Content uploaded by Dave A. Higgs
Author content
All content in this area was uploaded by Dave A. Higgs on Feb 16, 2018
Content may be subject to copyright.
Flesh Quality of Market-Size Farmed
and Wild British Columbia Salmon
M. G. IKONOMOU,*,† D. A. HIGGS,‡
M. GIBBS,†J. OAKES,‡B. SKURA,§
S. MCKINLEY,‡,§ S. K. BALFRY,‡,§
S. JONES,⊥R. WITHLER,⊥AND
C. DUBETZ†
Department of Fisheries and Oceans (DFO), Institute of Ocean
Sciences, Sidney, BC, Canada, DFO/University of British
Columbia, Centre for Aquaculture and Environmental
Research, West Vancouver, BC, Canada, Faculty of Land and
Food Systems, University of British Columbia,
Vancouver, BC, Canada, DFO, Pacific Biological Station,
Nanaimo, BC, Canada
This study compared the flesh quality of farmed and wild
sources of British Columbia (BC) salmon with respect
to concentrations of polychlorinated biphenyl compounds,
polychlorinated dibenzodioxins/dibenzofurans and their
associated toxic equivalents, total mercury (THg),
methylmercury (MeHg), and selected fatty acids of known
importance for human health viz., omega-3 (n-3) highly
unsaturated fatty acids (n-3 HUFAs) and (n-6) fatty acids.
Skinned fillets from known sources of farmed Atlantic, coho,
and chinook salmon (
n
)110) and wild coho, chinook,
chum, sockeye, and pink salmon (
n
)91) were examined.
Atlantic salmon contained higher PCB concentrations
(means, 28-38 ng/g) than farmed coho or chinook salmon,
and levels in these latter species were similar to those
in wild counterparts (means, 2.8-13.7 ng/g). PCB levels in
Atlantic salmon flesh were, nevertheless, 53-71-fold
less than the level of concern for human consumption of
fish, i.e., 2000 ng/g as established by Health Canada and the
U.S. Food and Drug Administration (US-FDA). Similarly,
THg and MeHg levels in all samples were well below the
Health Canada guideline (0.5 µg/g) and the US-FDA
action level (1.0 µg/g). On average, THg in farmed salmon
(0.021 µg/g) was similar to or lower than wild salmon (0.013-
0.077 µg/g). Atlantic salmon were a richer source (mean, 2.34
g/100 g fillet) of n-3 HUFAs than the other farmed and
wild sources of salmon examined (means, 0.39-1.17 g/100
g). The present findings support the recommended
weekly consumption guidelines for oily fish species (includes
all BC salmon sources) for cardio-protective benefits as
made by the American Heart Association and the UK Food
Standards Agency.
Introduction
Canadian farmed salmon and trout represented approxi-
mately 7.4% of global production for these species in 2002,
with 4.7% stemming from British Columbia (BC; 1,2). BC is
ranked as the fourth largest producer of farmed salmon in
the world (Government of BC, Ministry of Agriculture, Food
and Fisheries, 2002 statistics). A recent report of polyhalo-
genated arylhydrocarbons (PHAHs) or organohalogen con-
taminant levels in farmed Atlantic salmon stocks suggested
that their frequent consumption poses a greater risk of
causing cancer than that of wild stocks (3). However, the
risks of consuming farmed salmon were not compared to
the considerable human health benefits that may accrue from
eating salmon regardless of origin. These benefits arise
primarily from their flesh content of omega-3 (n-3) highly
unsaturated fatty acids (n-3 HUFAs), especially eicosapen-
taenoic acid (EPA) and docosahexaenoic acid (DHA), and, to
a lesser degree, from concentrations of linolenic acid, and
monounsaturated fatty acids. The potential health benefits
related to adequate weekly ingestion of oily fish such as
salmon include reduced risk of cardiovascular disease (CVD)
such as coronary heart disease (CHD), some cancers, and
various inflammatory responses and conditions, and en-
hanced brain, cognitive, and ocular development and func-
tion (4-11). Some of the preceding conditions such as CVD
account for considerable health care costs in nations with
low annual per capita fish consumption such as the United
States and Canada and many European countries versus those
where frequent fish consumption is a mainstay of life, e.g.,
Japan (6,12,13).
The benefits of an oily fish diet can be counteracted when
methylmercury (MeHg) levels exceed human health stan-
dards (14). MeHg is found principally in the low oxygen,
sub-thermocline ocean region (15), and is present in nearly
all marine fish in trace amounts (16). Accumulation in fish
is proportional to age, size, and trophic level (17). Hence,
MeHg levels in large, long-lived, predatory fish such as tuna,
shark, or swordfish, can exceed both the Health Canada
guideline (0.5µg/g) and the United States Food and Drug
Administration (US-FDA) action level (1µg/g), while smaller
fish at lower trophic levels, e.g., salmon, pollock, and hake,
contain low MeHg levels (18).
Clearly, any agency or government that sets recommended
consumption levels of oily fish must also weigh the risks
associated with potential organohalogen intake with the
known benefits derived especially from the n-3 HUFAs. The
consumption levels for both wild and farmed salmon species
recommended in the Hites et al. (3) study only considered
the U.S. Environmental Protection Agency (US-EPA) guide-
lines for organohalogen intake. The US-EPA and Agency for
Toxic Substances and Disease Registry (ATSDR) daily con-
taminant concentration threshold intake levels are highly
conservative when compared to the guidelines for an array
of other agencies/countries, e.g., Health Canada (19). In fact,
the US-EPA draft value for tolerable daily intake of poly-
chlorinated dibenzodioxins (PCDDs) is not yet US-EPA
policy, as it is under review by the National Academy of
Sciences (20).
Comparisons of the contaminant and nutrient levels in
farmed and wild salmon have been complicated by physi-
ological and geographical variations that exist between and
within species. BC has five species of wild salmon, namely,
chinook (Oncorhynchus tshawytscha), coho (O. kisutch),
sockeye (O. nerka), chum (O. keta), and pink (O. gorbuscha).
Three salmon species are currently farmed in BC, namely,
Atlantic (Salmo salar), chinook, and coho and these are
ranked respectively, first, second, and third in order of
production volume/value (1). Previous studies (3,21) have
included farmed coho and chinook in the broad category of
farmed salmon without considering possible differences in
* Corresponding author phone: (250) 363-6804; fax: (250) 363-
6807; e-mail: IkonomouM@dfo-mpo.gc.ca.
†Institute of Ocean Sciences.
‡Centre for Aquaculture and Environmental Research.
§University of British Columbia.
⊥Pacific Biological Station.
Environ. Sci. Technol.
2007,
41,
437-443
10.1021/es060409+ CCC: $37.00 2007 American Chemical Society VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9437
Published on Web 12/13/2006
their flesh contaminant concentrations and nutrients relative
to farmed Atlantic salmon. Differences in diet, geographic
origin, maturation stage, and harvest season of both farmed
and wild salmon can affect contaminant levels measured
between and within species as revealed by the large
confidence limits noted in previous studies (3) and as reported
by O’Neill et al. (22) for Puget Sound wild chinook and coho
salmon. Hence, when the sample size is small, there may be
under- or over-estimation of mean flesh contaminant levels
and difficulty in ascertaining whether there are real (sig-
nificant) differences in contaminant levels between dissimilar
sources of salmon.
Additionally, within a given fish fillet, regional differences
exist in lipid content (23) and consequently in levels of the
lipophilic organohalogen contaminants. Since contaminant
and nutrient concentration data (e.g., n-3 HUFAs) set
recommended human consumption levels, it is important
that the portion of the fish that is analyzed represents that
which is consumed. Some studies have analyzed only a
salmon “steak” or the epaxial muscle. To avoid regional
differences in contaminants and nutrients, the whole fillet
should be analyzed. Moreover, because the skin of the fillet
contains adhering lipid and associated organohalogen com-
pounds, and the skin is typically not consumed, skinless fillets
should be analyzed to avoid overestimation of contaminant
loads. Further, the lack of lipid concentration determinations
in some of the previous investigations has made comparisons
of contaminant findings with this study difficult. Thus,
standardization of methodology enables more accurate
comparisons between studies.
A final point of concern related to some previous studies
that have reported organohalogen concentrations in the flesh
of aquatic species, in particular polychlorinated biphenyl
compounds (PCBs), is the frequent omission of many
congeners. This has resulted in underestimation of con-
taminant concentrations, and consequently, in values for
toxic equivalents (TEQs) of these compounds for humans.
If TEQs are included in reports, it is essential to assign
appropriate toxic equivalent factors (TEFs) for all congeners.
In this study, we addressed some of the limitations of
past studies to present the data in an unbiased manner.
Accordingly, we collected many market-size farmed salmon
from eight farm sites (included all three salmon species
farmed in BC). Also, we sampled eight wild salmon popula-
tions that were representative of the five salmon species found
in coastal BC waters. The fish were sampled across a range
of sampling dates and geographical locations mainly for
determinations of fillet (flesh) contaminant and nutritional
analyses. Contaminant analysis focused on full congener
PCDD/Fs, full congener PCBs, total mercury (THg), and
MeHg, while the nutritional analyses reported herein are
specific to lipid concentrations and compositions. Our
collection procedures were unique among all of the other
studies on this theme since we had complete knowledge of
the sample origins, handling procedures, and storage condi-
tions. Also, the identities of all of the salmon flesh samples
were unknown to the analysts (blinded study). In previous
studies noted above, the fish originated mainly from retail
outlets. Further, in this study, extra samples that were
representative of each species from farmed and wild sources
were analyzed both with and without the skin to compare
the percent decrease in lipid levels due to removal of skin
and associated fat, and thereby facilitate interpretation of
the contaminant findings found in the present study with
those of Hites et al. (3) (samples included the skin).
The aim of our study was to examine salmon flesh quality
from a human health perspective. The aforementioned flesh
quality parameters were measured in a total of 201 wild and
farmed salmon and eight farmed salmon feed samples from
BC. MeHg was measured in a subset of salmon (n)22) and
all eight farmed salmon diet samples. The findings for salmon
species from all sites were used to evaluate safe consumption
rates of these species using available consumption and
threshold concentration guidelines from several regulatory
agencies for THg, PCDD/Fs, and PCBs on a wet weight, lipid-
normalized, and TEQ basis. Lipid-normalization of the data
allowed for relative comparisons to be made between species
for the lipophilic organohalogen contaminants. Further, since
absolute concentrations of fatty acids and organohalogen
contaminants are directly related to flesh lipid contents, lipid-
normalization of the data permitted examination of the
relative relationship between the beneficial EPA and DHA
content and the contaminant content to facilitate improved
understanding of the relative benefits and risks of consuming
farmed and wild salmon.
Experimental Section
Sample Collection, Storage, Preparation, and Analysis.
Three farmed salmon species (i.e., Atlantic, coho, and
chinook), and five wild salmon species (i.e., coho, chinook,
pink, chum, and sockeye) were sampled. Tables S1 and S2
and Figure S1 in the Supporting Information (SI) collectively
show the species and number of farmed and wild fish sampled
and their respective mean sizes and dates of sampling in
relation to their geographical origins. Complete details of of
the protocols used for sample collection and handling and
of preparation and analysis are given in the SI.
Results and Discussion
PCBs and PCDD/F Levels in the Flesh of Farmed and Wild
Salmon. In relation to farmed (F) salmon fillets, Atlantic
salmon from a location in the Broughton Archipelago
exhibited the highest mean values for PCBs, and Atlantic
salmon from a location at Quadra Island had the highest
mean values for PCDD/F concentrations (i.e., 38.3 (2.3 ng/g
and 3.07 (1.27 pg/g, respectively). By contrast, the lowest
mean concentrations noted for the preceding contaminants
were found in farmed coho salmon from Jervis and Sechelt
Inlet (i.e., 9.7 (2.0 ng/g and 0.73 (0.25 pg/g respectively)
(Figure 1).
With respect to wild (W) salmon fillets, chinook salmon
from Barkley Sound had the highest mean values for PCB
and PCDD/F concentrations (i.e., 13.7 (4.6 ng/g and 1.25
(0.33 pg/g, respectively). The lowest average concentrations
found for PCBs and PCDD/Fs were measured respectively,
in chum salmon from Johnstone St. (1.7 (1.0 ng/g) and
coho salmon from Prince Rupert (0.47 (0.21 pg/g).
The overall ranking for average wet weight PCB con-
taminant levels found within all of the preceding sources of
salmon was: F-Atlantic >W-chinook >F-chinook >F-coho
>W-sockeye>W-coho >W-pink >W-chum. The ranking
found for PCDD/F contaminant levels was: F-Atlantic >
F-chinook >F-coho >W-sockeye >W-chinook >W-chum
>W-coho >W-pink.
Although the wet weight PCB levels were generally found
to be lower in wild than in farmed salmon species, it should
be stressed that the highest mean concentrations of PCBs
found in this study, respectively, were 147-52 fold lower
than the level of concern for human consumption of fish as
established by Health Canada and the US-FDA, (i.e., 2000
ng/g) (24). Furthermore, the highest levels of PCBs found
among salmon species on Canada’s Pacific Coast were well
below levels of concern for PCBs in infant foods (Figure 1).
Specific comparisons of the flesh PCB concentrations
found for the different wild Pacific salmon species in this
study with those obtained in previous studies revealed
important similarities and differences. In this study, the fillets
from W chinook from Barkley Sound and the West coast of
Vancouver Island (WCVI) troll fishery had mean PCB
438 9ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007
concentrations of 13.7 (4.6 (n)10) and 11.1 (3.9 ng/g (n
)10) respectively, and the mean for all chinook samples
from the two sources was 12.3 (3.0 ng/g. The latter overall
mean for total PCBs in our wild chinook samples agreed
closely with that obtained by Hites et al. (3) for wild BC
chinook salmon, namely, 12.0 ng/g. By contrast, Jackson et
al. (25) reported a markedly higher mean value of 1941 (200
ng/g ((1 SE) for 78 PCB congeners in the flesh of wild Lake
Michigan chinook salmon. In the present study, wild coho
salmon from Prince Rupert, Nootka Sound, and the WCVI
Troll fishery had mean flesh PCB levels of 2.8 (1.7 (n)12),
4.2 (1.6 (n)10), and 6.2 (1.4 ng/g (n)13), respectively,
with a mean value of 4.4 (1.5 ng/g. Once again, our overall
mean value agreed well with that found by Hites et al. (3)
who noted a mean value of 5.0 ng/g in wild BC coho salmon
fillets. The comparable value found by Jackson et al. (25) for
wild Lake Michigan coho salmon was 1268 (77 ng/g. Fillets
from wild Johnstone St. sockeye, pink, and chum salmon in
this study, had mean PCB concentrations of 6.5 (2.3 (n)
11), 2.0 (1.2 (n)10), and 1.7 (1.0 ng/g (n)12), respectively,
and values reported by Hites et al. (3) for these wild BC species
were respectively, 7.0, 3.0, and 1.5 ng/g. Thus, all of our mean
PCB concentration values found for the aforementioned wild
BC salmon species were similar to those obtained by Hites
et al. (3). It is noteworthy that the mean PCB level observed
in wild chinook salmon from Lake Michigan was 157 times
greater than that of wild BC chinook salmon and the
concentration observed for wild coho salmon from Lake
Michigan was 288-fold more than that found for wild BC
coho salmon.
The highest mean flesh PCB concentrations found in this
study (i.e., 38 ng/g) for farmed Atlantic salmon compared
slightly lower than the highest levels reported by Hites et al.
(3) (i.e., 51 ng/g, full-congeners) and was similar to levels
reported by Jacobs et al. (26) (i.e., 32ng/g, 59 congeners).
The above-mentioned differences in flesh concentrations
of contaminants between the farmed and wild BC salmon
species likely reflect their respective differences in (1)
nutritional status (ration and feed or food composition,
especially the ratio of digestible protein to lipid), (2) fish size
and age and marine residency period at the time of sexual
maturation or harvesting, (3) innate ability to deposit lipid,
and consequently, the lipophilic organohalogen compounds,
into their flesh at different sizes (ages), (4) state of sexual
maturity at the time of capture and proportions of males to
females in the samples, or (5) a combination of these factors
(23,27).
Between-species variation in wild salmon contaminant
levels can be explained by several factors including differences
in the marine residency periods, which for chinook, coho,
sockeye, pink, and chum salmon may be as great as 4-5, 3,
3-4, 2, and 6 years, respectively (27,28). Fish size at the time
of maturity is greatest for chinook salmon and least for pink
salmon and this is reflected in the amount of lipid deposited
into their bodies and flesh since fish size is generally directly
related to lipid deposition in salmonids when food is
unrestricted (29,30). Moreover, wild chinook and coho
salmon, during the marine residency period, feed on small
fish, higher in the trophic chain, whereas amphipods and/or
euphuasiids are emphasized more in the diets of sockeye,
chum, and pink salmon (27). The generally longer lifespan
(increased size at maturity) of chinook salmon and their
predominately piscivorous dietary preferences, likely account
for their higher flesh concentrations of contaminants com-
pared to other salmon species. In relation to pink, sockeye,
and chum salmon, the relative importance of copepods,
amphipods, euphuasiids, and other invertebrate species in
their diet versus fish, likely had the most influence on their
flesh contaminant levels at the time of harvesting.
Interestingly, PCB concentrations in farmed salmon fillets
were less variable than in fillets from wild salmon both on
a wet weight and lipid-normalized basis (see below). For
instance, the mean values obtained for percent relative
standard deviations (%RSD) of PCBs on a wet weight basis
were 18.3% for farmed salmon and 78.6% for wild. The
increased variability in the PCB data for the wild salmon
species probably reflected greater variation in their dietary
input of contaminants from their prey items than occurred
for farmed salmon. If differences in feeding habits and
lifespan are considered, the five wild Pacific salmon species
can be ranked in order of their susceptibility to contaminant
bioaccumulation, with chinook and coho salmon being more
susceptible than sockeye, pink, and chum salmon. Both wet
weight and lipid-normalized PCB contaminant trends were
observed to follow this sequence, with some exceptions noted
between the wild coho and sockeye salmon that probably
FIGURE 1. Wet weight concentrations of PCBs (colored bars) and PCDD/Fs (black bars) in various sources of farmed and wild BC salmon
with 95% confidence intervals shown in relation to the USFDA tolerances for residues of PCBs in human and animal feed (
24
); PCBs in
farmed and wild fish are represented by red and blue, respectively). Numbers in parentheses reflect the number of individual fish analyzed
from each sampling location (see Tables S1 and S2 for additional information)
VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9439
can be ascribed to differences in the average masses of these
latter species obtained for this study, their state of sexual
maturity, and the proportion of females to males in the
samples. Some of these points will be considered in more
detail below.
The flesh concentrations of PCBs and PCDD/Fs in fish
sampled at the farm sites and their respective processing
plants (data not included) were not different. Therefore, our
findings suggest that the marine transportation procedures
used in BC had no influence on the preceding contaminant
levels in farmed salmon.
Relationship Between Lipid Content and PCBs Levels
in Farmed and Wild Salmon. The skin and its associated
subcutaneous fatty tissue, dark muscle, contribute a sub-
stantial amount to the total lipid content of salmon. Analyses
of 24 samples showed that the lipid content in the skinless
homogenates was 25-89% less than noted for the homo-
genates that included the skin (see Figure S2). Since PCBs,
PCDD/Fs, and many other organohalogen contaminants are
lipophilic, the inclusion of skin with its associated subcu-
taneous fatty tissue, dark muscle, and belly flaps in the
analysis may result in a significant overestimation of
contaminant concentrations in those tissues commonly
consumed. For all of the samples examined in this study, a
clear linear relationship was found between their lipid content
and PCB concentrations
In this regard, we noted that the PCB concentrations in
both wild and farmed salmon species were generally
positively correlated with their flesh lipid content and those
that had the most flesh lipid deposition, namely, farmed
Atlantic salmon, had the highest mean PCB concentrations
(ranged from 26-38 ng/g; see Figure S3). Alternatively, the
species that had the lowest lipid concentrations, namely,
wild chum salmon, had the lowest PCB concentrations.
Moreover, the PCB accumulation rate, as a function of
lipid content, was observed to be greater in wild salmon
than in farmed salmon. In fact, farmed chinook and coho
salmon followed a nearly horizontal trend in this regard, as
seen in Figure S3, such that increases in their flesh lipid
content due to chronic consumption of formulated diets rich
in lipid content was nearly independent of any attendant
rises in PCB concentration. This was not true, however, in
the farmed Atlantic salmon that ingested high energy (lipid-
rich) formulated diets during the culture period. Interestingly,
the flesh PCB concentrations in wild coho and chinook
salmon were found to increase in direct relation to their
flesh lipid content. This likely can be attributed to several
related factors. The increased fitness of wild stocks (feeding,
predation, migration, etc.) is taxing on lipid stores when
compared to the situation for pen-reared farmed salmon
stocks. This, coupled with the greater age of some wild stocks
(farmed stocks are harvested before sexual maturation),
results in a longer time period for PCBs to accumulate in
wild species even though natural prey may be lower in
contaminants than formulated salmon diets where fishmeal
and fish oil are the main sources of dietary protein and lipid.
These foregoing factors also help to explain why there was
greater variation of PCB concentrations found in the flesh
of the various sources of wild salmon relative to farmed
salmon, i.e., increased variations in ages, diets, and stages
of sexual development most probably accounted for the
variations seen in the lipid and contaminant contents in the
wild fish. Conversely, the farmed salmon are harvested before
energy from lipid is diverted toward reproductive develop-
ment and they are fed a more uniform diet.
The aforementioned observations can be used to design
future research projects aimed at reducing flesh PCB
concentrations in farmed salmon to concentrations present
in wild salmon. In addition to extensive replacement of
fishmeal protein and fish oil in formulated diets with plant
protein and lipid sources that are lower in organohalogen
content (31,32), another strategy would be to raise the dietary
ratio of digestible protein to lipid during the latter part of the
production period before the fish are marketed since this
strategy would reduce lipid deposition (27). Work is presently
underway in our laboratory and elsewhere in the world to
demonstrate the efficacy of these strategies for future
production of farmed Atlantic salmon. This could also be
done for farmed coho and chinook salmon, but as our findings
above indicate, there is little need with these species since
their flesh organohalogen concentrations were already found
to be similar to those of their wild counterparts.
When PCB data are presented on a lipid-normalized basis,
the PCB concentrations between the dissimilar farmed and
wild BC salmon sources were less obvious than on a wet
weight basis and the ranking was very different as well. The
trends observed and the significance of presenting the data
on a lipid-normalized basis is discussed in the corresponding
Supporting Information section.
PCDD/Fs and PCBs TEQ Values and Regulatory Levels
in Foods Destined for Human Consumption. The calculated
wet weight TEQ values for PCBs and PCDD/Fs in the various
sources of BC salmon in relation to the tolerable daily intakes
of PCDD/Fs as established by various national and inter-
national agencies are presented in Figure 2.
Among the farmed salmon, Atlantic salmon from the
Broughton Archipelago had the highest PCB and PCDD/F
TEQ concentrations. Coho salmon from Jervis inlet, by
contrast, had the lowest concentrations. With respect to the
wild salmon, chinook salmon from the WCVI troll fishery
exhibited the highest PCB concentration, and chinook salmon
from Barkley Sound exhibited the highest PCDD/Fs con-
centration. By contrast, the lowest PCB and PCDD/F
concentrations were found in chum and pink salmon from
Johnstone Strait. Overall, the flesh wet weight PCB TEQs for
the different salmon species and sources ranked as follows:
F-Atlantic >F-chinook >F-coho >W-chinook >W-sockeye
>W-coho >W-pink >W-chum. The ranking for the PCDD/F
TEQ values was identical except for the reverse in order of
W-chum and W-pink. Estimated mean values for farmed
salmon feed were 2.40 (0.62 and 1.34 (0.67pg TEQ/g (n
)7) for PCBs and PCDD/Fs, respectively.
The preceding findings indicate that the combined TEQ
values for PCDD/Fs and PCBs did not exceed 1.85 pg/g flesh
regardless of the origin of the salmon. Since TEQ limits for
PCDD/Fs are given in pg TEQ/kg bw/day, and not pg TEQ/g
of food, the data in Figure 2 can be better interpreted by
assuming a body weight of 70 kg for an individual consuming
a 100 g fish-flesh portion (39). This allowed a similar
comparison of the data as presented above but in relation
to the recommended thresholds for the daily consumption
of 100 g of fish by a 70 kg person. In this latter case, aside
from the draft value set by the US-EPA, all TEQ levels found
for both PCDD/Fs and PCBs were below all agency/country
guidelines on a kg body weight/day basis, except for one
farm that surpassed the US-ATSDR threshold for PCDD/Fs.
Hites et al. (3) found that the highest TEQs (for PCDD/Fs
and dioxin-like PCBs) in farmed Scottish Atlantic salmon
flesh were about 3 pg TEQ/g (full congener) whereas the
highest TEQs in the present study were nearly half of this
amount (i.e., 1.85 (0.27pg TEQ/g, n)11).
Mercury and Methylmercury Levels in the Flesh Of
Farmed and Wild Salmon. On average, the THg concentra-
tions measured in farmed and wild salmon were 0.021 µg/g
and 0.013-0.077 µg/g, respectively, Figure S5. The salmon
samples that had the highest THg levels (n)22) were also
analyzed for MeHg to estimate the proportion of MeHg to
THg. Results from the MeHg analysis indicated that, on
average, 97.1% ((7.4%) of the THg was in the form of MeHg
440 9ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007
(Figure S6) which is similar to previous studies at 93% for
salmon and sea trout flesh (40).
Based upon the 97.1% value, average MeHg levels
measured across farmed BC salmon species (0.021 µg/g
(0.004) and wild salmon species (0.038 µg/g (0.022) were
found to be well below the current Health Canada guideline
of 0.5 µg/g and the US-FDA action level of 1.0 µg/g and
therefore do not pose a human health concern. Relative to
other, high trophic level marine fish, such as halibut (0.25
µg/g (0.23), canned albacore tuna (0.35 µg/g (0.13), and
swordfish (0.98 µg/g (0.51) (18), our data show that farmed
and wild BC salmon contain very low Hg levels (Figure S5).
Additionally, our study revealed that average THg concen-
trations in farmed BC salmon were 73% lower than W-chinook
(0.077 µg/g (0.019), 49% below that of W-sockeye (0.041
µg/g (0.016), 45% lower than W-coho (0.038 µg/g (0.013)
and similar to W-chum and W-pink viz., 0.021 µg/g (0.006
and 0.013 µg/g (0.002, respectively. This distinct THg
concentration difference between farmed and wild salmon
has not been reported before in studies of similar scope (21,
41) perhaps due to the low number of samples and/or limited
species tested.
Since THg accumulates in fish over time and increases
trophically (17), long-lived pisciverous salmon (i.e., chinook)
would be expected to exhibit the highest THg levels for salmon
species while smaller, shorter-lived, low trophic level salmon
such as pink and chum would be anticipated to accumulate
the least THg. The data of our study (Figure S5) clearly support
this hypothesis.
Low THg levels measured in farmed salmon (mean, 0.021
µg/g (0.004) can likely be attributed to the relatively low
levels of THg measured in the farmed salmon diets (averaged
0.022 µg/g (0.005 (n)8) of which 75.1% on average was
determined to be MeHg; Figures S5 and S6). Low MeHg levels
in BC farmed salmon diets were likely due to their composi-
tions containing South American origin fishmeals and oils
that were based on the processing of small marine pelagic
fish (42) with low MeHg levels as reported by the US-FDA
for anchovy; THg )0.043 µg/g (n)40) (18), and plant and
/or animal protein and lipid sources as partial replacements
for fishmeal and oil with little or no MeHg content, e.g., canola
meal and oil, soybean meal, corn gluten meal, poultry by-
product meal, and poultry fat (42). The increased use of such
products to replace fish-based ingredients will further
decrease the already low MeHg levels in the feeds.
Benefits and Risks of Wild versus Farmed Salmon
Consumption. The reduced flesh lipid content in wild versus
farmed salmon not only depresses the concentrations of
contaminants in the former fish on a wet weight basis but
also their total n-3 fatty acid content/100 g portion, see Figure
S7. For instance, the averages for flesh lipid and n-3 fatty
acid content in wild chum salmon from Johnstone St. were,
respectively, 1.6% and 0.50 (0.08 g/100 g (n)12), whereas
those for farmed Atlantic salmon from a Quadra Island
processing plant were 13.9% and 3.43 (0.68 g/100 g (n)12).
Further, the high overall mean concentrations of selected
n-3 HUFAs (i.e., EPA and DHA) that we observed in the flesh
of farmed BC Atlantic salmon sources (Figure S8) are
potentially of greater significance because of the many human
health benefits that have been reported for these fatty acids
(44).
In this regard, strongest evidence has been seen so far for
their cardio-protective effects, but there is also mounting
evidence from recent studies related to their benefits
pertaining to prevention of some forms of cancer, enhance-
ment of cognitive and visual function, and attenuation of
various inflammatory indices and conditions (4-11). These
benefits have been observed most consistently when the daily
intake of EPA and/or DHA, particularly from oily species of
fish, that have been adequately prepared (e.g., baked or
broiled but not deep-fried), and/or from marine fish oil
capsules, has been sufficient to exert positive effects on the
aforementioned human health parameter(s), condition and/
or disease under study. For instance, Mozaffarian and Rimm
(6) recently presented evidence which shows that an intake
of 250 mg of EPA and DHA per day can decrease the risk of
CHD death by 36%.
FIGURE 2. TEQs for PCBs (gray bars) and PCDD/Fs (red, blue, and yellow bars for farmed, wild, and farmed feed, respectively) in the flesh
of the different sources of BC salmon (with 95% confidence intervals) in relation to the recommended tolerable daily intake for PCDD/Fs
as suggested by the U.S. Agency for Toxic Substances and Disease Registry (ATSDR) (
34
), European Commission Scientific Committee
on Food (
35
), UN Food and Agriculture Organization/World Health Organization Joint Expert Committee on Food Additives (UN/WHO JECFA)
(
36
), Ministerial Council on Dioxin Policy of Japan (
37
), Nordic (
38
), and Health Canada (
39
). The U.S. EPA guidelines are being reviewed,
but the draft value is at 0.001 pg TEQ/g (
20
). See Tables S1 and S2 for additional information.
VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9441
Furthermore, positive outcomes have been affected by
other factors such as the relative contributions of the n-6
fatty acids (i.e., linoleic acid and especially arachidonic acid),
linolenic acid, saturated fatty acids, monounsaturated fatty
acids, total fat, and trans fatty acids to the total daily energy
intake and by other major factors that are known to influence
human health such as smoking, obesity, extent of physical
exercise, fruit, fiber, and alcohol intake, etc. (12,13,23,41,
43). Foran et al. (44) performed a risk analysis for salmon
contaminated with dioxins and dioxin-like compounds and
concluded that the maximum benefit of n-3 fatty acids can
be gained by choosing fish that have low concentrations of
dioxins and dioxin-like compounds.
If one follows the recommended daily intake of 0.5 g of
n-3 HUFAs for cardio-protective effects in adults without
CHD as made by the American Heart Association (AHA) and
several other agencies listed in Figure S8 for adults without
documented CHD, then according to our results, only two
100 g portions of farmed Atlantic salmon would be required
per week. By contrast three to five servings per week would
be needed for all of the other sources of BC salmon except
for wild chum salmon where nine servings would be required.
The AHA recommends at least two servings of fish (especially
fatty fish) per week as well as 1.5-3 g/d of alpha-linolenic
acid from various sources such as flaxseed, canola, or soybean
oils, or walnuts for cardio-protective effects (45). Moreover,
the United Kingdom Foods Standards Agency (46) recom-
mends consumption of at least two 140 g portions of fish per
week including one of oily fish for promotion of good health.
Further, the European Food Safety Authority recently con-
cluded that one to two 130 g servings of preferably fatty fish
per week would decrease the risk of CVD and stroke, and
improve neurodevelopment and perinatal growth in infants.
All sources of BC salmon examined in this study had
combined TEQ values for PCDD/Fs and PCBs that did not
exceed 1.85 pg/g flesh. To reiterate, this level was well below
the recommended tolerable daily intake for PCDD/Fs as
provided by all of the world agencies shown in Figure 2,
except for the U.S. Agency for Toxic Substances and Disease
Registry, where the level of concern was set at 1 pg/g flesh.
Hence, because of these findings and the importance of
adequate intake of EPA and DHA from salmon and other
fatty fish for prevention of CVD, which in 2003 accounted for
almost 30% of global deaths (13), and possibly for other major
health benefits, we concur with the aforementioned fish
consumption recommendations of the AHA and UK Food
Standards Agency. In support of this viewpoint, Mozaffarian
and Rimm (6) presented convincing evidence from a
quantitative risk-benefit analysis that daily intake of 250 mg
of EPA and DHA from farmed or wild salmon over a 70-year
lifetime would result in 7125 fewer CHD deaths per 100 000
individuals while estimated lifetime cancer risk was six and
two cancer deaths per 100 000 lifetimes when ingesting
farmed and wild salmon, respectively.
Finally, brief mention should be made of the increased
concentrations of n-6 fatty acids noted in the flesh of farmed
salmon (mostly as linoleic acid) relative to the concentrations
observed in the wild salmon (Figure S8), and whether these
higher levels of n-6 fatty acids could potentially reduce the
beneficial effects of the EPA and DHA from these food sources.
In this regard, we consider this possibility unlikely since the
intake of n-6 fatty acids (mainly as linoleic acid) in a serving
of farmed salmon would contribute less than 13.5% of the
estimated total daily amount of linoleic acid consumed from
other food sources in the North American diet (47).
Thus, we conclude that all of the salmon sources examined
in this study had high nutritional value based upon their
projected low overall contributions to the body burdens of
individuals for PCDD/Fs and PCBs and Hg, and the fact that
they are excellent sources of EPA and DHA as well as linolenic
acid and monounsaturated fatty acids, especially in the case
of farmed salmon servings.
Acknowledgments
We gratefully acknowledge the financial assistance provided
by AquaNet, DFO-ACRDP, the BC Salmon Farmers Associa-
tion, the BC Science Council, and the former federal Office
of the Commissioner for Aquaculture Development. Also,
we appreciate the assistance during fish sample collection
of Jack Smith, Kevin Butterworth, Erin Friesen, Jill Sutton,
Dionne, Sakhrani, Nassim Ghani, Jennifer McDonald, Beth
Piercey, Mahmoud Rowshandeli, Nancy Richardson, Dr. Bob
Devlin, and numerous personnel at participating BC salmon
farms and fish processing plants as well as DFO Area
Managers. The assistance of all the chemists and support
staff of the DFO laboratories at IOS and FWI who processed
and analyzed all of the samples is much appreciated.
Supporting Information Available
More details about the sample locations, collection, and
analysis. This material is available free of charge via the
Internet at http://pubs.acs.org.
Literature Cited
(1) Statistics Canada, Aquaculture Statistics 2003.(available at
http://www.statcan.ca/english/freepub/23-222-XIE/23-222-
XIE2003000.pdf)
(2) FAO. World Review of Fisheries and Aquaculture.(available at
http://www.fao.org/documents/show_cdr.asp?url_file )//do-
crep/007/y5600e/y5600e04.htm.
(3) Hites, R. A.; Foran, J. A.; Carpenter, D. O.; Hamilton, M. C.;
Knuth, B. A.; Schwager, S. J. Global assessment of organic
contaminants in farmed salmon. Science 2004,303, 226-229.
(4) Connor, W. E. Importance of n-3 fatty acids in health and disease.
Am. J. Clin. Nutr. 2000, 71, 171S-175S.
(5) Horrocks, L. A.; Yeo, Y. K. Health benefits of Docosahexaenoic
acid (DHA). Pharmacol. Res. 1999,40, 211-225.
(6) Mozaffarian, D.; Rimm, E. B. Fish intake, contaminants, and
human health; Evaluating the risks and benefits. J. Am. Med.
Assoc. 2006,296, 1885-1899.
(7) Shahidi, F.; Miraliakbari, H. Omega-3 (n-3) fatty acids in health
and disease: Part 1-Cardiovascular disease and cancer. J. Med.
Food 2004,7, 387-401.
(8) Siddiqui, R. A.; Shaikh S. R.; Sech, L. A.; Yount, H. R.; Stillwell,
W.; Zaloga, G. P. Omega-3 fatty acids: Health benefits and
cellular mechanisms of action. Mini-Rev. Med. Chem. 2004,4,
859-871.
(9) Sidhu, K. S. Health benefits and potential risks related to
consumption of fish or fish oil. Regul. Toxicol. Pharmacol. 2003,
38, 336-344.
(10) Terry, P.; Wolk, A.; Vainio, H.; Weiderpass, E. Fatty fish
consumption lowers the risk of endometrial cancer: A nation-
wide case-control study in Sweden. Cancer Epidemiol., Biom-
arkers Prev. 2002,11, 143-145.
(11) Whelan, J.; McEntee, M. F. Dietary (n-6) PUFA and intestinal
tumorigenesis. J. Nutr. 2004,134, 3421S-3426S.
(12) World Health Organization. In Food and Health in Europe: A
New Basis for Action, Chapter 1. Diet and Disease, European
Series, no. 96; Robertson, A., Tirado, C., Lobstein, T., Jermini,
M., Knai, C., Jensen, J.J., Ferro-Luzzi, A., James, W. P. T., Eds.;
WHO, Regional Publications: Copenhagen, Denmark, 2004; pp
7-90.
(13) World Health Organization, Global Strategy on Diet, Physical
Activity and Health 2005 (available at http://www.who.int/
dietphysicalactivity/publications/facts/cvd/en/).
(14) Amercian Heart Association. Fish, Levels of Mercury and
Omega-3 Fatty Acids. 2006. (available at http://www.ameri-
canheart.org/presenter.jhtml?identifier )3013797).
(15) Mason, R. P., Fitzgerald, W. F., The distribution and bio-
geochemical cycling of mercury in the equatorial Pacific Ocean.
Deep-Sea Res., Part I. 1993,40(9), 1897-1924.
(16) US-EPA. What You Need to Know about Mercury in Fish and
Shellfish. 2004 EPA and FDA Advice. (available at http://
www.epa.gov/waterscience/fishadvice/advice.html).
(17) Burger, J., Gochfeld, M. Mercury in Fish Available in super-
markets in Illinois: Are there regional differences. Sci. Total
Environ. 2006,367, 1010-1016.
442 9ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 2, 2007
(18) Food and Drug Administration (U.S. FDA). Mercury Levels in
Commercial Fish and Shellfish. 2006 (Available at http://
www.cfsan.fda.gov/∼frf/sea-mehg.html).
(19) Health Canada, Health Based Tolerable Daily Intakes/Con-
centrations and Tumorigenic Doses/Concentrations for Priority
Substances (available at http://www.hc-sc.gc.ca/ewh-semt/
pubs/contaminants/existsub/hbct-jact/non_carcinogenic-non-
_cancerogenes_e.html
(20) United States Environmental Protection Agency. Exposure and
Human Health Reassessment of 2,3,7,8-Tetrachlorodibenzo-
p-Dioxin (TCDD) and Related Compounds, National Academy
Sciences (NAS) Review Draft 2005 (available at http://www.
epa.gov/ncea/pdfs/dioxin/nas-review/)
(21) Easton, M.; Luszniak, D.; Von der Geest, E. Preliminary
examination of contaminant loadings in farmed salmon, wild
salmon and commercial salmon feed. Chemosphere 2002, 1053-
1074.
(22) O’Neill, S. M.; West, J. E.; Hoeman, J. C. Spatial trends in the
concentration of polychlorinated biphenyls (PCBs) in chinook
(Oncorhynchus tshawytscha) and coho salmon (O. kisutch)in
Puget Sound and factors affecting PCB accumulation: Results
from the Puget Sound ambient monitoring program. Puget
Sound Res. 1998, 212-328.
(23) European Food Safety Authority. Opinion of the scientific panel
on contaminants in the food chain on a request from the
european parliament related to the safety assessment of wild
and farmed fish. EFSA J. 2005,236,1-118.
(24) United States Food and Drug Administration. Title 21 Chapter
1 Subchapter B Part 109- Unavoidable contaminants in food
for human consumption and food-packaging material. (available
at http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/search/
search.cfm?db )CFR&ID )109.30).
(25) Jackson, L. J.; Carpenter, S. R.; Manchester-Neesvic, J.; Stow, C.
A. PCB congeners in Lake Michigan coho (Oncorhynchus kisutch)
and chinook (Oncorhynchus tshawytscha) salmon. Environ. Sci.
Technol.,2001,35, 856-862.
(26) Jacobs, M. N.; Covaci, A.; Schepens, P. Investigation of selected
persistent organic pollutants in farmed Atlantic salmon (Salmo
salar), salmon aquaculture feed, and fish oil components of the
feed. Environ. Sci. Technol. 2002, 36, 2797-2805.
(27) Higgs, D. A.; Macdonald, J. S.; Levings, C. D.; Dosanjh, B. S.
Nutrition and Feeding Habits in Relation to Life History Stage.
In Physiological Ecology of Pacific Salmon; Groot, C., Margolis,
L., Clarke, W. C., Eds.; UBC Press: Vancouver, BC; 1995, pp
159-315.
(28) Groot, C.; Margolis, L. Pacific Salmon Life Histories; Groot, C.,
Margolis, L., Eds.; UBC Press: Vancouver, BC; 1991.
(29) Shearer, K. D. Factors affecting the proximate composition of
cultured fishes with emphasis on salmonids. Aquaculture 1994,
119,63-88.
(30) Trudel, M.; Tucker, S.; Morris, J. F. T.; Higgs, D. A.; Welch, D.
W. Indicators of energetic status in juvenile coho salmon and
chinook salmon. North Am. J. Fish. Manage. 2005,25, 374-390.
(31) Bell, J. G.; McGhee, F.; Dick, J. R.; Tocher, D. R. Dioxin and
dioxin-like polychlorinated biphenyls (PCBs) in Scottish farmed
salmon (Salmo salar): effects of replacement of dietary marine
fish oil with vegetable oils. Aquaculture 2004,243, 305-314.
(32) Lundebye, A. K.; Berntssen, M. H. G.; Lie, Ø.; Ritchie, G.; Isosaari,
P.; Kiviranta, H.; Vartiainen, T. Dietary uptake of dioxins (PCDD/
PCDFs) and dioxin-like PCBs in Atlantic salmon (Salmo salar).
Aquacult. Nutr. 2004,10, 199-207.
(33) Asplund, L., Athanasiadou, M., Sjo¨din, A., Bergman, Å., Bo¨ rjeson,
H. Organohalogen substances in muscle, egg and blood from
healthy baltic salmon (Salmo salar) that produced offspring
with the M74 Syndrome. Ambio 1999,28,67-76.
(34) ATSDR, Minimal Risk Levels (MRLs) for Hazardous Substances.
(available at http://www.atsdr.cdc.gov/mrls.html).
(35) European Commission, Opinion of the Scientific Committee
on Food on the Risk Assessment of Dioxins and Dioxin-Like
PCBs in Food (available at http://europa.eu.int/comm/food/
fs/sc/scf/out90_en.pdf).
(36) World Health Organization. In Food and Health in Europe: A
New Basis for Action, Chapter 4. Policies and Strategies, European
Series, no. 96; Robertson, A., Tirado, C., Lobstein, T., Jermini,
M., Knai, C., Jensen, J. J., Ferro-Luzzi, A., James, W. P. T., Eds.;
WHO, Regional Publications: Copenhagen,Denmark, 2004; pp.
221-307.
(37) Environment Agency of Japan, Ministerial Council on Dioxin
Policy, Informational Brochure on Dioxins. (available at http://
www.env.go.jp/en/chemi/dioxins/brochure2003.pdf).
(38) Johansson, N.; Hanberg, A. Report from a Nordic meeting on
the 1998 WHO consultation on assessment of the health risks
of dioxins; re-evaluation of the tolerable daily intake (TDI).
Organohalogen Compd. 2000,48, 252-255.
(39) Health Canada, Canada’s Food Guide to Healthy Eating,
(available at http://www.hc-sc.gc.ca/ewh-semt/pubs/contami-
nants/existsub/hbct-jact/index_e.html
(40) Westoo, G. Methylmercury as percentage of total mercury in
flesh and viscera of salmon and sea trout of various ages. Science
1973,181 (1099), 567-568.
(41) Foran, A. F., Hites, R. A., Carpenter D. O., Hamilton, M. C.
Mathews-Amos, A., Schwager, S. J. A survey of metals in tissues
of farmed Atlantic and wild Pacific salmon. Environ. Toxicol.
Chem. 2004,23, 2108-2110.
(42) Tacon, A. State of Information on Salmon Aquaculture Feed
and the Environment. 2005. (available at http://worldwildlife-
.org/cci/pubs/Feed_final_resaved2.pdf).
(43) Ruxton, C. H. S.; Reed, S. C.; Simpson, J. A.; Millington, K. J. The
health benefits of omega-3 polyunsaturated fatty acids: a review
of the evidence. J. Hum. Nutr. Diet. 2004,17, 449-459.
(44) Foran, J. A.; Carpenter, D. O.; Hamilton, M. C.; Knuth, B. A.;
Schwager, S. J. Risk-based consumption advice for farmed
Atlantic and wild Pacific salmon contaminated with dioxins
and dioxin-like compounds. Environ. Health Perspect. 2005,
113, 552-556.
(45) Kris-Etherton, P. M.; Harris, W. S.; Appel, L. J. Omega-3 fatty
acids and cardiovascular disease. New recommendations from
the American Heart Association. Arterioscler., Thromb., Vasc.
Biol. 2003,23, 151-152.
(46) UK Food Standards Agency, (available at http://www.eatwell.go-
v.uk/healthydiet/nutritionessentials/fishandshellfish/ac-
cessed 5/26/2005).
(47) Kris-Etherton, P. M.; Taylor, D. S.; Yu-Poth, S.; Huth, P.; Moriarty,
K.; Fishell, V.; Hargrove, R. L.; Zhao, G.; Etherton, T. D.
Polyunsaturated fatty acids in the food chain in the United
States. Am. J. Clin. Nutr.2000,71, 179-188.
Received for review February 20, 2006. Revised manuscript
received November 2, 2006. Accepted November 5, 2006.
ES060409+
VOL. 41, NO. 2, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9443
