This Annex presents the methodology utilised
in the risk assessments performed on data gathered as part of the
bioaccumulation assessment. Included
in this Annex are the detailed
results of the Human Health Risk Assessment and the Ecological Risk Assessment.
Risk
assessment can be divided into four major steps:
·
hazard
identification;
·
dose-response
evaluation;
·
exposure
assessment;
·
risk
characterisation.
Each
is discussed in the following sections.
Hazard
identification is the process of determining whether exposure to a chemical
could cause an increase in adverse health effects. It involves characterising the nature and quantity of
possible contaminant releases to the environment, selecting a set of
Contaminants of Concern (COC), gathering and evaluating data on the types of
health injury or disease that may be produced by a contaminant, and gathering
and evaluating data on the conditions of exposure under which injury or disease
is produced.
This
section presents a framework for the evaluation of the potential human health
and ecological effects resulting from ingestion of contaminants contained
within the edible portion of organisms.
The estimation of contaminant levels within the edible portion of
organisms has been conducted as part of the bioaccumulation assessment, which
is detailed separately in Annex B.
Some
of the COCs are known carcinogens, whereas, others are not considered to be
carcinogenic but cause other toxic effects. There are also COCs that cause both toxic responses and are
known to be carcinogenic.
Assessment criteria have been developed for each type of toxicological
effect and are discussed in later sections.
The
contaminants of concern adopted for use in this study are those included in
ETWBTCW 34/2002. Information on the
toxic effects of each of the COCs can be found at the following sources.
·
EVS
(1996b) Classification and Testing of Sediments for Marine Disposal. Prepared for CED.
·
EVS
(1996c) Contaminated Mud Disposal at East of Sha Chau: Comparative Integrated
Risk Assessment. Prepared for CED.
·
Aspinwall
Clouston Ltd (1998) A Study of Tributyltin Contamination of the Marine
Environment of Hong Kong. Prepared
for EPD.
·
Irwin
RJ, M VanMouwerik, L Stevens, MD Seese & W Basham (1998) Environmental
Contaminants Encyclopaedia. National Park Service, Water Resources Division,
Water Operations Branch, Colorado.
·
Integrated
Risk Information System (IRIS), US EPA.
·
ERM
(2002) Environmental Monitoring and Audit
for Contaminated Mud Pit IV at East of Sha Chau. Final Report submitted to the Civil
Engineering Department.
Dose-response
evaluation involves quantifying the relationship between the degree of exposure
to a substance and the extent of toxic injury or disease. The majority of data are derived from
animal studies in the laboratory or, less frequently, from studies in exposed
human populations. There may be
many different dose-response relationships for a substance if it produces
different toxic effects under different conditions of exposure. The risks of a substance cannot be
ascertained with any degree of confidence unless dose-response relationships
are quantified, even if the substance is known to be "toxic". Such dose-response relationships have
been established for various COCs for exposures to humans but with varying
degrees of certainty. Exposures to
species such as Sousa chinensis are
less accurately quantified and few published dose-response relationships are
available for marine mammals.
For the
purpose of the assessment, the effects of the substances listed in Section 2.2.2 have been classified into
two categories, ie non-carcinogenic effects or carcinogenic effects to
humans. Substances are included
within both categories if they exhibit both types of effect.
One
of the fundamental principles of toxicology is the dose-response relationship.
For virtually all toxic substances, there is a direct relationship
between the exposure level (and duration) and the severity of the effects
produced. As the exposure level
(and/or duration period) is lowered, for the great majority of toxic effects, a
point is reached at which no detectable effect occurs. This is termed the threshold dose or No Adverse Effects Level (NOAEL).
In
laboratory experiments non-carcinogens display NOAELs as the animals under
testing can tolerate doses below a certain finite value, with only a limited
chance of the expression of toxic effects. NOAELs themselves are not directly used for human health
criteria as the NOAELs relate to toxicity observed in animal bioassays and may
not adequately protect the most sensitive receivers in human populations (eg
embryos). In order to develop
criteria for human health Uncertainty
Factors (UFs) (USEPA 1989) are applied to the NOAEL data in order to insure
that risks are over-estimated rather than underestimated. For example, extrapolation of animal
toxicity response doses to humans utilises two safety factors of ten, the first
for animal-to-human extrapolation and the second for variation of sensitivities
within the human population.
The human health criteria developed after
application of the UFs are referred to as Reference
Doses (RfDs). The RfD, promulgated
by the US EPA, is an estimate of the daily exposure which appears to present a
low risk of adverse effects during an exposure to the most sensitive members of
the receiving population. The
purpose of the RfD is to provide a benchmark against which other doses might be
compared. Doses which are less
than the RfD are not likely to be of concern. Doses which are significantly greater (ie at least one order
of magnitude) than the RfD may indicate that inadequate margins of safety could
exist for exposure to that chemical.
The RfD is an approximate number, and while doses higher than the RfD
have a higher probability of producing an adverse effect, it should not be
inferred that such doses are, by definition, unacceptable or of concern. For the ingestion route, the RfD is
expressed in units of mg kg (body weight)-1 day-1, ie mg
kg-1 day -1.
A summary of RfDs for the COCs is presented in Table 2.3a. Table 2.3a also indicates the
carcinogenic class of each COC according to the US EPA classification system
which comprises the following categories:
·
Class
A human carcinogen
·
Class
B probable human carcinogen:
B1 indicates limited human
evidence;
B2 indicates
sufficient evidence in animals and inadequate or no evidence in humans
·
Class
C possible human carcinogen
·
Class
D evidence of non-carcinogenicity for
humans
Table 2.3a Toxicity
Information Taken from Integrated Risk Information System (IRIS)
|
Substance |
Oral RfD Mg kg-1 day-1 |
Oral Slope Factor mg kg-1 day-1 |
US EPA Carcinogenic Class |
||
|
Arsenic(a) |
0.0003 |
1.5 |
Class
A, human carcinogen |
||
|
Cadmium(b) |
0.001 |
|
Class
B1, probable human carcinogen |
||
|
Chromium(c) Chromium(d) |
0.003 1.5 |
|
Class
D, not classifiable as to human carcinogenicity for oral exposure of Cr (VI),
Class D also for Cr (III) |
||
|
Copper(e) |
0.043 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Lead |
0.00143 |
0.0085 |
Class
B2, probable human carcinogen for lead and compounds (inorganic) |
||
|
Mercury(f) |
0.00022 |
|
Class
C for methyl mercury and mercuric chloride, Class D for elemental mercury |
||
|
Nickel(g) |
0.02 |
0.91 |
Class
A for nickel refinery dust and nickel subsulphide via inhalation, Class B2
for nickel carbonyl. Slope
factor is derived from a draft value for inhalation and oral exposure from
Californian/EPA database and is not endorsed by USEPA. |
||
|
Silver |
0.005 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Zinc |
0.3 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Acenaphthene |
0.06 |
|
No
information on carcinogenicity available on IRIS |
||
|
Acenaphthylene |
|
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Anthracene |
0.3 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Benzo(a)anthracene |
|
1.1 |
Class
B2, probable human carcinogen |
||
|
Benzo(a)pyrene |
|
7.3 |
Class
B2, probable human carcinogen |
||
|
Chrysene |
|
0.032 |
Class
B2, probable human carcinogen |
||
|
Dibenzo(ah)anthracene |
|
8.1 |
Class
B2, probable human carcinogen |
||
|
Fluoranthene |
0.04 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Fluorene |
0.04 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Pyrene |
0.03 |
|
Class
D, not classifiable as to human carcinogenicity |
||
|
Phenanthrene |
No
information available |
Class
D, not classifiable as to human carcinogenicity |
|
||
|
Naphthalene |
0.02 |
|
Class
C, possible human carcinogen |
||
|
DDT |
0.0005 |
0.34 |
Class
B2, probable human carcinogen |
||
|
4,4-DDE |
|
0.34 |
Class
B2, probable human carcinogen |
||
|
|
|
|
|
||
|
PCBs |
|
2.0 |
Class
B2, probable human carcinogen |
||
|
Tributyltin(h) |
0.0003 |
|
Class
D, not classifiable as to human carcinogenicity |
||
Source: Integrated
Risk Information System, USEPA (www.epa.gov/ngispgm3/iris).
Notes: (a)
as inorganic arsenic, (b) specific RfD for food intake, (c) Cr (VI) was used in
the risk assessment, (d) Cr (III), (e) value derived from HEAST reported water
quality criteria, (f) no IRIS or HEAST for Hg, converted 0.0003 for HgCl2
by * 0.739, RfD for MeHg is 0.0001, (g) as soluble salts, (h) as tributyltin
oxide.
For
carcinogenic contaminants there are theoretical grounds for presuming that
there may not be a true NOAEL. A carcinogenic
health effect can be produced through the mechanisms of initiation or
promotion. Genotoxic substances
induce cancers by causing mutations in DNA, whereas non-genotoxic substances
cause initiated cells to proliferate or differentiate. The two mechanisms differ in that their
modes of action lead to fundamentally different techniques of risk
assessment. On the one hand,
genotoxic substances are generally treated as carcinogens for which there is no
threshold below which carcinogenic effects are not manifested; in other words,
zero risk is only associated with zero exposure. However, non-genotoxic substances are treated as substances
which can be tolerated by the receptor up to some finite concentration or dose,
beyond which toxic effects are then manifested. In this study, we have assumed a non-threshold approach for
all carcinogens, ie all carcinogens are considered to be genotoxic. This is a conservative assumption.
Where
a no effect level cannot be demonstrated experimentally, mathematical models
have been developed, particularly in the US, to enable a worst case
extrapolation from high doses to much lower exposures to be made. Using such calculations, the US
Environmental Protection Agency (US EPA) has also ranked substances causing
cancer in animals using so called Slope
Factors (SF) (formerly known as Cancer Potency Factors).
The
SFs can be used to estimate the excess lifetime cancer risks associated with
various levels of exposure to potential human carcinogens. The SF is a number which when
multiplied by the lifetime average daily dose per kilogram body weight of a
potential carcinogen, yields the lifetime cancer risk resulting from exposure
at that dose. In practice, slope
factors are derived from the results of human epidemiological studies or
chronic animal bioassays. The data
from animal studies are fitted to linearised multistage models and a
dose-response curve is obtained.
The slope in the low dose range is subjected to various adjustments, and
an interspecies scaling factor is applied to derive the slope factor for
humans. The SF is used to
determine the number of tumours likely to occur at low doses below which
experimental data do not exist.
The extrapolation is forced through the origin since for carcinogens
NOAELs are not predicted to occur, ie only zero exposure equals zero risk.
Among
the potential contaminants of concern are several substances that exhibit
route-specific toxicity.
Inhalation of cadmium, chromiumVI and nickel has been associated with increased
incidence of cancer in animals and/or humans. There is no adequate evidence, however of systematic
carcinogenic effects following oral exposure to these compounds, because the
substances may not be available for absorption through the gastrointestinal
tract, or may cause lung cancer by a mechanism which has no parallel in the
gastrointestinal tract. In this
assessment we are mainly concerned with evaluating risks associated with the
ingestion of seafood and hence only the oral SFs are of interest. Oral SFs are summarised above in Table 2.3a.
Previous
reports (EVS 1996 b and c, ERM 2002) have summarised the risks to marine
organisms from exposure to several heavy metals. In general, the toxic effects of metals in marine organisms
may include mortality, carcinogenicity, growth retardation, reduced
reproduction, effects on blood chemistry, neurological and developmental
effects, and behavioural effects.
Various organic contaminants may cause reproductive impairment, systemic
pathology, and cancer in cetaceans, including Sousa chinensis (Leland and Kuwabara 1985; Marsili et al 1997).
Although some of the metals (arsenic,
cadmium, chromium, and nickel) in some forms and DDT and PCBs are considered
possible human carcinogens, information is not available for deriving non-human
carcinogenicity factors (SFs).
Therefore, this assessment is based on risks of systemic toxicity,
including reproductive effects.
Estimated doses from the ingestion of contaminated prey species were
compared to Toxicity Reference Values
(TRV) to determine the potential risk to Indo-Pacific Humpback Dolphins
associated with the consumption of contaminated prey. The TRV is a maximum acceptable ingestion rate in mg kg-1
day-1 of a chemical in food of the species of concern, in this case,
the Indo-Pacific Humpback Dolphin.
To derive a TRV, it is necessary to perform a feeding study in which
food containing different concentrations of the contaminant of concern (the
doses) is fed to large numbers of test animals, usually mice or rats. Alternatively, a TRV can be estimated
from a food chain model if the absorption efficiency of the chemical from the
food is known and the critical body residue (the concentration in tissues
associated with adverse effects) of the chemical is known or can be estimated.
Although
it would be ideal to use TRVs derived for the specific species being evaluated
(ie the Indo-Pacific Humpback Dolphin), there are presently no available
feeding studies on cetaceans from which to estimate a TRV. In addition, only limited data are
available on the concentrations of 22 metals and several organochlorine
compounds (PCBs and chlorinated pesticides) in tissues of Indo-Pacific Humpback
Dolphins from Hong Kong waters (Appendix C-1).
There
is a large published scientific literature on the concentrations of several
metals and organic contaminants in tissues of cetaceans throughout the
world. In a few cases, the
concentrations of contaminants in cetacean tissues are related to various
pathological conditions. However,
nearly always, the cetaceans with pathological conditions contain several
contaminants at high concentrations in their tissues. Thus, it is not possible to derive a cetacean-specific TRV
for chemicals in cetacean tissues, based on tissue residue data alone. The TRV values are adjusted for weight
and metabolic rate differences between the species of concern and the test
species by a scaling factor (see below) following the standard approach used to
derive the oral reference doses (RfDs) for toxic chemicals in human food. In essence the TRV values act as RfDs
for marine mammals but have been derived using the body weight scaling factor
instead of the uncertainty factors used in the human health assessment.
In general,
when selecting toxicity studies for use in TRV derivation, the most important
information to evaluate (in addition to the overall quality and reliability of
the study) is: mode of exposure
(ie ingestion vs inhalation or gavage); endpoint evaluated (ie reproductive
effects vs behavioural effects); duration of study (ie chronic vs acute); and
lifestage of test organism evaluated.
It should be noted that the TRVs have been derived to take into account
chronic lifetime exposure to contaminants. The TRVs also take into account the potential for
bioaccumulation of contaminants (such as mercury, PCBs, DDT) by marine
mammals. Other factors, such as
the specific species evaluated is less important to the overall conclusions
regarding toxicity because it is assumed that most chemicals follow a similar
mode of action in all mammalian species.
Typically, laboratory toxicological studies are conducted using
relatively small mammals such as mice, rats, or mink due to the space
limitations associated with larger animals. Although as noted, differences in body weight can result in
differences in toxic response to chemicals, it has been demonstrated that these
differences can be accounted for by using a body weight scaling factor as
follows (Sample et al 1996):
TRVr
= NOAELt (Bwt/Bwr)1/4
where,
TRVr = Toxicity reference
value for receptor species (mg kg-1 wet wt day-1)
NOAELt = No observed adverse
effect level for test species (mg kg-1 wet wt day-1)
Bwr = Body weight of the
receptor species (kg wet wt)
Bwt = Body weight of the test
species (kg wet wt)
Using
this scaling factor, TRVs were derived for the Indo-Pacific Humpback Dolphin
based on NOAELs from mammalian species used as surrogates (Table 2.3b). Sample et al (1996) conducted an extensive review of the available
mammalian literature, carefully evaluating both the overall quality and
reliability of the study as well as the parameters described above. Therefore, the NOAEL values provided
are representative and appropriately conservative for the purpose of deriving
TRVs.
Table 2.3b Derivation of
toxicity reference values (TRV) for the Indo-Pacific Humpback Dolphin. The TRV is derived by scaling the toxic
dose from the test mammal to the dolphin.
The unit for NOAELs and TRVs are mg kg-1 wet wt day-1.
|
Chemical |
NOAEL |
Test
Species |
Test
Species wt (kg) |
TRV |
Reference |
|
|
|
Arsenic |
0.13 |
Mouse |
0.03 |
0.01 |
Schroeder
& Mitchner 1971 |
|
|
|
Cadmium |
1.00 |
Rat |
0.303 |
0.20 |
Sutou
et al 1980 |
|
|
|
Chromium
(Cr3+) |
2737.00 |
Rat |
0.35 |
570.82 |
Sample
et al 1996 |
|
|
|
Copper |
11.70 |
Mink |
1 |
3.17 |
Aulerich
et al 1982 |
|
|
|
Lead |
8.00 |
Rat |
0.35 |
1.67 |
Azar
et al 1973 |
|
|
|
Mercury |
1.00 |
Mink |
1 |
0.27 |
Aulerich
et al 1974 |
|
|
|
Nickel |
40.00 |
Rat |
0.35 |
8.34 |
Ambrose
et al 1976 |
|
|
|
Silvera |
0.01 |
Human |
70 |
0.004 |
USEPA
1999b |
|
|
|
Zinc |
160.00 |
Rat |
0.35 |
33.37 |
Schlicker
& Cox 1968 |
|
|
|
Naphthalene
a |
0.04 |
Human |
70 |
0.03 |
USEPA
1999a |
|
|
|
Total
PCB |
0.14 |
Mink |
1 |
0.04 |
Aulerich
& Ringer 1977 |
|
|
|
a.
A human health RfD was used as the
basis for the TRV in the absence of a mammalian NOAEL. This value was applied to both Low
and High MW PAHs b.
In the absence of data for DDE, values
for DDT were applied. c.
In the absence of chemical-specific
data, values for tributyltin were applied. |
||||||
The
NOAEL values of Sample et al (1996)
are conservative enough that additional uncertainty factors were not
applied. Typically, uncertainty
factors are applied to provide a more conservative toxicity estimate when
essential processes or toxicodynamic factors are not understood. Uncertainty factors can be applied for
various reasons, such as deriving no-observed-adverse-effect levels (NOAEL)
from less conservative toxicity endpoints such as
lowest-observed-adverse-effect levels (LOAEL) and acute toxicity values. An uncertainty factor can be applied to
a TRV if toxicity data for one species (the test species) is used to evaluate
effects in a second species (the wildlife receptor of concern). Specific values of uncertainty factors
applied to TRVs generally are not based on science, but are chosen because they
are simple (ie usually integer values) and result in conservative risk
assessments. The most recent
national EPA guidelines for ecological risk assessment (US EPA 1998)
qualitatively discuss empirical approaches to the use of uncertainty factors,
but do not propose a specific approach for uncertainty factor application. The national guidelines also note that
"uncertainty factors can be misused, especially when used in an overly
conservative fashion, as when chains of factors are multiplied together without
sufficient justification" (US EPA 1998).
In
deriving the TRV values used to evaluate risk to the Indo-Pacific Humpback
Dolphin, the focus is on studies in which a chronic NOAEL value was
reported. In the event that a
chronic NOAEL was not available, a chronic LOAEL was selected, and an
uncertainty factor of 10 was applied as discussed by Sample et al (1996). No acute values were considered, therefore, an additional
uncertainty factor is not required.
In addition, a body-weight scaling factor was applied (Sample et al 1996) to account for interspecies
differences. Application of
an additional uncertainty factor would assume that the Indo-Pacific Humpback
Dolphin is always more sensitive to the chemical of concern than the test
species for which the TRV was derived.
However, there are no empirical data available to support this
assumption. In fact, there is
evidence that cetaceans are more tolerant than terrestrial mammals to some
metals, such as mercury and cadmium (1)
(2) (3) (4). These and some other metals (e.g.
silver) accumulated from food are sequestered in the tissues (mostly liver for
mercury and silver and kidney for cadmium) as insoluble, inert particles that
are not toxic. Only when the sequestration capacity of the tissues is exceeded
do the metals accumulate in toxic forms in tissues. Therefore,
the approach as described is appropriately conservative to be protective of
potential adverse effects.
Measurement
endpoints for the human health risk assessment will include:
·
Incidence
of cancer in humans (for carcinogenic substances); and,
·
Incidence
of chronic conditions in humans (for non-carcinogenic substances).
In
this case, Sousa chinensis has been
identified as the ecological receptor of concern. As it is an endangered species the assessment must be
focused on evaluating impacts to individual organisms. Using the criteria presented, two
assessment endpoints have been identified for this ecological risk assessment:
·
Health
of individual Indo-Pacific Humpback Dolphins frequenting the East of Sha Chau
Area; and,
·
Reproductive
viability of the Indo-Pacific Humpback Dolphins inhabiting the East of Sha Chau
Area.
For
the purpose of this assessment, exposure parameters representing the “typical”
or “average” individual were selected.
It is assumed that values protective of this individual will be
protective of the majority of the exposed population. Assessment endpoints can be evaluated through either direct
or indirect measurements. These
measurements are referred to as measures of effect. Measures of effect are measurable responses to stressors
that may affect the characteristic component of the assessment endpoint (Suter
1990; Suter 1993). For this
assessment, the health and reproductive viability are the specific
characteristics of the dolphin that are potentially at risk. While some contaminants may influence
both characteristics, other contaminants may affect only health or only
reproductive viability. By
assessing the risk associated with each of the contaminants of concern both
endpoints are addressed.
The
purpose of an exposure assessment is to determine the intake of each COC by
potentially exposed individuals. In
this study, this will involve characterisation of the major pathways for
contaminant transport leading from the CMPs to the points of exposure. Exposure evaluation considers various
routes of contaminant release and migration from the CMPs to targeted populations
by:
·
evaluating
fate and transport processes for the contaminants;
·
establishing
likely exposure scenarios for each medium (eg water, diet, etc);
·
determining
the concentrations of the contaminants in each medium;
·
determining
exposures to potentially affected populations; and,
·
calculating
maximum short-term or average lifetime doses and resultant intakes.
The
resultant doses to and intakes by potentially exposed populations are
calculated once exposure concentrations in all relevant media have been
determined. Dose is defined as the
amount of chemical contacting body boundaries (skin, lungs, or gastrointestinal
tract) and intake is the amount of chemical absorbed by the body. When the extent of intake from a given
dose is unknown, or cannot be estimated defensibly, dose and intake are taken
to be the same (ie 100 percent absorption from contact). This is a highly conservative approach
and there are very few instances in which 100% of a chemical is absorbed in
this manner.
ERM
has developed a conceptual model to aid the assessment of contaminant exposures
to humans and dolphins (Figure 2.4a). The model is used to illustrate the
relationship between the stressors (contaminants of concern), and the receptors
of concern (humans and Sousa chinensis). The conceptual model integrates the
available information to identify exposure pathways. Each exposure pathway will include the stressor source
(dredged material disposal activities), the stressor of concern (COCs), the
exposure route (ingestion), and the receptor of concern (humans and Sousa chinensis). The basic premise of the model is to
evaluate the toxicological effects of the contaminants of concern associated
with disposal activities at East of Sha Chau.
Substances
potentially migrating from the pit into the marine environment will be
dispersed into the ambient environment and may potentially impact on human and
dolphin populations through ingestion of contaminated sediment, ingestion of
dissolved and suspended contaminants in water, ingestion of organisms with
contaminant residues in their edible portions and through contact with
water. Of these four pathways the
primary pathway of concern is considered to be that of the ingestion of
contaminants contained within the edible portion of marine organisms.
The
impact hypotheses for the assessment of human health risks are thus defined as
follows:
IH1: Risks to human health from consumption of commercial
species captured adjacent to the proposed contaminated mud disposal facility
are no greater than risks associated with consumption of species remote from
the proposed facility;
AND
IH2: Risks to human health from consumption of commercial
species captured adjacent to the proposed contaminated mud disposal facility
are below the screening risk criterion.
The
impact hypotheses for the assessment of ecological risks are defined as
follows:
IH1: Risks to dolphins from consumption of prey species
captured adjacent to the proposed contaminated mud disposal facility are no
greater than risks associated with consumption of species remote from the
proposed facility;
AND
IH2: Risks to dolphins from consumption of prey species
captured adjacent to the proposed facility are below the screening risk
criterion.
The
general equation used to estimate exposure is presented below:
Intake (mg kg-1
day-1) = CF ´ IR ´ FI ´ EF ´ ED
BW ´ AT
Where:
CF = Contaminant Concentration in Fish
and Shellfish (mg kg-1 ww)
IR = Ingestion Rate (kg day-1)
FI = Fraction Ingested from Contaminated
Source (unitless)
EF = Exposure Frequency (day year-1 )
ED = Exposure Duration (years)
BW = Body Weight (kg)
AT = Averaging Time (period over which
exposure is averaged - days)
The
relative contributions of each dietary item to the total intake are then
included in the calculation to give an indication of the overall exposure via
fish and shellfish ingestion.
Input values have been calculated to reflect local conditions and are
discussed below.
The
data incorporated into this assessment are the tissue contaminant
concentrations obtained in the bioaccumulation assessment. As discussed in Annex B these
values represent the high end of the range as they are determined from worse
case assumptions and are consequently expected to result in high-end estimates
of risk. Reference concentrations
are also used in the assessment for comparison purpose.
The
rate of ingestion of seafood is a key exposure variable for use in this risk
assessment. Seafood is known to be
an important component of the diet of Hong Kong residents and it is estimated
that the amount consumed daily is an order of magnitude higher than that
consumed in other countries such as the US. The seafood consumed in Hong Kong is derived from a wide
variety of sources:
·
Imported
from overseas in live, fresh, chilled, frozen, canned, preserved, salted,
smoked or dried forms;
·
Landed
by the Hong Kong fishing fleet but caught outside of Hong Kong waters; and,
·
Landed
by the Hong Kong fishing fleet and caught within Hong Kong waters.
According to AFCD's Annual Report (AFD
1998a) and information provided by AFCD the amount of fisheries and seafood
products consumed by the Hong Kong populace is 43 kg yr-1 capita-1. Of this amount, 6.6 kg are freshwater
fish which can be eliminated from the marine consumption total for this
analysis, consequently the seafood consumption per capita is 36.4 kg yr-1
or 0.104 kg day-1 (36.4 ¸ 350 days). It is assumed that this figure is based on the amount
ingested (0.104 kg day-1) comprising the entire seafood
product. This figure is used to
represent the average consumer of fish products. For sectors of the population that consume comparatively
more fisheries products, eg fishermen, the USEPA recommends using a gross
consumption rate of 0.3kg day-1. This rate is considered to be upper bound and is not
expected to occur in reality.
Consequently the maximum consumption rate has only been applied to East
of Sha Chau Fishermen for scenario using all 3 years of data.
The values above are likely to be an
overestimate as the amount actually ingested will be lower due to molluscs,
crustaceans and fish having shells, viscera and skeletal structures. Conversion factors that can be used to
convert gross seafood ingestion rates into tissue specific ingestion rates were
presented in Shaw (1995). These
values were higher than those suggested for use by the US National Marine
Fisheries Service (NMFS 1987) because it was considered that in eastern
cultures more of the seafood product is eaten, such as internal organs (eg swim
bladder or crab hepatopancreas) that are not usually part of the western
diet. For the purposes of this
risk assessment the following factors have been applied to calculate net
ingestion rates for each dietary item:
·
Prawns
= 0.88 (maximum value used by the NMFS 1987)
·
Swimming
Crab = 0.22 (NMFS 1987)
·
All
fish = 0.5 (NMFS 1987)
·
Bivalve
= 1.0
The
risk assessment calculations for ingestion rate were proportioned into the
different dietary items. It was
assumed that the proportion of each dietary item in catches in Hong Kong would
reflect the proportion in the diet of Hong Kong people. The composition of the catch from the
East of Sha Chau area was identified using data from AFCD's Fisheries Study
(ERM 1998) presented below in Table 2.4a. Values are also presented below for the
composition of landings at Tuen Mun Port (the main port in the Study Area) and
for the composition of catches taken in Hong Kong waters for comparison. As can be seen from Table 2.4a the composition of catches
from East of Sha Chau are broadly similar to those from the whole of Hong Kong
and those landed at Tuen Mun Port.
Table 2.4a Composition of
Catches (%) from Hong Kong, Tuen Mun Port & East of Sha Chau (ERM 1998)
|
Type |
Hong
Kong Catch |
Catch
Landed at Tuen Mun Port |
Catch
from East of Sha Chau Area |
|
Pelagic Fish |
41.7 |
43.0 |
41.6 |
|
Predatory Fish |
46.8 |
44.8 |
44.7 |
|
Crab |
3.0 |
3.1 |
4.0 |
|
Prawn |
6.1 |
8.4 |
8.8 |
|
Mollusc |
2.4 |
0.7 |
0.9 |
After application of the conversion
factor data and the catch composition/dietary fraction information presented
above to the gross seafood consumption estimate of 0.104 kg day-1, individual
ingestion rates can be calculated for each of the dietary items in terms of net
consumption in kg day-1.
The resultant total net seafood consumption rate after application of
the conversion factors is 0.0548 kg day-1. Application of the conservation factors and catch fraction
information to the maximum consumption rate of 0.3 kg day-1 results
in a net consumption of 0.1580 kg day –1 (Table 2.4b).
Table 2.4b Ingestion Rates
(kg day-1) for Each Dietary Item (for an average consumer) – Average
Consumer and Maximum Consumer (East of Sha Chau Fishermen)
|
Type |
Average Net Consumption (kg day-1) |
Maximum Net Consumption (kg day-1) |
|
|
Pelagic
Fish |
0.021660 |
39.5 |
0.062480 |
|
Predatory
Fish |
0.024362 |
44.5 |
0.070276 |
|
Crab |
0.000692 |
1.3 |
0.001997 |
|
Prawn |
0.005544 |
10.1 |
0.015991 |
|
Mollusc |
0.002510 |
4.6 |
0.007242 |
|
TOTAL |
0.054768 |
|
0.157985 |
It is
unlikely that 100% of the seafood consumed by an individual will be from the
same source. The Fraction Ingested
(FI) value represents the fraction of total seafood ingested from the
contaminated region of interest (ie the East of Sha Chau area).
The
catch from the old AFCD fishing zones in the Study Area (0017, 0018, 0019,
0020, 0032, 0033, 0040, 0041, 0042, 0043, 0044, 0045) amounts to a total of
1,894 tonnes per year (AFD 1998a).
The total amount of seafood products consumed in Hong Kong per year was
reported in AFCD's (AFCD 1999) information to ERM at 243,440 tonnes per year.
The fraction of this amount obtained from
the East of Sha Chau area is therefore 1,894 ¸ 243,440 = 0.0078. This value is lower than that
used by Shaw (1995) who based the fraction ingested on the amount caught in the
East of Sha Chau area divided by the total landings (ie 1,894 ¸ 186,000 = 0.01). This number appears to be an
overestimate because the consumption rate of 36.4 kg yr-1 is based
on all seafood products not just that landed by the Hong Kong fleet. The AFCD Annual Report (AFD 1998a) has
indicated that the total catch landed in Hong Kong is 186,000 tonnes per year
of which 17,681 tonnes per year has been estimated to have been caught in Hong
Kong waters (ERM 1998). Estimates
of the FI have been prepared for three exposure populations of concern, which
are as follows:
Hong Kong People: It is assumed that this population experience the average
exposure to COCs in seafood. The
FI for this population is represented by the value derived above, ie 0.0078. This indicates that 0.78% of the seafood consumed by Hong
Kong people is obtained in the East of Sha Chau area. Information on the contribution of seafood to the total diet
of Hong Kong People is not needed in this risk assessment as the methodology is
concerned with the effects of contaminants in the edible portion of seafood on
human health. This population is
comparable to the Central Tendency used in previous risk assessments (Shaw
1995; EVS 1996a) and follows the method used during the CMP IV EM&A
Programme (ERM 2002).
Hong Kong Fishermen: Calculating the values for this population is more
speculative due to uncertainties over the amount of a fisherman's diet that is
composed of seafood. The US EPA estimate
that 75% of a fishermen's diet will originate from within local waters (defined
as the whole of Hong Kong). 10.7%
of the Hong Kong catch comes from East of Sha Chau (1,894t/17,681t) the FI is
set at 0.08 (10.7% ´ 75%). This indicates that 8% of the seafood
consumed by Hong Kong Fishermen is obtained in the East of Sha Chau area. This population is comparable to the
Reasonable Maximum Exposure used in previous risk assessments (Shaw 1995; EVS
1996a).
East of Sha Chau Fishermen: For this population it is assumed again that 75% of the diet
is obtained in local waters, but this time local refers to catches landed at
the home port within the East of Sha Chau area (Tuen Mun). The fishing fleet that operate from
Tuen Mun obtain 65% of their catch within the East of Sha Chau area. Hence the FI for these fishermen is
estimated at 0.49 (65% x 75%). This
indicates that 49% of the seafood consumed by East of Sha Chau Fishermen is
obtained in the East of Sha Chau area. This population is comparable to the Sensitive
Subpopulation used in previous risk assessments (Shaw 1995; EVS 1996a).
Combining
the FI values for each population of concern with the information on catch
breakdown provides FI estimates for each food type. These values are presented below in Table 2.4c.
Table 2.4c Fraction Ingested
from the East of Sha Chau Area for Three Populations of Concern
|
Type |
%
of Catch |
HK
people FI
= 0.0078 |
HK
Fishermen FI
= 0.08 |
Tuen
Mun Fishermen FI
= 0.49 |
|
Pelagic Fish |
41.63 |
0.003246816 |
0.033301 |
0.203967 |
|
Predatory Fish |
44.74 |
0.00348936 |
0.035788 |
0.219203 |
|
Crab |
3.97 |
0.000309663 |
0.003176 |
0.019453 |
|
Prawn |
8.79 |
0.000685247 |
0.007028 |
0.043048 |
|
Mollusc |
0.88 |
0.00006891 |
0.000707 |
0.004329 |
The
exposure frequency is the average number of days per year over which an individual
is exposed to one or more COCs via ingestion of seafood. A value of 350 days, as specified by
the USEPA (USEPA 1991) for long term average contact, has been assumed for this
assessment.
The
exposure duration is the time period in years over which an individual is
exposed to one or more contaminants in seafood from East of Sha Chau. For the purposes of this assessment we
have adopted the lifetime of the facility, ie 8 years.
US EPA
guidelines for risk assessment (US EPA 1989) indicate that the default value
recommended for body weight (BW) is 70 kg. However, Asians are in general smaller in stature than their
Caucasian counterparts, so it is considered that the US EPA default value would
not be representative of the Hong Kong population. A value of 60 kg was assumed for body weight to represent
the local Hong Kong population as determined by Shaw (1995).
The
averaging time (AT) is another important parameter of the intake equation. The AT selected will depend on the type
of constituent being evaluated, for example, to assess long term or chronic
effects associated with exposure to noncarcinogens, the intake is averaged over
the exposure duration (expressed in days). Exposure to carcinogens, however, is averaged over a
lifetime in order to be consistent with the approach used to develop Slope
Factors (SFs). A value of 70 years
was assumed for mean life expectancy according to the default value used by the
US EPA.
A
summary of the values incorporated into the human health risk assessment are
presented below in Table 2.4d.
Table 2.4d Summary of Input
Parameters for Intake Equation
|
Variable |
Values |
|
Contaminant
Concentration in Seafood (CF) |
Presented
in Section 3.2 |
|
Ingestion
Rate (IR) |
0.0548
kg day-1 |
|
Fraction
Ingested from East of Sha Chau (FI) |
Values
for each population presented in Table
2.4c |
|
Exposure
Frequency (EF) |
350
days yr-1 |
|
Exposure Duration (ED) |
·
8 years |
|
Body
Weight (BW) |
60
kg |
|
Averaging
Time (AT) |
·
2920 days (8 years x 365 days = 2920
days) non-carcinogens ·
25,550 days (carcinogen - assuming a 70
year life expectancy) |
The data
from the bioaccumulation assessment of COCs in potential prey species of the
Indo-Pacific Humpback Dolphin were used to estimate doses received via the
dolphin diet. An average dose from
the total diet was estimated by determining the fraction of the total dolphin
diet derived of each category of food (eg prawns, crabs, predatory fish,
pelagic fish) and summing the tissue concentration values for each category
multiplied by the fraction of that category in the dolphin diet. As previously discussed, the intent of
this evaluation is to provide a determination of the potential risks to the
Indo-Pacific Humpback Dolphin population in the North Lantau waters of Hong
Kong, resulting from dredged material disposal in the proposed contaminated mud
disposal facilities. The exposure
pathway is assumed to be consumption of contaminated food by dolphins residing
in potentially impacted areas near the mud pits, and in reference areas. The methodology is designed to provide
a conservative estimate of the risks to Indo-Pacific Humpback Dolphins. For the purpose of this assessment,
dose estimates were derived for the Indo-Pacific Humpback Dolphin according to
the following equation:
Dose = (PC x IR x SRT x FI x ED) / BW x
AT
Where:
Dose = Chemical-specific
ingested dose (mg kg-1 day-1)
PC = Concentration
of chemical in prey item (mg kg-1)
IR = Ingestion
Rate (kg day-1)
BW =
Body weight
of dolphin (kg)
SRT =
Site
Residency Time (day year-1)
FI = Fraction
Ingested (unitless)
ED =
Exposure
Duration (years)
AT = Averaging
Time (period over which exposure is averaged - days)
Due
to lack of data previous risk assessments have assumed that the dolphins spend
100% of their time feeding at the mud pits throughout their lifespan. Information presented in the Baseline
Conditions section of this EIA (Part 1, Section 4) would indicate that
the two proposed mud pit areas are not as frequently used as reference areas to
the north around Lung Kwu Chau.
Consequently we have adopted values as follows:
·
Reference
Area site residency time = 100 % = 365 days (FI = 1)
·
Airport
East site residency time = 10 % = 36.5 days (FI = 0.1)
·
East
of Sha Chau site residency time = 50 % = 182.5 days (FI = 0.5)
The
rationale for the selection of the body weight and ingestion rate parameter
values is presented below.
Concentrations of contaminants in the prey items are presented in Section 3.
Available
data on the body weight of the Indo-Pacific Humpback Dolphin is variable.
Zongguo (1996) reported adult body weights ranging from 120 to 240 kg for
females, and from 110 to 230 for males.
These data were based on 36 dolphins collected in Xiamen Harbour in
1961. In southern African waters,
average adult body weights for humpback dolphins range from 170 kg for females
to 260 kg for males (Cockroft 1996).
Based on these data, an average body weight of 185 kg was assumed for
the purpose of this assessment.
This weight represents a high estimate of the average body weight of all
age classes in the East of Sha Chau dolphin population.
Ingestion Rate
For
the purpose of this evaluation, the ingestion rate of the Indo-Pacific Humpback
Dolphin was assumed to be similar to that of humpback and bottlenose
dolphins. Data for these species indicate
that they consume approximately four percent to six percent of their body
weight per day (Parsons 1996). An
ingestion rate of 9 kg day-1 was used for this assessment, assuming
a body weight of 185 kg and an average ingestion rate of five percent of body
weight per day. The values for the
ingestion rate and body weight were selected based on the available
literature. It is important to
note that the risk assessment methodology is designed to evaluate potential
risks to a representative individual of an affected population. For the purpose of this assessment,
exposure parameters representing the ‘typical’ or ‘average’ individual were
selected. It is assumed that
values protective of this individual will be protective of the majority of the
exposed population.
The
averaging time (AT) is another important parameter of the intake equation. The AT is expressed in days, ie 8 years
for the lifetime of the facility multiplied by the days in the year, ie 8 x 365
= 2920 days). Exposure to
carcinogens, however, is averaged over a lifetime in order to be consistent
with the approach used to develop Slope Factors (SFs). A value of 70 years was assumed for
mean life expectancy according to the default value used by the US EPA.
The dose calculations have been modified
to account for the level of organic Arsenic present in seafood. The RfD and TRV values for Arsenic are
based on the toxic effect of inorganic arsenic. Arsenic in marine cephalopod, crustacean, and fish tissues
is, however, predominantly in the form of organo-arsenic compounds, primarily
arsenobetaine (Neff 1997). These
organo-arsenic compounds are not accumulated in tissues of mammalian consumers,
including dolphins and humans, and are not toxic. Arsenobetaine was excreted unmetabolized in the urine of
male mice (Kaise and Fukui 1992).
The median lethal dose (LD50) of arsenobetaine in the mice
was greater than 10 g kg-1 body wt (10,000 ppm). Other organo-arsenic compounds evaluated
had LD50 values ranging from 1.2 to 10.6 g kg-1. By comparison, the acute toxicity of
arsenic trioxide (the form of arsenic used to derive both the Human Health RfD
and the Marine Mammal TRV) was 34.5 mg kg-1.
Therefore,
the naturally high concentrations of Arsenic in the tissues of marine organisms
do not pose a risk to either humans or Indo-pacific Humpbacked dolphins. It is rapidly excreted unchanged in the
urine of mammals and so does not bioaccumulate. Arsenobetaine is not easily converted to the inorganic
arsenite form which is of concern due to cancer risk. It can therefore be considered that the results of the risk
assessment for Arsenic may be an overestimation of the likely risks associated
with the consumption of seafood given that the Arsenic consumed is in a toxic
form.
Estimations
of the inorganic Arsenic fraction of seafood components of the risk assessment
have previously been determined during the monitoring works at CMP IV ([6]). The data were obtained by chemical
analysis of samples collected January and February 2000. The mean percentage of total Arsenic
that is represented by the inorganic fraction was calculated for each of the
human health risk assessment groupings.
At that time no tissue samples were collected for prawns and hence the
ratio from mantis shrimps was used.
This is considered to be an appropriate assumption given the ecological
and taxonomic similarity between the two organisms. The following ratios were applied to the total Arsenic data:
·
Prawns
and Mantis Shrimps = Total Arsenic (mg kg-1) x 0.535 %
·
Swimming
Crabs = Total Arsenic (mg kg-1) x 0.285 %
·
Flatfish
= Total Arsenic (mg kg-1) x 0.265 %
·
Burrowing
Fish = Total Arsenic (mg kg-1) x 1.895 %
·
Demersal/Pelagic
Fish = Total Arsenic (mg kg-1) x 0.650 %
·
Gastropod
= Total Arsenic (mg kg-1) x 5.215 %
For the purposes of this risk assessment the highest value
5% from the gastropod has been applied to the Arsenic values from the
Bioaccumulation Assessment (Annex B). The corrected data were then used in the risk assessment.
Risk
characterisation generally involves the integration of the information and
analysis of the first three components of the assessment, as discussed in Sections 2.2, 2.3 and 2.4. Risk is generally characterised as follows:
·
For
non-carcinogens, and for the non-carcinogenic effects of carcinogens, the
margin of exposure is estimated by dividing an estimated daily dose by a
derived "safe" dose to form a ratio. This ratio is referred to as a Hazard Quotient and if it is
greater than one there is sufficient concern for further analysis.
·
For
carcinogens, risk is estimated by multiplying the estimated dose by the risk
per unit of dose. A range of risks
might be produced, using different models and assumptions about dose-response
curves and the relative susceptibilities of humans and animals.
Although
this step can be more complex than is indicated above, especially if issues of
the timing and duration of exposure are introduced, the hazard quotient and the
carcinogenic risk are the ultimate measures of the likelihood of injury or
disease from a given exposure or range of exposures. This section describes the approach used to assess the overall
risks of fish and shellfish ingestion to humans and dolphins. The approaches used are independent of
each other to a large degree, and are presented separately.
The
intakes, calculated using the data presented in Table 2.4c and the equation in Section
2.4.2, will be compared with the Reference Doses (RfD) (see Table 2.3a) as a means of calculating
non-carcinogenic hazards, which are expressed as the Hazard Quotient (HQ).
Hazard Quotient = Intake
Reference Dose
HQs
can be summed to provide an estimate of the cumulative non-carcinogenic hazard
which is known as the Hazard Index (HI).
This is a conservative approach and assumes that all of the COCs exert
an effect on the same target organ.
Carcinogenic
risks will be calculated using the following equation:
Risk = Intake x Slope Factor
This
equation will provide an estimate of the lifetime carcinogenic risk associated
with the estimated intake.
Concern
is often expressed about the hazard to health from exposure to mixtures of
substances, rather than individual substances. There is no agreed procedure among toxicologists for
estimating such a hazard. The
toxic effects of two substances in combination may be the sum of the individual
toxicities (ie additive), more than the sum (ie synergistic), or less than the
sum (ie antagonistic). Synergism
appears to be, in practice, a very much less common phenomenon than a noticeable
combined effect or an additive effect. However, since there is a lack of direct data on most
chemical combinations, the most reasonable strategy is to assume that chemicals
which affect the same target organisms, in a similar manner, will have additive
toxicities.
The
available literature on such effects is very limited and, where it does exist,
is largely restricted to the behaviour of metals in experimental animals. The application of such data to human
studies is, at best, questionable.
In the absence of any reasonable scientific basis for predicting
antagonistic or synergistic reactions in complex mixtures, only examination of
an additive model of toxicity is considered to be justified.
There
are two related methods of making some quantitative assessment of the toxic
impact of a mixture. The first,
that is recommended by the UK Health and Safety Executive (HSE), is to use the
following equation:
C1 + C2 + C3 + ... Cn
=
X
L1 L2 L3 Ln
Where C1, C2, C3...Cn
are the concentrations of each contaminant in food and L1, L2,
L3...Ln = the "safe levels" of each, ie the
reference dose RfD. If the total X
is less than one, the mixture is considered not to represent a health hazard;
whereas if X is greater than one, steps should be taken to reduce the concentrations
of one or more of the contaminants.
For
carcinogens, a conservative approach is achieved using the
"response-addition" process, which simply sums the individual
lifetime risks linearly to reflect the combined potential of cancer should a person
be exposed to all of the substances over a lifetime.
Total Excess Cancer Risk = Risk 1 + Risk 2 +
Risk 3 + ... Risk" n"
Where:
Risk
1 = Individual excess cancer risk from a lifetime exposure from the first
substance;
Risk
"n" = Individual risk of additional substances.
While
the "response-addition" process is encouraged as a
"first-cut" or screen to indicate that a cancer may occur from the
exposure to multiple substances, it should be remembered that the conservative
nature of risk assessments for individual substances can be exaggerated by this
additive approach.
For
each contaminant, a hazard quotient will be calculated using the following
ratios (US EPA 1997):
HQ = Dose/TRV
where,
HQ hazard quotient for individual chemicals
Dose estimated contaminant concentration ingested through consumption of
prey items (mg contaminant kg wet body weight-1 day-1);
and,
TRV the toxicity reference value (defined in Section 2.3.2, Table 2.3b) mg kg-1 wet weight day-1
The
risk estimates generated in this investigation are based on a considerable
number of assumptions, uncertainties and variability associated with each step
in the risk assessment process. According
to US EPA guidelines these assumptions and uncertainties should be presented
along with the results so that a fully informed picture is given to decision
makers (US EPA 1989; LaGrega et al
1994).
Hazard Identification: This stage is based on data for which detection,
identification and quantification limits could introduce errors. The selection of COCs in this
assessment was made according to the list presented in Study Brief which,
though not an exhaustive list appears sufficiently comprehensive for the
purposes of this assessment. Other
chemicals may pose a threat to human and/or dolphin health and exclusion from
this investigation does not infer that they are not of concern.
Dose-Response Evaluation: The toxicity assessment stage has a very high degree of
uncertainty associated with the slope factors and reference doses. In future assessments the toxicological
information should be revisited and updated using the latest available
information. For example, the
slope factor for Nickel was formulated by the Californian EPA. The slope factor is draft and not
endorsed by the USEPA and represents both oral and inhalation exposures. At present there is considerable
uncertainty as to the elements carcinogenicity through this exposure pathway. Any estimate is therefore conservative
and may be overly protective as for most metals inhalation slope factors are
generally an order of magnitude higher than oral slope factors.
Exposure Assessment: This stage depends heavily on the assumptions made about the
pathways, frequency and duration of exposure to COCs.
Risk Characterisation: The computation of screening-level risk is an exercise in
applied probability of extremely rare events, therefore not every conceivable
outcome can be evaluated. This
introduces an inherent conservatism which often results in assessing a scenario
that will never be experienced.
In
summary, risk assessment by design is very protective of human and ecological
health by ensuring that potential exposures and risks are not understated. Despite varying degrees on uncertainty
surrounding risk assessments, they represent the most useful tool that can be
used to determine and protectively manage the risk to human and ecological health.
Annex
C – Appendix A International
Literature Review on Marine Mammals
(1) Caurant, F. and C. Amiard-Triquet (1995) Cadmium contamination in pilot whales Globicephala melas: source and potential hazard to the species. Mar.Pollut. Bull. 30:207-210.
(2) Caurant, F., M. Navarro, and J.-C. Amiard (1996) Mercury in pilot whales: possible limits to the detoxification process. Sci. Tot. Environ. 186:95-104.
(4) Palmisano, F., N. Cardellicchio, and P.G. Zambonin (1995) Speciation of mercury in dolphin liver: a two-stage mechanism for the demethylation accumulation process and role of selenium. Mar. Environ. Res. 40:109-121.
(5)