effects are available”.
Neither the terms cumulative, synergistic or potentiating, nor
Pesticides residues in food and feed are specifically addressed by
the need for mixture toxicity assessments is mentioned in the food
Reg. 396/2005 on Maximum Residue levels (MRLs), which aims at
additives regulation (Reg. 1333/2008), although the previous
ensuring that those residues are not present in food and feed
regulation provided a basis for mixture toxicity assessments
products at levels presenting an unacceptable risk to humans and
(
Kortenkamp et al., 2009). Besides, established procedures for the
animals. This regulation establishes the maximum quantities of
safety assessment of food additives on the basis of ADI values for
pesticide residues permitted (MRLs) in products of animal or
single substances do not specifically consider interactions between
vegetable origin intended for human or animal consumption; MRLs
additives and food consumption
(Groten et al., 2000). However,
link to page 13 link to page 13 link to page 13 link to page 13 link to page 14 link to page 13 link to page 13 link to page 13 link to page 12 link to page 14
326
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
Regulation (EC) No 178/2002 laying down the general principles
interactions of the substances contained in the cosmetic product.
and requirements of food law states that “regard shall be had (…) to
This regulation does not address potential environmental concerns
the probable cumulative toxic effects” for human health, without
of cosmetic products specifically, as they are considered to be
however defining the term “cumulative toxic effects”, which could
assessed under REACH, which addresses the assessment of envi-
either mean a toxic effect resulting from repeated exposure to a
ronmental safety of the individual substances in a cross-sectorial
single toxicant or a toxic effect resulting from simultaneous or
manner.
sequential exposure to different toxicants and thus be used as a
synonym for mixture toxicity. Thus, European Regulation (EC) 1333/
3.6. Water framework directive (WFD, Dir 2000/60/EC) and marine
2008 neither excludes nor explicitly defines the need for mixture
strategy framework directive (Dir 2008/56/EC)
toxicity assessments for food additives (
Kortenkamp et al., 2009).
Regarding feed additives (Reg 1331/2008), consideration should be
The Water Framework Directive aims to establish the basic
given to the cumulative effects in case of additives with multiple
principles of sustainable water policy in the European Union, and to
components; however, there is no consideration of mixtures
assess, maintain or improve the chemical and biological status of
assessment from different sources.
European waters
(EC, 2000). Thus, this regulation does not address
a particular type of chemicals but all of those that could be of
3.4. REACH (Reg 1907/2006)
concern in surface water. It aims at identifying priority hazardous
substances for the aquatic environments on the basis of scientific
The REACH Regulation aims at ensuring the chemical safety
RA carried out under sectorial regulation (PPP, biocide, pharma-
assessment (CSA) of all chemicals unless they are specifically
ceuticals …), and sets common environmental quality standards
covered by other sectorial regulations (
EC, 2006). The REACH
(EQS) and emission limit values for chemicals or groups of pollut-
registration requirements apply to each of the individual sub-
ants (
EC, 2011). However, this directive does not mention chemical
stances in a preparation, but not to the preparation itself. REACH
mixtures or mixture effects, although the EQS guidance document
defines a “chemical mixture” as a deliberate combination of two or
recognises that in some circumstances (i.e. release of known and
more individual substances; however, the legal definition of “sub-
constant composition mixtures or other mixtures with a partly
stance” in REACH can contain up to 20% arbitrary by-products
unknown, reasonably constant composition, that both change after
without the need for specific consideration. It also includes
entry into environment) an EQS for mixtures may be preferable to
Multi-Constituent Substances (MCS) which are substances result-
deriving EQSs for the individual constituent substances
(EC, 2011).
ing from a chemical reaction in which several constituents are
Thus, this guidance document briefly outlines how to estimate EQS
present at >10%, and UVCB (substances of Unknown or Variable
for mixtures, using the toxic unit (TU) approach (
Table S1) for well-
composition, Complex reaction products or Biological materials)
defined mixtures, and the hydrocarbon blocks and the use of non-
which are mixtures that cannot be completely identified by their
testing methods such as PETROTOX6 for the derivation of EQS for
chemical composition. MCS and UVCB are generally treated as a
petrochemical mixtures of unknown or variable composition.
single substance under REACH, and the testing of hazard and fate
The Marine Strategy Framework Directive adopts an ecosystem-
properties is therefore made on the mixture itself.
based approach, aiming at a Good Environmental Status (GES)
Although there might be multiple sources of exposure to the
focusing on 11 descriptors related to ecosystem features, human
same substance in real life (i.e. aggregate exposure), a registrant is
drivers and pressures (
Berg et al., 2015). Descriptor 8 is formulated
not obliged to take into account an exposure to the same substance
as “concentrations of contaminants are at levels not giving rise to
from activities from other producers or importers when doing the
pollution effects”, referring to substances or groups of substances
exposure assessment
(ECHA, 2013), and no specific hazard assess-
that give rise to a level of concern. Where possible, this should also
ment is required for chemical mixtures, preparations, MCS or
include effects which may be caused by synergistic or cumulative
UVCBs, unless they have persistent, bioaccumulative and toxic or
interactions between different contaminants. The list of com-
very persistent/very bioaccumulative (PBT/vPvB) properties; i.e. if
pounds mentioned is not restrictive, but rather takes the list as
they contain more than 80% of a substance with PBT/vPvB prop-
indicative. However, it is recognized that the causal relationships
erties
(ECHA, 2012). Thus, REACH is a typical substance-oriented
between levels of contaminants and observed effects are not well
regulation.
understood, and that there is rarely a direct relationship between
tissue levels of contaminants and their effects
(Law et al., 2010). The
3.5. Cosmetics (Reg 1223/2009)
understanding of the effects of mixtures of contaminants and of
interactions between contaminants and other environmental
Cosmetic products are typically a mixture, composed of multiple
stressors is even more limited. Therefore, it was decided that before
substances. According to Reg 1223/2009, the assessment of
implementing the Directive, more research was needed for assess-
cosmetic products should take into account the anticipated sys-
ing good environmental status in a coherent and holistic manner to
temic exposure to individual ingredients in a final formulation. It
support the ecosystem-based approach (Commission Decision 2010/
includes a toxicological profile of the substances that should take
477/EU, Article 3). It was still emphasized however, that it is
into account all significant toxicological routes of absorption ac-
important to consider cumulative and synergistic effects, not only
cording to the intended use, as well as possible impacts on the
by compounds listed in the WFD and others, but also compounds
toxicological profile due to interactions of substances.
that may entail significant risks to the marine environment from
The safety assessment of substances as individual ingredients
past and present pollution.
should consider the overall exposure to such substances stemming
from all sources, which implies the development of a harmonized
3.7. Drinking water directive (Directive 98/83/EC)
approach to the use of such overall exposure estimates. However,
the regulation does not specify if only sources of exposure linked to
This directive aims at ensuring a good quality of water intended
cosmetics uses are meant or if other uses (e.g. as pharmaceuticals)
for human consumption, by setting individual parametric values for
are included; although the latest guidance document only men-
substances that are of health concern at a level strict enough to
tions cosmetic use
(SCCS, 2015). Additionally, the safety of the
ensure human health protection on a life-long basis. Member States
cosmetic product itself must also be assessed, including possible
are in charge of ensuring that drinking water respects the
link to page 13 link to page 14 link to page 14 link to page 14 link to page 14 link to page 14 link to page 14 link to page 13 link to page 14 link to page 14
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
327
minimum requirements of the Directive, of setting values for those
chloroacetanilides (
USEPA, 2007, 2006a, 2006b, 2002b). It identifies
parameters which shall not be less stringent than those set out in
several groups of chemicals that are considered to induce a com-
the Directive, and of setting values for additional parameters not
mon toxic effect by a common mechanism, a so-called common
included in this Directive where required by human health pro-
mechanism group (CMG). Pesticides that contribute to exposures
tection. They shall also ensure the efficiency of the disinfection
by minor pathways are excluded, forming a subset called cumula-
treatment applied, while keeping any contamination from disin-
tive assessment group (CAG). For each CAG member, dose response
fection by-products as low as possible and ensure that regular
analyses are performed to determine its toxic potency for the
monitoring of the quality of drinking water is carried out. Thus, this
common effect. The concept of CA is normally used to estimate the
regulation requires the monitoring of individual parameters but
combined risks in the CAG, and the relative potencies of the CAG
does not address chemical RA or any mixture issue, although
members to one selected index chemical are defined for the stan-
disinfection products or by-products might be of concern for hu-
dardization of their common toxicity in terms of relative potency
man health
(Jeong et al., 2012; Nieuwenhuijsen et al., 2009).
factors (RPF). Exposure assessment is made through detailed
exposure scenarios, including all relevant pathways, durations and
3.8. Conclusions on mixture assessment under current EU
routes where simultaneous exposure may occur, as well as
legislation
sequential exposures. The output of this analysis is an aggregation
of exposures via all routes and pathways, for each chemical, which
Overall, chemical RA requirement in Europe is most of the time
is then expressed in terms of an equivalent exposure of the index
substance-driven and rather sector specific. Clear regulatory re-
chemical, by using RPFs. The risk contributions from each pathway
quirements for RA of mixtures of chemicals within a given regu-
and route should be evaluated both individually and in combina-
latory framework are rare, except for intentional mixtures such as
tion, in order to identify risk contributors. This risk characterisation
formulated products, and are most of the time prospective. Regu-
step also includes descriptions of variability and major areas of
latory requirements for RA of mixtures across various regulatory
uncertainty and the need for uncertainty and safety factors is
frameworks is scarce, although for aggregate exposure it is
determined.
important to acknowledge that numerous chemicals are concerned
The RA of drinking water (mainly focusing on disinfection by-
by more than one regulatory framework. Regarding exposure to
products) and air pollutants also requires consideration of chemi-
multiples substances, it has to be kept in mind that substances
cal mixtures, however the approaches developed for such an
regulated under different regulations can elicit similar effects or
assessment took minimal considerations of synergistic or antago-
follow the same mode of action so that combined effects cannot be
nistic effects, nor were non-chemical stressors taken into account
excluded.
(
Kortenkamp et al., 2009).
Regulations on unintentional mixtures in the environment, like
In 2003, the US EPA also published a Framework for Cumulative
the WFD and the MSFD, recognize the importance of mixture ef-
Risk Assessment (
USEPA, 2003), which provides starting principles
fects, but do not provide specific details on how this should be
for EPA’s CRA, for the future development of a comprehensive and
assessed. In these cases, assessment is hampered by a lack of in-
detailed guidance on methods for evaluating cumulative risk. This
formation on pollutants levels, identity and effects, as well as a
report emphasizes chemical risks to human health including the
limited understanding of the causal relationship between level of
effects from a variety of stressors, including non-chemical stressors.
contaminants and biological effects.
This was further developed in the 2006 publication on the “Con-
siderations for developing alternative health risk assessment ap-
3.9. Mixtures risk assessment in the US
proaches for addressing multiple chemicals, exposures and effects”
(
USEPA, 2006c), which presents concepts that could assist the
Also in the US; exposure to multiple chemicals is considered in
development of detailed guidance and provides explicit approaches
various regulatory frameworks. The CRA of contaminated site
for addressing some of the complicating “multiples” in CRA. These
specifically requires mixture RA for the evaluation of risks stem-
approaches include new methods and the extension of existing
ming from hazardous waste sites and chemical accidents
(USEPA,
methods to address health risk from multiple chemicals and mul-
1989, 1987). Exposure assessments are made on “reasonably
tiple exposure pathways and times.
maximally exposed” people, and the toxicity assessment is based
on reference toxicological value for each chemical. For carcinogens,
4. Chemical mixture risk assessment approaches: experiences
it is assumed that there is no dose threshold, and that the dose-
from case studies and an expert survey
response function is essentially linear. For non-carcinogenic
chemicals, a HI is calculated using the CA methods. If HI 1, it is
4.1. Case studies focusing on the assessment of mixtures
assumed that there is unlikely to be a risk; if HI > 1, further analysis
may be performed to determine whether application of dose
Case studies on chemical mixture RA are numerous in the
additivity to all the chemicals simultaneously is justifiable. In 2000,
literature and help in assessing the applicability of approaches and
the US EPA published a Supplementary Guidance for Health Risk
identifying data gaps and hurdles in the context of mixture RA.
Assessments for Mixtures, which introduces an Interaction Hazard
Recent case studies on a broad range of chemical classes and
Index
(USEPA, 2000).
exposure scenarios have been selected and reviewed (methodology
Pesticide RA requires the estimation of health risks from com-
being used, data gaps identified, outcome) and are reported
binations of pesticides with a common MOA. In order to do so, the
hereafter.
US EPA developed guidelines to determine which pesticides should
qualify for inclusion in common mechanism groups
(USEPA, 1999),
4.1.1. Chemical class based examples
and a guidance document concerning the application of the HI
4.1.1.1. Pesticides and environmental risk assessment. Because pes-
principle to pesticides (
USEPA, 2002a), which deals with simulta-
ticides are designed to be biologically active, directly emitted into
neous exposures from food, drinking water and residential (non-
the environment and used at fairly high volume, there are clear
occupational) use of pesticides for the general population. This RA
data requirements regarding the toxicity on target and non-target
procedure was used to extensively assess the risk linked to mix-
species. Therefore, pesticides are amongst the more data rich
tures
of
organophosphates,
carbamates,
triazines
and
chemicals regarding toxicity and they are frequently included in
link to page 13 link to page 13 link to page 13 link to page 14 link to page 14 link to page 13 link to page 13 link to page 14 link to page 14 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13
328
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
environmental monitoring programs and case studies on the effects
and agronomic parameters that are thought to represent a realistic
of co-occurring pesticides.
worst-case situation for the environmental context in which the
Junghans et al. (2006) showed that pesticides in mixtures are
model is to be run
(EFSA, 2014b).
clearly more toxic to algae than any individual component, which
clearly highlights the limit of current prospective RA process based
4.1.1.2. Pesticides and human health. Aggregate exposure assess-
on individual chemicals. Overall, CA showed a good predictive
ment combining dietary and non-dietary sources for a single sub-
quality over the complete range of effects considered, irrespective of
stance allows identifying the relative contributions to exposure,
the similarity or dissimilarity of their mechanisms of actions.
which can differ between particular scenarios and populations. For
Gregorio and Ch
evre (2014) also used the CA model and the risk
instance,
Kennedy et al. (2015) identified inhalation as the main
quotient methodology to retrospectively assess the risk posed by
route of exposure to pesticides for spray users, and dermal expo-
mixtures of chemicals (mainly pesticides) in the Geneva Lake and
sure for operators. For child bystanders, non-dietary (dermal)
the Rhone River, and identified the most problematic substances
exposure is estimated to be small compared to dietary exposure.
demanding risk reduction. The authors showed that the risk levels
However, data are lacking on realistic frequency or use of plant
associated with mixtures of compounds can rapidly exceed critical
protection products by amateurs and more generally on non-
aquatic thresholds, and that therefore, it is the sum of the substances
dietary exposure, which would be essential for chronic RA.
that is problematic. However, when the risk quotient is greater than
Regarding combined exposure to multiple pesticides, EFSA
1, it is often due to only a few chemicals (1e4 in this case).
published a guidance document on probabilistic modelling of di-
The pesticide toxicity index methodology has also been used as
etary exposure, including an optimistic and a pessimistic model,
a screening tool to assess potential aquatic toxicity of complex
which were applied to pesticide residue mixtures from the triazole
pesticide mixtures by combining measures of pesticide exposure
group
(Boon et al., 2015). The grouping of the chemicals was based
and acute toxicity in an additive toxic-unit model (
Nowell et al.,
on the toxicological effect. In the optimistic model run, none of the
2014), but this methodology is a relative ranking system that in-
simulated acute nor chronic exposures exceeded the reference
dicates that one sample is likely to be more or less toxic than
toxicological value (i.e. respectively Acute Reference DoseeARfD-
another sample, without indicating that toxicity will necessarily
for acute and ADI for chronic); however in the pessimistic model
occur.
run, which takes into account animal commodities including cattle
Moreover, those methodologies are limited because they do not
milk and meat at the level of the MRLs, an exceedance of the ARfD
consider synergistic effects, which are known to be possible with
or ADI was frequently observed, and the model was judged to result
pesticides. As an example, the combination of pyrethroid in-
in unrealistic conclusions regarding the contribution of animal
secticides and azoles fungicides such as deltamethrin and pro-
commodities to the dietary exposure. The authors conclude that the
chloraz, is known to be much more toxic to bees than the chemicals
pessimistic model runs, besides being laborious, could provide re-
individually with a ratio ranging from 366 to 1786 fold (
Colin and
sults that are too far from reality, and that the optimistic model
Belzunces, 1992; Sammataro and Yoder, 2011). The proposed
runs would likely give results underestimating the real exposure.
mechanism is that those fungicides, by inhibiting ergosterol
Some kind of intermediate ‘realistic’ scenario is therefore needed,
biosynthesis via the inhibition of cytochromes P450 also involved
which would result in more realistic acute and chronic exposures,
in detoxification, decrease the capacity of the organisms to detoxify
conservative enough (precautionary principle) without being over-
other chemicals. Similar interaction has been found between mi-
conservative (
Boon et al., 2015).
ticides and pyrethroids, or between miticides
(Sammataro and
In the case of retrospective RA using biomonitoring data, all the
Yoder, 2011). Synergism has also been shown to occur between
sources of exposure are by default included in the RA, but when
organophosphates and carbamates pesticides in salmon
(Laetz
doing a prospective RA, this is not the case, and residues of pesti-
et al., 2009).
cides from sources other than current plant protection uses of
Additionally, linking toxic effects to monitoring data can only
active substances are usually not taken into account; except for
consider chemical substances that are identified; substances that
biocides and veterinary drug uses in the settings of MRLs. More-
are not analysed nor detected because they are present at con-
over, there were no case studies identified on Human Health
centrations below the limit of detection are not taken into account,
regarding combined exposure to several active substances
although they might be biologically active. Other methodology
including both dietary and non-dietary exposure, nor were case
could be used, such as a two-step model approach mixing CA for
studies found considering aggregate exposure to pesticide active
modelling mixture toxicity of individual MoA, and IA to combine
substances across regulatory frameworks, which could be of
the toxicity of different MoA
(De Zwart and Posthuma, 2005); or the
interest.
use of CA or IA on species sensitivity distributions, which can be
much more robust, but requires a huge quantity of ecotoxicity data,
4.1.1.3. Biocides. The methodology published by ECHA for HRA
which are often not available (
Gregorio et al., 2013).
addressing combined exposure to multiple substances within a
Most of the case studies on pesticides are carried out retro-
single biocidal product (
ECHA, 2015) can theoretically be applied to
spectively, based on monitoring data; however, such types of RAs
assess aggregate exposure to multiple biocidal product types con-
could also be carried out prospectively, prior to placing a product on
taining the same AS, by combining the exposure estimates from
the market, and based on calculated Predicted Environmental
uses/releases from the different product types, although this would
Concentration (PEC) data, in order to screen and detect the com-
require lots of data. It could also theoretically be applied to com-
binations that could be of concern. One way of addressing com-
bined exposure to multiple substances coming from different
bined environmental risk from pesticide co-exposure could be to
sources of release and/or uses, provided that the various exposure
base the selection of co-occurring pesticides on their use patterns
scenarios and cumulative effects are taken into account, and that
in specific crops or based on common tank mixes. Data collections
sufficient data are available to do so. However, to our knowledge
on use patterns have been performed throughout Europe that could
this has not been put into practice so far.
serve as a basis
(Garthwaite et al., 2015).
Moreover, prospective RA of pesticides could be improved by
4.1.2. Groups of compounds or matrix-based examples
the development of environmental scenarios for mechanistic effect
4.1.2.1. Environmental risk assessment. Several studies focus on
modelling of pesticides, defined as a combination of abiotic, biotic
relating effects to the compounds present, including many effects
link to page 14 link to page 14 link to page 14 link to page 14 link to page 12 link to page 13 link to page 13 link to page 13 link to page 13 link to page 14 link to page 13 link to page 12 link to page 12 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 14 link to page 14 link to page 14 link to page 14
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
329
currently not covered by the respective regulations. E.g.
Tang et al.
of anti-androgenic compounds should be performed. Moreover,
(2014) investigated waste water and recycled water samples,
those case studies do not take into account synergistic effects,
considering 299 chemicals present at concentrations below the
although synergism has been observed with a mixture of anti-
regulatory safety limit. Artificial mixtures of those chemicals were
androgens with diverse MOA
(Christiansen et al., 2009). Further-
found to explain less than 3 and 1% of the observed effluent cyto-
more, it is often neglected that the effective internal dose of some
toxicity and oxidative stress response respectively, showing that
anti-androgenic chemicals (i.e p,p0-DDE and BDE 99) may be higher
the identified compounds do not explain the observed toxic effect.
than suggested due to their highly lipophilic nature. A different
This large proportion of unknown toxicity, which could either be
dose metric and tissue concentrations should be used, although the
due to other non-monitored chemicals or to mixture effects, calls
data necessary for such calculations are currently not available
for effect-based monitoring complementary to chemical moni-
(
Kortenkamp and Faust, 2010).
toring (
Ohe et al., 2004; Tang et al., 2014).
Another study has shown that butyl paraben makes up 50% of
Pesticides, followed by pharmaceuticals and personal care
the HI of the highly exposed population group to anti-androgens,
products seem to dominate the observed mixture effects on the
and that the percentile of the population of children from 0 to 3
environment (
Tang et al., 2014), and the mixture risk quotient of
year old with an exposure probability to propyl- and butylparaben
pharmaceuticals in sewage treatment plant effluents has been
above the assumed ‘‘safe’’ level, was estimated to be 13% and 7%,
shown to regularly exceed 1, which points out the fact that those
respectively. Further refinement of the exposure calculations is
mixtures can be of concern
(Backhaus and Karlsson, 2014).
therefore necessary
(Gosens et al., 2013; Kortenkamp and Faust,
However, it has to be highlighted that many case studies do not
2010), although hampered by the scarcity of detailed data on the
take into account the degradation products and metabolites of the
use of personal care products, especially for children. Furthermore,
chemicals in the environment; nor the bioconcentration potential
those chemicals are also used in other types of products such as
of the mixture. In order to assess the risk of complex effluents based
pharmaceuticals and food additives and those uses have never been
on both acute toxicity and the bioconcentration potential of the
assessed together, as they are regulated under different legal
mixture, a methodology has been developed, combining the esti-
frameworks. More exposure data regarding these products would
mation of Kow of a mixture by RP-HPLC on one hand, and the
therefore be needed to obtain a more accurate estimate of the
extraction, fractionation and ecotoxicity testing of the fraction on
aggregate exposure to parabens.
the other hand (Effect Directed Analysis) (
Guti
errez et al., 2008).
A case study focusing on dioxins using biomonitoring data of 26
Finally, a relative hazard index (RHI) for any particular mixture is
dioxin-like compounds based on toxic equivalency factors (TEFs)
estimated, ranging from 1 to 10, which takes into account the
found similar MCR values in two occupationally exposed groups
bioconcentration potential. Instead of further analysing the whole
and in the general public, although the two occupational groups
sample, the efforts can be focussed on the more toxic fractions to
have higher total toxicity equivalence (TEQ) levels. MCR values
identify relevant toxic compounds, which reduces analysis costs.
indicated that only 2e5 of the 26 chemicals make significant con-
This method could add value for whole effluent assessment and
tributions to total TEQ values. This was also the case for human
help to refine PNEC values, however it has not been applied to real
exposure to environmental mixtures in surface water usually
effluents so far
(Gutierrez et al., 2008).
dominated by a relatively small number of components
(Han and
Price, 2011).
4.1.2.2. Human risk assessment. Investigating food contact mate-
When looking at human exposure via surface water or effluents
rials, olefins and saturated hydrocarbons for the Non Intentionally
from wastewater treatment plants, 2% of the considered mixtures
Added Substances and ethyl-4-ethoxybenzoate for water bottles
were of concern for human health effects (HI > 1), although those
were identified as main contributors to toxicity from multiple
HH effects would have been sufficiently addressed by chemical-by-
substances released to food
(Price et al., 2014), although the risk of
chemical approaches and showed little need for an assessment of
adverse effects to individuals were found to be low (HI < 1).
the combined exposure (individual HQ > 1) (
Price et al., 2012).
However, the study did not consider many inorganics, due to a lack
However, the assumption made of a 10-fold dilution of the effluents
of available reference values. Potential ED effects of five phthalates
could be wrong for small rivers under low-flow condition; and for
were evaluated based on human urinary biomonitoring data. The
rivers receiving multiple discharges the receiving water might
HI of the mixture exceeded the “safe” level for 6.2% of adults and
already contain some of the compounds from upstream discharges,
25% of children (
Dewalque et al., 2014). This means that even when
which would increase the risk.
focusing on a small subset of compounds, safe levels might be
Another relevant route for human exposure to unintentional
exceeded. DEHP was the only phthalate studied for which the main
mixtures relates to indoor air, since people spend approximately
pathway of exposure was the dietary intake; for all other, it seemed
90% of their time indoors, of which 2/3 would be spent at home
to be a minor pathway, highlighting the importance to take into
(
WHO, 2014a). This makes this type of exposure of concern, espe-
account all pathways of exposure to make a reliable RA. This wide
cially for subpopulations such as young children since their lung
exposure to phthalates is confirmed by
Becker et al. (2009), who
structure and immune system is not yet fully developed. Lead is the
detected 12 phthalate metabolites in urine samples of German
most studied indoor pollutant, while VOCs (Volatiles Organic
children, with contamination levels 3e5 times higher than in adults
Compounds) and SVOCs (Semi Volatiles Organic Compounds) are
(
Becker et al., 2009). This might be a situation of concern, as anti-
still of concern and could be correlated with allergic effects and
androgenic effects of phthalates on reproductive health could
respiratory symptoms in children (
Le Cann et al., 2011). For carci-
occur at all life stages and because phthalates are not the only anti-
nogenic VOCs the estimated carcinogenic risks were up to three
androgenic chemicals to which humans are exposed. This was
orders of magnitude higher than the one proposed as acceptable by
confirmed by
Kortenkamp and Faust (2010), which have assessed
risk management bodies (
Sarigiannis et al., 2011), whereas con-
cumulative anti-androgenic effects of 15 chemicals including
servative exposure limits were not exceeded for non-carcinogenic
phthalates and concluded that the cumulative risk exceeds
effects, except for formaldehyde. However, the RA evaluation pro-
acceptable levels for people on the upper end of the exposure
cess faces difficulties, either due to the relative paucity of indoor air
levels. The results suggest that combined exposure to anti-
quality measurements in many EU countries, or by the lack of
androgens have reached levels of concern, and that larger human
sampling consistency in the already existing studies, indicating the
biomonitoring studies including pertinent biomarkers of exposure
need for additional measurements of indoor air quality following a
link to page 14 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 10 link to page 13 link to page 13 link to page 11 link to page 12 link to page 12 link to page 14 link to page 14 link to page 14 link to page 14 link to page 13
330
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
harmonized sampling and analytical protocol. Some hazardous
VOCs are also directly link to the use of flame retardants; and
halogenated organophosphates releasing chlorinated degradation
products (
Salthammer et al., 2003). Organophosphate flame re-
tardants’ exposure has been shown to be widespread, with hand-
to-mouth contact or dermal absorption being important path-
ways of exposure
(Hoffman and Stapleton, 2015); However, the RA
of poly-brominated diphenyl ether in the indoor environment does
not seem to be of concern, except in the US (
Fromme et al., 2015;
Lim et al., 2014), although according to some authors the high
exposure to these substances indoor calls for better risk assess-
ments that include mixtures effects
(de Boer et al., 2016).
A high variability has been found in the proportion of samples of
concern for mixture toxicity in residential indoor air with the MCR
methodology, this variability being due both to the variation in
indoor air contaminant levels across the studies but also to other
factors such as differences in number and type of substances
monitored, analytical performance, and choice of RVs
(De Brouwere
et al., 2014).
Fig. 2. 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
4.2. Expert survey on approaches, experiences and future directions
account?” divided by stakeholder group. Chemicals were further specified in the sur-
in assessing human and environmental health risks from chemical
vey as “multiconstituent or UVCB substances under REACH”. Other mixtures of
mixtures
importance mentioned were those present in human tissues and container systems.
In order to gain an overview of current practices and experi-
The most used tools in the RA of mixtures were QSARs, Read-across
ences with assessing the effects and risks from combined exposure,
and in-vitro tools, both for HRA and ERA (
Fig. 3). TTC approach is
an online survey was performed among experts in the field in the
also often used for HRA of single chemical and chemical mixture;
period of January to March 2015, addressing both, human health
this approach is not frequently used in ERA but an eco-TTC
and environmental RA. Fifty-eight experts from 21 countries,
approach is currently being developed to assist ERA (
Belanger
different stakeholder groups and sectors of legislation participated
et al., 2015).
in the survey. The main sectors where most experience is already
Experts were also asked in the survey about their experience
gained in assessing mixtures are in the area of plant protection
with the three most widespread international frameworks devel-
products and chemicals regulated under REACH. These were also
oped for addressing combined exposure to chemical mixtures, i.e.
rated highest regarding the priority for performing mixture as-
the WHO/IPCS framework
(Meek et al., 2011), Proposal by the three
sessments (
Fig. 2), followed by pharmaceuticals, biocides and food
non-food scientific committees of the European Commission
or feed contaminants.
(
SCHER et al., 2012), and the proposal by CEFIC MIAT (
Price et al.,
Experts mainly used CA based prediction tools and less IA based
2012). 73% of the experts were familiar with at least one of those
approaches. Some experts mentioned IA based approaches as a
frameworks, mainly with the WHO/IPCS framework, which was
method they had abandoned due to the large amount of data
rated as an easy and transparent approach. However, it was found
needed for IA based predictions. In order to perform IA assess-
to be rather general and lacks criteria when refinement should be
ments, usually the full dose response curve for each mixture
stopped. The data available usually allow only to perform Tier 1 and
component is needed while in CA assessments the reference values
2 assessments and not to go to higher tiers. The SCHER, SCENIHR,
are sufficient
(Kortenkamp et al., 2009). When experts were asked
SCCS framework is considered useful for organizing data and
about the need for addressing interactions in mixture RA, the vast
deciding how to perform the assessment, but the main limitation
majority (65 answers) agreed that interactions should be addressed
mentioned for those two frameworks is that they provide a more
on a case-by-case basis if there is specific evidence from which
conceptual framework and less practical guidance. The CEFIC MIAT
interactions could be expected. Only a small part of the experts (13
framework was judged as useful since it comprises practical tools;
answers) thought that interactions do not need to be addressed
however, most input received on this framework was from experts
specifically since they are either covered by CA based conservative
involved in its development.
approaches, or since they are anyway rare at relevant concentra-
tions. Only 3 experts agreed to the statement that a conservative
default safety factor should be applied to cover potential in-
5. Regulatory challenges and future perspectives
teractions in a non-case-specific way.
Experts were then asked about the use of novel tools in the RA of
5.1. Legal requirements
mixtures, such as in vitro methods, omics, (Q)SARs, read-across,
toxicokinetic modelling, TTC approaches, Adverse Outcome Path-
As described above, chemical RA and mixture RA is most of the
ways (AOPs), or Integrated Approach to Testing and Assessment
time substance-driven and sector-specific (
Kienzler et al., 2014).
(IATA). These methodologies were selected based on their potential
Clear regulatory requirements for RA of mixtures of chemicals
to contribute to the improved assessment of combined effects and
within a given regulatory framework are rare, except for intentional
unravelling modes of action
(Bopp et al., 2015). Expert opinions
mixtures such as formulated products. Moreover, regulatory re-
were split between those applying them (often in a research
quirements for RA of mixtures across various regulatory frame-
context) and those that generally think these tools are valuable but
works is scarce, even though numerous chemicals are subject to the
their use is currently limited because of a lack of guidance, lack of
provisions of more than one regulatory framework. Thus, aggregate
data, or lack of expertise. A general need for clear guidance for
exposure to one chemical regulated under different legislation as
combined exposure assessments was highlighted by many experts.
well as combined exposure to different chemicals with similar toxic
link to page 13 link to page 13 link to page 13 link to page 13 link to page 13 link to page 12 link to page 12 link to page 13 link to page 13 link to page 13 link to page 12 link to page 13 link to page 14 link to page 14
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
331
Fig. 3. Replies to the question “Do you apply in vitro tools/omics approaches/(quantitative) structure activity relationships ((Q)SARs)/read-across/physiologically based pharma-
cokinetic (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?”.
MoA or similar effects need to be further addressed
(Evans et al.,
again highlights the necessity to consider chemicals exposure
2015).
beyond regulatory framework boundaries.
It has to be stressed that for both environmental and human
5.2. Retrospective versus prospective risk assessment of mixtures
exposure, none of the case studies reviewed takes synergistic ef-
fects or bioaccumulation (except when biomonitoring data are
The review of case studies identified only retrospective mixture
used) into account, which could underestimate the risks. More
assessments, although several environmental scenarios have been
knowledge could be gained from additional case studies covering
proposed for prospective assessment of pesticides. These scenarios
different sectors to further analyse this source of uncertainty. In the
allow characterizing exposure, direct and indirect effects and re-
survey, most experts stated that interactions should be considered
covery of aquatic non-target species in order to assess individual,
if there is specific evidence for interactions and on a case-by-case
population and/or community-level effects and recovery under
basis.
realistic worst-case condition. A conceptual framework for the
To predict and address interactions, toxicokinetic and tox-
development of such scenario has been developed
(EFSA, 2014b;
icodynamic modelling are valuable tools. Toxicokinetic and tox-
Rico et al., 2015). Although the proposed scenarios still focus on
icodynamic information to feed into these models, can be gained
the assessment of single compounds, in principle they can be used
e.g. from in vitro studies or using QSAR models. Adverse Outcome
for a prospective assessments of mixtures, taking into account the
Pathway (AOP) (
Ankley et al., 2010), might also help to identify
already authorized uses of other chemicals to identify possible
potential nodes between those different pathways which might
situations of concern. These data can be assessed using the recently
trigger interaction and potentiation by giving information on the
developed cumulative assessment groups for PPPs, based on com-
key events of the different toxicity pathways triggered by the
pounds with similar effects or MoA
(EFSA, 2014a). Comprehensive
different chemicals within a mixture.
approaches are needed, however, as the implementation of risk
Also read-across information from similar mixtures can be used
management measures becomes a challenge if risks from combined
to identify mixtures where interactions could play a role and should
exposure to multiple chemicals under different legislation are
be further investigated (
Bopp et al., 2015). A way forward, as
identified.
chemical characterisation of mixtures might be difficult, would be
to further investigate read-across based on biological similarity
5.3. Considering interactions
from in vitro screening (
Reif et al., 2013, 2010).
Current evidence in the literature suggests that interactions
5.4. Underlying exposure and toxicity data
(synergistic or antagonistic effects) at lower concentration levels
such as environmental concentrations are rare and, if observed,
For exposure assessments, usually exposure concentrations are
lead to relatively small deviations from CA predictions (
Boobis
estimated based on volume of production or use or on product use
et al., 2011; Cedergreen, 2014; Cedergreen et al., 2012). However,
surveys, can be modelled using appropriate scenarios or can be
interactions are frequently reported for pesticides (
Colin and
based on (bio)monitoring data. Monitoring data are essential as
Belzunces, 1992; Laetz et al., 2009; Sammataro and Yoder, 2011)
they can provide information on magnitude, duration, frequency
and can also occur between chemicals from different regulatory
and/or timing of real exposure, and allow to assess the co-exposure
silos: it has been shown that a pharmaceutical oestrogen and a
patterns to chemicals
(Qian et al., 2015), both for human and
persistent organochlorine pesticide, both exhibiting low efficacy
environmental RA. Several authors from the selected case studies
when studied separately, were leading to synergistic activation by
highlighted the problem of large data gaps regarding the avail-
cooperatively binding to the pregnane X receptor. In this case, the
ability of measured exposure concentrations and appropriate
binary mixture induces a substantial biological response at doses at
modelling approaches. An important aspect for assessing combined
which each chemical individually is inactive, each ligand enhancing
effects is also the temporal profile of co-exposure. Ideally, internal
the binding affinity of the other
(Delfosse et al., 2015). This example
or target organ exposure concentrations over time should be
link to page 13 link to page 12 link to page 13 link to page 13 link to page 14 link to page 14
332
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
available. Toxicokinetic modelling plays a prominent role here in
assessment of mixtures, more guidance on their use is needed to
gaining further insight, although it is currently rarely considered at
facilitate a more widespread application. In the survey, a lack of
higher tier assessments.
guidance, lack of data, and lack of expertise were frequently cited as
For hazard assessment mostly toxicological data from published
main reasons hampering the application of novel concepts and
databases are used (based on peer-reviewed literature or substance
tools.
authorisation dossiers). Data gaps were often identified in the
reviewed case studies. Approaches to fill these are e.g. the use of the
6. Conclusion
Threshold of Toxicological Concern (TTC) or in silico approaches
(QSAR and read-across) (
Bopp et al., 2015). Another issue is also the
Model specifications, exposure and toxic reference values used
combination of toxicity data based on reference values that are
can greatly influence the outcome of a mixture RA. Therefore, it is
often derived from different endpoints. For the combined effect
crucial to properly document and justify the choices that have been
assessment, usually CA based predictions for groups of chemicals
made, and to carefully interpret the results considering the un-
eliciting similar effects or showing a similar MoA are used. The
derlying hypothesis, the related uncertainties, and the degree of
available reference values from substance authorisation dossiers
conservatism that has been chosen. Two guidance documents have
used in single substance RA are often based on the most sensitive
recently been published on characterisation of uncertainties in RA
endpoint measured. This is however not necessarily the one for the
(
EFSA SC, 2015; WHO, 2014b), however the assessment of uncer-
effect under consideration. To use such reference values in lower
tainty in the hazard characterisation for mixtures and for cumula-
tier assessments is acceptable; however, more detailed information
tive exposure to multiple stressors still need to be further
on specific effects is needed for further refinements and is often not
investigated
(WHO, 2014b).
available.
Several frameworks for the assessment of chemical mixtures
More data on the underlying MoA are also needed to improve
have been developed by international bodies in recent years, i.e.
grouping approaches for mixture components. The AOP concept
WHO/IPCS (
Meek et al., 2011; Price et al., 2012; SCHER et al., 2012).
has been shown to provide a valuable framework to map available
These frameworks provide high-level guidance as well as tiered
toxicity data for mixture components to key events in AOP net-
approaches for screening level assessments and further re-
works relevant for grouping as well as to identify data gaps and
finements. A limitation in their application arises however due to
tailored testing strategies (
Ankley et al., 2010).
the lack of data for performing higher tier assessments. Therefore,
The limitation in the availability of appropriate exposure and
there are still many open issues and more detailed guidance is
toxicity data has a direct impact on the uncertainty of the RA
needed, that harmonises approaches used across different legisla-
outcome. As a matter of fact, data gaps are identified as the major
tive sectors. There is now a need to build on all these frameworks to
issue when it comes to RA of chemical mixtures, especially when
develop a robust and transparent approach not only for conducting,
dealing with particular uses or population subgroups (i.e. amateur
but also reporting a chemical mixture RA.
uses of pesticides, frequency of use of cosmetic for children). The
integration of data that were originally derived under different
Appendix A. Supplementary data
scope and that are not always directly comparable, is linked to an
increased uncertainty in the assessment of combined effects.
Supplementary data related to this article can be found at
http://
dx.doi.org/10.1016/j.yrtph.2016.05.020.
5.5. Most commonly studied compound groups in mixture risk
assessment.
Transparency document
Pesticides followed by pharmaceuticals and personal care
Transparency document related to this article can be found
products dominated the observed mixture effects in the case
online at
http://dx.doi.org/10.1016/j.yrtph.2016.05.020.
studies, whereas chemicals regulated under REACH and plant
protection products were the areas where most experts partici-
References
pating in the survey had experience. Tributyltin, polycyclic aro-
matic hydrocarbons, and brominated flame retardants are also
Ankley, G.T., Bennett, R.S., Erickson, R.J., Hoff, D.J., Hornung, M.W., Johnson, R.D.,
major contributors to the environmental chemical risk of the
Mount, D.R., Nichols, J.W., Russom, C.L., Schmieder, P.K., Serrrano, J.A.,
monitored chemicals. Human exposure to parabens, phthalates and
Tietge, J.E., Villeneuve, D.L., 2010. Adverse outcome pathways: a conceptual
framework to support ecotoxicology research and risk assessment. Environ.
more generally anti- androgenic chemicals seems to be of concern,
Toxicol. Chem. 29, 730e741.
http://dx.doi.org/10.1002/etc.34.
particularly for highly exposed or more sensitive population
Backhaus, T., Arrhenius, A., Blanck, H., 2004. Toxicity of dissimilarly acting sub-
subgroups.
stances to natural algal communities: predictive power and limitations of IA
and CA. Environ. Sci. Technol. 6363e6370.
Backhaus, T., Faust, M., 2012. Predictive environmental assessment of chemical
5.6. Use of novel concepts and tools for the hazard assessment of
mixtures: a conceptual framework. Environ. Sci. Technol. 2564e2573.
mixtures.
Backhaus, T., Karlsson, M., 2014. Screening level mixture risk assessment of phar-
maceuticals in STP effluents. Water Res. 49, 157e165.
http://dx.doi.org/10.1016/
j.watres.2013.11.005.
A high potential in applying novel tools and scientific method-
Becker, K., G€
oen, T., Seiwert, M., Conrad, A., Pick-Fuss, H., Müller, J., Wittassek, M.,
ologies (e.g. AOPs, in vitro methods, omics, in silico approaches,
Schulz, C., Kolossa-Gehring, M., 2009. GerES IV: phthalate metabolites and
toxicokinetic modelling, TTC approach, IATA) for the assessment of
bisphenol A in urine of German children. Int. J. Hyg. Environ. Health 212,
685e692.
http://dx.doi.org/10.1016/j.ijheh.2009.08.002.
chemical mixtures was identified based on a literature review and
Belanger, S.E., Sanderson, H., Embry, M.R., Coady, K., DeZwart, D., Farr, B.A.,
the expert survey
(Bopp et al., 2015). These tools allow meaningful
Gutsell, S., Halder, M., Sternberg, R., Wilson, P., 2015. It is time to develop
information to be obtained on individual mixture components or
ecological thresholds of toxicological concern to assist environmental hazard
assessment. Environ. Toxicol. Chem. 34, 2864e2869.
http://dx.doi.org/10.1002/
whole mixtures, enabling a better understanding of the underlying
etc.3132.
mechanisms of mixture effects. Their main strengths lie in their
Berg, T., Fürhaupter, K., Teixeira, H., Uusitalo, L., Zampoukas, N., 2015. The marine
integrated use and ability to put different aspects regarding the
strategy framework directive and the ecosystem-based approach e pitfalls and
solutions.
Mar.
Pollut.
Bull.
96,
18e28.
http://dx.doi.org/10.1016/
hazard from combined exposure to multiple chemicals into
j.marpolbul.2015.04.050.
context. In order to benefit from these tools in the hazard
Boobis, A., Budinsky, R., Collie, S., Crofton, K., Embry, M., Felter, S., Hertzberg, R.,
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
333
Kopp, D., Mihlan, G., Mumtaz, M., Price, P., Solomon, K., Teuschler, L., Yang, R.,
j.efsa.2014.3589.
Zaleski, R., 2011. Critical analysis of literature on low-dose synergy for use in
EFSA, 2012. Guidance on the Use of Probabilistic Methodology for Modelling Di-
screening chemical mixtures for risk assessment. Crit. Rev. Toxicol. 41, 369e383.
etary Exposure to Pesticide Residues. European Food Safety Authority, Parma,
http://dx.doi.org/10.3109/10408444.2010.543655.
Italy.
http://dx.doi.org/10.2903/j.efsa.2012.2839.
Boon, P.E., van Donkersgoed, G., Christodoulou, D., Cr
epet, A., D’Addezio, L.,
EFSA SC, 2015. Guidance on Uncertainty in EFSA Scientific Assessment. European
Desvignes, V., Ericsson, B.-G., Galimberti, F., Ioannou-Kakouri, E., Jensen, B.H.,
Food and Safety Authority. Scientific Opinion, Parma, Italy.
http://dx.doi.org/
Rehurkova, I., Rety, J., Ruprich, J., Sand, S., Stephenson, C., Str€
omberg, A.,
10.2903/j.efsa.20YY.NNNN.
Turrini, A., van der Voet, H., Ziegler, P., Hamey, P., van Klaveren, J.D., 2015. Cu-
EMEA, 2006. Guideline on the Environmental Risk Assessment of Medicinal. Eu-
mulative dietary exposure to a selected group of pesticides of the triazole group
ropean Medicines Agency, London, UK.
in different European countries according to the EFSA guidance on probabilistic
EU, 2013a. Regulation (EU) No 283/2013 Setting Out the Data Requirements for
modelling.
Food
Chem.
Toxicol.
79,
13e31.
http://dx.doi.org/10.1016/
Active Substances, in Accordance with Regulation (EC) No 1107/2009 of the
j.fct.2014.08.004.
European Parliament and of the Council Concerning the Placing of Plant Pro-
Bopp, S., Berggren, E., Kienzler, A., van der Linden, S., Worth, A., 2015. Scientific
tection Products on the Market. OJ L93.
Methodologies for the Assessment of Combined Effects of Chemicals e a Survey
EU, 2013b. Regulation (EU) No 284/2013 Setting Out the Data Requirements for
and Literature Review.
Plant Protection Products, in Accordance with Regulation (EC) No 1107/2009 of
Bunke, D., Groß, R., Kalberlah, F., Oltmanns, J., Schwarz, M., Reihlen, A., Reineke, N.,
the European Parliament and of the Council Concerning the Placing of Plant
2013. 4M: Mixtures under REACH. Concepts and Options to Act. Mixtures in the
Protection Products on the Market. OJ L93.
Environment e Development of Assessment Strategies for the Regulation of
Evans, R.M., Martin, O.V., Faust, M., Kortenkamp, A., 2015. Should the scope of
Chemicals under REACH.
human mixture risk assessment span legislative/regulatory silos for chemicals?
Cedergreen, N., 2014. Quantifying synergy: a systematic review of mixture toxicity
Sci. Total Environ.
http://dx.doi.org/10.1016/j.scitotenv.2015.10.162.
studies within environmental toxicology. PLoS One 9, e96580.
http://dx.doi.org/
Fromme, H., Becher, G., Hilger, B., V€
olkel, W., 2015. Brominated flame retardants e
10.1371/journal.pone.0096580.
exposure and risk assessment for the general population. Int. J. Hyg. Environ.
Cedergreen, N., Sørensen, H., Svendsen, C., 2012. Can the joint effect of ternary
Health 219, 1e23.
http://dx.doi.org/10.1016/j.ijheh.2015.08.004.
mixtures be predicted from binary mixture toxicity results? Sci. Total Environ.
Garthwaite, D., Sinclair, C., Glass, R., Pote, A., Trevisan, M., Sacchettini, G.,
427e428, 229e237.
http://dx.doi.org/10.1016/j.scitotenv.2012.03.086.
Spanoghe, P., Ngoc, K.D., Fevery, D., Charistou, A., Nikolopoulou, D., Arapaki, N.,
Christiansen, S., Scholze, M., Dalgaard, M., Vinggaard, A.M., Axelstad, M.,
Gerritsen-ebben, R., Spaan, S., Gonz
alez, F.E., Stobiecki, S.,
Sliwi
nski, W.,
Kortenkamp, A., Hass, U., 2009. Synergistic disruption of external male sex or-
Stobiecki, T., Hakaite, P., 2015. Collection of Pesticide Application Data in View of
gan development by a mixture of four antiandrogens. Environ. Health Perspect.
Performing Environmental Risk Assessments for Pesticides. EFSA supporting
117, 1839e1846.
http://dx.doi.org/10.1289/ehp.0900689.
publication, 2015:EN-846.
Colin, M.-E., Belzunces, L.P., 1992. Evidence of synergy between prochloraz and
Gosens, I., Delmaar, C.J.E., Burg, W., Heer, C., De Schuur, A.G., 2013. Aggregate
deltamethrin in apis mellifera L.: a convenient biological approach. Pest Manag.
exposure approaches for parabens in personal care products;: a case assess-
Sci. 36, 115e119.
http://dx.doi.org/10.1002/ps.2780360206.
ment for children between 0 and 3 years old, 24, 208e214.
http://dx.doi.org/
de Boer, J., Ballesteros-G
omez, A., Leslie, H.A., Brandsma, S.H., Leonards, P.E.G., 2016.
10.1038/jes.2013.33.
Flame retardants: dust e and not food e might be the risk. Chemosphere 2014,
Gregorio, V., Ch
evre, N., 2014. Assessing the risks posed by mixtures of chemicals in
10e13.
http://dx.doi.org/10.1016/j.chemosphere.2015.12.124.
freshwater environments: case study of Lake Geneva. Switz. Wiley Interdiscip.
De Brouwere, K., Cornelis, C., Arvanitis, A., Brown, T., Crump, D., Harrison, P.,
Rev. Water 1, 229e247.
http://dx.doi.org/10.1002/wat2.1018.
Jantunen, M., Price, P., Torfs, R., 2014. Application of the maximum cumulative
Gregorio, V., Ch
evre, N., Junghans, M., 2013. Critical issues in using the common
ratio (MCR) as a screening tool for the evaluation of mixtures in residential
mixture toxicity models concentration addition or response addition on species
indoor air. Sci. Total Environ. 479e480, 267e276.
http://dx.doi.org/10.1016/
sensitivity distributions: a theoretical approach. Environ. Toxicol. Chem. 32,
j.scitotenv.2014.01.083.
2387e2395.
http://dx.doi.org/10.1002/etc.2320.
De Zwart, D., Posthuma, L., 2005. Complex mixture toxicity for single and multiple
Groten, J.P., Butler, W., Feron, V.J., Kozianowski, G., Renwick, a G., Walker, R., 2000.
species: proposed methodologies. Environ. Toxicol. Chem. 24, 2665e2676.
An analysis of the possibility for health implications of joint actions and in-
http://dx.doi.org/10.1897/04-639r.1.
teractions between food additives. Regul. Toxicol. Pharmacol. 31, 77e91.
http://
Delfosse, V., Dendele, B., Huet, T., Grimaldi, M., Boulahtouf, A., Gerbal-Chaloin, S.,
dx.doi.org/10.1006/rtph.1999.1356.
Beucher, B., Roecklin, D., Muller, C., Rahmani, R., Cavaill
es, V., Daujat-
Guti
errez, S., Fern
andez, C., Escher, B.I., Tarazona, J.V., 2008. A new hazard index of
Chavanieu, M., Vivat, V., Pascussi, J.-M., Balaguer, P., Bourguet, W., 2015. Syn-
complex mixtures integrates bioconcentration and toxicity to refine the envi-
ergistic activation of human pregnane X receptor by binary cocktails of phar-
ronmental risk assessment of effluents. Environ. Int. 34, 773e781.
http://
maceutical and environmental compounds. Nat. Commun. 6, 8089.
http://
dx.doi.org/10.1016/j.envint.2008.01.002.
dx.doi.org/10.1038/ncomms9089.
Han, X., Price, P.S., 2013. Applying the maximum cumulative ratio methodology to
Delmaar, C., Bokkers, B., Ter Burg, W., Schuur, G., 2015. Validation of an aggregate
biomonitoring data on dioxin-like compounds in the general public and two
exposure model for substances in consumer products: a case study of diethyl
occupationally exposed populations. J. Expo. Sci. Environ. Epidemiol. 23,
phthalate in personal care products. J. Expo. Sci. Environ. Epidemiol. 25,
343e349.
http://dx.doi.org/10.1038/jes.2012.74.
317e323.
http://dx.doi.org/10.1038/jes.2014.68.
Han, X., Price, P.S., 2011. Determining the maximum cumulative ratios for mixtures
Dewalque, L., Charlier, C., Pirard, C., 2014. Estimated daily intake and cumulative
observed in ground water wells used as drinking water supplies in the United
risk assessment of phthalate diesters in a Belgian general population. Toxicol.
States. Int. J. Environ. Res. Public Health 8, 4729e4745.
http://dx.doi.org/
Lett. 231, 161e168.
http://dx.doi.org/10.1016/j.toxlet.2014.06.028.
10.3390/ijerph8124729.
EC, 2012. Communication from the Commission to the Council - the combination
Hoffman, K., Stapleton, H., 2015. Exposure to organophosphate flame retardants:
effects of chemicals. Chem. Mix. Com. 252 (final. Brussels).
hand wipes and house dust. Environ. Health Perspect. 123, 160e165.
http://
EC, 2011. Common Implementation Strategy for the Water Framework Directive (
dx.doi.org/10.1289/ehp.1408669.
2000/60/EC ) Guidance Document No.27-Technical Guidance for Deriving
Jeong, C.H., Wagner, E.D., Siebert, V.R., Anduri, S., Richardson, S.D., Daiber, E.J.,
Environmental Quality Standards.
http://dx.doi.org/10.2779/43816.
Mckague, A.B., Kogevinas, M., Villanueva, C.M., Goslan, E.H., Luo, W.,
EC, 2006. Regulation (EC) No 1907/2006 of the European Parliament and of the
Isabelle, L.M., Pankow, J.F., Grazuleviciene, R., Cordier, S., Edwards, S.C., Righi, E.,
Council Concerning the Registration, Evaluation, Authorisation and Restriction
Nieuwenhuijsen, M.J., Plewa, M.J., 2012. Occurrence and toxicity of disinfection
of Chemicals (REACH). OJ L396.
byproducts in european drinking waters in relation with the HIWATE epide-
EC, 2000. Directive 2000/60/EC of the European Parliament and of the Council of 23
miology study. Environ. Sci. Technol. 46, 12120e12128.
October 2000 Establishing a Framework for Community Action in the Field of
Junghans, M., Backhaus, T., Faust, M., Scholze, M., Grimme, L.H., 2006. Application
Water Policy.
http://dx.doi.org/10.1039/ap9842100196. OJ L327.
and validation of approaches for the predictive hazard assessment of realistic
ECHA, 2015. Guidance on the Biocidal Products Regulation - Volume III Human
pesticide mixtures. Aquat. Toxicol. 76, 93e110.
http://dx.doi.org/10.1016/
Health - Part B Risk Assessment. European Chemicals Agency, Helsinki, Finland.
j.aquatox.2005.10.001.
ECHA, 2014. Transitional Guidance on the Biocidal Products Regulation - Transi-
Kennedy, M.C., Glass, C.R., Bokkers, B., Hart, A.D.M., Hamey, P.Y., Kruisselbrink, J.W.,
tional Guidance on Mixture Toxicity Assess Emtn for Biocidal Products for the
de Boer, W.J., van der Voet, H., Garthwaite, D.G., van Klaveren, J.D., 2015.
Environment. European Chemicals Agency, Helsinki, Finland.
A European model and case studies for aggregate exposure assessment of
ECHA, 2013. Guidance for the Implementation of REACH - Guidance on Information
pesticides.
Food
Chem.
Toxicol.
79,
32e44.
http://dx.doi.org/10.1016/
Requirements and Chemical Safety Assessment - Chapter R. 20: Table of Terms
j.fct.2014.09.009.
and Abbreviations. European Chemicals Agency, Helsinki, Finland.
Kienzler, A., Berggren, E., Bessems, J., Bopp, S., Van Der Linden, S., Worth, A., 2014.
ECHA, 2012. Guidance on Information Requirements and Chemical Safety Assess-
Assessment of Mixtures - Review of Regulatory Requirements and Guidance.
ment Chapter R. 11: PBT Assessment. European Chemicals Agency, Helsinki,
JRC Science and Policy Report. European Commission, Joint Research Center,
Finland.
Ispra, Italy.
http://dx.doi.org/10.2788/84264.
EFSA, 2014a. Scientific Opinion on the identification of pesticides to be included in
Kortenkamp, A., Backhaus, T., Faust, M., 2009. State of the Art Report on Mixture
cumulative assessment groups on the basis of their toxicological profile (2014
Toxicity - Final Report, 391 pp.
update). EFSA J. Eur. Food Saf. Auth. Sci. Opin. Parma, Italy.
http://dx.doi.org/
Kortenkamp, A., Faust, M., 2010. Combined exposures to anti-androgenic chemicals:
10.2903/j.efsa.2013.3293.
steps towards cumulative risk assessment. Int. J. Androl. 33, 463e474.
http://
EFSA, 2014b. Scientific Opinion on good modelling practice in the context of
dx.doi.org/10.1111/j.1365-2605.2009.01047.x.
mechanistic effect models for risk assessment of plant protection products.
Laetz, C.A., Baldwin, D.H., Collier, T.K., Hebert, V., Stark, J.D., Scholz, N.L., 2009. The
EFSA
J.
Eur.
Food
Saf.
Auth.
Parma,
Italy.
http://dx.doi.org/10.2903/
synergistic toxicity of pesticide mixtures: implications for risk assessment and
334
A. Kienzler et al. / Regulatory Toxicology and Pharmacology 80 (2016) 321e334
the conservation of endangered Pacific Salmon. Environ. Health Perspect. 117,
1714e1720.
http://dx.doi.org/10.1289/ehp.1002180.
348e353.
http://dx.doi.org/10.1289/ehp.0800096.
Reif, D.M., Sypa, M., Lock, E.F., Wright, F.A., Wilson, A., Cathey, T., Judson, R.R.,
Law, R., Hanke, G., Angelidis, M., Batty, J., Bignert, A., Dachs, J., Davies, I., Denga, Y.,
Rusyn, I., 2013. ToxPi GUI: an interactive visualization tool for transparent
Duffek, A., Herut, B., Hylland, K., Lepom, P., Leonards, P., Mehtonen, J., Piha, H.,
integration of data from diverse sources of evidence. Bioinformatics 29,
Roose, P., Tronczynski, J., Velikova, V., Vethaak, D., 2010. JRC Scientific and
402e403.
http://dx.doi.org/10.1093/bioinformatics/bts686.
Technical Reports - Marine Strategy Framework Directive. Task Group 8 Report
Rico, A., Van den Brink, P.J., Gylstra, R., Focks, A., Brock, T.C.M., 2015. Developing
Contaminants and Pollution Effects. Joint Research Centre, Office for Official
ecological scenarios for the prospective aquatic risk assessment of pesticides.
Publications of the European Communities.
http://dx.doi.org/10.2788/85887.
Integr. Environ. Assess. Manag.
http://dx.doi.org/10.1002/ieam.1718 n/aen/a.
Le Cann, P., Bonvallot, N., Glorennec, P., Deguen, S., Goeury, C., Le Bot, B., 2011. In-
Salthammer, T., Fuhrmann, F., Uhde, E., 2003. Flame retardants in the indoor
door environment and children’s health: recent developments in chemical,
environment - Part II: release of VOCs (triethylphosphate and halogenated
biological, physical and social aspects. Int. J. Hyg. Environ. Health 215, 1e18.
degradation products) from polyurethane. Indoor Air 13, 49e52.
http://
http://dx.doi.org/10.1016/j.ijheh.2011.07.008.
dx.doi.org/10.1034/j.1600-0668.2003.01150.x.
Lim, Y.-W., Kim, H.-H., Lee, C.-S., Shin, D.-C., Chang, Y.-S., Yang, J.-Y., 2014. Exposure
Sammataro, D., Yoder, J.A., 2011. Honey Bee Colony Health: Challenges and Sus-
assessment and health risk of poly-brominated diphenyl ether (PBDE) flame
tainable Solutions. CRC Press, 320 pp.
retardants in the indoor environment of elementary school students in Korea.
Sarigiannis, D.A., Karakitsios, S.P., Gotti, A., Liakos, I.L., Katsoyiannis, A., 2011.
Sci.
Total
Environ.
470e471,
1376e1389.
http://dx.doi.org/10.1016/
Exposure to major volatile organic compounds and carbonyls in European in-
j.scitotenv.2013.09.013.
door environments and associated health risk. Environ. Int. 37, 743e765.
http://
Malaj, E., von der Ohe, P.C., Grote, M., Kühne, R., Mondy, C.P., Usseglio-Polatera, P.,
dx.doi.org/10.1016/j.envint.2011.01.005.
Brack, W., Sch€
afer, R.B., 2014. Organic chemicals jeopardize the health of
SCCS, 2015. SCCS Notes of Guidance for the Testing of Cosmetic Ingredients, 9th
freshwater ecosystems on the continental scale. Proc. Natl. Acad. Sci. U. S. A. 111,
revision. SCCS/1564/15. SCCS/1564/15.
9549e9554.
http://dx.doi.org/10.1073/pnas.1321082111.
SCHER, SCCS, SCENIHR, 2012. SCHER, SCCS, SCENIHR, Opinion on the Toxicity and
Meek, M.E.B., Boobis, A.R., Crofton, K.M., Heinemeyer, G., Raaij, M. Van, Vickers, C.,
Assessment of Chemical Mixtures. European Commission, Brussels.
2011. Risk assessment of combined exposure to multiple chemicals: a WHO/
Tang, J.Y.M., Busetti, F., Charrois, J.W.A., Escher, B.I., 2014. Which chemicals drive
IPCS framework. Regul. Toxicol. Pharmacol. 60, S1eS14.
http://dx.doi.org/
biological effects in wastewater and recycled water? Water Res. 60, 289e299.
10.1016/j.yrtph.2011.03.010.
http://dx.doi.org/10.1016/j.watres.2014.04.043.
Nieuwenhuijsen, M., Smith, R., Golfinopoulos, S., Best, N., Bennett, J., Aggazzotti, G.,
USEPA, 2007. Revised N -Methyl Carbamate Cumulative Risk Assessment (Wash-
Righi, E., Fantuzzi, G., Bucchini, L., Cordier, S., Villanueva, C., Moreno, V., La
ington DC).
Vecchia, C., Bosetti, C., Vartiainen, T., Rautiu, R., Toledano, M., Iszatt, N.,
USEPA, 2006a. Cumulative Risk from Chloroacetanilides Pesticides (Washington
Grazuleviciene, R., Kogevinas, M., 2009. Health impacts of long-term exposure
DC).
to disinfection by-products in drinking water in Europe: HIWATE. J. Water Heal
USEPA, 2006b. Triazine Cumulative Risk Assessment (Washington DC).
7, 185e207.
USEPA, 2006c. Considerations for Developing Alternative Health Risk Assessment
Nowell, L.H., Norman, J.E., Moran, P.W., Martin, J.D., Stone, W.W., 2014. Pesticide
Approaches for Addressing Multiple Chemicals, Exposures and Effects - EPA/
Toxicity Index-a tool for assessing potential toxicity of pesticide mixtures to
600/R-06/013A (Washington DC).
freshwater aquatic organisms. Sci. Total Environ. 476e477, 144e157.
http://
USEPA, 2003. Framework for Cumulative Risk Assessment - EPA/630/P-2/001F
dx.doi.org/10.1016/j.scitotenv.2013.12.088.
(Washington DC).
OECD, 2011. In: WHO OECD ILSI/HESI International Workshop on Risk Assessment
USEPA, 2002a. Guidance on Cumulative Risk Assessment of Pesticide Chemicals that
of Combined Exposures to Multiple Chemicals. Series on Testing and Assess-
Have a Common Mechanism of Toxicity (Washington DC).
ment No 140. Organisation for Economic Co-operation and Development. 77pp.
USEPA, 2002b. Organophosphate Pesticides: Revised Cumulative Risk Assessment
Ohe, T., Watanabe, T., Wakabayashi, K., 2004. Mutagens in surface waters: a review.
(Washington DC).
Mutat. Res. Rev. Mutat. Res. 567, 109e149.
USEPA, 2000. Supplementary Guidance for Conducting Health Risk Assessment of
Price, P., Han, X., Junghans, M., Kunz, P., Watts, C., Leverett, D., 2012. An application
Chemical Mixtures (Washington DC).
of a decision tree for assessing effects from exposures to multiple substances to
USEPA, 1999. Guidance for Identifying Pesticide Chemicals and Other Substances
the assessment of human and ecological effects from combined exposures to
that Have a Common Mechanism of Toxicity (Washington DC).
chemicals observed in surface waters and waste water effluents. Environ. Sci.
USEPA, 1989. Risk Assessment Guidance for Superfund Volume I Human Health
Eur. 24, 34.
http://dx.doi.org/10.1186/2190-4715-24-34.
Evaluation Manual-part A-EPA/540/1-89/002 (Washington, DC).
Price, P., Zaleski, R., Hollnagel, H., Ketelslegers, H., Han, X., 2014. Assessing the safety
USEPA, 1987. The Risk Assessment Guidelines of 1986-EPA/600/8-87/045 (Wash-
of co-exposure to food packaging migrants in food and water using the
ington, DC).
maximum cumulative ratio and an established decision tree. Food Addit.
WHO, 2014a. Combined or Multiple Exposure to Health Stressors in Indoor Built
Contam. Part A. Chem. Anal. control. Expo. Risk Assess. 31, 414e421.
http://
Environments - An Evidence-based Review Prepared for the WHO Training
dx.doi.org/10.1080/19440049.2013.865145.
Workshop “Multiple Environmental Exposures and risks.”. World Health Or-
Qian, H., Chen, M., Kransler, K.M., Zaleski, R.T., 2015. Assessment of chemical
ganization, 82 pp.
coexposure patterns based upon phthalate biomonitoring data within the
WHO, 2014b. Guidance Document on Evaluating and Expressing Uncertainty in
2007/2008 national health and nutrition examination survey. J. Expo. Sci. En-
Hazard Characterization. World Health Organization. IPCS Harmonization
viron. Epidemiol. 25, 249e255.
http://dx.doi.org/10.1038/jes.2014.24.
project document no 11. 181 pp.
Reif, D.M., Martin, M.T., Tan, S.W., Houck, K. a., Judson, R.S., Richard, A.M.,
WHO IPCS, 2009. Assessment of Combined Exposures to Multiple Chemicals:
Knudsen, T.B., Dix, D.J., Kavlock, R.J., 2010. Endocrine profling and prioritization
Report of a WHO/IPCS International Workshop on Aggregate/cumulative Risk.
of environmental chemicals using toxcast data. Environ. Health Perspect. 118,
World Health Organization. IPCS harmonization project document n 7. 83 pp.
Document Outline