The emerging farmed fish species meagre (Argyrosomus regius): How culinary treatment affects nutrients and contaminants concentration and associated benefit-risk balance

Page 1

Food and Chemical Toxicology 60 (2013) 277–285

Contents lists available at ScienceDirect

Food and Chemical Toxicology

journal homepage: www.elsevier .com/locate / foodchemtox

Review

The emerging farmed fish species meagre (Argyrosomus regius):How culinary treatment affects nutrients and contaminantsconcentration and associated benefit-risk balance

0278-6915/$ - see front matter � 2013 Published by Elsevier Ltd.http://dx.doi.org/10.1016/j.fct.2013.07.050

⇑ Corresponding author. Tel.: +351 21 302 7000; fax: +351 21 301 5948.E-mail addresses: [email protected], [email protected] (S. Costa).

Sara Costa a,⇑, Cláudia Afonso a, Narcisa Maria Bandarra a, Sandra Gueifão c, Isabel Castanheira c,Maria Luísa Carvalho b, Carlos Cardoso a, Maria Leonor Nunes a

a Division of Aquaculture and Upgrading, Portuguese Institute of the Sea and Atmosphere, IPMA, Avenida de Brasília, 1449-006 Lisboa, Portugalb Centre of Atomic Physics, Faculty of Sciences, University of Lisbon, Avenida Professor Gama Pinto 2, 1649-003 Lisboa, Portugalc National Health Institute of Dr. Ricardo Jorge (INSA), Avenida Padre Cruz, 1649-016 Lisboa, Portugal

a r t i c l e i n f o

Article history:Received 12 March 2013Accepted 19 July 2013Available online 27 July 2013

Keywords:Benefit/risk assessmentCulinary treatments in meagreNutritional/chemical compositionEPA + DHAContaminants

a b s t r a c t

The effect of cooking methods (boiling, grilling, and roasting) on the proximate and mineral composition,contaminants concentration and fatty acids profile was evaluated aiming to understand the benefits andrisks associated to the consumption of the emerging farmed fish meagre (Argyrosomus regius).

All the treatments led to lower moisture content. After grilling and roasting, the SFA, MUFA and PUFAcontents increased. There was no degradation of EPA and DHA during the culinary processes. Significantretention of minerals in grilled and roasted meagre samples was registered. For Pb and Cd there were noconcentration differences between culinary treatments and regarding raw fish. Whereas As level washigher in grilled meagre, total Hg and Me-Hg values were augmented in grilled and roasted meagre.

The consumption of meagre is advisable due to the low and healthy fat, high selenium and protein con-tent. Grilling would be the best culinary treatment due to the retention of protein, EPA, DHA and miner-als. But as the risk of ingestion of Me-Hg content also increases, based on the risk assessment, intakeshould not exceed two weekly meals, provided that no other important Me-Hg food source is presentin the diet. Otherwise, even this maximum threshold should be lower.

� 2013 Published by Elsevier Ltd.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2782. Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

2.1. Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2782.2. Cooking methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2782.3. Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278

2.3.1. Proximate composition and energy value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2782.3.2. Fatty acids profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2792.3.3. Mineral elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2792.3.4. Contaminants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2792.3.5. Mercury and methylmercury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2802.3.6. Quality control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2802.3.7. Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2802.3.8. Nutritional and contaminant contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2802.3.9. Risk and benefit assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

3. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

3.1. Proximate composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2803.2. Fatty acid composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2803.3. Mineral composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2813.4. Potential benefits and hazards of meagre consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

4. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

Page 2

278 S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285

Conflict of Interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284

1. Introduction

In the recent years, the demand for seafood has been increasing.Such fact is, in part, the result of the perception about the diversityof species, the pleasant sensory attributes and nutritional valuethat contribute to good health and well being. It may also be linkedwith increases in the variety of options, since new species, newpresentations (e.g. presented under vacuum or in modified atmo-sphere packing) and new products (e.g. fish loins, fishburgers,whole fish fillets, new crispy fish products and frozen spiced fish)are available. Nevertheless, this rising has resulted in an extensiveover fishing in wild fisheries which in turn led to an increase in theglobal demand for aquaculture products, specially salmon, shrimp,tilapia, carp and cat fish species (FAO, 2010).

Aquaculture has increased in recent years, remaining an impor-tant sector for high-protein food supply (FAO, 2010). The currentreal trend in aquaculture development is the increasing productionof marine species, both molluscs and finfish (FAO, 2010). Thegrowth is mainly due to the mastering of seed production tech-niques for many species, namely gilthead seabass and seabreamand to the application of new farming technologies. These emerg-ing technologies could also allow the introduction of new fish spe-cies, like meagre, native and non-native, to reduce imports andincrease freshness of the product for consumers. A native or indig-enous species, accordingly to Manchester and Bullock (2000), is aspecies that occurs naturally in a geographical area without humanintervention whereas a non-native species dispersal into an area ismediated by human intervention (UKINC, 1979; IUCN, 1987 andHolmes and Simons, 1996 cited in Manchester and Bullock, 2000).

Within the new farmed fish species, meagre (Argyrosomus re-gius) could be a suitable candidate species for the diversificationof aquaculture. This species was scored at the eighth positionout of a total of 27 species evaluated (Quéméner et al., 2002).Rearing interest is mostly supported by its excellent biologicalcharacteristics and relatively high rates of growth, feed conver-sion, and fertility. On the other hand, meagre is a lean specieshaving a high yield and low fat. It shows good marketing poten-tial and can be processed into portions (fillets, loins) to supplythe growing segment for portion sized ready-to-cook products,provided commercial size is attained (>2 kg) (Hernández et al.,2009). In addition, meagre is a species with regional importancein the Mediterranean and the Black Sea, as well as on the Atlan-tic European and East African coasts (Griffiths and Heemstra,1995).

Fish are usually submitted to a culinary treatment to enhanceits flavor and taste and to make it digestible and microbiologicallysafe (Weber et al., 2008). The fish muscle undergoes many physicaland chemical changes during cooking. These changes includeweight loss, modifications of water-holding capacity, texture, mus-cle fiber shrinkage, color and aroma development, which in turnare strongly dependent on protein denaturation and water loss.The effects of different processing or cooking methods on nutritivevalues of some fish species have been previously studied (Türkkanet al., 2008; Marimuthu et al., 2011; Bandarra et al., 2009; Perellóet al., 2008; Weber et al., 2008; Moradi et al., 2011). According tothese authors, in most cases cooking affects fish nutritional value

since some nutrients are concentrated or lost. In what regardsmeagre, available data is still scarce.

According to FAO (2010), it is necessary to assess the healthbenefits and risks associated with the consumption of fish and bal-ance them in order to enjoy the benefits while minimizing therisks. Particularly, for emergent farmed fish species, which maylack important chemical data, there is an even more importantneed to assess and compare any possible risks and benefits. Thougha thorough benefit-risk assessment in seafood cannot necessarilybe confined to a single fish species, this study does not portendto give an overall assessment of the benefits and risks incurredby seafood consumption in Portugal. Indeed, the current study onlyintends to use the meagre (A. regius) as an example of the new risksand benefits that arise to Portuguese consumers through the con-sumption of such an emergent farmed fish species.

Therefore, due to the important contribution foreseen for thisspecies, this work aims to characterize the proximate composition,mineral elements content and fatty acid profile of the edible part inone serving portion sized raw and cooked (boiling, grilling androasting) meagre. Based on these results, it is also aimed to assessthe benefits and hazards associated to the consumption of thisspecies.

2. Materials and methods

2.1. Samples

Farmed meagre (A. regius) were cultivated in earth ponds from an Aquacultureunit located in Olhão, Portugal. Fishes were slaughtered by immersion in an ice:water slurry. After slaughtering, a total of 60 fish were packed and transported inice to the laboratory in Lisbon, Portugal, within 12 h. The weight and length of fishwere recorded (weight: 601 ± 44 g and length: 40 ± 1 cm). Then fresh fish werescaled, gutted, beheaded and washed by hand. In Portugal, legislation allows thecapture of wild meagre fish with 420 mm size (Portaria no. 402/2002 from 18thApril 2002) but no imposition in terms of the size is made for aquaculture fish spe-cies. Also, since meagre is an emergent species, producers differentiate smaller fish(0.600 to 1 kg) that are sold whole fresh from the larger fish (1 to 3–5 kg) that aresold in the form of fillet or slice (Monfort, 2010).

2.2. Cooking methods

In each traditional cooking treatment 15 individuals were used. In the boilingprocess (1:2 fish/water ratio, water contained 2 g of salt per 100 g of fish) the por-tions corresponding to each individual fish were cooked for 10 min. For roasting,fish were spiced over 15 min with garlic cloves (4 per fish), onion (15 g/100 g)and salt (2 g per 100 g of fish). Just before the introduction of fish in the oven, oliveoil (2 mL/100 g fish), water (10 mL/100 g fish) and white wine (5 mL/100 g fish) wasadded. Roasting took place in a professional kitchen oven (Gico – mod. FC10 – E1 N,16,00 w Ac 400-3N-50 Hz) at 175 �C for 40 min. The grilling process was carried outin a domestic griller (Flama Sketch 230 V, 50 Hz, 2000 W) operated at about 180 �C.Each side of the salted fish (1.5 g/100 g) was grilled for about 10 min. After cooking,the skin and bones were removed and the edible part was subsequently homoge-nized. The homogenized material was separated into two sub-samples: one wasfrozen stored at �80 �C and the other frozen at �20 �C and then freeze dried (for48 h at �45 �C and low pressure, approximately 10�4 atm) and afterwards storedat �80 �C until further analysis.

2.3. Analyses

2.3.1. Proximate composition and energy valueThe moisture and ash contents were determined according to AOAC methods

(2005), respectively by drying at 105 �C and combustion of dried samples over16 h at 500 �C until constant weight. The protein level was quantified using a

Page 3

S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285 279

combustion method of analysis with the FP-528 DSP LECO nitrogen analyzer (LECO,St. Joseph, USA) calibrated with EDTA according to the Dumas method (Saint-Denisand Goupy, 2004). Total lipid was determined following the Smedes extractionmethod (Smedes, 1999) using propan-2-ol and cyclohexane as solvent.

The energy value, expressed as kcal/100 g edible part, was estimated using FAO(1989) factors: proteins, 4.27 kcal g�1 wet weight; lipids, 9.02 kcal g�1 wet weight;carbohydrates, 4.11 kcal g�1 wet weight. Since carbohydrate values in fish are verylower (<0.3%) (Belitz et al., 2004) they were not quantified so the factor of4.11 kcal g�1 for the carbohydrates was not used in the calculation.

2.3.2. Fatty acids profileFatty acid methyl esters (FAME’s) of non-polar and polar lipids were prepared

by acid-catalyzed transesterification using Lepage & Roy method (Lepage andRoy, 1986), modified by Cohen (Cohen et al., 1988). Samples were injected into aVarian Star 3800 Cp gas chromatograph (Walnut Creek, CA, USA), equipped withan auto sampler with a flame ionization detector at 250 �C. FAME were identifiedby comparing their retention time with those of Sigma–Aldrich standards (PUFA-3, Menhaden oil and PUFA-1, Marine source from Supelco Analytical). Data inmg/100 g of edible part were calculated using the peak area ratio (% of total fattyacids) and the lipid conversion factors set by Weihrauch et al. (1977). The athero-genic and thrombogenic index (AI and TI, respectively) were calculated accordinglyto Ulbricht and Southgate (1991) for evaluation of the predisposition for incidenceof coronary heart disease.

2.3.3. Mineral elementsBromine (Br), calcium (Ca), chloride (Cl), iron (Fe), potassium (K), rubidium (Rb),

selenium (Se), strontium (Sr), sulfur (S) and zinc (Zn) were determined by EnergyDispersive X-ray Fluorescence method (EDXRF, EXTRA II A, Atomika Instruments,Temple, Arizona, USA). The equipment has a source consisting of an X-ray tube (Phi-lips, PW 1140; 100 kV, 80 mA), a secondary target (in molybdenum) and a detector

Table 1Results of analysis (n P 4) for some certified reference material.

Technique DL Certified

Proximate composition (g/100 g)Moisture Drying nd SMRD 20Ash Combustion 0.1Protein FP-528 ndLipids Smedes nd

Fatty acids (mg/100 g)14:0 GC 0.4–1 BCR-16316:0 GC 0.4–116:1 GC 0.4–118:0 GC 0.4–118:1 GC 0.4–118:2 GC 0.4–118:3 GC 0.4–1Macro, trace and toxic elements (mg kg�1)Sodium FAAS 0.37 SMRD 20Magnesium FAAS 0.02 LUTS-1Chloride EDXRF 10 SRM 156Sulfur EDXRF 10Potassium EDXRF 10Calcium EDXRF 20Manganese FAAS 0.01 LUTS-1

TORT-2Iron EDXRF 3 TORT-2Nickel FAAS 0.02 TORT-2

LUTS-1Copper FAAS 0.02 LUTS-1Zinc EDXRF 1 TORT-2Selenium EDXRF 0.6Arsenic ICP-MS 0.01 DORM-3Bromine EDXRF 1 IAEA-A-1Rubidium EDXRF 1.1 SRM 157Strontium EDXRF 0.5 TORT-2Cadmium ETAAS 0.04 � 10�3 TORT-2Lead ETAAS 0.6 � 10�3

Mercury AAS 0.02Methly-mercury AAS 0.02 TORT-2

Values are presented as average ± standard deviation. (*) - Uncertified values presented bmatrix meat (Swedish Meats R&D and Scan Foods/National Food Administration, SwedenBelgium), LUTS-1 – Nondefatted lobster hepatopancreas, SRM 1566 – Oyster tissue (Uni(International Atomic Energy, Austria), Tort-2 – Lobster hepatopancreas (National ResearX-ray Fluorescence), FAAS (Flame Atomic-absorption spectroscopy, nd (not determined)

(in lithium drifted silicon [Si (Li)] with a 30 mm2 active area and 8 lm berylliumwindow). The energy resolution was 135 eV at 5.9 keV and the acquisition systemwas Nucleus PCA card. The X-ray generator was operated at 50 kV, 20 mA andacquisition time of 1000 s. The concentration of each element was determinedusing an EDXRF spectrometer, according to Custódio et al. (2003).

The elements copper (Cu), magnesium (Mg), manganese (Mn), sodium (Na) andnickel (Ni) were measured by atomic absorption spectrophotometry (Spectr AA55B, Varian, Palo Alto, CA, US) with a background deuterium correction, accordingto official analytical methods (Jorhem, 2000). The concentrations were determinedthrough linear calibration obtained from absorbance measurements of, at least, fivedifferent concentrations of standard solutions: NaNO3, Mg(NO3)2, Mn(NO3)2,Cu(NO3)2 and Ni(NO3)2 (dissolved in 0.5 M HNO3).

2.3.4. ContaminantsCadmium (Cd) and lead (Pb) were determined by graphite furnace atomic

absorption spectrometry, using a Varian apparatus Spectr 220Z with a Zeeman cor-rection (k = 283.3 and 228.8 nm for Pb and Cd, respectively). The methodology fol-lowed was based on the European Standard EN 14084 (CEN, 2003). Theconcentrations were determined through linear calibration obtained from absor-bance measurements of, at least, five different concentrations of standard solutions:Pb(NO3)2 and Cd(NO3)2 (1 g L�1 dissolved in 0.5 M HNO3).

For arsenic determination, approximately 0.25 g of each sample was weighedinto PTFE vessels and 8 mL of a nitric acid mixture (65%, Merck), previously purifiedwith a sub boiling distillation system (Milestone, SubPUR), hydrogen peroxide (30%,suprapur, Merck) and ultra pure water (ratio 4:1:3). Samples were digested induplicate using a microwave oven (Milestone, Ethos1). The analysis was performedby a quadrupole inductively-coupled plasma mass spectrometry unit (ICP-MS;Thermo Elemental, X-series 2, UK), according to the EN 15763:2009 (CEN, 2009).Calibrations were performed using certified standard solutions of 1000 mg L�1 ofAs (dissolved in 5% HNO3; Merck). ICP-MS tuning was performed on a daily basis

reference biological material Certified Present work

00 68.5 ± 0.1 68.6 ± 0.02.65 ± 0.09 2.71 ± 0.001.63 ± 0.05 1.60 ± 0.00nd nd

2.29 ± 0.04 2.23 ± 0.0425.96 ± 0.30 25.39 ± 0.172.58 ± 0.16 2.22 ± 0.0318.29 ± 0.16 17.65 ± 0.0638.34 ± 0.36 38.65 ± 0.167.05 ± 0.17 7.19 ± 0.030.86 ± 0.14 0.81 ± 0.01

00 8533 ± 281 8346 ± 28089.5 ± 4.1 90.9 ± 2.2

6 10,000* 10,200 ± 5007600* 8000 ± 5009690 ± 50 10,000 ± 801500 ± 200 1350 ± 501.20 ± 0.13 1.28 ± 0.0313.6 ± 1.2 13.1 ± 0.1105 ± 13 109 ± 32.5 ± 0.2 2.4 ± 0.50.200 ± 0.034 0.195 ± 0.00915.9 ± 1.2 15.4 ± 0.2180 ± 6 177 ± 65.63 ± 0.67 5.53 ± 0.396.88 ± 0.30 6.88 ± 0.15

3 22 ± 3 22 ± 21 12 ± 1 11.5 ± 0.4

45.2 ± 1.9 45.0 ± 1.026.7 ± 0.6 22.4 ± 2.40.27 ± 0.06 0.28 ± 0.000.35 ± 0.13 0.28 ± 0.010.152 ± 0.013 0.151 ± 0.008

y United States National Bureau of Standards. Abbreviations: SRMD2000 – Canned), BRC-163 – Beef pork fat blend (Institute for Reference Materials and measurement,ted States National Bureau of Standards, ESA), IAEA-13 – Freeze-dried animal bloodch Council of Canada, Canada). GC (gas Chromatography), EDXRF (Energy dispersive, DL (detection limit) ICP-MS, inductively coupled plasma mass spectroscopy.

Page 4

Table 2Proximate composition (g/100 g) of raw and cooked meagre (n = 15).

Proximatecomposition

Cooking method

Raw Boiled Grilled Roasted

Moisture 76.3 ± 0.2c 75.4 ± 0.3c 66.0 ± 1.1a 70.2 ± 0.4b

Ash 1.29 ± 0.02a 1.26 ± 004 a 2.65 ± 0.08b 2.62 ± 0.06b

Protein 21.0 ± 0.6a 22.2 ± 0.6b 28.9 ± 0.9c 25.4 ± 0.7d

Fat 1.4 ± 0.3a 1.6 ± 0.2ab 2.6 ± 0.2c 2.2 ± 0.2bc

Energy value (kcal/100 g)

102 ± 3a 109 ± 3ab 147 ± 5c 128 ± 4bc

Values are presented as average ± standard deviation. Different letters within a rowcorrespond to statistical differences (p < 0.05).

280 S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285

with a diluted 10 mg L�1 multi-element solution (Analytika, UNICAM). Indium andRhodium (1000 mg L�1; Merck) were chosen as internal standards to correct forinstrumental drift.

2.3.5. Mercury and methylmercuryTotal mercury (Hg) and methylmercury (Me-Hg) were quantified using a mer-

cury analyzer spectrophotometer (AMA 254, Leco, St. Joseph, Michigan, USA) (EPA,1998). For the extraction of Me-Hg from fish samples, the method referred by Afon-so et al. (2013) was used. The concentration of Hg was determined from a linear cal-ibration plot obtained from the measurement of a 1 g L�1 Hg standard solutiondiluted in 0.5 M HNO3.

2.3.6. Quality controlA minimum of two replicates were performed for each sample and analysis.

Accuracy was performed through analysis of certified reference biological material.The detection limit was determined by: (a) EDXRF—the mean between two valueswith the signal-to-noise approach, where the equipment compares the signal ofeach element with blank samples and establishes the minimum concentration atwhich the element is reliably detected; (b) FAAS—with the residual standard devi-ation (RSD) of the response and the slope (S) of the calibration curve of each stan-dard solution used (DL = 3.3 � RSD/S); (c) ICP-MS — the lowest concentration of theanalyte detected by 10 independent sample blanks fortified. Detection limits wereset at three times the standard deviation of the blanks.

The detection limits, certified reference biological materials and results ob-tained for each analysis are presented in Table 1.

2.3.7. StatisticsAll data were analyzed using STATISTICA 6 software (Statsoft, Inc., Tulsa,

OK74104, USA). Outlier detection was performed using Grubbs method (Grubbs,1969). A general linear model, one-way ANOVA, was used to determine significantdifferences (p < 0.05) between the type of culinary treatment in meagre, followedby a multiple comparison test (Tukey HSD). When data could not satisfy normaldistribution and homoscedasticity requirements, differences were analyzed withnon-parametric analysis of variance (Kruskal–Wallis) followed by non-parametricmultiple comparisons test (Mann–Whitney) (Zar, 1999).

2.3.8. Nutritional and contaminant contributionsThe nutritional contribution of meagre consumption was estimated taking into

consideration the concentrations of EPA and DHA and of macro, trace and ultra-trace elements, a meal of 160 g/day and the dietary reference intake values (DRIs)referred by the European Food Safety Authority (EFSA, 2010), the American HeartAssociation (AHA, 2010) and by the Institute of Medicine (IOM, 2012). Portugal isthe European country with the highest per capita fish consumption, around 58 kgper year (FAO, 2007), which represents more than 160 g of fish per capita per day.

In the case of EPA and DHA, the recommended intake for the prevention of car-diovascular disease is between 250 and 500 mg/day. Therefore a value of 250 mgwas used for the calculations because seems to be adequate for a primary preven-tion in healthy subjects (EFSA, 2010).

The contaminant levels were compared with the maximum permissible concen-tration limits (MPC) established by the European commission (EC, 2008) and thenprocessed in order to calculate associated intakes, which were compared with theProvisional Tolerable Weekly/Monthly Intakes (PTWI/PTMI) established by ExpertCommittee on Food Additives (JECFA) of the FAO/WHO (2010a,b). It was assumedan adult with a body weight of 60 kg and the other assumptions that were usedfor nutritional contributions.

The formulas used for calculation of the mineral and contaminants contribu-tions were:

Nutritional contribution ð%Þ ¼ ½ðC �MÞ=DRI ðmgÞ� � 100

Contribution ð%Þ ¼ ½ðC �MÞ=ðBW � PTWI or PTMIÞ� � 100

where C is the mean concentration of the mineral or contaminant (mg kg�1); M is themeal portion consumed (kg); DRI is the dietary reference intake value (mg); BW isthe body weight (kg); PTWI or PTMI is the provisional tolerable weekly or monthlyintake (mg kg�1).

2.3.9. Risk and benefit assessmentBenefit and risk assessment was done accordingly to Cardoso et al. (2010). For

the fitting of distributions to raw data and the exposure distribution generation, thesoftware @RISK�, advanced risk analysis for spreadsheets, from Palisade Corpora-tion (Ithaca, NY, USA), version 4.5, 2005, was used. For the calculus of probabilityof exceeding threshold values, the extreme value theory with excel macros wasused (Tressou, 2005).

Only, EPA + DHA, Se and Me-Hg were chosen for this assessment analysis sinceEPA + DHA and Se are the benefit components that are primarily associated to fishconsumption and Me-Hg was the only contaminant that presents a real threat. So,the benefit/risks were quantified through the probability of exceeding the referencedaily intake of the nutrients or the provisional tolerable weekly intake (PTWI) of thecontaminant, and was identified as P(Xi > PTWI or DRI). There are two alternatives

for estimating the risk: the plug-in (PI) and the tail estimation (TE) based estima-tors. Both were used, the PI estimator for large probabilities and the TE estimatorfor the other situations (especially, for very low probabilities). The statistical meth-odology based on the extreme value theory (which has found application in variousscientific fields, such as hydrology) was applied to the EPA + DHA, selenium andMe-Hg data of meagre. This analysis is described in detail in Cardoso et al. (2010).

3. Results and discussion

3.1. Proximate composition

The proximate composition (g/100 g) of meagre edible part be-fore and after each cooking treatment is shown in Table 2. Rawmeagre is a lean species with high protein content (>20%). Suchcomposition is similar to that obtained by Grigorakis et al.(2011). Energy value of raw meagre was estimated to be102 kcal/100 g, indicating that this species is low in calories.

Except for boiled meagre, the moisture content decreased sig-nificantly after cooking and such decline was more evident forgrilled fish. Similar results were obtained by Weber et al. (2008)for boiled, baked and grilled silver catfish, Ersoy and Özeren(2009) for grilled and baked African catfish, Bandarra et al.(2009) in grilled black scabbard fish and Marimuthu et al. (2011)for boiled, baked and grilled striped snakehead fish. The water lossmay be due to evaporation and is also probably due to heat-in-duced protein denaturation during cooking, which causes lesswater to be entrapped within the protein structures held by capil-lary forces. Therefore, in general, the subsequent increases in pro-tein, ash and fat may be related to that water loss, modifying theirrelative contents.

The fat content increased in grilled and roasted meagre but notfor boiled products. These results are in agreement with those ob-served by Türkkan et al. (2008) for baked seabass. Nevertheless,these results have not been observed for instance in grilled, boiled,oven-baked silver catfish (Weber et al., 2008), in grilled and bakedAfrican catfish (Ersoy and Özeren, 2009), in grilled black scabbardfish (Bandarra et al., 2009) and in boiled and baked striped snake-head fish (Marimuthu et al., 2011). Similar results to thosereported for fat were also found for ash content in the case ofgrilled and roasted treatment.

3.2. Fatty acid composition

The fatty acid composition (mg/100 g) in raw, boiled, grilled androasted meagre is shown in Table 3. The most abundant fatty acidsin the lipids extracted from meagre were oleic and palmitic acids(18:1x9, and 16:0, respectively) followed by docosahexaenoicand linoleic acids (22:6x3 or DHA and 18:2x6, respectively). Thisis similar to the results obtained by Grigorakis et al. (2011), Türk-kan et al. (2008) and Bandarra et al. (2009), respectively for rawmeagre, baked and microwave cooked seabass and grilled blackscabbard fish. These major four fatty acids represent around 64%

Page 5

Table 3Fatty acids profile of raw and cooked meagre (mg/100 g).

Fatty acid Cooking method

Raw Boiled Grilled Roasted

SaturatedMiristic 14:0 33.18 ± 4.61a 38.52 ± 7.16a 73.16 ± 2.39c 48.10 ± 5.03b

Pentadecilic 15:0 6.44 ± 0.82a 7.27 ± 1.24a 13.98 ± 0.49c 9.26 ± 0.79b

Palmitic 16:0 212.74 ± 25.32a 242.82 ± 34.17a 423.23 ± 12.37c 333.47 ± 27.17b

Stearic 18:0 61.63 ± 6.90a 71.87 ± 7.56a 115.39 ± 4.84c 94.68 ± 8.57b

R SFAA 331.66 ± 39.31a 381.19 ± 53.96a 659.84 ± 15.53c 514.49 ± 40.66b

MonounsaturatedHexadecenoic C16:1 51.66 ± 6.69a 61.14 ± 11.81a 115.41 ± 3.16c 82.43 ± 7.79b

Oleic 18:1x9 183.00 ± 22.44a 221.41 ± 35.27a 392.77 ± 14.98b 430.19 ± 58.40b

Eicosenoic C20:1 41.70 ± 6.29a 50.43 ± 8.33a 89.18 ± 1.13c 66.90 ± 9.65b

Docosenoic C22:1 36.97 ± 6.60a 45.89 ± 8.65a 82.10 ± 8.53c 60.42 ± 9.12b

MUFAB 349.44 ± 48.74a 430.58 ± 72.53a 772.86 ± 30.12b 709.68 ± 89.02b

PolyunsaturatedHexadecadienoic 16:2x4 4.06 ± 1.40a 5.25 ± 0.41ab 9.28 ± 0.77c 6.76 ± 0.57b

Hexadecatrienoic C16:3 6.91 ± 1.61a 8.69 ± 1.01ab 13.75 ± 0.54c 10.97 ± 1.02bc

Hexadecatetraenoic 16:4x3 1.65 ± 0.22a 1.95 ± 0.53ab 3.87 ± 0.57b 2.58 ± 0.14b

Linoleic 18:2 x6 111.28 ± 13.68a 128.25 ± 23.09a 217.96 ± 3.55c 176.94 ± 16.95b

a-Linolenic 18:3x3 12.90 ± 1.67a 15.43 ± 3.24a 27.63 ± 0.67b 21.58 ± 2.44c

Stearidonic 18:4x3 9.91 ± 1.34a 11.99 ± 2.77a 23.26 ± 1.09c 15.58 ± 1.51b

Eicosadienoic C20:2 (x9 + x6) 3.08 ± 0.56a 4.16 ± 0.89a 6.99 ± 0.50b 5.54 ± 1.00b

Arachidonic 20:4x6 14.74 ± 1.56a 17.46 ± 1.24ab 27.68 ± 1.96b 21.46 ± 3.37b

Eicosapentaenoic 20:5x3 53.32 ± 5.77a 66.33 ± 12.42a 121.55 ± 6.06c 85.04 ± 10.50b

Heneicosapentaenoic 21:5x3 1.95 ± 0.33a 2.48 ± 0.60a 4.48 ± 0.18c 3.21 ± 0.90b

Adrenic 22:4x6 2.27 ± 0.34a 2.84 ± 0.69a 4.78 ± 0.42b 3.90 ± 0.76b

Docosapentaenoic 22:5x6 6.44 ± 0.67a 7.72 ± 0.73a 12.32 ± 0.97c 9.70 ± 1.55b

Docosapentaenoic 22:5x3 16.60 ± 1.70a 21.14 ± 3.61a 35.31 ± 3.93c 26.73 ± 4.00b

Docosahexaenoic 22:6x3 161.77 ± 15.40a 194.08 ± 18.42a 298.65 ± 24.93c 234.86 ± 29.66b

PUFAC 416.29 ± 43.60a 499.44 ± 69.97a 829.19 ± 37.64c 639.88 ± 72.80b

EPA + DHA 215.08 ± 20.65a 260.41 ± 30.58a 420.20 ± 30.95c 319.90 ± 39.57b

Rx3 268.23 ± 26.64a 326.24 ± 42.59a 535.73 ± 35.95c 405.50 ± 49.37b

R x6 138.42 ± 15.04a 162.28 ± 26.14a 272.33 ± 3.72c 219.28 ± 23.06b

x3/x6 1.94 ± 0.07ab 2.02 ± 0.09b 1.97 ± 0.13ab 1.85 ± 0.06a

Atherogenic index AI 0.457 ± 0.018a 0.432 ± 0.017ab 0.454 ± 0.018a 0.397 ± 0.035b

Thrombogenic index TI 0.293 ± 0.011a 0.277 ± 0.011a 0.288 ± 0.012a 0.285 ± 0.021a

Only the most important fatty acids are listed. Values are presented as average ± standard deviation. Different letters within a row correspond to statistical differences(p < 0.05).

A 11:0, 12:0, 13:0, 17:0, 19:0, 20:0 and 22:0 are included.B 17:1, 18:1x7, 18:1x5, 19:1w8 and 24:1w9 are included.C 18:3x6, 18:3x4, 20:3x3, 20:4x3.

S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285 281

of the total fatty acids; as was observed by Grigorakis et al. (2011)for meagre.

Though significant amounts of monounsaturated fatty acids(MUFA) and saturated fatty acids (SFA) were found, the polyunsat-urated fatty acids (PUFA) were the main fraction in all treatmentswith except of roasting. The majority of fatty acids amounts wereincreased by grilling and roasting, probably as consequence ofthe water losses during the cooking process, similarly to the resultsobserved by Bandarra et al. (2009) for grilled black scabbard fishand by Türkkan et al. (2008) for baked seabass. In Weber et al.(2008) boiling, grilling and baking marginally affected the fattyacid contents for silver catfish. The fatty acids profiles of raw andboiled samples were identical with no significant differences forall the fatty acids.

In what concerns PUFAs, the level of x3 (omega3) was higherthan x6 (omega6) and, with the exception of roasted meagre, thex3/x6 ratio was higher when compared to that (�1.85) referredby Grigorakis et al. (2011) for raw meagre. Regarding the x3/x6ratio, no differences were found between boiled, roasted andgrilled, relatively to raw meagre. Culinary treatments did not pro-mote the loss of EPA and DHA content, two of the most importantfatty acids for health in fish lipids, even in the most aggressive one,the grilling process. Similar results were obtained by other authors(Weber et al., 2008 and Bandarra et al., 2009).

Culinary treatments did not affect the atherogenic (AI) andthrombogenic indices (TI), calculated according to Ulbricht and

Southgate (1991), with the exception of AI in roasted meagre.These indices were found to be higher than the ones reported byGrigorakis et al. (2011), also for meagre. Since AI is the ratio be-tween some SFA and the MUFA and PUFA (x3 and x6), the addi-tion of olive oil in roasted fish samples resulted in a decrease ofAI values. As AI and TI represent factors of promotion and protec-tion against coronary heart diseases, the low values obtained inmeagre suggest a high cardio-protective effect (Valfré, 2008).

3.3. Mineral composition

The mean concentration of the macro, trace and ultra-trace ele-ments are presented in Table 4. For raw fish, the most abundantelements were K, S, Cl and Na (above 500 mg kg�1). Within theanalyzed trace elements, Fe, Zn and Br were the dominants. Thecomparison of these values with those published for other farmedand wild species (Ersoy and Özeren, 2009; Bandarra et al., 2009;Marimuthu et al., 2011; Nuray and Özden, 2007) allow to concludethat mineral contents are in the same range of the literature. How-ever the values between individuals of the same species maywidely vary. It is known that a variation in the mineral composi-tion of fish species can be related to the species, seasonal andbiological differences (i.e. species, size, dark/white muscle, age,sex and sexual maturity), feed composition and environmentalconditions (i.e. water chemistry, salinity and temperature) (Lall,1995).

Page 6

Table 4Concentration of mineral elements and contaminants (mg kg�1) in raw and cookedmeagre (boiled, grilled and roasted).

Raw Boiled Grilled Roasted

MacroelementsK 4396 ± 437a 3349 ± 156c 4704 ± 384a 3955 ± 283b

S 1587 ± 150a 2017 ± 300b 2414 ± 306c 2519 ± 309c

Cl 725 ± 88a 1663 ± 91b 6367 ± 695bc 8912 ± 566c

Na 552 ± 42a 1323 ± 163b 4342 ± 423 c 5409 ± 549d

Mg 284 ± 7a 274 ± 27a 340 ± 20b 265 ± 37a

Ca 112 ± 14a 150 ± 17b 156 ± 27b 189 ± 39b

Trace and ultra-trace elementsFe 5.83 ± 1.36a 3.84 ± 0.32b 4.25 ± 1.26ab 4.02 ± 0.35b

Zn 5.32 ± 0.35a 5.96 ± 0.31ab 7.08 ± 1.15c 6.69 ± 0.31bc

Br 4.87 ± 0.39a 6.22 ± 0.44b 10.93 ± 1.01c 10.24 ± 0.67c

Rb 0.69 ± 0.09a 0.51 ± 0.10a 0.87 ± 0.11c 0.52 ± 0.12b

Sr 0.52 ± 0.69a 1.12 ± 0.81ab 1.32 ± 0.99b 1.80 ± 1.41b

Se 0.27 ± 0.10ab 0.26 ± 0.13ab 0.37 ± 0.07b 0.24 ± 0.06a

Cu 0.19 ± 0.02a 0.25 ± 0.03b 0.33 ± 0.03c 0.27 ± 0.02b

Mn 0.09 ± 0.02a 0.10 ± 0.02a 0.13 ± 0.02a 0.13 ± 0.05a

Ni <DL <DL 0.14 ± 0.15 <DL

ContaminantsAs 0.54 ± 0.05a 0.52 ± 0.05a 0.80 ± 0.11b 0.56 ± 0.07a

Hg 0.21 ± 0.02a 0.22 ± 0.01ab 0.28 ± 0.01c 0.25 ± 0.03bc

Me-Hg 0.16 ± 0.01a 0.18 ± 0.01ab 0.20 ± 0.01b 0.19 ± 0.02b

Cd 0.003 ± 0.001a 0.003 ± 0.001a 0.002 ± 0.000a 0.003 ± 0.001a

Pb 0.009 ± 0.004a 0.007 ± 0.010a 0.010 ± 0.006a 0.012 ± 0.005a

Values are presented as average ± standard deviation. Different letters within a rowcorrespond to statistical differences (p < 0.05). Abbreviations: DL – Detection limit.

Table 5Nutritional contribution (%) of raw and cooked meagre in terms of mineral elementsand EPA and DHA, taking into account a meal of 160 g.

DRI (mg/d) Nutritional Contribution (%)

Raw Culinary treatment

Adult Boiled Grilled Roasted

MacroelementsCa Male/Female 1000 12 17 17 21Cl Male/Female 2300 5 12 44 62K Male/Female 4700 5 11 16 3Mg Male 400–420 11 11 13 10

Female 310–320 14 14 17 13Na Male/female 1500 6 14 46 58

Trace and ultra-trace elementsCu Male/Female 0.9 24 32 41 34Fe Male 8 82 54 59 56

Female 18 36 24 26 25Mn Male 1.8 6 6 8 8

Female 2.3 5 5 7 6Se Male/Female 0.055 79 77 108 71Zn Male 11 54 61 72 68

Female 8 75 83 99 94x3 fatty acidsEPA + DHA Male/Female 250 138 167 269 205

The dietary reference intake values (DRI) for macro and trace and ultra-trace ele-ments are presented for adults. When the DRI values presented a range, a meanvalue was used for the nutritional contribution (%) calculation.

282 S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285

The mean concentration of total Hg in raw meagre was0.21 ± 0.02 mg kg�1(range 0.19–0.23), the As was around0.54 ± 0.05 mg kg�1(range 0.49–0.59). Concerning Cd and Pb theobserved levels were 0.003 ± 0.001 and 0.009 ± 0.004 mg/kg,respectively (Table 4). Regarding Hg, the mean concentrationsfound were higher than those frequently described for otherfarmed fish species, as in sea bass (0.14 ± 0.04 mg/kg), sea bream(0.11 ± 0.04 mg/kg), rainbow trout (0.06 ± 0.02 mg/kg) and turbot(0.06 ± 0.03 mg/kg) (Lourenço et al., 2012). Cd values were lowerthan those found in several studies for farmed seabass, 0.20–0.27 mg kg�1(Alasalvar et al., 2002). Pb concentrations were alsolower than those reported in other studies 0.83–1.23 mg kg�1 (Ala-salvar et al., 2002) and 0.269–0.287 mg kg�1 (Ersoy et al., 2006).Arsenic levels are within the range indicated in literature for fishproducts, 0.330–0.414 mg kg�1in sea bass (Ersoy et al., 2006),3.53–4.47 mg kg�1in gilthead seabass and 5.29–8.11 mg kg�1 inhake (Anacleto et al., 2009).

Generally, the cooking treatments influenced the levels of min-eral composition. For K, the most abundant element in meagre, nodifferences were observed between raw and grilled samples, but asignificant decrease was observed both in boiled and roasted prod-ucts. In other fish species, K content increased in all the cookingmethods (Ersoy and Özeren, 2009). For grilled and roasted stripedsnakehead fish and grilled black scabbard fish, the content of thismineral also increased significantly (Marimuthu et al., 2011 andBandarra et al., 2009). Due to the addition of salt during the culi-nary process, Na values augmented. Although salt was added inthe same proportions in all culinary treatments, it is difficult to as-sess if the culinary treatments per se resulted in the increasing inthe Na content. Similar results were obtained by other authors inother fish species (Marimuthu et al., 2011; Bandarra et al., 2009;Ersoy and Özeren, 2009).

The S, Cl, Ca, Br and Cu, contents increased in all three treat-ments. In relation to Cl content, Bandarra et al. (2009) also ob-tained a significant increase in grilled treatment. For Ca, similarbehavior was obtained by Bandarra et al. (2009) and Ersoy and

Özeren (2009), respectively for black scabbard fish and African cat-fish. For Cu, Ersoy and Özeren (2009) observed a significant de-crease in all culinary treatments whereas Marimuthu et al.(2011) did not observe any difference.

For Mg there was a concentration phenomenon for grilled sam-ples, as observed by Bandarra et al. (2009) in black scabbard fish.

Fe contents in cooked samples ranged from 3.84 to 4.25 mg/kg.It was only observed a decrease in Fe values for boiled and roastedmeagre. It was reported that Fe levels did not changed in grilled,microwave and baked samples (Marimuthu et al., 2011 and Ersoyand Özeren, 2009).

With respect to Zn, an increase was observed for grilled androasted samples. Similar results were obtained for roasted catfish(Ersoy and Özeren, 2009), but not for grilled catfish (Ersoy and Öz-eren, 2009) and black scabbard fish (Bandarra et al., 2009).

Grilling had a concentrating effect on Br, Se, Rb, Sr and Ni con-tent. No references were found about the influence of culinarytreatments on concentrations of these elements. The increase inNi concentration in grilled fish samples was possibly due to theuse of a grill that has nickel in its metallic alloy. Nevertheless,the Ni concentration was inferior to the one obtained by Ersoyet al., 2006 (0.213 mg/kg) for sea bass and by Carvalho et al.(2005) (0.09–0.32 mg/kg) for meagre.

As described in the literature (Ersoy and Özeren, 2009 andMarimuthu et al., 2011), the culinary treatments had no influenceon Mn content after boiling, grilling and roasting.

Arsenic concentration ranged from 0.52 to 0.80 mg kg�1 andonly the grilled sample was different from the raw. However, grill-ing did not yield differences in other studies (Ersoy et al., 2006).

Mean Hg concentration ranged from 0.22 to 0.28 mg/kg, respec-tively in boiled and grilled meagre. It was observed an increase inHg concentration in grilled and roasted fish, similar to the resultsobtained by Perelló et al. (2008). Identical behavior was observedfor Me-Hg.

The Pb and Cd concentrations were similar between cooked(ranged between 0.007–0.012 mg kg�1 and 0.002–0.003 mg/kg,respectively) and raw samples. Similar results were observed forPb between raw and grilled African catfish (Ersoy et al., 2006).

Page 7

S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285 283

3.4. Potential benefits and hazards of meagre consumption

Attending to the DRIs, the values obtained indicate that meagreis a particularly good Se source (Table 5). This is more noteworthyfor grilled meagre, where the calculated intake fulfils the DRI(108%). For EPA + DHA, the estimated contribution overlaps therecommended value, ranging from 138% in raw to 269% in grilledfish.

For contaminants, all values were below the limits establishedby the European community (EC, 2008). The estimated weekly in-take values for Cd and Pb found in this study (Table 6) achieve only0.1% of the established PTMI or PTWI, respectively. Concerning Me-Hg, one meal of 160 g contributes to attain around 30% of the PTWIvalue in both raw and cooked products.

Therefore, a risk and benefit assessment was performed (Ta-ble 7). From the results obtained, it can be inferred that the con-sumption of raw meagre until a frequency of three meals perweek presents a very low risk (0.34%), considering the Me-Hg,which is overcome by the benefit associated with the ingestionof selenium (0.84%). When meagre is cooked, the risk increasesconsiderably when the consumption frequency exceeds two mealsper week, especially in the case of grilled (86.98%) and roasted(29.32%), since the distribution of the intake of Me-Hg may inter-sect the PTWI. In this case, although there is an increase in the ben-efit associated to the ingestion of Se (2.9 � 10�3 and 6.0 � 10�3%,respectively) and EPA + DHA (100.00% and 10.15%, respectively),the consumption of more than two meals per week becomes unde-sirable because of Me-Hg risk. Even in the case of boiled meagre,the probability of exceeding the Me-Hg PTWI becomes non-negli-gible (1.10%) when a consumption frequency scenario of threeweekly meals is projected. Thus, the weekly consumption shouldnot exceed 2 meals.

Table 7Probability of exceeding the Me-Hg PTWI and EPA + DHA and Se DRI, P(Xi > PTWI or DRI)

1 Meal (year) 1

Raw Me-Hg P(Xi > PTWI or DRI) (%) <1.0 � 10�8 <1Se <1.0 � 10�8 4.EPA + DHA <1.0 � 10�8 <1

Boiled Me-Hg <1.0 � 10�8 <1Se <1.0 � 10�8 <1EPA + DHA <1.0 � 10�8 <1

Grilled Me-Hg <1.0 � 10�8 <1Se <1.0 � 10�8 <1EPA + DHA <1.0 � 10�8 <1

Roasted Me-Hg <1.0 � 10�8 <1Se <1.0 � 10�8 <1EPA + DHA <1.0 � 10�8 <1

P(Xi > PTWI or DRI)—probability of exceeding PTWI or DRI; values of probability estimaNote: Values of probability (%) that were inferior to 1.0 � 10�8 were presented as <1.0occurrence in ten billions, more than the world population.

Table 6Contribution (%) of raw and cooked meagre in terms of contaminants, taking intoaccount a meal of 160 g and a consumer with a body weight of 60 kg.

Contribution (%) MPC(mg/kg)

PTWI(mg/kgbw)

PTMI(mg/kgbw)

Raw Boiled Grilled Roasted

Me- Hg 26.2 29.2 33.8 30.8 0.500.0016 –

Cd 0.1 0.1 0.1 0.1 0.050 – 0.025Pb 0.1 0.1 0.1 0.1 0.30 0.025

MPC – maximum permissible concentration; PTWI – provisional tolerable weeklyintake; PTMI – provisional tolerable monthly intake.

However, besides this farmed fish species, methyl-Hg is foundin other foods, particularly in large predator wild fish species, suchas the swordfish, because of the long food chains in the sea andbiomagnification phenomena (Afonso et al., 2007). Hence, pre-sented estimates only give a fraction of the overall probability ofexceeding the methyl-Hg PTWI, the Se or the EPA + DHA DRIsthrough food consumption. Nevertheless, this very partial assess-ment is enough to raise concerns regarding exposure of consumersto methyl-Hg. A meagre consumption frequency not exceeding twoweekly meals may be desirable due to these compounding effects.Other studies have already highlighted this issue (Tressou et al.,2004; Cardoso et al., 2010). In particular, a probabilistic exposureassessment to food chemicals also based on the EVT was carriedout in the French population (Tressou et al., 2004). It was foundthat the probability of exceeding the PTWI of methyl-Hg was, atleast, in the most favorable situation, 7.40% (PI) or 9.26% (TE) (Tres-sou et al., 2004). Of course, these higher probabilities are due to thefact that this study considered all food items, differently from thiswork, which was focused on assessing the risks and benefits of asingle fish species.

Of course, the attained probability results stand on assumptionsthat are not beyond doubt and may receive further improvement infuture works, particularly, if more information is available. First,the utilization of the methyl-Hg PTWI, the Se and the EPA + DHADRIs as sole references is disputable, since these threshold valuesincorporate some uncertainty. Particularly, it is difficult to set a ref-erence for a whole population, which has necessarily a great vari-ability of health conditions and genetic traits. In doing so, safetyfactors are generally incorporated, which, for instance, lower thetolerable intake level for methyl-Hg. Nonetheless, these are thevalues generally recognized as the reference. Other assumptionswere the utilization of an average body weight of 60 kg and a mealsize of 160 g. In all this discussion of the results, another assump-tion should not be forgotten, namely, one hypothesis underlyingthe utilization of available data is that individuals are subjectedto a constant exposure over time and keep the same consumptionbehavior over their lifetime. This is a strong assumption that can-not be avoided. However, it could be weakened by combining theused methods with some ideas proposed by other authors (Wallaceet al., 1994; Nusser et al., 1996), namely, by the study of consump-tion patterns over time. These methods were compared and dis-cussed by Hoffmann et al. (2002). Accordingly, the parameter ofinterest could be viewed as the probability of occasional short-term excursions above the DRI rather than a true probability todevelop a benefit/risk as a result of mineral ingestion or to theexposure to contaminants.

(%), as a result of raw and cooked meagre consumption.

Meal (month) 1 Meal (week) 2 Meals (week) 3 Meals (week)

.0 � 10�8 <1.0 � 10�8 1.8 � 10�5 0.342 � 10�8 2.6 � 10�2 0.12 0.84.0 � 10�8 <1.0 � 10�8 <1.0 � 10�8 <1.0 � 10�8

.0 � 10�8 <1.0 � 10�8 9.2 � 10�7 1.10

.0 � 10�8 2.3 � 10�8 1.5 � 10�4 1.9 � 10�3

.0 � 10�8 <1.0 � 10�8 9.8 � 10�4 0.36

.0 � 10�8 <1.0 � 10�8 2.9 � 10�3 86.98

.0 � 10�8 <1.0 � 10�8 4.1 � 10�5 2.9 � 10�3

.0 � 10�8 1.9 � 10�7 0.11 100.00

.0 � 10�8 3.0 � 10�5 0.17 29.32

.0 � 10�8 4.6 � 10�6 7.5 � 10�4 6.0 � 10�3

.0 � 10�8 1.2 � 10�4 0.10 10.15

ted using the plug-in (PI) estimators are presented in bold.� 10�8 and may be considered negligible since this probability is lower than one

Page 8

284 S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285

Finally, the used statistical models also involve assumptions,mainly regarding the application of the extreme value theory(Tressou et al., 2004).

It must be emphasized that this study does not pretend to givean overall assessment of the benefits and risks incurred by seafoodconsumption in Portugal. The current study only intends to assessthe benefits and risks of the consumption of meagre (A. regius), anemergent farmed fish species. Indeed, the methyl-Hg risk can bebetter assessed by a more thorough and exhaustive review of allpossible foods in a diet that would require the analysis of mainsources of methyl-Hg to the diet, such as, large predator wild fishspecies. Hence, presented estimates only give a fraction of theoverall probability of exceeding the methyl-Hg PTWI, the Se orthe EPA + DHA DRIs through food consumption.

4. Conclusions

Culinary treatments affected the proximate composition as wellas the levels of EPA and DHA of meagre since the reduction inmoisture content resulted in an increase of the other constituentscontent due to the dehydration, especially in the grilled androasted samples.

As for the fatty acid profile, in general, culinary treatments re-sulted in an increase in the amounts of the majority of fatty acids,including EPA and DHA, as consequence of the water losses duringthe cooking process, displaying raw and boiled samples identicalprofiles.

Generally, the macro, trace and ultra-trace elements contentswere different in grilled and roasted meagre samples comparedwith raw meagre. Na and Cl contents were higher in cooked prod-ucts due to the addition of salt. Mg, Fe and K were the elements lessretained in boiled and roasted samples.

Concerning contaminants, a significant retention was observedfor As in grilled meagre and for Hg and Me-Hg in grilled androasted products.

Overall, grilled treatment would be the best culinary treatmentdue to the retention of protein, EPA, DHA, macro, trace and ultra-trace elements content, but as the risk of ingestion of Me-Hg con-tent also increases, the intake should not exceed two meals perweek based on the risk assessment.

Conflict of Interest

The authors declare that there are no conflicts of interest.

Acknowledgements

This work was supported by the project ‘‘GOODFISH’’, Ref.PTDC/SAU-ESA/103825/2008 and the individual Post DoctoralGrant, Ref.: SFRH/BPD/64951/2009 both of ‘‘Fundação para a Ciên-cia e a Tecnologia’’ (FCT). We are also grateful to all members of thephysics laboratory/FCUL for technical assistance with the EDXRFtechnique.

References

Afonso, C., Lourenço, M.H., Cardoso, C., Bandara, N.M., Carvalho, M.L., Nunes, M.L.,2013. From fish chemical characterisation to the benefit-risk assessment – PartA. Food Chem. 137, 99–107.

Afonso, C., Lourenço, M.H., Dias, A., Nunes, M.L., Cardoso, M., 2007. Contaminantmetals in black scabbard fish (Aphanopus carbo) caught off Madeira and theAzores. Food Chem. 101, 120–125.

AHA, 2010. 2010 Dietary Guidelines for Americans. <http://www.heart.org/idc/groups/heart-public/@wcm/@adv/documents/downloadable/ucm_312853.pdf>(accessed on January 2013).

Alasalvar, C., Taylor, K.D.A., Zubcov, E., Shahidi, F., Alexis, M., 2002. Differentiation ofcultured and wild sea bass (Dicentrarchus labrax): total lipid content, fatty acidand trace mineral composition. Food Chem. 79 (2), 145–150.

Anacleto, P., Lourenço, H.M., Ferraria, V., Afonso, C., Carvalho, M.L., Martins, M.F.,Nunes, M.L., 2009. Total arsenic content in seafood consumed in Portugal. J.Aquat. Food Prod. Technol. 18, 32–45.

AOAC, 2005. Official Methods of Analysis of the ASSOCIATION of Official AnalyticalCommunities International, 18th ed., Washington DC, USA.

Bandarra, N.M., Batista, I., Nunes, M.L., 2009. Chemical composition and nutritionalvalue of raw and cooked black scabbard fish (Aphanopus carbo). Sci. Mar. 73S2,105–113.

Belitz, H.-D., Grosch, W., Schieberle, P., 2004. Food Chemistry. Springer-Verlag,Berlin.

Cardoso, C., Bandarra, N., Lourenço, H., Afonso, C., Nunes, M.L., 2010. Methylmercuryrisks and EPA + DHA benefits associated with seafood consumption in Europe.Risk Anal. 30, 827–840.

Carvalho, M.L., Santiago, S., Nunes, M.L., 2005. Assessment of the essential elementand heavy metal content of edible fish muscle. Ana. Bioanal. Chem. 382, 426–432.

CEN, 2003. European Standard EN 14084. Foodstuffs – Determination of traceelements Determination of lead, cadmium, zinc, copper and iron by atomicabsorption spectrometry (AAS) after microwave digestion. CEN-EuropeanCommittee for Standardization, 16 p.

CEN, 2009. European Standard EN 15763. Foodstuffs – Determination of TraceElements – Determination of Arsenic, Cadmium, Mercury and Lead inFoodstuffs by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) afterpressure digestion. CEN-European Committee for Standardization, 18p.

Cohen, Z., Von Shak, A., Richmond, A., 1988. Effect of environmental conditions onfatty acid composition of the red algae Porphyridium cruentum: correlation togrowth rate. J. Phycol. 24, 328–332.

Custódio, P.J., Carvalho, M.L., Nunes, F., 2003. Trace elements determination byenergy dispersive X-ray fluorescence (EDXRF) in human placenta andmembrane: a comparative study. Ana. Bioanal. Chem. 375, 1101–1106.

EC, 2008. Commission Regulation (EC) No. 629/2008 of 2 July 2008 amendingRegulation (EC) No 1881/2006 setting maximum levels for certaincontaminants in foodstuffs. JO L 173 03-07-2008.

EFSA, 2010. EFSA panel on dietetic products, nutrition, and allergies (NDA);scientific opinion on dietary reference values for fats, including saturated fattyacids, polyunsaturated fatty acids, monounsaturated fatty acids, trans fattyacids, and cholesterol. EFSA J. 8, 1461.

EPA, 1998. Test method 7473: Mercury in Solids and Solutions by ThermalDecomposition, Amalgamation and Atomic Absorption Spectrometry. SW-846,Environment Protection Agency, USA, 14p.

Ersoy, B., Özeren, A., 2009. The effect of cooking methods on the mineral andvitamin content of African catfish. Food Chem. 115, 419–422.

Ersoy, B., Yanar, Y., Küçükgülmez, A., Çelik, M., 2006. Effects of four cookingmethods on the heavy metal concentration of sea bass fillets (Dicentrarchuslabrax Linnaeus, 1758). Food Chem. 99, 748–751.

FAO, 1989. Yield and Nutritional Value of the Commercially More Important FishSpecies. FAO Fisheries Technical Paper 309 PART I, pp. 1–187.

FAO, 2007. FAO Yearbook. Fishery Statistics. Commodities/FAO, vol. 101. Food andAgriculture Organization of the United Nations, Rome, Italy, 237p.

FAO, 2010. The State of World Fisheries and Aquaculture 2010 (SOFIA). Food andAgriculture Organization of the United Nations, Rome.

FAO/WHO, 2010a. Joint FAO/WHO Expert Committee on Food Additives. Seventy-second meeting, Rome, 16–25 February 2010. Summary and Conclusions.Issued 16th March 2010. <http://www.who.int/foodsafety/chem/summary72_rev.pdf>.

FAO/WHO, 2010b. Joint FAO/WHO Expert Committee on Food Additives. Seventy-third meeting, Geneva, 8–17 June 2010. Summary and Conclusions (Issued24.06.10).

Griffiths, M.H., Heemstra, P.C., 1995. A contribution to the taxonomy of the marinefish genus Argyrosomus (Perciformes: Sciaenidae), with description of two newspecies from Southern Africa. Smithiana Bull. 65, 1–40.

Grigorakis, K., Fountoulaki, E., Vasilaki, A., Mittakos, I., Nathanailides, C., 2011. Lipidquality and filleting yield of reared meagre (Argyrosomus regius). Int. J. Food Sci.Technol. 46, 711–716.

Grubbs, F., 1969. Procedures for deleting outlying observations in samples.Technometric 11, 1–21.

Hernández, M.D., López, M.B., Álvarez, A., Ferrandini, E., García Garcia, B., Garrido,M.D., 2009. Sensory, physical, chemical and microbiological changes inaquacultured meagre (Argyrosomus regius) fillets during ice storage. FoodChem. 24, 237–245.

Hoffmann, K., Boeingand, H., Dufour, A., Volatier, J.L., Telman, J., Virtanen, M.,Becker, W., Henauw, S.D., 2002. Estimating the distribution of usual dietaryintake by short-term measurements. Eur. J. Clin. Nutr. 56, 53–62.

Holmes, J.S., Simons, J.R., 1996. The Introduction and Naturalisation of Birds. HMSO,London, UK.

IOM, 2012. Dietary Reference Intakes. Recommended Intakes for Individuals.National Academy of Sciences, Institute of Medicine, Food and NutritionBoard. <http://www.iom.edu/Activities/Nutrition/SummaryDRIs/~/media/Files/Activity%20Files/Nutrition/DRIs/5_Summary%20Table%20Tables%201-4.pdf>.

IUCN, (International Union for the Conservation of Nature) 1987. The IUCN PositionStatement on Translocation of Living Organisms. IUCN, Gland, Switzerland.

Jorhem, L., 2000. Determination of metals in food by atomic absorption spectrometryafter dry ashing: NMKL collaborative study. J. AOAC Int. 83, 1204–1211.

Lall, S.P., 1995. Macro and trace elements in fish and shellfish. In: Ruiter, A. (Ed.),Fish and fishery products. Composition, Nutritive Properties and Stability. CabInternational, Wallingford, UK, pp. 187–213.

Page 9

S. Costa et al. / Food and Chemical Toxicology 60 (2013) 277–285 285

Lepage, G., Roy, C.C., 1986. Direct transesterification of all classes of lipids in a one-step reaction. J. Lipid Res. 27, 114–119.

Lourenço, H.M., Afonso, C., Anacleto, P., Martins, M.F., Nunes, M.L., Lino, A.R., 2012.Elemental composition of four farmed fish produced in Portugal. Int. J. Food Sci.Nutr. 67, 853–859.

Manchester, Sarah.J., Bullock, James.M., 2000. The impacts of non-native species onUK biodiversity and the effectiveness of control. J. Appl. Ecol. 37, 845–864.

Marimuthu, K., Thilaga, M., Kathiresan, S., Xacier, R., 2011. Effect of differentcooking treatments on the proximate and mineral composition of stripedsnakehead fish (Channa striatus, Bloch). J. Food Sci. Tech. (communication).http://dx.doi.org/10.1007/s13197-011-0418-9.

Monfort, M.C., 2010. Present market situation and prospects of meagre(Argyrosomus regius), as an emerging species in Mediterranean aquaculture.Studies and Reviews. General Fisheries Commission for the Mediterranean. No.89. Rome, FAO, pp. 28.

Moradi, Y., Bakar, J., Motalebi, A.A., Syed Muhamad, S.H., Che Man, Y., 2011. Areview on fish lipid: composition and changes during cooking methods. J.Aquat. Food Prod. Technol. 20, 379–390.

Nuray, E., Özden, Ö., 2007. Proximate composition and mineral contents in aquacultured sea bass (Dicentrarchus labrax), sea bream (Sparus aurata) analyzed byICP-MS. Food Chem. 102, 721–725.

Nusser, S., Carriquiry, A.L., Dodd, K., Fuller, W., 1996. A semiparametrictransformation approach to estimating usual intake distributions. J. Am. Stat.Assoc. 91, 1440–1449.

Perelló, G., Martí-Cid, R., Llobet, J.M., Domingo, J.L., 2008. Effects of various cookingprocesses on the concentration of arsenic, cadmium, mercury and lead in foods.J. Agric. Food Chem. 56, 11262–11269.

Portaria no. 402/2002 from 18th April 2002. Diário Da República N� 91 — I Série-B.Ministério Da Agricultura,Do Desenvolvimento Rural e das Pescas Lisboa,3851p. <http://www.dgrm.minagricultura.pt/xeo/attachfileu.jsp?look_parentBoui=894561&att_display=n&att_download=y>.

Quéméner, L., Suquet, M., Mero, D., Gaignon, J.L., 2002. Selection method of newcandidates for finfish aquaculture: the case of the French Atlantic, the Channeland the North Sea coasts. Aquat. Living Resour. 15, 293–302.

Saint-Denis, T., Goupy, J., 2004. Optimization of a nitrogen analyser based on theDumas method. Ana. Chim. Acta 515, 191–198.

Smedes, F., 1999. Determination of total lipid using non-chlorinated solvents.Analyst 124, 1711–1718.

Tressou, J., 2005. Méthodes statistiques pour l’évaluation du risque alimentaire.Paris: Ph.D. Thesis, Univ. Paris X.

Tressou, J., Crépet, A., Bertail, P., Feinberg, M.H., Leblanc, J.C., 2004. Probabilisticexposure assessment to food chemicals based on extreme value theory.Application to heavy metals from fish and sea products. Food Chem. Toxicol.42, 1349–1358.

Türkkan, A.U., Cakli, S., Kilinc, B., 2008. Effects of cooking methods on the proximatecomposition and fatty acid profile composition of seabass (Dicentrarchus labrax,Linnaeus, 1758). Food Bioprod. Process. 86, 163–166.

Ulbricht, T.L., Southgate, D.A.T., 1991. Coronary heart disease: seven dietary factors.Lancet 338 (8773), 985–992.

UKINC, (UK Committee for International Nature Conservation) 1979. WildlifeIntroductions to Great Britain. Nature Conservancy Council, Peterborough, UK.

Valfré, F., 2008. The control of risk in animal husbandry primary production. Vet.Res. Commun. 32, 3–9.

Wallace, L.A., Duan, N., Ziegenfus, R., 1994. Can long-term exposure distributions bepredicted from short-term measurements. Risk Anal. 14, 75–85.

Weber, J., Bochi, V.C., Ribeiro, Cristiane.P., Victório, A.M., Emanuelli, T., 2008. Effectof different cooking methods on the oxidation, proximate and fatty acidcomposition of silver catfish (Rhamdia quelen) fillets. Food Chem. 106, 140–146.

Weihrauch, J.L., Posati, L.P., Anderson, B.A., Exler, J., 1977. Lipid conversion factorsfor calculating fatty acids contents in foods. J. Am. Oil Chem. Soc. 54, 36–40.

Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice-Hall, Inc., Upper SaddleRiver, New Jersey, pp. 616–663.