This article was downloaded by: [Canadian Research Knowledge Network] On: 30 September 2008 Access details: Access Details: [subscription number 783016891] Publisher Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Food Additives & Contaminants: Part A Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713599661

Balancing the risks and the benefits of local consumption in Bermuda É. Dewailly ab; P. Rouja c; R. Dallaire a; D. Pereg a; T. Tucker d; J. Ward c; J. P. Weber b; J. S. Maguire a; P. Julien a a Public Health Research Unit, Laval University Research Centre-CHUL-CHUQ, Québec, Canada b Institut National de Santé Publique du Québec, Québec, Canada c Department of Conservation Services, Government of Bermuda, Bermuda d Bermuda Underwater Exploration Institute, Bermuda

First Published on: 25 September 2008

To cite this Article Dewailly, É., Rouja, P., Dallaire, R., Pereg, D., Tucker, T., Ward, J., Weber, J. P., Maguire, J. S. and Julien, P.(2008)'Balancing the risks and the benefits of local fish consumption in Bermuda',Food Additives & Contaminants: Part A, To link to this Article: DOI: 10.1080/02652030802175285 URL: http://dx.doi.org/10.1080/02652030802175285

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf This article may be used for research, teaching and private study purposes. Any substantial or systematic , re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Food Additives and Contaminants 2008, 1–11, iFirst

Balancing the risks and the benefits of local fish consumption in Bermuda E´ . Dewaillyad*, P. Roujab, R. Dallairea, D. Perega, T. Tuckerc, J. Wardb, J.P. Weberd, J.S. Maguirea and P. Juliena aPublic Health Research Unit, Laval University Research Centre-CHUL-CHUQ, Que´bec, Canada; bDepartment of Conservation Services, Government of Bermuda, Bermuda; cBermuda Underwater Exploration Institute, Bermuda; dInstitut National de Sante´ Publique du Que´bec, Que´bec, Canada (Received 14 February 2008; final version received 1 May 2008)

Fish consumption today is widely recognized as highly beneficial since it constitutes a good source of several essential nutrients, such as selenium and polyunsaturated fatty acids (n-3 PUFA). However, fish can also contain contaminants such as mercury, which make the consumer, especially pregnant women, confused about the risk–benefit balance associated with fish consumption. This is particularly true for tropical fish for which little information is available. We have previously reported that some Bermudian neonates had elevated mercury in their umbilical blood compared with international guidelines. The objective of this study was to give precise and balanced information on the content of mercury, selenium and PUFA in the most consumed fish species in Bermuda. In 2003 and 2006, a total of 307 fish were collected from 43 fish species and 351 samples were analysed (305 flesh samples, 44 liver samples, one roe and one fat sample) by inductively coupled plasma-mass spectrometry (ICP-MS) (metals) and high-resolution gas chromatography (HRGC) (fatty acids). Results show 1 that mercury varies among species from 0.03 to 3.3 mggÀ and that it is possible for at-risk groups such as pregnant women to make informed choices concerning fish consumption, e.g. maximizing fish species rich in nutrients and low in mercury. Keywords: fish consumption; tropical; mercury; selenium; omega-3

Introduction Despite the fact that n-3 fatty acids can be found in Fish is one of the last remaining ‘wild foods’ in the other foods, fish remains the best and richest source diets of Bermuda residents and represents not only a of these nutrients and hence is an important part of a rich source of several unique dietary nutrients, but also healthy diet. Coastal populations have a generally high forms an important part of Bermuda’s cultural identity intake of long-chain PUFAs, the most important as an island nation. Highlighting the benefits and compounds being eicosapentanoic acid (EPA) and minimizing the risks from seafood consumption is docosahexanoic acid (DHA). During pregnancy, fish an important part of maintaining this resource’s consumption provides DHA to the mother and the important nutritional and economic contribution to foetus, which is critical in brain and retina develop- Bermuda’s residents. ment (Lauritzen et al. 2004; Hibbeln et al. 2007). Fish consumption today is widely recognized as In addition, fish and seafood are a very good natural Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 highly beneficial since it constitutes a good source source of selenium (Svensson et al. 1992; Leung et al. of several essential nutrients, such as selenium and 2007). Concentrations of this element in foods depend polyunsaturated fatty acids (n-3 PUFA). Several on the local geological composition. It has been health organizations recommend eating fish twice a proposed that selenium may exert an antagonistic week for the general population (Kris-Etherton et al. effect on mercury (Hg) toxicity (World Health 2002; Harris 2004). In fact, fish consumption is Organization (WHO) 1990) and have a mediating recognized as beneficial for brain development (Uauy effect on certain cancers, specifically prostate cancer. et al. 2001) in the foetus and newborn and has been On the other hand, fish may also contain varying demonstrated to be protective against cardiovascular amount of contaminants such as Hg, most of which diseases (He et al. 2004; Wang et al. 2006), mental occurs in the form of methylmercury (MeHg) (over disorders (Logan 2003; Sinclair et al. 2007), and 90%). This phenomenon occurs because fish feed on various inflammatory conditions such as bowel dis- aquatic organisms that contain MeHg originating from ease, asthma, and arthritis (Ruxton et al. 2004). biomethylation of inorganic Hg by microorganisms.

*Corresponding author. Email: [email protected]

ISSN 0265–203X print/ISSN 1464–5122 online ß 2008 Taylor & Francis DOI: 10.1080/02652030802175285 http://www.informaworld.com 2 E´. Dewailly et al.

Hg has always been present in all the world’s Hg measured was in the form of MeHg, pointing as a result of river run off, undersea volcanoes, air toward seafood as the main source of Hg exposure deposition, etc. and has always built up in fish at the during pregnancy in Bermuda. In contrast, persistent top of the food chain. However, in the last century the organic pollutants (POPs) such as polychlorinated industrialization of many countries has resulted in a biphenyls (PCBs) and chlorinated pesticides along with net increase in the load of Hg in the environment, lead were also measured in cord blood, but concen- which consequently is manifest in the world’s fish and trations were very low (Dewailly and Pereg 2004). in those people who consume them. This is not a local For that reason, only Hg was considered of public Bermuda problem but a global one. Coastal people health concern. living in some of the world’s most remote places are Fish consumption was also measured and the in fact finding that their fish are as contaminated as association of Hg levels with fish consumption was fish in industrialized coastal areas. tested using Pearson’s correlation coefficient. Fish The amount of MeHg in fish usually correlates with consumption was estimated by the number of meals several factors including the weight and age of the fish, per month that each patient declared eating on a its trophic level and the species. regular basis. Species that were most often consumed Numerous studies have reported that MeHg include wahoo, snapper, tuna and salmon; the present in predator represents a potential consumption of other species seemed anecdotal. health threat for the developing foetus. As this seafood While wahoo and snapper are clearly of local origin, borne contaminant affects the development of the tuna is most often imported. The consumption of local cerebral architecture by perturbing neural cell division fish was analysed separately from that of total fish and migration, the developing nervous system is consumption. Local fish consumption was estimated recognized as the main target (Castoldi et al. 2001; by pooling the number of meals/month of wahoo and Counter and Buchanan 2004). Several prospective snapper that each woman declared eating. Separate cohort studies have reported effects of prenatal analyses were also conducted on each fish specie exposure to MeHg on different domains of cognition separately (salmon, wahoo, snapper, tuna), despite (attention, memory, visuo-spatial performances and the low statistical power of some of these analyses. language) and gross and fine motor development, some A significant association was seen between Hg of which persist well beyond the first years of life concentrations and the consumption of local fish (Grandjean et al. 1997; Steuerwald et al. 2000; Debes (Spearman’s r 0.584, p 0.001, n 40), mainly et al. 2006). In contrast, a large epidemiological study 5 wahoo (r 0.502,¼ p 0.001, n 40)¼ and snapper from the Seychelles Islands did not find any effects on (r 0.499, ¼p 0.001, ¼n 40). The¼ consumption of infant neurodevelopment from chronic MeHg in utero other¼ fish species¼ was not¼ related to Hg content of cord exposure (Myers et al. 2003; Davidson et al. 2006). Several explanations have been proposed to explain blood. this discrepancy such as chronic versus episodic The globalization of contaminants and the strong exposure, age at neurobehavioral assessment, differ- media attention regarding the presence of these ential confounders adjustment, etc. (National Research contaminants in fish in all the world’s oceans has Council 2000). made some consumers wary of fish consumption. In 2003, as part of the Atlantis Mission, we This has been counterbalanced by Western scientific

Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 conducted a study titled ‘Prenatal Exposure of discoveries of the last 20 years highlighting the unique the Bermudian Population to Environmental health benefits to be gained from the consumption of Contaminants’ that aimed to provide baseline data on fish. The informed consumer is today extremely prenatal exposure to a number of environmental sensitive in their consumption choice. contaminants, including Hg. The Atlantis Mobile Given the fact that wild fish consumption is an Laboratory was created as a novel, unique infrastruc- important part of the lifestyle of many Bermuda ture to meet emerging environmental and health residents and because this habit presents not only challenges in remote settings. It is a completely health benefits, but also risks for pregnant women and autonomous facility specifically designed for field- their unborn children, Atlantis researchers and the based environmental and health assessments. It com- Ministry of the Environment and the Ministry of prises six modules built in standard size containers. Health undertook the task of substantiating the Hg, Three modules are laboratories and three are dedicated selenium and fatty acid profiles of fish in Bermuda. to support the facility. Such an assessment is necessary in to provide a Mean Hg concentration measured in umbilical cord baseline for guidelines regarding the number of meals blood from 42 healthy newborns was higher than and the type of fish to be preferentially consumed in 1 expected, with a mean concentration of 8.3 mglÀ , order to maintain the benefits associated with a fish- 1 ranging from 1 to 32.1 mglÀ (Dewailly and Pereg based diet while avoiding excessive exposure to MeHg, 2004). Furthermore, it was estimated that 85% of total particularly during pregnancy. Food Additives and Contaminants 3

Materials and methods within acceptable limits for both CRMs. The laboratory Sampling participates successfully in several external quality assessment schemes for heavy metals, including sele- Fish samples were collected in fall 2003 and summer nium and Hg, in biological fluids and tissues. 2006 from species generally caught at the top of the The total fatty acid composition of fish fillets was food chain (e.g., predatory fish, pelagic fish) and measured at Que´ bec Lipid Research Centre (CHUQ) those that people might commonly consume. Fish were after total lipid extraction using a chloroform/metha- mostly caught by fishermen and given to the research- nol mixture (2 : 1 v/v) spiked with phosphatidylcholine ers, who sampled them onboard the boat or at the laboratory. Some samples of local fresh fish were C:15 used as internal standard according to a modified sampled at restaurants or bought at the store. A total Folch method (Shaikh and Downar 1981). Fatty acids of 307 fish were collected from 43 fish species and 351 were then methylated (Lepage and Roy 1986) and fatty samples were analysed (305 flesh samples, 44 liver acid profiles obtained by capillary gas–liquid chroma- samples, one roe and one fat sample). Fish samples were tography using a HP5890 gas chromatograph (Hewlett Packard, Toronto, Canada) equipped with a HP-88 packaged in plastic bags and stored at 20C until laboratory analyses. À capillary column (100 m 0.25 mm i.d. 0.20 mm film thickness) coupled with a flame ionization detector. Helium was used as carrier gas (split ratio 1 : 80). ¼ Laboratory analyses Fatty acids were identified according to their retention time, using a standard mixture of 43 methylated fatty Fish fillets were analysed for their Hg, selenium and acids (Supelco Inc., Bellefonte, PA, USA). Results fatty acid concentrations, whereas livers were analysed were expressed as per cent of total fatty acids. The for their Hg and selenium contents only. Determination detection limit was 0.1% and the intra-assay coefficient of Hg concentrations was performed in the Atlantis of variation ranged from 0.87% for total omega-3 fatty mobile laboratory and at the reference laboratory of the acids to 6.8% for DHA. This standard method is Que´ bec Toxicology Center (INSPQ) by cold vapour currently used at the Que´ bec Lipid Research Centre. atomic absorption mass spectrometry. Samples were Data are presented as percent of total n-3 PUFA or as digested in nitric acid in pressurized vials. Digests were total n-3 PUFA concentrations (mg) per g of fresh analysed by cold vapour atomic absorption mass spectrometry in a Pharmacia mercury monitor, model flesh. Analyses of PUFA content were replicated for 16 100. In this instrument, samples are vaporized through a fishes of different species in order to assess the cell that is crossed by a light beam and Hg concentration potential variability between samples. is calculated by comparing the absorbance at 254 nm with that obtained with a Hg standard curve. Results Statistical analysis are expressed on a wet weight basis and the limit of 1 For computational purposes, samples with undetected detection using this method is 0.05 mggÀ wet weight. levels of Hg or selenium were attributed a value equal Selenium concentrations of local Bermuda fish to half the detection limit, which was 0.025 and were determined in flesh and liver samples at the 1 0.005 mggÀ , respectively. Arithmetic means, standard INSPQ by inductively coupled plasma mass spectro- deviations and range of Hg, selenium and fatty acid metry (ICP-MS). Samples were diluted in ammonium concentrations in flesh and/or liver were computed for Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 hydroxide and elements were brought to their elemen- species with more than one sample available. n-3 tary form by passing through argon plasma before PUFA concentrations (mg) were reported for a portion being identified by mass spectrometry. All samples of 220 g of fish. The 220 g mean portion comes from were analysed on a Perkin Elmer Sciex Elan 6000 ICP- the US Environmental Protection Agency (USEPA) MS (DRC II for Hg) instrument. Detection limits for 1 (2000) using a postal survey (with pictures of fish selenium were 0.01 ggÀ . m portion size) of 2600 adult consumers. For metals, sample analyses were done using Correlations between fish Hg content and fish methods developed and validated by the Quebec weight were carried out. For these analyses data were Toxicology Center (CTQ). The CTQ laboratory is available only for some of the samples and were there- accredited under ISO 17025 by the Standard Council of fore carried out by species. Correlation analyses were Canada (SCC). Appropriate analyses of method blanks, carried out using Pearson’s correlation coefficients. spiked samples and replicates were used to validate the data. To ensure accuracy, certified reference materials (CRMs) DORM-2 (Dogfish Muscle Certified Results Reference Material for Trace Metals) and DOLT-2 (Dogfish Liver Certified Reference Material for Trace Mercury (Hg) Metals) from the Canadian National Research Council Table 1 shows the Hg concentrations found in all fish were analysed with each series of samples. Results were flesh samples analysed (wet weight). These mean Hg 4 E´. Dewailly et al.

1 Table 1. Mercury and selenium concentrations (mggÀ ) in the flesh of local fish and species.

Mercury Selenium

Common name Scientific name n Mean SD Range n Mean SD Range Æ Æ Amberjack Seriola dumerili 8 0.55 0.46 0.20–1.55 7 1.33 2.24 0.42–6.40 Anchovy Sardinella anchovia 1 0.07Æ 1 1.50Æ Bar jack/green jack ruber 1 1.0 1 0.41 Barber (creole fish) Paranthias furcifer 12 0.035 0.028 0.025–0.12 11 0.68 0.12 0.47–0.86 Barracuda Sphyraena barracuda 7 0.73 Æ 0.38 0.18–1.40 4 0.37 Æ 0.12 0.29–0.54 Bermuda chub Kyphosus sectatrix 15 0.038 Æ 0.039 0.025–0.17 14 0.24 Æ 0.20 0.11–0.68 Bermuda dolphinfish Coryphaena hipparus 1 0.05Æ Æ Bermuda mackerel Euthynnus alletteratus 15 0.68 0.69 0.17–2.30 13 0.50 0.22 0.33–0.50 Bermuda spiny lobster Panulirus argus 4 0.21 Æ 0.08 0.15–0.30 Æ Black grouper Mycteroperca bonaci 8 0.87 Æ 0.46 0.35–1.55 3 0.45 0.08 0.37–0.52 Blackfin tuna Thunnus atlanticus 2 0.20 Æ 0.00 0.20–0.20 Æ Blue Makaira nigricans 3 3.10 Æ 0.17 3.00–3.30 3 2.10 0.35 1.90–2.50 Caranx fusus 10 0.98 Æ 0.33 0.39–1.30 10 0.40 Æ 0.07 0.28–0.54 Blue-striped grunt Haemulon sciurus 14 0.33 Æ 0.09 0.20–0.51 14 0.35 Æ 0.05 0.23–0.47 Bone fish Albula vulpes 1 0.50Æ 1 0.68Æ Bonita (alamaco jack) Seriola rivoliana 17 0.170 0.139 0.025–0.53 14 0.48 0.09 0.32–0.66 Caesar grunt Haemulon carbonarum 1 0.09Æ 1 0.46Æ Coney Cephalopholis fulva 29 0.21 0.10 0.10–0.52 17 0.55 0.17 0.27–0.78 Deep water red snapper Etelis oculatus 1 0.52Æ Æ Dusky (Galapagos) Carcharhinus galapagensis 2 0.56 0.24 0.39–0.73 Flounder sp. Bothus sp. 1 0.06 1 0.42Æ Flying fish Cypselurus heterurus 6 0.035 0.016 0.025–0.057 6 0.31 0.04 0.26–0.38 Frys Anchoa choerostoma 2 0.12 Æ 0.01 0.05–0.19 2 0.54 Æ 0.09 0.47–0.60 Grasby Epinephelus cruentatus 1 0.16Æ 1 0.62Æ Gray snapper Lutjanus griseus 13 0.28 0.17 0.09–0.68 13 0.33 0.05 0.25–0.42 Gummy shark (dogfish) Mustelus antarcticus 1 1.4Æ 1 0.47Æ Gwelly Pseudocaranx dentex 1 0.37 1 0.58 Hogfish Lachnolaimus maximus 8 0.067 0.044 0.025–0.15 5 0.32 0.09 0.21–0.42 Hybrid coney-barber Cephalopholis fulva 1 0.12Æ 1 0.54Æ Longfin jack/bar jack Caranx ruber 1 0.48 1 0.28 robin Decapterus macarellus 17 0.23 0.09 0.07–0.35 13 0.69 0.19 0.41–0.95 Rainbow runner Elagatis bipinnulata 2 0.16 Æ 0.06 0.11–0.20 2 0.43 Æ 0.00 0.43–0.43 Red hind Epinephelus guttatus 29 0.26 Æ 0.08 0.14–0.41 Æ Red snapper Lutjanus sp. 2 0.10 Æ 0.01 0.09–0.10 2 0.46 0.00 0.46–0.46 Sixgill shark Hexanchus griseus 1 0.92Æ Æ Swordfish Xiphias gladius 1 3.31 Tapioca fish Ruvettus nuvettus 1 0.15 Tuna Thunnus sp. 5 0.35 0.21 0.21–0.70 1 0.63 Turbot Balistes capriscus 8 0.072 Æ 0.024 0.025–0.10 7 0.36 0.07 0.25–0.45 Wahoo Acanthocybium solandri 14 0.34 Æ 0.27 0.06–1.00 7 0.86 Æ 0.36 0.38–1.40 White water (lane) snapper Lutjanus synagris 16 0.43 Æ 0.12 0.17–0.60 16 0.39 Æ 0.03 0.35–0.45 Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 Yellowfin tuna Thunnus albacares 17 0.37 Æ 0.18 0.18–1.10 2 0.64 Æ 0.08 0.58–0.69 Yellowtail snapper Ocyurus chrysurus 6 0.35 Æ 0.23 0.05–0.65 1 0.46Æ Æ

1 concentrations can be compared with action limits above the 0.5 mggÀ limit are also indicated in bold. determined by various international agencies (e.g., Gummy and sixgill are the two species with the 1 0.5–1.0 mggÀ ) for the importation of commercial fish. highest Hg concentrations in liver, followed by gray Swordfish and blue marlin are the two species with the snapper and coney. Hg concentrations in the flesh of highest Hg content, followed by bar jack/green jack, fish were highly correlated with Hg liver levels blue runner and black grouper. For some species, the (r 0.60). The concentration of Hg in liver was mean Hg concentration is below that threshold, but the always¼ higher than in flesh except for black grouper, range shows that some individuals may have values bonita and dusky shark. that are above. This is particularly the case for wahoo Figure 1 tabulates local fish species according to and yellowfin tuna. The lowest mean concentrations of the increasing concentrations of Hg in their flesh. Hg were found in flying fish, barber and Bermuda In order to help protect the developing foetus against chub. the effects of MeHg, the number of fish portions Hg concentrations in the liver, fat and roe of some (220 g) that can be consumed per week by a pregnant species are found in Table 2. Mean Hg concentrations women have been calculated. These calculations are Food Additives and Contaminants 5

1 Table 2. Mercury and selenium concentrations (mggÀ ) in the liver, fat and roe of some local fish species.

Mercury Selenium

Common name Scientific name Tissue n Mean SD Range n Mean SD Range Æ Æ Barber (Creole fish) Paranthias furcifer Liver 7 0.21 0.08 0.09–0.33 7 8.84 2.57 6.20–14.00 Bermuda chub Kyphosus sectator Liver 8 0.21 Æ 0.08 0.07–0.32 8 2.75 Æ 2.72 1.20–8.40 Black grouper Mycteroperca bonaci Liver 1 0.025Æ 1 Æ Bonita (almaco jack) Seriola rivoliana Liver 3 0.04 0.03 0.03–0.07 2 4.45 0.35 4.20–4.70 Coney Cephalopholis fulva Liver 13 1.07 Æ 0.31 0.66–1.60 13 3.52 Æ 0.85 2.40–4.90 Dusky shark (Galapagos) Carcharhinus galapagensis Liver 1Æ 1 5.20Æ Galapagos shark Carcharhinus galapagensis Liver 1 0.65 1 Gray snapper Lutjanus griseus Liver 2 1.14 1.22 0.28–2.00 2 2.25 1.06 1.50–3.00 Gummy shark (dogfish) Mustelus antarcticus Liver 1 2.70Æ 1 4.30Æ Gwelly Pseudocaranx dentex Liver 1 0.98 1 0.98 Hogfish Lachnolaimus maximus Liver 2 0.23 0.11 0.15–0.30 2 Red hind Epinephelus guttatus Liver 1 0.63Æ 1 Sixgill shark Hexanchus griseus Liver 1 19.10 1 Hogfish Lachnolaimus maximus Roe 1 0.025 1 Coney Cephalopholis fulva Fat 1 0.25 1

Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 Figure 1. Mercury concentrations in fish and recommendations for pregnant women.

based on the provisional tolerable weekly intake pregnant women, in order to help ensure that the 1 (PTWI) for MeHg of 1.6 mg kgÀ bodyweight recom- PTWI will not be reached: mended by the Food and Agricultural Organization/ Without restriction: species that can be World Health Organization (FAO/WHO) (FAO/ . consumed every day. WHO Expert Committee on Food Additives 2006), Limited or little restriction: species that can these recommendations for fish consumption assume . be consumed between once per day and once that no others species of fish are consumed during the per week. same period. They apply to pregnant women and Strong or severe restriction: species that can be infants and children aged up to about 17 years. For . consumed once per week to once per month. infants, the number of fish portion is similar if one Do not consume: species that should not be considers that the portion size is proportional to the . consumed during pregnancy and by children. bodyweight. Based on the Hg concentrations measured in the Based on these categories, only four species for flesh of local fish species, we have come up with four which Hg content was measured (barber, flying human consumption categories for fish, each category fish, Bermuda chub and Bermuda dolphin fish (Mahi representing a suggested consumption frequency for Mahi)) can be consumed every day without risk of 6 E´. Dewailly et al.

reaching the PTWI. Most species investigated (25 out Correlations between fish weight and Hg content of 43) deserve very little restriction and it is recom- were determined in species for which sample size was mended that not more than one portion of these fish be sufficient (n 6) and in which the Hg concentration of  1 eaten per day. These include wahoo and yellowfin tuna. most samples was above the detection limit of 0.5mggÀ . However, twelve species should be restricted in their The relationship between fish weight and Hg concen- consumption to one portion a week or less. This tration in the flesh of four species are presented in category includes all shark species investigated in this Figure 2. Morphological features such as weight were a study. Finally, the consumption of blue marlin and good predictor of flesh Hg content in bonita (r 0.89, ¼ swordfish should be avoided by women during their p 5 0.0001), ocean robin (r 0.90, p 5 0.0001) and ¼ pregnancy as only one portion of these species will turbot (r 0.85, p 5 0.016). A moderate association exceed the tolerable dose for at least one month. between fish¼ weight and Hg levels was observed for blue It is important to note again that these guidelines runner (r 0.69, p 0.026), wahoo (r 0.67, p 0.212) presume that no other fish (local or imported) is eaten and white¼ water snapper¼ (r 0.68, p ¼0.004). Whereas¼ in the same period. Therefore, one meal per week of fish an absence of association was¼ observed¼ for most species, from category ‘Strong or severe restriction’ would likely blue-striped grunt (r 0.24, p 0.400), coney (r 0.08, ¼ ¼ ¼ keep a pregnant woman or a child below the recom- p 5 0.752), gray snapper (r 0.22, p 0.472), Red hind mended PTWI. However, if any other fish meal, even (r 0.15, p 0.469) and yellowfin¼ tuna.¼ from the ‘Limited or little restriction’ category, is ¼ ¼ consumed it would possibly push her above that limit. Pregnant and nursing mothers as well as children need Selenium to assess their seafood consumption in a holistic manner when choosing which fish to consume, considering all Table 1 shows the selenium concentrations for 36 fish the fish and seafood they eat both local and foreign. species in all available flesh samples. Amberjack, It is therefore recommended that pregnant women and anchovy and blue marlin had the highest selenium children who are regular fish consumers consume only levels whereas Bermuda chub. Longfin jack and flying those species in the ‘Without restriction’ or ‘Limited or fish had the lowest concentrations. Correlations little restriction’ categories, and if so inclined, selecting between Hg and selenium content for all species was those that are higher in omega-3 fatty acids. moderate (r 0.29). ¼ Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008

Figure 2. Relation between fish species’ weight and mercury content. Food Additives and Contaminants 7

Selenium concentrations in the livers of eight local in the flesh of local species. The precise fatty acid species are reported in Table 2. The highest concentra- profile of all fish species is presented in the annex to tions of selenium were found in barbers and bonitas this report. Omega-3 FAs represent between 20 and the lowest in gwelly liver. Selenium concentrations and 49% of total fatty acids. n-3 PUFA concentration in fish flesh were highly correlated with selenium (mg/220 g) was moderately but significantly correlated liver levels (r 0.60, p 5 0.0001). The concentration with Hg (r 0.33) and selenium (r 0.24) concentra- of selenium in¼ liver was always higher than levels tions in flesh¼ of the species studied.¼ measured in flesh. Percentages of variation in the percentage of total Figure 3 shows Hg and selenium concentrations in n-3 PUFA ranged between 0.4% to 13% whereas those the flesh of 34 fish species in Bermuda. A small for total n-3 PUFA (in mg/220 g) varied from less than proportion of species had Hg concentrations higher 1–207%. The largest variation in total PUFA was than selenium levels. This is particularly the case for found in a sample of dusky shark. Variations in n-3 shark species, bar jack, blue runner, barracuda and PUFA measurements may be attributable to the black grouper. However, most species (25 out of 34) presence of residual pieces of skin and bones or to showed higher selenium than Hg concentrations. more or less fat in the analysed samples. Among these, Anchovy and barber have 21 and 19 Figure 4 shows Hg and n-3 PUFA flesh concentra- times more selenium than Hg in their flesh, respectively. tions in 35 fish species from Bermuda. While the flesh of some species like marlin and shark contain high concentrations of Hg with low n-3 PUFA concentra- Omega (n-3) PUFA tions, others like blue runner, longfin jack and Bermuda The percentage of n-3 PUFA related to total fatty acid mackerel show relatively high to moderate Hg content content and the concentration of n-3 PUFA (mg/220 g with elevated n-3 PUFA concentrations. Wahoo flesh of fish) in flesh of 34 local fish species is presented in contains the highest n-3 PUFA concentration, while Hg Table 3. Flying fish, amberjack and the hybrid of levels appeared to be relatively moderate. coney-barber had the highest percentage of n-3 PUFA related to total lipid or fat content whereas the Caesar grunt had the lowest percentage. Ocean robins had the Discussion greatest variation in the percentage of n-3 PUFA This study determined the Hg content as well as between fish samples. The highest concentration of selenium and n-3 PUFA concentrations in a number of n-3 PUFA was measured in blue runner, wahoo and Bermudian fish species to evaluate the risk and benefit Bermuda mackerel. These species were also those that of their consumption during pregnancy in order to had the largest range of n-3 PUFA concentrations (mg/ protect the developing foetus and during childhood. 220 g of fish) between individuals of the same species. As found in other studies (Storelli and Hogfish, red hind and grasby had the lowest n-3 PUFA Marcotrigiano 2000; Mahaffey et al. 2004), high-end concentrations in their flesh. There is a very strong predatory fish such as swordfish and marlin have very correlation (r 0.98), shown in Figure 5, between high Hg content in their flesh. Shark livers also appear concentrations¼ of n-3 PUFA and total lipids (fatness) to retain Hg. Therefore, as recommended by other Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008

Figure 3. Mercury and selenium content in flesh of local fish species. 8 E´. Dewailly et al.

Table 3. Percentage of n-3 polyunsaturated fatty acid (PUFA) concentrations in relation to total fatty acid contents and n-3 PUFA concentrations (mg/220 g of flesh) in the flesh of local fish species.

n-3 PUFA n-3 PUFA mg/220 g

Common name Scientific name n Mean SD (%) Range (%) Mean SD Range Æ Æ Amberjack Seriola dumerili 6 46.2 5.5 38.5–50.5 467 73 373–590 Anchovy Sardinella anchovia 1 43.5Æ 1159Æ Bar jack/green jack Caranx Ruber 1 32.5 1422 Barber (Creole fish) Paranthias furcifer 11 29.4 2.9 24.9–33.8 681 192 454–1003 Barracuda Sphyraena barracuda 4 34.0 Æ 1.4 32.8–35.8 335 Æ 108 176–415 Bermuda chub Kyphosus sectator 13 25.0 Æ 2.9 18.4–29.0 403 Æ 347 162–1283 Bermuda mackerel Euthynnus alletteratus 12 31.5 Æ 5.0 20.9–31.5 2803 Æ 3008 354–9385 Black grouper Mycteroperca bonaci 3 24.6 Æ 11.7 18.1–45.4 355 Æ 218 150–720 Blue marlin Makaira nigricans 3 34.9 Æ 4.1 29.4–38.1 419 Æ 481 142–1138 Blue Runner Caranx fusus 10 28.0 Æ 5.2 21.6–40.6 3801 Æ 3043 492–12 287 Blue-striped grunt Haemulon sciurus 14 27.0 Æ 5.6 15.9–35.5 337 Æ 113 155–593 Bone fish Albula vulpes 1 27.7Æ 876Æ Bonita (almaco jack) Seriola rivoliana 14 42.8 6.3 30.3–49.6 537 246 302–1183 Coney Cephalopholis fulva 17 31.9 Æ 5.8 21.9–40.1 520 Æ 536 160–2097 Dusky shark (Galapagos) Carcharhinus galapagensis 2 27.4 Æ 1.5 25.5–28.6 283 Æ 62 240–373 Flounder sp. Bothus sp. 1 31.1Æ 386Æ Flying fish Cypselurus heterurus 5 48.8 2.0 44.9–50.4 829 233 587–1148 Fry Anchoa choerostoma 2 44.3 Æ 3.6 41.8–46.9 1130 Æ 8 1124–1136 Grasby Epinephelus cruentatus 1 31.5Æ 244Æ Gummy shark (dogfish) Mustelus antarcticus 1 31.0 315 Gwelly Pseudocaranx dentex 1 33.7 295 Hogfish Lachnolaimus maximus 5 23.2 6.5 15.0–32.4 201 17 182–227 Hybrid coney-barber Cephalopholis fulva 1 45.2Æ 390Æ Longfin jack Caranx ruber 1 24.5 2239 Ocean robin Decapterus macarellus 13 37.2 9.0 14.1–50.7 534 344 36–1391 Rainbow runner Elagatis bipinnulata 2 44.2 Æ 1.9 42.9–45.5 544 Æ 24 527–561 Red hind Epinephelus guttatus 26 20.4 Æ 4.3 11.5–28.2 201 Æ 65 123–353 Red snapper Lutjanus sp. 2 39.9 Æ 9.7 31.2–49.0 328 Æ 63 251–406 Caesar grunt Haemulon carbonarum 1 11.7Æ 865Æ Tuna Thunnus sp. 1 29.1 1006 Turbot Balistes capriscus 7 39.4 4.7 31.6–44.3 299 39 253–361 Wahoo Acanthocybium solandri 7 39.1 Æ 3.3 34.8–43.2 3816 Æ 4219 264–9858 Whitewater (lane) snapper Lutjanus synagris 15 28.9 Æ 5.3 16.9–38.7 412 Æ 114 216–713 Yellowfin tuna Thunnus albacares 2 40.7 Æ 5.0 36.0–45.9 321 Æ 217 190–572 Yellowtail snapper Ocyurus chrysurus 1 35.9Æ 519Æ

agencies (US Department of Health and Human the study revealed that 19%, 17% and 12% of them Services and US Environmental Protection Agency), had consumed, on regular basis, locally caught wahoo,

Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 these species as well as king mackerel and tilefish, which snapper and tuna, respectively. Furthermore, a positive were not included in the present study, should not be correlation (r 0.50) was found between wahoo and consumed by women planning to become pregnant, snapper consumption¼ with cord blood Hg content. In those that are pregnant or breastfeeding, and by the present study, all snapper species, except the deep children. Moreover, analysis of selenium and n-3 water red snapper (Etelis oculatus), wahoo and tuna, PUFA content did not reveal any particular nutritional were categorized in the ‘limited restrictions’ frequency benefit associated with their consumption. A number of consumption category based on their flesh Hg con- other species should also be consumed with moderation, centration. Therefore, consumption of any one of i.e. between once per week and once per month, as their those species during pregnancy and childhood should flesh contained moderately high concentrations of Hg. be restricted from one fish meal per day to one meal We previously conducted a study aiming to assess per week in order to reduce the exposure of the foetus umbilical cord blood concentrations of environmental to methylmercury in the Bermudian population. contaminants, including Hg (Dewailly and Pereg Selenium concentrations were found to be high 2004). Results showed a higher than expected mean in all fish species with a relative narrow variation 1 1 Hg concentration of 41.3 nmol lÀ , eight times the US (0.24–2.1 mggÀ ). The daily nutrient intake for sele- 1 average, measured among women of childbearing age nium is between 50 and 70 mg dayÀ . Significantly, a in 2000. An accompanying fish consumption frequency single 220 g meal of Bermuda fish will provide between 1 questionnaire administrated to the pregnant women in 53 and 460 mg dayÀ . Food Additives and Contaminants 9

Figure 4. Mercury and n-3 polyunsaturated fatty acids content in flesh of local fish species.

Figure 5. Relationship between n-3 polyunsaturated fatty acids and total lipids content in flesh of local fish species.

Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 While this sampling covers a wide range of fish reveal the mechanisms that result in such variability species, it is of importance to note that the sample size (water temperature, season, physiological state within each category of species is limited, often to a (spawning), etc.). single sample. Ranges of concentrations were some- The investigation of Hg levels in Bermuda fish has times very wide, particularly for n-3 PUFA and smaller been carried out along with an evaluation of two key sample sizes offer less confidence. nutrients, omega-3 fatty acids and the mineral sele- Tropical fish are often considered to be lean with nium, in order to help in the preparation of balanced little omega-3 content when compared with cold water Bermuda fish consumption advice. Pregnant and fish. However, if we consider an optimal dietary intake nursing mothers and Bermuda’s obstetricians can use 1 of 500 mg dayÀ recommended by the American Heart such a guide to decide which local fish to consume and Association (AHA) and International Society for the which to avoid during pregnancy. Study of Fatty Acids and Lipids (ISSFAL) (Kris- These data on balancing risk and benefits from Etherton et al. 2002; Cunnane et al. 2004), most Bermuda fish consumption need to be released to the Bermudian fish analysed in this study will provide the Bermuda population, particularly to pregnant women recommended targets. Some species are particularly and children. It is now possible to minimize the risk high in omega-3 content (Bermuda mackerel, wahoo, due to Hg exposure while maintaining the nutritional blue runner). However, these fish also show the highest benefit of fish consumption. However, for the general variation in content. Further studies will hopefully population, it is not so simple and past experiences 10 E´. Dewailly et al.

reported that consumer advisory, which is a major tool Castoldi AF, Coccini T, Ceccatelli S, Manzo L. 2001. for risk management, has a minimal effect under Neurotoxicity and molecular effects of methylmercury. experimental conditions to reduce the exposure of Brain Res Bull. 55:197–203. groups at risk (Verger et al. 2007). Other results suggest Counter SA, Buchanan LH. 2004. Mercury exposure in children: a review. Toxicol Appl Pharmacol. 198: that strategies to reduce Hg exposure by reducing 209–230. the amount of fish consumed are more effective at Cunnane S, Drevon CA, Harris B, Sinclair A, Spector A. eliminating the high end of the exposure distribution 2004. ISSFAL Board Statement No. 3: Recommendations than are strategies intended to change the types of fish for intake of polyunsaturated fatty acids in healthy adults. consumed (Carrington et al. 2004). However, such International Society for the Study of Fatty Acids advice in a more isolated community such as the Faroe and Lipids (ISSFAL). 28 June 2004. Available from: Islands was very efficient in reducing exposure from http://www.issfal.org.uk/Welcome/PolicyStatement3.asp/ pilot whale consumption (Weihe et al. 2005). Davidson PW, Myers GJ, Cox C, Wilding GE, Shamlaye CF, In any case, while recommending eating more or Huang LS, Cernichiari E, Sloane-Reeves J, Palumbo D, Clarkson TW. 2006. Methylmercury and neurodevelop- less of specific Bermuda fish, their individual sustain- ment: longitudinal analysis of the Seychelles child develop- ability in the Bermuda environment should be assessed. ment cohort. Neurotoxicol Teratol. 28:529–535. The data used to make the above recommendations Debes F, Budtz-Jorgensen E, Weihe P, White RF, are part of ongoing studies, both internationally and Grandjean P. 2006. Impact of prenatal methylmercury locally, that aim to characterize Hg in the marine exposure on neurobehavioral function at age 14 years. environment and its effects on human health. It is Neurotoxicol Teratol. 28:363–375. important to note that there is natural variability Dewailly E´ , Pereg D, Knap A, Rouja P, Galvin J, Owen R. within species groups and sometimes between speci- 2008. Exposure and effects of sea food-borne contami- mens from the same fish. There are important and nants in maritime populations. In: Walsh PJ, Smith S, interesting questions that still need to be answered Fleming L, Solo-Gabriele H, Gerwick WH, eds. Oceans and human health: risks and remedies from the seas. about the variability and potential seasonality of both Burlington, MA: Elsevier. p. 181–197. Hg and fat levels in local fish that require more FAO/WHO Expert Committee on Food Additives. 2006. research. In addition, the sample size for some species Summary and conclusions of the Sixty-seventh Meeting of has been limited and some species still need to be the Joint FAO/WHO Expert Committee on Food sampled. As more data become available the recom- Additives (JECFA), 20–29 June 2006. Geneva mendations made here should be assessed and updated (Switzerland): FAO/WHO. JECFA 67/SC. when necessary. Grandjean P, Weihe P, White RF, Debes F, Araki S, Yokoyama K, Murata K, Sorensen N, Dahl R, Jorgensen PJ. 1997. Cognitive deficit in 7-year-old children with prenatal exposure to methylmercury. Neurotoxicol Acknowledgements Teratol. 19:417–428. This work would not have been possible without the financial Harris WS. 2004. Are omega-3 fatty acids the most support of the Lepercq Foundation and the authors would important nutritional modulators of coronary heart like to thank Mrs Franc¸ oise Lepercq for long standing disease risk? Current Atherosclerosis Rep. 6:447–452. support. The Bermuda Government provided ship and boat He K, Song Y, Daviglus ML, Liu K, Van Horn L, Dyer AR, time and the authors are very grateful to the Department of Greenland P. 2004. Accumulated evidence on fish con- Conservation Services and its staff for their collaboration sumption and coronary heart disease mortality: a meta- Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008 and scientific support. The authors would also like to analysis of cohort studies. Circulation. 109:2705–2711. recognize the support and expertise of the Bermuda Hibbeln JR, Davis JM, Steer C, Emmett P, Rogers I, Department (Dr Tom Sleeter, Dr Tammy Trott, Williams C, Golding J. 2007. Maternal seafood consump- Dr Brian Luckhurst, Mr John Edmunds and Mr Craig tion in pregnancy and neurodevelopmental outcomes in Trott). The Bermuda Underwater Exploration Institute childhood (ALSPAC study): an observational cohort (BUEI) with Mr Teddy Tucker also provided boat support study. Lancet. 369:578–585. and meeting facilities; and the Bermuda Institute for Ocean Science (BIOS) provided accommodation and freezer storage Kris-Etherton PM, Harris WS, Appel LJ. 2002. Fish for visiting scientists and fieldworkers. Finally, all the consumption, fish oil, omega-3 fatty acids, and cardiovas- Bermudian commercial fishermen who helped in collecting cular disease. Circulation. 106:2747–2757. fish samples are thanked. Lauritzen L, Jorgensen MH, Mikkelsen TB, Skovgaard M, Straarup EM, Olsen SF, Hoy CE, Michaelsen KF. 2004. Maternal fish oil supplementation in lactation: effect on visual acuity and n-3 fatty acid content of infant References erythrocytes. Lipids. 39:195–206. Lepage G, Roy CC. 1986. Direct transesterification of all classes Carrington C, Montwill B, Bolger P. 2004. An intervention of lipids in a one-step reaction. J Lipid Res. 27:114–120. analysis for the reduction of exposure to methylmercury Leung AM, Braverman LE, Pearce EN. 2007. A dietary from the consumption of seafood by women of child- iodine questionnaire: correlation with urinary iodine and bearing age. Regul Toxicol Pharmacol. 40:272–280. food diaries. Thyroid. 17:755–762. Food Additives and Contaminants 11

Logan AC. 2003. Neurobehavioral aspects of omega-3 fatty Storelli MM, Marcotrigiano GO. 2000. Fish for human acids: possible mechanisms and therapeutic value in major consumption: risk of contamination by mercury. Food depression. Altern Med Rev. 8:410–425. Addit Contam. 17:1007–1011. Mahaffey KR, Clickner RP, Bodurow CC. 2004. Blood Svensson BG, Schutz A, Nilsson A, Akesson I, Akesson B, organic mercury and dietary mercury intake: National Skerfving S. 1992. Fish as a source of exposure to mercury Health and Nutrition Examination Survey, 1999 and 2000. and selenium. Sci Total Environ. 126:61–74. Environ Health Perspect. 112:562–570. Uauy R, Hoffman DR, Peirano P, Birch DG, Birch EE. Myers GJ, Davidson PW, Cox C, Shamlaye CF, Palumbo D, 2001. Essential fatty acids in visual and brain development. Cernichiari E, Sloane-Reeves J, Wilding GE, Kost J, Lipids. 36:885–895. Huang LS, et al. 2003. Prenatal methylmercury exposure US Environmental Protection Agency (USEPA). 2000. Risk from ocean fish consumption in the Seychelles child assessment methods. In: Guidance for assessing chemical development study. Lancet. 361:1686–1692. contaminant data for use in fish advisories, Vol. 2: Risk National Research Council. 2000. Toxicological effects of assessment and fish consumption limits. 3rd Edn. methylmercury. Washington, DC: National Academy Washington, DC: USEPA. Available from: http://www.epa. Press. gov/waterscience/fishadvice/volume2/v2ch2.pdf/. p. 2–30. Ruxton CH, Reed SC, Simpson MJ, Millington KJ. 2004. Verger PH, Blanchemanche S, Roosen J, Marette S. 2007. The health benefits of omega-3 polyunsaturated fatty Impact of a risk–benefit advisory on fish consumption and acids: a review of the evidence. J Hum Nutr Diet. 17: dietary exposure to methylmercury in France. Regul 449–459. Toxicol Pharmacol. 48:259–269. Shaikh NA, Downar E. 1981. Time course of changes in Wang C, Harris WS, Chung M, Lichtenstein AH, Balk EM, porcine myocardial phospholipid levels during ischemia. Kupelnick B, Jordan HS, Lau J. 2006. n-3 Fatty acids from A reassessment of the lysolipid hypothesis. Circ Res. fish or fish-oil supplements, but not alpha-linolenic acid, 49:316–325. benefit cardiovascular disease outcomes in primary- and Sinclair AJ, Begg D, Mathai M, Weizinger RS. 2007. Omega secondary-prevention studies: a systematic review. Am J 3 fatty acids and the brain: review of studies in depression. Clin Nutr. 84:5–17. Asia Pacific J Clin Nutr. 16(Suppl. 1):391–397. Weihe P, Grandjean P, Jørgensen PJ. 2005. Application of Steuerwald U, Weihe P, Jorgensen PJ, Bjerve K, Brock J, hair-mercury analysis to determine the impact of a seafood Heinzow B, Budtz-Jorgensen E, Grandjean P. 2000. advisory. Environ Res. 97:200–207. Maternal seafood diet, methylmercury exposure, World Health Organization (WHO). 1990. Environmental and neonatal neurologic function. J Pediatr. 136: Health Criteria No. 101: methyl mercury. Geneva 599–605. (Switzerland): WHO. Downloaded By: [Canadian Research Knowledge Network] At: 18:33 30 September 2008