RSC_CITE_Ch005.

qxd 11/15/2005 6:16 PM Page 84

Environmental Risk Assessment

LORRAINE MALTBY

1 What is Environmental Risk Assessment?

Risk assessment is the process of estimating the likelihood that a particular event
will occur under a given set of circumstances. Hence, environmental risk assessment
involves an analysis of information on the environmental fate and behaviour of
chemicals in the environment (i.e. air, water and land) integrated with an analysis of
information on their effects on human beings and ecological systems.1 Ecological
risk assessment is that component of environmental risk assessment which is con-
cerned with the effects of chemicals on non-human populations, communities and
ecosystems.2 Risk assessment is an important decision-making tool that can be used
to identify existing problems and to predict potential risks of planned actions. It is
useful both for prioritising management and regulatory efforts and for evaluating the
effectiveness of management actions that are implemented.
Central to any risk assessment process is the distinction between hazard and risk.
Whereas hazard is the ability of a chemical to harm organisms, risk is the probabil-
ity that harm will occur under a particular set of circumstances. A chemical may be
extremely hazardous, but if there is no environmental exposure, it will not present an
environmental risk. Risk assessment brings together information on exposure and
effects, and both are of equal importance. The fate and behaviour of chemicals in the
environment is the focus of Chapter 7 3 and is therefore not discussed in detail here.
Rather, the emphasis of this chapter will be on the effects of chemicals on ecosys-
tems and their components and the use of this information to assess risk. The risks
that chemicals in the environment pose to human health are discussed in Chapter 4.4

1.1 Prospective and Retrospective Risk Assessment

Prospective (or predictive) risk assessment predicts the likely consequences of releas-
ing a chemical substance into the environment. It is used to help the risk manager and

Issues in Environmental Science and Technology, No. 22

Chemicals in the Environment: Assessing and Managing Risk
© The Royal Society of Chemistry, 2006
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 85

Environmental Risk Assessment 85

regulator decide whether a chemical substance should be registered for use, and if so,
under what circumstances and with what control measures. Prospective risk assess-
ments are usually generic rather than site-specific and, consequently, require the iden-
tification of the potential effects of a chemical on a large number of organisms in a
variety of ecosystems. As it is not possible to study all potentially affected species and
habitats, generic prospective risk assessments are based on indicator organisms and
defined environmental compartments. For example, the standard exposure scenario
used in the aquatic risk assessment of plant protection products in Europe is a static
ditch containing a 30-cm depth of water overlying a 5-cm depth of sediment and the
standard test organisms are an alga, an invertebrate and a fish.5
Retrospective risk assessments assess the effects of chemicals once they have been
released into the environment and may be regional, local or site-specific. The focus
of retrospective risk assessments is usually the effect of a single chemical or chem-
ical source on selected ecosystem components. However, ecosystems are comprised
of a large number of interacting species and the structure and functioning of ecosys-
tems are the consequence of a variety of physical, chemical and biological factors
acting sequentially or concurrently. Teasing out the effects of a single chemical from
the multitude of natural and anthropogenic stressors operating on ecosystems is one
of the major challenges of retrospective risk assessment.

1.2 The Risk Assessment Process

Assessing the risks that chemicals pose to the environment is a three-phase process
(Figure 1). The process begins with a problem formulation phase in which the haz-
ard is identified and the study is planned. This is followed by an analysis phase in
which the exposure to, and effects of, the chemical are assessed. Finally, exposure
and effects information are brought together to characterise risk. Risk assessments
may be performed for different environmental compartments (i.e. aquatic, terrestrial,
atmospheric) and a tiered approach is usually adopted. Lower tier risk assessments
are based on limited data and reasonable worst-case assumptions, whereas higher-
tier assessments are based on more realistic, but complex data sets. Moving through
the tiers refines the assessment of exposure and effects and hence reduces uncer-
tainty in the risk characterisation.
The purpose of the problem formulation stage is to determine whether a particu-
lar danger exists, and if so, whether the associated effects warrant further study or
management action. If further studies are necessary, then the types of data required
to characterize the risk are identified and appropriate assessment and measurement
endpoints selected. Information used for hazard identification includes short-term or
screening toxicity tests and reviews of existing information that characterise the
potentially affected ecosystems and contaminants in question. In retrospective risk
assessments, this phase should include the development of a conceptual model that
identifies contaminant sources, biological receptors and the processes that link them.
It should also include the identification of assessment endpoints (what is to be pro-
tected?), the selection of measurement endpoints (what is to be measured to deter-
mine exposure and effects?) and the determination of level of effect to be detected
(how large a change is of concern?).
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 86

86 Lorraine Maltby

Figure 1 Environmental risk assessment process

The analysis phase includes measurement or prediction of the emission, transport,

fate and behaviour of the chemical in the environment. The purpose of exposure
assessment is to determine exposure concentrations and identify key environmental
components for assessing risks. In prospective risk assessments, predicted environ-
mental concentrations (PECs) are based on predicted emissions and use data and are
derived using mathematical modelling. Mathematical modelling may also play a key
role in retrospective risk assessment, but it is usually supported by chemical analy-
sis of environmental media or ecological receptors. In addition to exposure assess-
ment, the analysis phase characterises the relationship between exposure and effects
in order to derive concentrations at which no adverse ecological effects occur (e.g.
predicted no effect concentration, PNEC).
Effect assessments can be based on information obtained from single-species
toxicity tests, microcosm or mesocosm studies, field studies, field surveys or pop-
ulation and ecosystem modelling. Lower-tier effects assessments are based on
short-term effects data derived from acute toxicity tests with standard species (i.e.
algae, Daphnia, fish, chironomid larvae). Acute toxicity is usually expressed as
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 87

Environmental Risk Assessment 87

median lethal (effective) concentration (L(E)C50), which is the concentration
required to kill (effect) 50% of the individuals exposed to the chemical in a given
time period. Higher-tier effects assessments use data from acute and chronic toxic-
ity tests conducted on a range of species. Chronic toxicity tests usually measure the
sublethal effects of chemicals (e.g. effects on growth, reproduction, development)
and toxicity is expressed as a no observed effect concentrations (NOEC). The
NOEC is the test concentration immediately below the test concentration causing a
statistically significant effect on the test endpoint (i.e. lowest observable effect con-
centration, LOEC). Higher-tier assessments move beyond the study of the effect of
a chemical on individual organisms to consider population-level and community-
level effects, including indirect toxic effects due to changes in trophic interactions
and secondary poisoning.6, 7
At the end of the analysis phase, information from the exposure and effects assess-
ments are brought together to describe the nature and magnitude of risks posed by
the chemical. Risk may be characterised as a quotient (i.e. toxicity exposure ratio
(TER), PEC/PNEC ratio) or as a probability, and should include a consideration of
the uncertainties inherent in the risk assessment process. As the purpose of risk char-
acterisation is to provide information to risk managers and decision makers in a
manner that is understandable and relevant to the decisions being made, the level of
risk that is acceptable or which triggers further study, should be clearly defined in
the problem formulation phase.

1.3 Uncertainty and Variability

Environmental risk assessments are subject to both variability and uncertainty.
Variability is an inherent property of the system being investigated and includes
interspecific, inter-population and inter-individual differences in exposure and
response as well as temporal and spatial differences in the biotic and abiotic com-
ponents of ecosystems. Variability cannot be reduced by increased information or
measurement, but it can be characterised and its influence minimised by careful
selection of study systems. Uncertainty is due to imprecise or incomplete knowledge
and is a property of the relationship between the study system and the risk assessor.
Uncertainty in risk assessment may be categorised as variable uncertainty or model
uncertainty. Variable uncertainty arises from imprecise, inaccurate or inappropriate
measurements and can be reduced by improving measurement techniques and study
design. Model variability arises from an incomplete mechanistic understanding of
the system under investigation and is a feature of both extrapolations and mathe-
matical or conceptual models.
Uncertainty and variability inherent in the risk assessment process may mean that
the true effects are larger than the estimated effects. Consequently, uncertainty and
variability must be incorporated into the analysis of risk. Risk assessments are often
based on information obtained on a small number of species, exposed to a single
chemical under a given set of conditions for a limited period of time. This informa-
tion is then used to estimate the generic risk of both short-term and long-term expo-
sure of complex ecosystems exposed to multiple stressors; a process that incorporates
many uncertainties. Risk assessment procedures take one of two approaches to
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 88

88 Lorraine Maltby
addressing uncertainty. The first, commonly applied to deterministic risk assess-
ments, is to apply uncertainty (or assessment) factors to either model parameters or
outputs; the second is to analyse uncertainty using probabilistic approaches.

1.4 Deterministic and Probabilistic Risk Assessments

Deterministic risk assessments use fixed values to estimate toxicity (e.g. L(E)C50,
NOEC) and exposure (e.g. PEC) and generate a single measure of risk, such as a risk
quotient. In contrast, probabilistic risk assessments express results of exposure
and/or effect assessments as probability distributions and it is these distributions that
are used to generate a probability distribution of risk estimates.8 Deterministic risk
assessments are simple to perform and require relatively little data. They generally
assume worst-case conditions and characterise risk using the lowest available toxic-
ity estimate and the highest available exposure estimate. Consequently, determinis-
tic risk assessments tend to be very conservative and may over-estimate risk. In
contrast, probabilistic risk assessments require quantitative information on exposure
and effects and therefore can only be applied to data-rich chemicals. However, prob-
abilistic approaches have two major advantages over deterministic approaches: first,
they provide quantitative information on the probability of an adverse effect occur-
ring; second, they can include estimates of uncertainty and variability.
Major sources of uncertainty in risk assessment include those associated with
extrapolating across: response endpoints (e.g. lethal, sublethal), individuals and
species, study systems (e.g. laboratory test, mesocosm study, field site), spatial
scales, temporal scales and geographical locations. For most chemicals there is lim-
ited data with which to predict ecosystem effects and in most cases only acute tox-
icity data are available. Extrapolating from single-species short-term toxicity data
to ecosystem effects involves extrapolating from few species to many species, from
short-term exposure and lethal endpoints to long-term exposure and sublethal end-
points, and from laboratory exposures to field situations. Uncertainties associated
with these extrapolations are incorporated into deterministic risk assessments by
using assessment factors to either derive PNEC values from toxicity data or to set
trigger values for TERs. The size of the assessment factor used to adjust toxicity
data depends on the confidence with which the PNEC can be derived; the greater
the confidence, the lower the assessment factor. For instance, when only short-term
data are available, an assessment factor of 1000 is applied to the lowest L(E)C50,
but if long-term toxicity data are available for species from three trophic levels
(usually fish, Daphnia, algae), the assessment factor is reduced to 10. The assess-
ment factor maybe as low as 1 if data are available from field or model ecosystem
studies.6 Similarly, trigger values for TER are set at 100 when based on short-term
acute toxicity data, 10 when based on long-term toxicity data, and decided on a
case-by-case basis when the TER is based on data from microcosm or mesocosm
studies.5
Assessment factors used in deterministic risk assessments are empirically derived
values that have limited scientific validity and whose use neither identifies nor
quantifies sources of uncertainty. In contrast, probabilistic risk assessments enable
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 89

Environmental Risk Assessment 89

estimates of uncertainty to be incorporated into both exposure and effects assess-
ments. For example, species vary markedly in their sensitivity to environmental
contaminants and uncertainties arising from interspecific variation can be described
by constructing species sensitivity distributions (SSDs).9 The SSD is estimated from
single-species toxicity test data and visualised as a cumulative distribution function;
it is used to calculate the concentration at which a specified fraction of the species
pool will be affected by a chemical (i.e. hazardous concentration, HC). The most fre-
quently estimated hazardous concentrations are the HC5 (5% of species affected)
and HC10 (10% of species affected). Despite its widespread use in risk assessment,
the SSD approach has been criticized for assuming that test species used to derive
the SSD are random selections from the specified distribution and that they are rep-
resentative of the ecosystem to be protected.10 Recent studies with pesticides have
demonstrated that hazardous concentrations derived from SSDs are protective of
adverse ecological effects in freshwater systems and therefore do provide a useful
tool for assessing environmental risk.11, 12 Probabilistic methods can also be applied
to exposure estimates. They can incorporate spatial and temporal variation in the dis-
tribution and bioavailability of contaminants in the environment, as well as spatial
and temporal variation in factors that influence the uptake of contaminants by the
species of concern.13 Probabilistic environmental risk assessments have recently
been conducted for metals,14 industrial chemicals15 and pesticides.16 In all cases,
probabilistic risk assessments provided a higher degree of refinement than deter-
ministic risk assessments and enabled management strategies to be targeted at those
sites and species at greatest risk.

2 What are we Trying to Protect?

The purpose of environmental risk assessment is to inform the decision-making
process. Therefore the first and most fundamental question that must be addressed
when embarking on an environmental risk assessment is ‘What are the protection
goals?’. In other words, what are we trying to protect and over what temporal and
spatial scales? For human risk assessment, the protection goal is the individual
human being. However, the focus of an environmental risk assessment may range
from the survival of individual members of an endangered species to the productiv-
ity of a community or the biodiversity of a region. Similarly, whereas it is relatively
straightforward to identify the spatial and temporal scales of key determinants of the
health and well-being of human populations, this is less obvious for ecosystems. For
instance, whereas the appropriate spatial and temporal scales for microorganisms are
measured in nanometers and seconds, for migratory fish or birds, they are measured
in hundreds of kilometres and years.
Deciding what an environmental risk assessment is designed to protect requires
input from stakeholders as well as scientists, and includes consideration of ecological
values as well as ecological principles. Ecosystems provide society with many goods
and services.17 Essential ecosystem goods include the supply of oxygen, food, water,
medicines and raw materials, and essential ecosystem services include nutrient
cycling, water purification, waste removal, flood control, pollination and pest control.
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 90

90 Lorraine Maltby
Whereas it is relatively easy to value goods and services that are bought and sold (e.g.
minerals, food, drinking water), it is much more difficult to value non-market goods
and services such as biodiversity, nutrient cycling and climate regulation.18 Despite
these difficulties, the economic value of global ecosystem goods and services has
been estimated to be about $US180 trillion based on 2000 prices;19 a value which
exceeds the Gross World Product by a factor of 4.5. Understanding how species inter-
act to provide these goods and services and understanding how stressors, including
chemicals, affect these interactions, is fundamental to ensuring the sustainable use of
ecosystems.
Brock20 identified three approaches for perceiving risk and setting protection goals;
the pollution prevention principle, the carrying capacity principle and the functional
redundancy principle. The pollution prevention principle adopts a precautionary
approach by assuming that all environmental stressors are potentially harmful. The
justification for adopting a conservative approach to risk assessment is the uncertainty
with which the transport, fate, behaviour and effects of chemicals in the environment
can be predicted. Environmental contaminants can move between media, be trans-
ported vast distances and rarely, if ever, occur in isolation.21–23 Consequently, there is
uncertainty in predicting which ecosystems will be exposed to which contaminants,
at what concentrations and in what combinations. Furthermore, even if exposure
could be predicted accurately and the direct impact of multiple stressors was known,
there is uncertainty associated with predicting the long-term ecological consequences
of those impacts. This includes understanding the indirect (i.e. trophic) effects of
chemicals on species abundances and community structure24 and the consequences of
biodiversity loss for ecosystem functioning.25
The carrying capacity principle and the functional redundancy principle differ
from the pollution prevention principle in that they assume that ecosystems can tol-
erate a certain degree of chemical stress. Where they differ from each other, is in the
amount of stress and hence level of effect that is considered acceptable. The focus of
the carrying capacity principle is the ability of ecosystems to resist and recovery
from chemical perturbation (i.e. ecosystem resistance and resilience). Resistance is
a measure of the capacity of a system to resist change, and resilience is either defined
as the speed at which a system returns to equilibrium following a perturbation26 or
as the size of the disturbance from which a system cannot recover.27 As sustainabil-
ity of a system is dependent on its ability to withstand disturbances, it has been
argued that resilience (sensu Holling) is an appropriate measure of sustainability28
and hence is an ecologically relevant protection goal. The ability of an ecosystem to
recover from chemical perturbation is currently considered as part of the European
aquatic risk assessment of pesticides,5 but it is not considered in the risk assessment
of other chemicals.
The functional redundancy principle considers changes in community structure
acceptable as long as key ecosystem functions are sustained. The application of this
principle assumes that ecological roles can be performed by more than one species
and therefore loss of species from a community does not necessarily result in loss of
function. Whereas there is some evidence of functional redundancy within ecosys-
tems, resulting in ecosystem processes being less sensitive to stressors than commu-
nity structure,29 this is not always the case,30 and species that are ‘redundant’ under
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 91

Environmental Risk Assessment 91

one set of environmental conditions may not be redundant under another set of con-
ditions. Consequently, there is an argument for maintaining high biodiversity even in
communities with high functional redundancy, in order to maximise the chance that
the system will be able to resist future stressors. The maintenance of high diversity
to protect ecosystems against unpredictable and unknown stressors is known as the
insurance hypothesis,31 and has been likened to maintaining a diverse stock portfo-
lio, that is, spreading risk by investing in many and varied companies. Moreover,
there is increasing evidence to suggest that knowing which species are lost from a
community is at least as important as knowing how many species are lost, when try-
ing to predict the consequences of biodiversity loss on ecosystem function.32
The identification and adoption of different principles for setting protection goals,
means that it is possible to establish an approach to risk assessment in which not all
patches of habitat are treated equally. Areas with high ecological value (e.g. nature
reserves, conservation areas) can be afforded a high degree of protection by apply-
ing the pollution prevention principle, whereas less pristine habitats in areas of inten-
sive land use (e.g. agricultural land) could be adequately protected by adopting the
carrying capacity or functional redundancy principle. The development of a targeted
approach to protection goals setting and its potential application to environmental
risk assessment is discussed by Brock et al.33

2.1 Assessment Endpoints

Once the protection goals have been established, the next task is to determine the
most appropriate assessment endpoints. Ecosystems are comprised of many levels of
organisation (e.g. individuals, populations, communities), multiple species and
numerous processes and, except in the case of the protection of endangered species,
it is rarely clear which species, process and level of organisation are the most criti-
cal for a risk assessment. Three criteria have been proposed to aid the selection of
assessment endpoints: (1) ecological relevance; (2) susceptibility to known or poten-
tial stressors; and (3) relevance to protection goals.34 Ecological relevance can be
ascribed to endpoints at any level or organisation that help sustain the structure or
function of the ecosystem under consideration. Ecologically relevant endpoints may
relate to the provision of energy (e.g. primary production, decomposition) or habitat
(e.g. vegetation structure) or to the structure of the community, ecosystem or land-
scape (e.g. species richness, habitat distribution). Susceptibility to environmental
contaminants will depend on the mode of toxic action of the chemical, the exposure
pathway and the life stage of the exposed organisms. Juvenile stages are often more
sensitive than adults and vulnerability to stressors may be enhanced during physio-
logically demanding events such as migration, moulting and reproduction. Although
ecological relevance and susceptibility are essential criteria for a risk assessment to
be scientifically defensible, the ultimate aim of the risk assessment is to improve
management decisions and therefore the assessment endpoints should include eco-
logical components or attributes that people care about protecting. These include
species and habitats with high conservation value (e.g. charismatic species) or
important ecosystem goods and services (e.g. flood defence, waste disposal, food,
timber, clean water, commercially or recreationally important species).
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 92

92 Lorraine Maltby

2.2 Impact and Causality

Once assessment endpoints have been measured the next questions to be addressed
are:

1. is the system adversely affected (i.e. impacted)?

2. what is causing the impact?

These questions are relatively trivial for a single species exposed to a single chem-
ical in a tightly controlled laboratory study, but they are less trivial when applied to nat-
ural ecosystems where multispecies assemblages are exposed to a variety of natural
and anthropogenic stressors. Whereas the impact of a chemical on an ecosystem may
be defined in terms of a deviation from a reference condition, a predefined change in
an assessment endpoint, or a comparison with predicted values, the challenge with all
these approaches is establishing the range of normal variation in the assessment end-
point so that extreme values can be identified. For instance, ecological quality is
defined in the European Water Framework Directive in terms of deviation from ‘undis-
turbed conditions’.35 What is not clear, however, is what defines the undisturbed (ref-
erence) condition. Should it be defined in terms of present day conditions? If so, this
raises the problem that in many European countries, including the UK, there are few
if any sites that are not impacted by anthropogenic factors, therefore characterising a
present-day baseline will be based on a limited number of potentially atypical sites.
Should it be defined in terms of an historical state? If so, how far back in time is the
baseline and what information do we use to characterise it? The definition of undis-
turbed (reference) condition is a non-trivial problem, as is the definition of ecological
quality, yet both are fundamental to the successful implementation of the Water
Framework Directive, which will apply to aquatic ecosystems across Europe.
A common approach to assessing causality is to compare ‘contaminated’ with
‘uncontaminated’ sites and to ascribed statistically significant differences in assess-
ment endpoints to the contamination under investigation. The problem with this
approach is that treatments are not randomly assigned to sites; sites are not true repli-
cates and samples within sites are pseudoreplicated.36 As a consequence, it is incor-
rect to infer that measured differences between ‘contaminated’ and ‘uncontaminated’
sites are due to the contaminant of interest. Causality can only be inferred by con-
ducting experimental studies that demonstrate the relationship between cause and
effect. Suter2 proposed the following modification of Koch’s postulates as a method
to provide evidence of causality:

1. The injury, dysfunction or other putative effects of the toxicant must be regularly
associated with exposure to the toxicant and any contributory causal factor.
2. Indicators of exposure to the toxicant must be found in the affected organisms.
3. The toxic effects must be seen when organisms or communities are exposed to
the toxicant under controlled conditions, and any contributory factors should
be manifested in the same way during controlled exposures.
4. The same indicators of exposure and effects must be identified in the con-
trolled exposures as in the field.
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 93

Environmental Risk Assessment 93

Whereas the co-occurrence of the effect and contaminant in both the laboratory
and field (postulates 1 and 3) is strong evidence for causation, proof is provided by
demonstrating that the exposure conditions are associated with the same types and
levels of effect in the field and laboratory (postulates 2 and 4).

3 What is the EU Legislative Framework?

Current EU chemicals legislation distinguishes between the 100,106 chemicals that
were placed in the market before 1981 (i.e. ‘existing’ chemicals) and appoximately
2700 chemicals introduced since 1981 (i.e. ‘new’ chemicals). Whereas ‘new’ chemi-
cals produced at volumes of 10 kg/year must be assessed for possible risks to the
environment, there are no such provisions for ‘existing’ chemicals. An assessment of
environmental risk is required only for those ‘existing’ chemicals that are identified as
priority substances under Regulation 793/93 (i.e. about 140 chemical substances). In
an attempt to rectify this anomaly, in February 2001, the European Commission issued
a White Paper detailing its strategy for future chemicals policy.37 The paper detailed a
new system for ensuring a high level of chemical safety whilst maintaining a compet-
itive chemicals industry, known as REACH (Registration, Evaluation, Authorisation,
and Registration of Chemicals). REACH will apply to both ‘new’ and ‘existing’
chemical substances manufactured or imported in quantities of >1 tonne/year and will
harmonise the environmental hazard and risk assessment of chemicals.
One of the major challenges posed by a harmonised scheme for ‘new’ and ‘exist-
ing’ chemicals is the backlog of ‘existing’ chemicals to be risk-assessed. Between
1993 and 2001, only 141 high-volume chemicals were identified for risk assessment,
and only a limited number of these have completed the process.38 Under REACH, all
substances manufactured or imported into the European Union at quantities
>1 tonne/year must be registered. Substances manufactured or imported in quanti-
ties >10 tonnes will require a hazard classification and an assessment of whether the
substance is persistent, bioaccumulative and toxic (PBT) or very persistent and very
bioaccumulative (vPvB). Exposure scenarios and risk management measures, which
ensure that the risks from the uses of the substance are adequately controlled, must
support substances classified as PBT or vPvB. Because of the large number of ‘exist-
ing’ chemicals requiring risk assessment, substances will be phased into REACH
starting with the high production volume substances (>1000 tonnes/year) and those
substances that are carcinogenic, mutagenic or toxic for reproduction.
Both representatives of the chemical industry and independent groups have sug-
gested modifications to the REACH process. For instance, industry representatives
have suggested a tiered approached based on existing EU risk assessment guidelines
and tools (i.e. Technical Guidance Document, TGD6), which considers hazards and
potential exposures simultaneously and which rapidly screens out chemicals and
uses of no immediate concern.39 The Royal Commission on Environmental
Pollution have recommended that environmental monitoring is used to identify
chemicals that require further investigation. Moreover, they suggest that synthetic
chemicals that are found in biological material (e.g. breast milk) at elevated con-
centrations should be removed from the market, irrespective of whether they are
known to cause harm.40
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 94

94 Lorraine Maltby
REACH will not apply to all chemicals and does not cover plant protection prod-
ucts, biocides, human medicines and veterinary medicines, which are regulated by
different European directives. Assessment of the potential risks to the environment
arising from veterinary medicines is required under Directive 2001/82/EC 41 and
assessment of risks from human medicines is required under Directive
2001/83/EC.42 Plant protection products (agricultural pesticides) are regulated under
Directive 91/414/EEC,43 whereas biocides (non-agricultural pesticides) are regulated
under Directive 98/8/EC.44 Biocides include products used as disinfectants or preser-
vatives as well as for pest control (e.g. rodenticides, antifouling agents). Some form
of environmental risk assessment is required for all chemicals. However, whereas the
registration of most chemicals is dependent on the outcome of this risk assessment,
this is not the case for human medicines. The environmental impact of human med-
icines must be assessed and arrangements made to limit them where appropriate, but
environmental impact cannot be used to refuse marketing authorization in the
European Union Member States (Directive 2004/27/EC).38
Environmental risk assessments for ‘new’ chemical substances, priority ‘existing’
chemical substances and biocidal products, are performed in accordance with the
approaches detailed in the TGD. This includes guidance on how to determine PECs
and PNECs, how to conduct a PBT assessment and how to decide on a testing strategy
if further tests are needed. The purpose of the environmental risk assessment is to
protect ecosystems and methodologies are specified for freshwater ecosystems, ter-
restrial ecosystems, marine ecosystems, air, top predators (secondary poisoning) and
microorganisms in sewage treatment systems. The PEC can be derived using moni-
toring data or model calculations and the PNEC is usually determined using infor-
mation from single-species studies, but could be based on information from model
ecosystems. As the PNEC is regarded as the concentration below which an unac-
ceptable effect is unlikely to occur, a PEC/PNEC >1 indicates unacceptable risk.
Assessment of the potential risks to the environment arising from medicinal prod-
ucts for human use is a step-wise procedure that emphasises the active substance
and/or its metabolite(s) and which draws heavily on the TGD.45 The first phase is a pre-
screening stage (Phase I) in which environmental exposure is estimated. If the PEC for
surface water is <0.01 µg L−1, and no further environmental concerns are apparent, it
is assumed that the product is unlikely to represent a risk to the environment following
its prescribed usage by patients. If the PEC is greater than 0.01 µg L−1 in surface water,
then the product enters a screening stage (Phase II) in which standard aquatic toxic-
ity and fate data are used to produce an initial assessment of risk based on
PEC/PNEC ratios and PBT, vPvB criteria. In contrast to many other chemical sub-
stances, the initially risk assessment for human medicines is based on chronic toxi-
city data rather than acute toxicity data. This is because the specific mode of action
of pharmaceuticals results in acute to chronic ratios being highly variable and often
greater than 1000. Consequently, the assessment factor of 1000 specified in the TGD
to be applied to acute data would not necessarily protect aquatic organism from
chronic exposure to human medicines.46 If the PEC/PNEC ratio exceeds 1, a refined
risk assessment follows based on an extended dataset containing emissions, fate and
effects information. This in turn may be followed by a site-specific risk assessment
as described in the TGD.
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 95

Environmental Risk Assessment 95

A similar approached to that described for human medicines is employed to assess
the environmental risk posed by veterinary medicinal products. However, in addition
to environmental release via excretory products, the risk assessment for veterinary
medicines also considers direct release of the product into the environment during
its use (e.g. fish medicines, sheep dips, fumigation). The Phase I trigger values for
veterinary medicines are an ‘environmental introduction concentration’ (e.g. con-
centration applied or in effluent) in water of 1 µg L−1 and a predicted environmental
concentration in soil of 100 µg kg−1; the exception being parasiticides administered
to aquatic or pasture animals which, if enter the environment, automatically proceed
to Phase II.47 The Phase I environmental concentration is predicted using informa-
tion on use (non-food animals and treatment of small flocks/herds are exempt),
metabolism by treated animal (metabolised products are exempt), dosage and treat-
ment regime, product type and manure production and disposal. If the substance
proceeds to Phase II, it undergoes a two-tiered risk assessment, the precise form of
which depends on whether the substance is used in aquaculture, administered to
intensively reared terrestrial animals or administered to pasture animals.48 For all
uses, a standard set of physico-chemical and fate information is used to derive pre-
dicted environmental concentrations. In Tier A, acute toxicity data for algae,
Crustacea and fish are used to derive PNEC data for aquatic systems and studies on
earthworms, terrestrial plants and nitrogen transformation are used to derived PNEC
values for terrestrial systems. In addition, if the substance is a parasiticide adminis-
tered to pasture animals, a toxicity test using dung fauna should be conducted. If Tier
A assessment results in a PEC/PNEC>1, then the substance moves on to Tier B,
where information from chronic toxicity tests and fish bioconcentration studies are
used to refine the risk assessment.48 If the Tier B PEC/PNEC is still >1, then the risk
assessment could possibly be refined further, by for example, testing metabolites or
conducting microcosm or mesocosm studies.49
The most extensive risk assessment procedure is performed for plant protection
products. A technical dossier on the active substance, formulation, and if appropri-
ate metabolites, must be submitted before a plant protection product can be autho-
rised for use in European Union Member States. The dossier contains information
used to evaluate the risk that the product may pose to human health and the envi-
ronment and includes, as a minimum, the following information: the identity, func-
tion, intended use and physical and chemical characteristics of the product; analysis
of residues in soil, water, air, body fluids and tissues, food and feed; mammalian
toxicity (acute toxicity, short-term toxicity, genotoxicity, reproductive toxicity,
long-term toxicity and carcinogenicity) and metabolism; environmental fate and
behaviour in soil, air and water; ecotoxicity to aquatic organisms and ecotoxicity to
birds, non-target arthropods, microorganisms where appropriate. Pesticide risk
assessment adopts a tiered approach with lower-tiers focusing on realistic worst-
case exposure scenarios and acute toxicity and higher-tiers using information for
more environmentally and ecologically realistic studies including mesocosm stud-
ies and field trials.7 Environmental risk assessments may be based on deterministic
or probabilistic approaches and studies used to support pesticide registration are
performed in accordance with guidance documents produced by the European
Commission.5, 50, 51
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 96

96 Lorraine Maltby

4 What are Some of the Challenges Associated with the

Environmental Risk Assessment of Chemicals?
4.1 ‘Data-Poor’ Chemicals
Approaches for assessing the environmental risks posed by plant protection products
and ‘new’ chemicals are well established and risk characterisation is informed by
robust datasets. However, these chemicals constitute a small fraction of the chemi-
cals released into the environment and for most chemicals there is limited or little
information on which to base the risk assessment. This ‘data-poor’ category includes
most ‘existing’ chemicals, human and veterinary medicines, biocides and nanomate-
rials. A system has been proposed for addressing the lack of information for ‘exist-
ing’ chemicals (i.e. REACH), but there are challenges associated with implementing
this system due to the large number of chemicals to be processed and the lack of
appropriate techniques for detecting, quantifying and assessing some of the chemi-
cals concerned. For many ‘data-poor’ chemicals not included in REACH, the chal-
lenges associated with assessing their environmental risk are even greater; this is
particularly the case for nanomaterials and pharmaceuticals.
Nanomaterials have at least one dimension less than 100 nm and have been pres-
ent in the environment for a long time. The adverse health effects of inhaling
nanoparticles (e.g. ultrafine particles in polluted air) have been documented,52 and
there is increasing concern over the potential environmental risks posed by advances
in nanotechnology. For example, there is concern that inhaled engineered nanoma-
terials will cause inflammatory reactions in the lungs similar to those resulting from
the inhalation of coal dust and asbestos.53 Engineered nanoparticles and nanotubes
that are neither embedded nor fixed in a bulk material are free to be transported
through environmental compartments and can enter organisms via ingestion, inhala-
tion or movement through external surfaces (e.g. skin).54 Moreover, once inside an
organism, nanoparticles and nanotubes can move throughout the body and pass
through biological membranes, including the blood–brain barrier.
A recent review of the potential health and environmental risks posed by nan-
otechnology55 has recommended that the release of nanoparticles and nanotubes to
the environment should be minimised until their risk can be properly evaluated.
Moreover, it recommends that free nanoparticles are treated as new chemicals, even
if they are comprised of materials that have already undergone a risk assessment.
The rationale for treating nanomaterials as new chemicals is that changing the phys-
ical properties of materials also changes their environmental behaviour and toxico-
logical properties.56 The fact that nanoparticles have a greater potential to enter
organisms than larger particles coupled with their toxicological potential57 has raised
the need for detailed studies of the environmental risks posed by these materials and
has even led to the suggestion that a new subcategory of toxicology, namely nan-
otoxicology, should be defined to specifically address these knowledge gaps.54
Veterinary medicines can be released to the environment during manufacture,
application or disposal and may be excreted by treated animals.58 Evidence that vet-
erinary medicines pose an environmental risk include the observation that many par-
asiticides are excreted in faeces in concentrations sufficient to disrupt the biology of
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 97

Environmental Risk Assessment 97

59
dung fauna, and the suggestion that population declines in vultures are caused by
scavenging on the carcasses of treated animals.60 The potential for the widespread
use of veterinary medicines such as the parasiticide, ivermectin, to result in biodi-
versity loss and hence threaten the sustainability of pastoral ecosystems has been
known for many years.61 However, for a large proportion of veterinary medicines in
current use, there is little or no information on which to base an assessment of the
environmental risk of long-term, low-level exposure.59 This information gap must be
filled if the sustainability of agricultural systems is to be ensured.
Human and veterinary medicines contain bioactive substances that may elicit
effects that are not detected in standard ecotoxicity studies; either because of inap-
propriate exposure durations, exposure concentration or response endpoints. The
risk assessment of both human and veterinary medicines follows a two-phase
process in which the first phase is an assessment of environmental exposure only. If
veterinary medicines are introduced to the aquatic environment at concentrations
<1 µg L−1 and human pharmaceuticals have a predicted environmental exposure to
aquatic organisms of < 0.01 µg L−1 (equivalent to an introductory concentration of
0.1 µg L−1, assuming a 10-fold dilution), the chemical is assumed not to pose an
environmental risk. Only if the environmental concentrations exceed these trigger
values, does the assessment proceed to Phase II, in which both exposure and effects
data are considered. These trigger values have been criticised for not being scientif-
ically validated 62 and for not considering low-concentration effects (e.g. endocrine
distruption) and mixture effects.63 Furthermore, Phase II assessments have been crit-
icized for their reliance on standard toxicity tests that do not take account of the fact
that many pharmaceuticals have highly specialized modes of action.64 The challenge,
therefore, is to refine the risk assessment procedure to either remove or validate
Phase I trigger values and to modify effects assessments to accommodate chemicals
with very specific modes of action. Issues surrounding chemical mixtures and sub-
stances with non-classical modes of action are discussed below.

4.2 Non-Classical Modes of Action

Current environmental effects assessment procedures are designed to detect effects
that occur during exposure and that can be measured in terms of changes in the sur-
vival, growth, reproduction or food consumption. They cannot be readily applied to
substances with reversible, latent or trans-generational effects caused by exposure
during a limited time window of sensitivity, nor to substances that result in effects
not detected using standard ecoxicological endpoints. Some chemicals may elicit
effects that occur after exposure, either because the chemical is slow-acting, or
because changes in the organism’s physiology makes the chemical more toxic, or
because the chemical disrupts a developmental process that only becomes apparent
later in the life cycle of the organism. Latent effects have been reported for a num-
ber of carcinogens, endocrine disruptors and some pesticides.65
The inability of standard risk assessments to identify the environmental impacts of
chemicals with non-classical modes of action has been highlighted by the discovery
that chemicals may disrupt the endocrine system of organisms at very low exposure
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 98

98 Lorraine Maltby
concentrations and that the consequences of endocrine disruption may not be mani-
fest until a considerable time after exposure. Although chemicals may disrupt the
endocrine system of biota by a wide variety of mechanisms, most research has
focused on the effect of oestrogenic compounds, which include natural and synthetic
(i.e. xeno-estrogens) substances that bind to the oestrogen receptor and elicit a femi-
nising effect in organisms. A recent study of the contamination of Dutch waters by
oestrogenic chemicals concluded that natural steroid hormones (e.g. estrone and 17ß-
estradiol) were present in all untreated wastewaters and xeno-estrogens (e.g. bisphe-
nol-A, alkylphenolethoxylates, phthalates) were present in almost all untreated
wastewaters. Although the concentrations of most of these substances were lower in
treated wastewater, estrone, bisphenol-A and several phthalates were each detected in
around 50% of surface water samples.66 Whereas discharges of municipal and indus-
trial wastewater are the major routes of emissions of estrogenic compounds into the
aquatic environment, in areas of intensive livestock farming, animal manure may also
be an important source of environmental estrogens.67
Endocrine disruption is undoubtedly an issue that requires further investigation
and is a mechanism of toxicity that must be picked up in the risk assessment proce-
dure. However, there are potentially other toxic mechanisms that are not being
detected and which may have important ecological consequences. The challenge is
to improve our understanding of the toxic mode of action of chemicals and to use
this understanding to develop more appropriate risk assessments.

4.3 Mixtures
Current regulatory risk assessments focus on individual substances acting on other-
wise unstressed ecosystems. In reality, chemical contaminants rarely if ever, occur
in isolation, and ecosystems are subject to a range of natural stressors. Two main
concepts are used to predict the toxicity of mixtures: concentration addition and
independent action. Concentration addition applies to mixtures of chemicals with the
same mode of action and assumes that the effect of the mixture is the sum of the rel-
ative toxicities (expressed as toxic units) of the individual components. Moreover,
concentration addition assumes that any chemical in the mixture can be substituted
for another chemical with the same mode of action with the same relative toxicity,
and that the overall mixture toxicity will remain the same. Independent action
assumes that chemicals in mixtures have different modes of action and hence act
independently. An important consequence of the difference between these two con-
cepts is that, whereas with independent action chemicals in mixtures will only have
an effect at concentrations above the NOEC, with concentration addition, chemicals
can contribute to the total mixture effect even if they are present at concentrations
below their individual NOEC.
Recent studies have demonstrated that the mixture toxicity of anti-inflammatory
drugs and the mixture toxicity of β -blockers can be accurately predicted using the
concept of concentration addition.68, 69 One implication of this finding is that mixtures
of pharmaceuticals may have environmental effects even when the individual chemi-
cals are present at concentrations below the Phase I trigger level. Other studies have
also demonstrated that mixtures containing chemicals at concentrations considerably
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 99

Environmental Risk Assessment 99

70
below NOEC values (e.g. EC1) exhibit considerable toxicity. These research find-
ings highlight the need to improve our understanding of mixture toxicity, especially
when dealing with chemicals present at low-effect concentrations, and to modify
environmental risk assessment procedures to take account of possible combined
effects resulting from multiple chemical exposures.

References
1. C.J. Van Leeuwen and J. Hermans, Risk Assessment of Chemicals: An
Introduction, Kluwer Academic Publishers, Dortrecht, The Netherlands, 1995.
2. G.W. Suter II, Ecological Risk Assessment, Lewis Publishers, Boca Raton. FL,
1993.
3. D. Mackay, E. Webster and T. Gouin, Issues Environ. Sci. Technol., 22, chapter 7,
this volume.
4. P. Harrison and P. Holmes, Issues Environ. Sci. Technol., 22, chapter 4, this volume.
5. EC, Guidance Document on Aquatic Toxicology in the Context of Directive
91/414/EEC. SANCO/3268/2001, Brussels, Belgium, 2002.
6. EC, Technical Guidance Document on Risk Assessment, European Commission,
Brussels, 2003.
7. P.J. Campbell, D.J.S. Arnold, T.C.M. Brock, N.J. Grandy, W. Heger, F. Heimbach,
S.J. Maund and M. Streloke, Guidance Document on Higher-Tier Aquatic Risk
Assessment for Pesticides (HARAP), SETAC-Europe, Brussels, 1999.
8. K. Solomon, J. Giesy and P. Jones, Crop Prot., 2000, 19, 649.
9. L. Posthuma, G.W. Suter II and T.P. Traas, Species Sensitivity Distributions in
Ecotoxicology, Lewis Publishers, Boca Raton, FL, 2002.
10. V. Forbes and P. Calow, Hum. Ecol. Risk Assess., 2002, 8, 473.
11. L. Maltby, N. Blake, T.C.M. Brock and P.J. Van den Brink, Environ. Toxicol.
Chem., 2005, 24, 379.
12. P.J. Van den Brink, N. Blake, T.C.M. Brock and L. Maltby, Hum. Ecol. Risk
Assess., in press.
13. A. Hart, G.C. Smith, R. Macarthur and M. Rose, Toxicol. Lett., 2003, 140–141, 437.
14. P.A. Van Sprang, F.A.M. Verdonck, P.A. Vanrolleghem, M.L. Vangheluwe and
C.R. Janssen, Environ. Toxicol. Chem., 2004, 23, 2993.
15. M. Zolezzi, C. Cattaneo and J.V. Tarazona, Environ. Sci. Technol., 2005, 39, 2920.
16. J.J. Johnston, W.C. Pitt, R.T. Sugihara, J.D. Eisemann, T.M. Primus, M.J.
Holmes, J. Crocker and A. Hart, Environ. Toxicol. Chem., 2005, 24, 1557.
17. R. Costanza, R. d’Arge, R. de Grrot, S. Farber, M. Grasso, B. Hannon,
K. Limburg, S. Maeem, R.V. O’Neill, J. Paruelo, R.G. Raskin, P. Sutton and
M. Van den Belt, Nature, 1997, 387, 253.
18. M.A. Wilson and S.R. Carpenter, Ecol. Appl., 1999, 9, 772.
19. R. Boumans, R. Costanza, J. Farley, M.A. Wilson, R. Portela, J. Rotmans,
F. Villa and M. Grasso, Ecol. Econ., 2002, 41, 529.
20. T.C.M. Brock, in Workshop on Risk Assessment and Risk Mitigation Measures
in the Context of the Authorization of Plant Protection Products (WORMM),
Vol. 383, R. Forster and M. Streloke (Eds.), Mitteilungen aus der Biologischen
Bundesanstalt für Land- und Forstwirtschaft, Berlin-Dahlem, 2001, 68–72.
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 100

100 Lorraine Maltby

21. F. Wania, Environ. Sci. Technol., 2003, 37, 1344.
22. D.C.G. Muir, C. Teixeira and F. Wania, Environ. Toxicol. Chem., 2004, 23, 2421.
23. K. Vorkamp, F. Riget, M. Glasius, M. Pécseli, M. Lebeuf and M. D, Sci. Total
Environ., 2004, 331, 157.
24. J.W. Fleeger, K.R. Carman and R.M. Nisbet, Sci. Total Environ., 2003, 317, 207.
25. M. Loreau, S. Naeem and P. Inchausti, Biodiversity and Ecosystem Functioning,
Oxford University Press, Oxford., 2002.
26. S.L. Pimm, Nature, 1984, 307, 321.
27. C.S. Holling, Ann. Rev. Ecol. Syst., 1973, 4, 1.
28. M.S. Common and C. Perrings, Ecol. Econ., 1992, 6, 7.
29. D.M.E. Slijkerman, D.J. Baird, A. Conrad, R.G. Jak and N.M. van Straalen,
Environ. Toxicol. Chem., 2004, 23, 455.
30. D.M. Carlisle and W.H. Clements, Freshwat. Biol., 2005, 50, 380.
31. S. Yachi and M. Loreau, PNAS, 1999, 96, 1463.
32. D.U. Hooper, F.S. Chapin III, J.J. Ewel, A. Hector, P. Inchausti, J. Lawton, D.M.
Lodge, M. Loreau, S. Naeem, B. Schmid, H. Setälä, A.J. Symstad,
J. Vandermeer and D.A. Wardle, Ecol. Monogr., 2005, 75, 3.
33. T.C.M. Brock, G.H.P. Artes, L. Maltby and P.J. van den Brink, Aquatic
Hazards of Pesticides, Ecological Protection Goals and Common Claims in
EU Regulation: An Attempt to Harmonise Scientific Methods to Derive
Permissible Concentrations, Alterra Report, Wageningen, The Netherlands, in
press.
34. EPA, Guidelines for Ecological Risk Assessment. EPA/630/R-95/002F, US
Environmental Protection Agency, Washington, DC, 1998.
35. EC, Official J. Euro. Commun., 2000, L 327, 1.
36. S.H. Hurlbert, Ecol. Monogr., 1984, 54, 187.
37. CEC, Strategy for a Future Chemicals Policy COM(2001) 88 final, Commission
of the European Communities, Brussels, 2001.
38. EC, REACH in Brief, European Commission, Brussels, Belgium, 2004.
39. T. Feijtel, M. Comber, W. De Wolf, M. Holt, V. Koch, A. Lecloux and A. Siebel-
Sauer, Environ. Toxicol. Chem., 2005, 24, 251.
40. RCEP, Chemicals in Products–Safeguarding the Environment and Human
Health, HMSO, London, 2003.
41. EC, Official J. Euro. Commun., 2001, L311, 1.
42. EC, Official J. Euro. Commun., 2001, L331, 67.
43. EC, Official J. Euro. Commun., 1991, L230, 1.
44. EC, Official J. Euro. Commun., 1998, L123, 1.
45. EMEA, Guideline on the Environmental Risk Assessment of Medicinal Products
for Human Use (CHMP/SWP/4447/00 draft), European Medicines Agency,
London, 2005.
46. B. Ferrari, R. Mons, B. Volleat, B. Fraysse, N. Paxéus, R. Lo Giudice, A. Pollio
and J. Garric, Environ. Toxicol. Chem., 2004, 23, 1344.
47. VICH, Environmental Impact Assessment (EIAs) for Veterinary Medicinal
Products (VMPs) – Phase I. VICH GL6 (Ecotoxicity Phase I), International
Cooperation on Harmonisation of Technical Requirements for Regulation of
Veterinary Medicinal Products, Brussels, Belgium, 2000.
RSC_CITE_Ch005.qxd 11/15/2005 6:16 PM Page 101

Environmental Risk Assessment 101

48. VICH, Environmental Impact Assessment for Veterinary Medicinal Products
Phase II Guidance. VICH GL38 (Ecotoxicity Phase II), International Cooperation
on Harmonisation of Technical Requirements for Regulation of Veterinary
Medicinal Products, Brussels, Belgium, 2004.
49. P.J. Van den Brink, J.V. Tarazona, K.R. Solomon, T. Knacker, N.W. Van den
Brink, T.C.M. Brock and J.P. Hoogland, Environ. Toxicol. Chem., 2005, 24, 820.
50. EC, Guidance Document for Risk Assessment for Birds and Mammals Under
Council Directive 91/414/EEC. SANCO/4145/2000, Brussels, Belgium, 2002.
51. EC, Guidance Document on Terrestrial Ecotoxicology Under Council Directive
91/414/EEC. SANCO/10329/2002, Brussels, Belgium, 2002.
52. G. Oberdorster, Int. Arch. Occup. Environ. Health, 2001, 74, 1.
53. R. Owen and M. Depledge, Mar. Pollut. Bull., 2005, 50, 609.
54. K. Donaldson, V. Stone, C.L. Tran, W. Kreyling and P.J.A. Borm, Occup.
Environ. Med., 2004, 61, 727.
55. Royal Society and Royal Academy of Engineering, Nanoscience and
Nanotechnologies: Opportunities and Uncertainties, [Link]/final
[Link], 2004.
56. K. Donaldson, V. Stone, P.S. Gilmour, D.M. Brown and W. MacNee, Phil.
Trans. R. Soc. Lond. Ser, A, 2000, 358, 2741.
57. V.L. Colvin, Nat. Biotechnol., 2003, 21, 1166.
58. A.B.A. Boxall, D.W. Kolpin, B. Halling-Sørensen and J. Tolls, Environ. Sci.
Technol., 2003, 37, 286A.
59. K.G. Wardhaugh, Environ. Toxicol. Chem., 2005, 24, 789.
60. J.L. Oaks, M. Gilbert, M.Z. Virani, R.T. Watson, C.U. Meteyer, B.A. Rideout,
H.L. Shivaprasad, S. Ahmed, M.J.I. Chaudhry, M. Arshad, S. Mahmood, A. Ali
and A.A. Khan, Nature, 2004, 427, 630.
61. R. Wall and L. Strong, Nature, 1987, 327, 418.
62. CSTEE, Opinion on Draft CPMO Discussion Paper on Environmental Risk
Assessment of Medicinal Products for Human Use [Non-Genetically Modified
Organism (Non-GMO) Containing], Scientific Committee on Toxicology,
Ecotoxicology and the Environment, Brussels, Begium, 2001.
63. J.P. Bound and N. Voulvoulis, Chemosphere, 2004, 56, 1143.
64. K.-P. Henschel, A. Wenzel, M. Diedrich and A. Fliedner, Reg. Toxicol. Pharm.,
1997, 25, 220.
65. M.K. Hurd, S.A. Perry and W.B. Perry, Environ. Toxicol. Chem., 1996, 15, 1344.
66. A.D. Vethaak, J. Lahr, S.M. Schrap, A.C. Belfroid, G.B.J. Rijs, A. Gerritsen,
J. de Boer, A.S. Bulder, G.C.M. Grinwis, R.V. Kuiper, J. Legler, T.A.J. Murk,
W. Peijnenberg, H.J.M. Verhaar and P. de Voogt, Chemosphere, 2005, 59, 511.
67. T.A. Hanselman, D.A. Graetz and C. Wilkie, Environ. Sci. Technol., 2003, 37,
5471.
68. M. Cleuvers, Ecotoxicol. Eviron. Safety, 2004, 59, 309.
69. M. Cleuvers, Chemosphere, 2005, 59, 199.
70. T. Backhaus, R. Altenburger, A. Arrhenius, H. Blank, M. Faust, A. Finizio,
P. Gramatica, M. Grote, M. Junghans, W. Meyer, M. Pavan, T. Porsbring,
M. Scholze, R. Todeschini, M. Vighi, H. Walter and L. Horst Grimme,
Continental Shelf Res., 2003, 23, 1757.