Evaluation of cold-pressed flaxseed oil as an alternative dietary lipid source for juvenile sablefish (Anoplopoma fimbria)

Abstract

This study was undertaken to determine the suitability of using cold-pressed flaxseed oil (FO) as a major source of lipid in place of anchovy oil (AO) in the diet of juvenile sablefish (Anoplopoma fimbria), a relatively new marine species to aquaculture. Sablefish were fed one of four diets twice daily to satiation for 15 weeks. The test diets were identical in composition, except for the source of supplemental lipid which was either 100% AO (100AO), or increasing replacement of AO with FO i.e., 75AO:25FO, 50AO:50 FO or 25AO:75FO. Sablefish growth parameters, whole body and fillet proximate constituent concentrations and apparent digestibility coefficients were uninfluenced by diet treatment. There were also no adverse effects of the diet treatments on fish health, as determined from analysis of various haematological and innate immunological parameters. Terminal fillet fatty acid compositions generally reflected the dietary fatty acid compositions, while flesh contaminant concentration decreased with increasing dietary flaxseed oil content. Results indicated that FO may comprise up to 75% of the supplemental lipid in a grower diet for sablefish, while still providing humans with a rich dietary source of highly unsaturated fatty acids.

Introduction

In recent years, there has been interest in diversifying finfish culture in various regions of the world to include economically valuable marine finfish species. In Canada, farmed salmon represent about 90% of total finfish production (Statistics Canada 2008) and the farming of marine finfish species could expand traditional markets. The sablefish (Anoplopoma fimbria) is a marine species found in the Pacific Ocean that has a range extending from Mexico to Alaska and from the West coast of North America to Japan. Juvenile sablefish inhabit surface waters, and as they approach weights of 1 kg, they seek water depths of up to 1500 m. Sablefish, more commonly known as black cod, is a highly valued finfish due to its high oil content and white flaky flesh. Japan is currently the world's largest importer of sablefish. However, there have been increased demands for sablefish in other Asian markets e.g., Korea and Hong Kong, and in North America as a delicacy item (Sonu 2000). Sablefish is a suitable alternative to the overfished Patagonian toothfish, Dissostichus eleginoides, which is also known in North America as the Chilean sea bass.

Research aimed at assessing the potential of sablefish as a species for intensive culture began in the late 1960's. Kennedy (1972), for example, indicated that sablefish can withstand high rearing densities and low dissolved oxygen concentrations. Juvenile sablefish have been found to have one of the fastest recorded growth rates of all teleost species (Shenker & Olla 1986; Sogard & Olla 2001), and they have also been found to grow well in marine net pens (Gores & Prentice 1984). While early research indicated great potential for sablefish aquaculture, the industry was limited mainly by the inability to produce marine fish larvae. In 1998, the first marine larvae were successfully reared from eggs (Clarke, Jensen, Klimek & Pakula 1999). There are now several commercial sablefish farms in operation; however, there is a lack of published literature on the nutritional requirements of sablefish. Feed represents the single largest operational expense in finfish aquaculture, and can account for 35–60% of the costs of fish production. Sablefish grow well on Atlantic salmon (Salmo salar) feeds that are based extensively on fishmeal and fish oil (Sogard & Olla 2001; Minkoff & Clarke 2003). Fish oil is a rich source of omega-3 (n-3) highly unsaturated fatty acids (n-3 HUFAs), especially eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), which are known to be essential fatty acids for the growth and health of marine finfish species (Kanazawa, Teshima & Ono 1979; Higgs & Dong 2000). However, the global supply of fish oil is finite, and rising demands for this commodity is escalating the price of fish oil. To improve the profitability of sablefish farming, there is a present need to develop more cost effective diet formulations.

Vegetable oils offer a good alternative to fish oil due to their higher availability and lower prices particularly in recent years and during major El Niño events off coastal South America. Use of vegetable oils can also be beneficial in lowering flesh levels of persistent organic pollutants (POPs) (Isosaari, Vartiainen, Hallikainen & Ruohonen 2002; Karl, Kuhlmann & Ruoff 2003; Serrano, Simal-Julian, Pitarch, Hernandez, Varo & Navarro 2003; Carline, Barry & Ketola 2004; Lundebye, Berntssen, Lie, Ritchie, Isosaari, Kiviranta & Vartiainen 2004). In aquaculture feeds, marine fish oils and, to a lesser extent, the residual, are considered to be the largest source of (POPs), and concentrations of polychlorinated biphenyls (PCBs) and polychlorinated dibenzo-p-dioxin/dibenzofurans (PCDD/Fs) are directly related to their levels in the feed. (Jacobs, Covaci & Schepens 2002).

Use of vegetable oils in marine fish feeds does, however, present a few challenges. These oils are rich in C18 polyunsaturated fatty acids, and they lack the n-3 HUFAs that are characteristically high in fish oil. As the fatty acid composition of edible fish flesh reflects that of the dietary lipid source (Dosanjh, Higgs, McKenzie, Randall, Eales, Rowshandeli, Rowshandeli & Deacon 1998; Higgs, Balfry, Oakes, Rowshandeli, Skura & Deacon 2006), changes in the composition of flesh fatty acid concentrations may be a concern from a human health perspective. EPA and DHA are important fatty acids for cardiovascular health, neural and ocular development, cognitive function, and prevention of various inflammatory conditions and types of cancer (Connor 2000; Shahidi & Miraliakbari 2004; Mozaffarian & Rimm 2006; Narayan, Miyashita & Hosakawa 2006).

As mentioned previously, marine fish have a nutritional requirement for n-3 HUFA's, and they have little or no ability to desaturate and elongate the parent acid of the n-3 family of fatty acids viz., linolenic acid (18:3n-3) to EPA (Mourente & Dick 2002; Rodriquez, Perez & Henderson 2002; Bell, Strachan, Good & Tocher 2006). Thus, the extent to which marine fish oil can be replaced by vegetable oil in the diet of a marine finfish species must be determined carefully to avoid negative impacts on fish growth and health. High substitution levels of flaxseed oil (FO) for fish oil in diets for several marine finfish species have resulted in reductions in their growth performance due to deficiencies in n-3 HUFAs (Bell, Tocher, Farndale, McVicar & Sargent 1999; Regost, Arzel, Robin, Rosenlund & Kaushik 2003; Izquierdo, Montero, Robaina, Caballero, Rosenlund & Ginés 2005; Montero, Robaina, Caballero, Gines & Izquierdo 2005). Also, the chronic feeding of diets based extensively on vegetable oils from single sources may compromise the immune system or alter the stress response of marine finfish species (Montero, Kalinowski, Obach, Robaina, Tort, Caballero & Izquierdo 2003; Ganga, Bell, Montero, Robaina, Caballero & Izquierdo 2005; Mourente, Good & Bell 2005a). Depression of immunocompetence in the case of excessive dietary concentrations of vegetable oils in the preceding studies with attendant deficiencies of n-3 HUFAs and possibly arachidonic acid (20:4n-6) have included alterations in the types and levels of eicosanoid compounds elaborated, and reductions in serum alternative complement pathway activity and head kidney phagocytic activity. Extreme cases of dietary essential fatty acid deficiency in juvenile turbot may also be accompanied using liver histopathology (Bell et al. 1999). Moreover, this situation may lead to depressions of neutrophil activity, the number of circulating lymphocytes, serum complement activity and erythrocyte volume as well as abnormal renal morphology in juvenile gilthead seabream (Montero, Socorro, Tort, Caballero, Robaina, Vergara & Izquierdo 2004).

At present, there is no knowledge of the n-3 HUFA requirements for sablefish, and it is assumed, because of good growth responses of sablefish fed salmon diets that are based upon fishmeal and oil, that the fatty acid composition of fish oil meets the needs of this species for essential fatty acids. However, it is unknown to what extent the fish oil concentration in diets for sablefish can be reduced through the use of a vegetable oil such as FO before there are adverse consequences on the growth performance and health of the fish. Hence, the objective in the present study was to determine how much supplemental anchovy oil (AO) could be replaced by cold-pressed FO in a diet for juvenile sablefish without compromising their growth, feed efficiency, proximate composition and health. In addition, as sablefish have the genetic propensity for the accumulation of lipids in their flesh and consequently lipophilic contaminants, this study examined the potential of the FO dietary treatments to reduce flesh POP concentrations while examining from a flesh quality perspective, the effect of the dissimilar dietary fatty acid compositions on the terminal fatty acid compositions of the muscle lipids.

Materials and methods

Fish husbandry

A total of 325 unvaccinated juvenile sablefish were purchased from Cluxewe Enterprises (Cedar, BC, Canada). Subsequently, the fish were transported to the Department of Fisheries and Oceans/University of British Columbia, Centre for Aquaculture and Environmental Research (CAER), West Vancouver, BC, Canada, where they were acclimated for 7 weeks to experimental conditions and fed a commercial non-pigmented salmon feed. Sablefish (23 per tank) were randomly distributed into one of 12 indoor 1100 L fibreglass tanks (range in mean initial weight, 153.4–155.6 g). The fish in each tank were subjected to a natural photoperiod (daylight fluorescent lights) and were provided with running (11–15 L min⁻¹), filtered and oxygenated sea water. During the experiment (November–March), the temperature, dissolved oxygen concentration and salinity of the sea water were measured daily at 12:00 h and these parameters ranged from 7.7°C to 11.1°C, 7.5 mg L⁻¹to 10.4 mg L⁻¹ and 28 L⁻¹ to 31 g L⁻¹ respectively.

Experimental diets and feeding protocol

Four highly palatable diets were prepared to contain 45% crude protein and 20% lipid. The basal diet was prepared at the beginning of the study and steam pelleted into both 4 and 6 mm pellets as described by Higgs, Market, MacQuarrie, McBride, Dosanjh, Nichols and Hoskins (1979). All diets were identical in composition except for the source of supplemental lipid, which, in all cases, provided 67.2% of the dietary lipid content or 134.4 g kg⁻¹ diet (Table 1). The control diet was supplemented with 100% South American AO (100AO), whereas the three experimental diets were supplemented with either 75% AO and 25% cold-pressed FO (75AO:25FO), an equal mixture of both lipid sources (50AO:50FO), or 25% AO and 75% FO (25AO:75FO). The FO was produced at CAER by cold pressing (Gusta 1 HP Model 11 laboratory-scale cold press equipped with a 7 mm die; Gusta Cold Press, St Andrews, MB, Canada) whole brown flaxseeds furnished using InfraReady Products, Saskatoon, SK, Canada. Thereafter, the oil was stabilized with 500 ppm of ethoxyquin (final concentration) and stored in 4 L brown bottles under nitrogen at 15°C. Each of the aforementioned sources of supplemental lipid was sprayed onto batches of the common pelleted base diet using an electrically operated sprayer and a cement mixer, and then each of the diets was stored at 4°C in an air-tight container. Chromic oxide (0.5% final concentration) was added to all the diets as an indigestible marker.

Table 1. Ingredients, proximate composition, gross energy content and contaminant concentration of the test diets

Ingredient (g kg^−1 dry weight)	Diet
	100AO	75AO:25FO	50AO:50FO	25AO:75FO
Proximate composition values (n = 3) are mean ± SD. ^a Supplemental levels of vitamins (amounts kg^−1 dry diet) were: vitamin A acetate, 5000 IU; vitamin D3, 2400 IU; vitamin E, 300 IU; inositol, 400.0 mg; niacin, 300.0 mg; pantothenate (as D-calcium pantothenate), 165.0 mg; riboflavin, 60.0 mg; pyridoxine (as pyridoxine HCl), 40.0 mg; thiamine (as thiamine mononitrate), 50.0 mg; menadione (as MSBC), 18.0 mg; folic acid, 15.0 mg; biotin, 1.5 mg; vitamin B12, 0.09 mg; and BHT, 22mg. ^b Supplemental levels of minerals (mg kg^−1 dry diet) were: potassium (as K2SO4 and K2CO3, 1:1), 2110; magnesium (as MgSO4·7H2O), 250; zinc (as ZnSO4·7H2O), 100; iron (as FeSO4·7H2O), 100; manganese (as MnSO4·H2O), 75; iodine (as KIO3 and KI, 1:1), 10.0; copper (as CuSO4·5H2O), 5.0; fluorine (as NaF), 5.0; cobalt (as CoCl2·6H2O), 3.0; and selenium (as Na2SeO3), 0.1. ^c Anchovy oil stabilized with 200–250 ppm BHA. ^d Flaxseed oil stabilized with 500 ppm ethoxyquin.
LT-anchovy meal	389.7	389.7	389.7	389.7
Blood flour; spray-dried	39.1	39.1	39.1	39.1
Squid meal	57.0	57.0	57.0	57.0
Krill meal	77.7	77.7	77.7	77.7
Wheat gluten meal	56.8	56.8	56.8	56.8
Wheat starch; pregelatinized	100.0	100.0	100.0	100.0
Vitamin supplement a	20.0	20.0	20.0	20.0
Mineral supplement b	30.0	30.0	30.0	30.0
Anchovy oilc	134.4	100.8	67.2	33.6
Flaxseed oil; cold pressed d	0.0	33.6	67.2	100.8
Soybean lecithin	10.0	10.0	10.0	10.0
Choline chloride (60%)	5.0	5.0	5.0	5.0
Vitamin C monophosphate (42%)	3.6	3.6	3.6	3.6
α-cellulose	59.8	59.8	59.8	59.8
Permapell (lignin sulphonate binder)	10.0	10.0	10.0	10.0
DL-methionine	2.0	2.0	2.0	2.0
Chromic oxide	5.0	5.0	5.0	5.0
Proximate constituents (g kg^−1 dry weight)
Dry matter	927 ± 1.8	927 ± 2.3	929 ± 2.2	929 ± 2.7
Protein	460 ± 8.1	460 ± 3.8	460 ± 5.4	460 ± 4.9
Lipid	196 ± 4.7	204 ± 5.1	197 ± 6.3	197 ± 5.2
Ash	95.8 ± 0.5	95.9 ± 0.5	95.9 ± 0.6	96.0 ± 0.5
Contaminant loadings
Total PCBs (ng g^−1)	23.63	19.54	13.90	10.52
Total PCDD/Fs (pg g^−1)	4.01	4.27	3.30	3.38
PCB TEQ (pg g^−1)	0.62	0.55	0.48	0.42
PCDD TEQ (pg g^−1)	0.03	0.02	0.02	0.01
PCDF TEQ (pg g^−1)	0.15	0.07	0.05	0.06
Total TEQ (pg g^−1)	0.79	0.63	0.55	0.48

Each of the four diets was fed to triplicate groups of fish using a randomized complete block design. Each group was fed their prescribed diet by hand twice daily to satiation (beginning at 08:00 and 13:00 hours). The fish were fed for approximately 20–30 min and then left for 20 min to ensure that the fish had time to consume pellets that had fallen to the bottom of the tank. At this point, all uneaten pellets were siphoned off the bottom of the tank into mesh buckets, and counted. Accurate estimates of the daily ration consumed by each group were subsequently derived by deducting the weight of the uneaten feed (number of pellets × mean pellet air-dry weight) from the total daily feed dispensed in each case.

Fish handling

Following 18 h of starvation, individual fish in each group were weighed and measured (fork length) on day 0 and every 5 weeks thereafter using a dual anaesthetic treatment. Clove oil (0.5 ppm; Hill Tech Canada Inc., Vankleek Hill, ON, Canada) was used to sedate the fish in their rearing tanks, immediately prior to their removal for sampling. The fish were then fully anaesthesized using 150 ppm tricainemethanesulphonate (MS 222; Syndel Laboratories, Vancouver, BC, Canada). To protect the fish from scale loss, a water conditioner (Vidalife, Syndel International Inc., Qualicum Beach, BC, Canada) was used on all surfaces that came into contact with the fish and this was added to the anaesthetic bath (50 ppm). Excess moisture was removed from each fish using an absorbent cloth before weighing. Following weight and length measurements, each fish was placed into an aerated recovery bath, and the entire group was returned to its respective experimental tank.

Sampling

On day 0 of the feeding trial, 12 fish, common to all groups were killed using a lethal dose of MS 222 (>1000 ppm) for determination of whole body proximate analysis (n = 6), fillet proximate composition and fatty acid analysis (n = 6, right fillets, skinned) and contaminant analysis (n = 3, six fish, composites of two left fillets, skinned). On day 105, fish from each tank were selected randomly and killed with a swift blow to the head for measurement of whole body proximate analysis (n = 5), fillet proximate analysis and fatty acid composition (n = 5, right fillets, skinned), contaminant analysis (n = 3, six fish, three composites of two left fillets, skinned) and various haematological and innate immunological parameters (n = 6, the same fish used for fillet proximate analysis and contaminant analysis).

To avoid cross contamination during sampling, each fish was handled with a fresh pair of nitrile gloves and filleted using solvent-rinsed equipment on hexane-rinsed aluminium foil. The fillet composite samples for contaminant analysis were wrapped in hexane-rinsed foil with the skin still on, and then, these were placed into contaminant-free bags and stored at –20°C pending analysis. Feed and oil samples for contaminant analysis were collected in solvent-rinsed glass Pyrex cups that were also stored at –20°C. Samples were frozen and transported on ice to the Institute of Ocean Sciences, Sidney BC, where the composite samples were homogenized using a Sorvall Omnimixer after removal of the skin, fins, belly flap and kidney. Knives and homogenization equipment were washed and solvent-rinsed between samples to prevent cross contamination. Details on the methodologies used for these measurements have been described previously (Ikonomou, Higgs, Gibbs, Oakes, Skura, McKinley, Balfry, Jones, Withler & Dubetz 2007).

Fish samples as well as diet samples for nutritional analysis were placed into 20.3 by 25.4 cm gold deli bags (Oxygen transmission, 0.7 cc m⁻² in 24 h at 23°C dry; West Coast FoodPak Systems) that were vacuum-sealed and immediately stored at –20°C until analysis. Faeces samples were also collected from the last 3 cm of the hindgut by dissection on day 105. Samples from each tank were pooled to obtain three faeces samples per diet treatment.

Chemical analysis

Samples from all the experimental diets were finely ground using a coffee grinder before analysis. Fish samples (fillets and whole bodies) were thawed overnight at 4°C prior to homogenization in a blender (Braun Type 3210-325; Braun, Cincinnati, OH, USA). Partially digested feed and faeces were removed from the intestinal tract of whole body fish prior to homogenization. Whole body, skinned fillets and feed were analysed in duplicate for proximate analysis according to the procedures of Higgs et al. (2006). A portion of the lipid/chloroform layer resulting from lipid extraction of each sample prepared according to Bligh and Dyer (1959) was collected and stored at –80°C in a 10 mL glass vial for subsequent determination of fatty acid composition. Fatty acid methyl esters (FAMEs) were obtained from concentrated lipid samples using base-catalysed transesterification (Christie 1973), and stored in 2 mL gas chromatography (GC) vials (Varian, Santa Clara, CA, USA) prior to GC analysis. Separation and analysis of FAMES were conducted using a Varian model 3400 GC equipped with a flame ionization detector and CP-Sil 88 fused silica column (Varian). The GC injector and detector temperatures were set at 250°C, and helium was used as the carrier gas. The column was initially set at a temperature of 60°C, and was raised to 160°C at a rate of 15°C min⁻¹. FAMEs were then eluted as the column increased in temperature at a rate of 4°C min⁻¹ to 220°C. The column was held at this final temperature for 15 min for a total run time of 38 min per sample. Individual FAME peaks were identified using external standards (FAME mix 37, and other individual standards; Supelco Inc., St. Louis, MO, USA), and concentrations were calculated as a percentage of the sum of the total identifiable fatty acids.

To assess the extent of dietary lipid oxidation, thiobarbituric acid reactive substances (TBARs; secondary products of lipid peroxidation) were measured in all four test diets following the completion of the feeding trial according the methods of Tarladgis, Watts and Younathan (1960) as modified by Sutton, Balfry, Higgs, Huang and Skura (2006).

Faecal samples collected by dissection were frozen at –20°C and freeze-dried. Samples were homogenized before analysis by grinding them using a mortar and pestle. Levels of chromic oxide in the diets and faeces were measured according to Fenton and Fenton (1979) and the moisture, protein and ash contents of the faeces were analysed according to Higgs et al. (2006). The gross energy contents of the diets and faecal samples were determined using adiabatic bomb calorimetry (IKA Calorimeter System C5000 duo control, IKA-WERKE, Staufen, Germany).

Contaminant analysis

Both oil samples, all four dietary treatments and 36 skinless fillet composites were analysed for PCDD/Fs and PCBs using gas chromatography/high resolution mass spectrometry (GC/HRMS)-based analytical methodology, and the details of these procedures have been provided elsewhere (Ikonomou et al. 2007; Ikonomou, Fraser, Crewe, Fischer, Rogers, He, Sather & Lamb 2001). The quality assurance/quality control (QA/QC) protocols used for the identification and quantification of PCDD/Fs were those of Environment Canada (Environment Canada 1992). The corresponding QA/QC protocols were also used for the identification and quantification of all the PCB congeners. Results were not blank subtracted, and congeners whose concentrations were below the limit of quantification were treated as zero. Levels of contaminants in the flesh of sablefish were compared both as the sum of the individual congeners and as toxic equivalents (TEQs) (Van den Berg, Birnbaum, Bosveld, Brunstrom, Cook, Feeley, Giesy, Hanberg, Hasegawa, Kennedy, Kubiak, Larsen, van Leeuwen, Liem, Nolt, Peterson, Poellinger, Safe, Schrenk, Tillitt, Tysklind, Younes, Waern & Zacharewski 1998). The results were also examined both on a wet basis (contaminant content per gram of wet sample) and on a lipid-corrected basis. To calculate the lipid-corrected results, the contaminant content determined for each sample was divided by the respective lipid content measured in each sample.

Fish health

On day 105, health assessments were conducted on six fish from each replicate group per diet treatment. Fish were killed in MS222 (as described above), and blood samples were taken from the caudal vessels using a sterile 22G syringe. Haematological measurements were performed using standard haematological methods as outlined by Klontz (1994). Two heparanized capillary tubes were filled with fresh blood from each fish. These tubes were kept chill for later measurement of haematocrit values and collection of peripheral blood leucocytes (PBLs) for the respiratory burst assay. Erythrocyte numbers were determined by placing 5 μL of whole blood into 995 μL of Hendrick's solution (Houston 1990) and later, by microscopic counts of the diluted blood (1:200) using a hemacytometer. Blood films were prepared from the fresh blood by smearing 5 μL of whole blood onto cleaned glass microscope slides. These films were air-dried and stained using a modified Wright-Giemsa stain. The slides were systematically examined microscopically under oil immersion (×1000), and the number of neutrophils, thrombocytes, lymphocytes, monocytes and erythrocytes were counted in ten fields. Differential leucocyte numbers were calculated using the ratio of each type of leucocyte: erythrocyte, and multiplying by the actual erythrocyte numbers determined from the hemacytometer counts. The remaining whole blood in the syringe was placed into a sterile tube, centrifuged (3000 g, 5 min, 5°C) and the plasma stored at –70°C for later measurements of lysozyme activity using a modification (modified for analysis using microtitre plates) of the method outlined by Litwack (1955).

Immediately after the haematocrit values were determined for each fish, the PBLs were collected from the leukocrit layer, by breaking the tubes just below this layer, and using a microcap, the PBLs and plasma were dispensed onto welled microscope slides. The PBLs from duplicate haematocrit tubes were combined into a single well for each fish. The respiratory burst activity of PBLs was determined using the nitroblue tetrazolium (NBT) assay described by Anderson, Moritomo and de Grooth (1992). Prepared slides were examined microscopically under oil immersion (×1000). The proportion of cells with blue halos associated with them, was considered to be NBT positive. The blue halos were the result of the NBT being reduced to the blue formazan due to the presence of O^2– (a byproduct of the respiratory burst reaction). A total of 200 cells were examined, and the per cent of NBT positive cells was calculated.

Data and statistical analyses

The effect of diet treatment on the growth performance of the fish was assessed by the following:

1. Wet weight gain (WG) (g) = (final mean wet weight (FW) (g) – initial mean wet weight (IW) (g))
2. Specific growth rate (SGR) (% body weight day⁻¹) = [(ln FW (g) – ln IW (g)) × time⁻¹ (days)] × 100
3. Dry feed intake (DFI) (g × fish⁻¹ × day⁻¹) = mean daily dry feed intake x fish⁻¹ over 105 days
4. Feed efficiency (FE) (g g⁻¹) = WG (g) × DFI_TOT (g fish⁻¹) where DFI_TOT is the total dry feed intake × fish⁻¹ consumed over 105 days
5. Protein efficiency ratio (PER) (g g⁻¹) = WG (g) × protein intake⁻¹ (g)
6. Per cent protein deposited (PPD) (%) = protein gain (g) × 100 protein intake⁻¹ (g)
7. Survival (%) = (number of fish in each group remaining on day 105 initial number of fish) × 100
8. Hepatosomatic index (HSI) = liver weight (g) × 100 × fish weight⁻¹ (g)
9. Condition factor (K) = fish weight (g) × 100 × fork length⁻¹ (cm)³
10. Apparent digestibility coefficients (ADC) (%) = [1–(F × D⁻¹ × Dcr × Fcr⁻¹)] × 100 were calculated for dietary protein (ADC_p), energy (ADC_en) and organic matter (ADC_orm), where F = % nutrient (p or orm) or energy content (MJ g⁻¹) of faeces, D = % nutrient (p or orm) or energy content (MJ g⁻¹) of diet, Dcr = % chromic oxide in diet and Fcr = % chromic oxide in faeces (Cho, Cowey & Watanabe 1985).
11. Concentrations (% of wet weight) of proximate constituents (protein, lipid, moisture and ash) in the whole body and fillet of the fish in relation to diet treatment.

Accumulation efficiencies (% retentions) for PCDDs, PCDFs and dioxin-like PCBs were calculated in two ways, as POP levels were measured only in the edible flesh rather than the whole body. The first method assumed that the fillet contaminant concentration would be similar to the whole body contaminant concentration in each case, and calculated AE_WB (whole body) using the equation:

• 

Where Wt was the average fish weight at the beginning _(i) and end _(f) of the 15-week trial in each tank, C was the concentration of the contaminant in the feed or fish (in this case fillet) (pg g⁻¹ wet wt) and Feed _(in) was the average total amount of feed consumed per fish (g wet wt) in each tank during the 15-week feeding trial. Accumulation efficiencies were also calculated as the per cent retention of the contaminants in the sablefish muscle . The fillet yield expressed as a percentage of body weight was calculated to be 40% and this value was similar to that determined previously for sablefish i.e., 41.5% (Thurston 1961). Consequently, the above equation was modified to correct for fillet yield:

• 

The results for each of the preceding parameters were analysed using randomized block ANOVA. Percentage data (e.g., proximate components, individual fatty acids, leucocyte respiratory burst activity, haematocrit) were arcsine square root-transformed, and the count data (e.g., leucocyte and erythrocyte numbers) were square root-transformed to achieve normalized distribution of the data and homogeneity of variance before statistical analysis. Tukey's test with P = 0.05 was used to detect significant differences among means, where appropriate.

Results

Diet composition

All test diets had almost identical concentrations of proximate constituents and gross energy content (Table 1). Contaminant levels in the feeds reflected the inclusion levels of the supplemental oil sources used in the experimental feeds. With regard to the two oil sources, the AO contained 130.8 ng g⁻¹ PCBs, 11.3 pg g⁻¹ PCDD/Fs and 3.75 pg g⁻¹ Toxic Equivalence (WHO_SUM-TEQ), and the FO had 9–75 times lower levels of contaminants. It contained 2.45 ng g⁻¹ PCBs, 1.2 pg g⁻¹ PCDD/F and 0.05 pg g⁻¹ WHO_SUM-TEQ. As the majority of lipid in the experimental fish feeds originated from the supplemental lipid, the measured levels of PCBs and PCDD/Fs in the diets decreased as more supplemental AO was replaced with FO (Table 1). PCDD/F concentrations in the feed recorded very low levels and were similar between diet treatments, ranging from 4.27 to 3.30 pg g⁻¹. These concentrations are close to laboratory background levels, and reflect concentrations measured in the procedural blanks, where concentrations of individual PCDD/F congeners ranged from 0.06 to 0.65 pg g⁻¹.

Similar to contaminant concentrations, the fatty acid composition of sablefish diets reflected the different inclusion levels of AO and cold-pressed FO and their respective fatty acid compositions (Table 2). Flaxseed oil had higher concentrations of 18:1n-9 (oleic acid), 18:2n-6, 18:3n-3 and totals for n-6 and n-3 fatty acids, and also had lower concentrations of 14:0, 16:0, total saturated fatty acids and 16:1n-7 as well as an absence of 18:4n-3, 20:5n-3, 22:5n-3, 22:6n-3, n-3 HUFAs and 20:4n-6. Higher inclusion of FO increased the dietary levels of polyunsaturated fatty acids, which often can increase the likelihood of dietary lipid oxidation. However, all diets at the end of the study had minimal levels of lipid peroxidation, as TBAR values were less than 10 μ moles kg⁻¹ in every case.

Table 2. Fatty acid composition (% of total fatty acids) of the lipid sources and the test diets

Fatty acid	Supplemental lipid sources		Diet
	AO	FO	100AO	75AO:25FO	50AO:50FO	25AO:75FO
Values (n = 3) are mean ± sd. AO, anchovy oil; FO, flaxseed oil.
12:0	0.17	0.00	0.15 ± 0.02	0.10 ± 0.01	0.09 ± 0.01	0.06 ± 0.00
14:0	9.27	0.06	6.99 ± 0.63	5.43 ± 0.11	4.51 ± 0.08	2.95 ± 0.04
15:0	0.65	0.02	0.57 ± 0.03	0.46 ± 0.02	0.38 ± 0.01	0.28 ± 0.02
16:0	20.43	5.27	20.63 ± 0.31	17.81 ± 1.28	15.04 ± 0.49	12.96 ± 1.29
17:0	0.98	0.00	0.90 ± 0.02	0.77 ± 0.08	0.64 ± 0.02	0.50 ± 0.03
18:0	3.65	3.27	3.93 ± 0.12	4.06 ± 0.55	3.66 ± 0.04	3.79 ± 0.28
20:0	0.31	0.21	0.25 ± 0.07	0.19 ± 0.12	0.15 ± 0.03	0.06 ± 0.02
Σ Saturated	35.47	8.84	33.42 ± 0.43	28.81 ± 1.16	24.48 ± 0.58	20.60 ± 1.21
16:1n-7	10.31	0.37	7.72 ± 0.74	5.98 ± 0.14	4.96 ± 0.06	3.13 ± 0.09
18:1n-9	9.11	22.20	9.48 ± 0.30	12.18 ± 1.35	14.60 ± 0.42	16.86 ± 1.46
18:1n-7	2.72	0.00	2.69 ± 0.13	2.07 ± 0.58	1.13 ± 0.05	0.81 ± 0.70
20:1n-9	1.04	0.00	1.80 ± 0.40	0.75 ± 0.09	0.54 ± 0.02	0.25 ± 0.13
22:1	1.75	0.00	1.43 ± 0.04	1 .07 ± 0.08	0.87 ± 0.01	0.60 ± 0.11
24:1n-9	0.38	0.00	0.16 ± 0.01	0.20 ± 0.09	0.11 ± 0.01	0.14 ± 0.06
Σ mono-unsaturated	25.31	22.57	23.28 ± 0.90	22.26 ± 0.41	22.22 ± 0.41	21.78 ± 0.62
18:2n-6	1.45	13.55	6.53 ± 1.42	8.41 ± 0.27	9.55 ± 0.44	11.50 ± 0.60
18:3n-6	0.20	0.00	0.15 ± 0.03	0.12 ± 0.11	0.03 ± 0.04	0.02 ± 0.03
18:3n-3	1.07	55.04	1.08 ± 0.06	11.76 ± 4.20	21.54 ± 1.14	29.54 ± 3.62
18:4n-3	3.08	0.00	2.50 ± 0.29	2.04 ± 0.04	1.85 ± 0.26	1.19 ± 0.08
20:2n-6	0.25	0.00	0.17 ± 0.03	0.09 ± 0.03	0.07 ± 0.00	0.03 ± 0.01
20:3n-6	0.25	0.00	0.14 ± 0.01	0.10 ± 0.01	0.09 ± 0.00	0.07 ± 0.03
20:4n-6	0.86	0.00	0.62 ± 0.04	0.38 ± 0.10	0.42 ± 0.04	0.27 ± 0.03
20:4n-3	0.97	0.00	0.74 ± 0.07	0.51 ± 0.09	0.46 ± 0.02	0.29 ± 0.01
20:5n-3	16.74	0.00	13.98 ± 0.70	10.91 ± 0.81	8.44 ± 0.15	5.82 ± 0.44
22:4n-6	0.53	0.00	0.52 ± 0.02	0.44 ± 0.09	0.31 ± 0.04	0.24 ± 0.06
22:5n-3	2.43	0.00	1.89 ± 0.12	1.36 ± 0.18	1.21 ± 0.03	0.79 ± 0.06
22:6n-3	11.40	0.00	14.99 ± 1.18	12.83 ± 2.33	9.34 ± 0.52	7.87 ± 1.87
Σ n-3	35.69	55.04	35.17 ± 0.05	39.39 ± 0.70	42.83 ± 0.69	45.49 ± 1.36
Σ n-6	3.54	13.55	8.13 ± 1.36	9.54 ± 0.35	10.47 ± 0.41	12.13 ± 0.74
n-3 n-6^–1	10.07	4.06	4.41 ± 0.79	4.13 ± 0.22	4.09 ± 0.20	3.75 ± 0.33
Σ polyunsaturated	39.23	68.59	43.30 ± 1.29	48.94 ± 0.80	53.30 ± 0.69	57.62 ± 0.68
Σ n-3 HUFAS	28.15	0.00	28.97 ± 1.18	23.73 ± 2.10	17.78 ± 0.64	13.69 ± 2.31

Influence of diet treatment on fish performance, diet digestibility and fish and fillet composition

Diet treatment did not significantly affect values for WG, SGR, CF, DFI, FE, PER, PPD fillet and liver weights, and values for HSI during the 15-week study (Table 3). On average, the sablefish exhibited a 2.7-fold increase in average weight during the study irrespective of diet treatment, and there were no case of fish mortality in any of the treatments. Furthermore, apparent digestibility coefficients values were unaffected by diet treatment and ranged between 77.2 81.1% for ADC_p, 67.7–70.1% for ADC_orm and 77.4–79.2% for ADC_en. Terminal proximate compositions in whole bodies (12.4–12.6% protein, 21.2–22.5% lipid, 63.4–64.5% moisture and 1.84–1.90% ash) and fillets (14.5–14.9% protein, 18.3–19.7% lipid and 65.1–66.1% moisture and 1.28–1.31% ash) of sablefish were not significantly affected by diet treatment.

Table 3. IBW, FBW, WG, SGR, DFI, FE, PER, PPD, S, fillet weight, HSI and CF of juvenile sablefish fed experimental diets.

Performance parameters	Diet
	100AO	75AO:25FO	50AO:50FO	25AO:75FO
Values for each parameter (n = 3) are mean ± SD. No significant differences were found (P < 0.05). IBW, initial body weight; FBW, final body weight; WG, weight gain; SGR, specific growth rate; DFI, dry feed intake; FE, feed efficiency; PER, protein efficiency ratio; PPD, per cent protein deposited; S, survival; HSI, hepatosomatic index; CF, condition factor; AO, anchovy oil; FO, flaxseed oil.
IBW (g)	154.5 ± 0.97	154.4 ± 0.57	154.1 ± 0.56	154.5 ± 0.64
FBW (g)	424.8 ± 6.72	423.5 ± 27.0	420.9 ± 11.6	430.4 ± 15.0
WG (g)	270.3 ± 6.83	269.1 ± 26.9	266.9 ± 11.1	275.9 ± 15.1
SGR (% day^−1)	0.96 ± 0.02	0.96 ± 0.06	0.96 ± 0.02	0.98 ± 0.03
DFI (g fish^−1 day^−1)	3.33 ± 0.06	3.45 ± 0.42	3.27 ± 0.09	3.40 ± 0.09
FE (g g^−1)	0.77 ± 0.03	0.74 ± 0.02	0.78 ± 0.02	0.77 ± 0.03
PER (g g^−1)	1.68 ± 0.06	1.62 ± 0.04	1.69 ± 0.05	1.68 ± 0.06
PPD (%)	21.3 ± 0.70	20.0 ± 0.41	20.9 ± 0.68	20.5 ± 0.78
S (%)	100.0 ± 0.00	100.0 ± 0.00	100.0 ± 0.00	100.0 ± 0.00
Fillet weight (g)	79.4 ± 3.51	75.1 ± 12.1	83.2 ± 6.01	80.8 ± 9.46
Liver weight (g)	13.8 ± 0.25	12.9 ± 1.77	13.8 ± 0.38	13.6 ± 1.63
HSI (%)	3.19 ± 0.12	3.17 ± 0.07	3.17 ± 0.04	3.14 ± 0.16
CF	2.02 ± 0.06	2.00 ± 0.07	2.02 ± 0.08	2.07 ± 0.05

Influence of diet treatment on flesh contaminant concentrations

There were significant differences in sablefish flesh contaminant concentrations. Total PCB and PCDD concentrations significantly decreased from 20.75 ng g⁻¹ and 0.89 pg g⁻¹,respectively, for fish that ingested diet 100AO to 13.54 ng g⁻¹ and 0.14 pg g⁻¹,respectively, for fish that consumed diet 75FO (Table 4). Flesh PCDF concentrations were observed to decrease from 0.74 pg g⁻¹ in fish fed diet 100AO to 0.46 pg g⁻¹ in fish fed diet 75FO, but these differences were not found to be significant (P = 0.27). Total flesh TEQ concentrations were found to significantly decrease from 0.68 pg g⁻¹ WHO_SUM-TEQ for fish fed diet 100AO to 0.46 pg g⁻¹ WHO_SUM-TEQ for those fed diet 75FO (Table 4). The differences in flesh contaminant concentrations paralleled the differences in dietary contaminant concentrations, and in all cases, the flesh TEQ levels were noted to be at or slightly below the dietary TEQ concentrations. On a wet weight basis, significant differences in flesh TEQ only occurred between sablefish fed diet 100AO and those fed diet 75FO. However, on a lipid-corrected basis, significant differences were found between sablefish fed each of the four test diets (Table 4).

Table 4. Contaminant levels in sample blanks, pre-trial flesh levels and flesh levels of sablefish fed on experimental diets

	Sample Blanks	Day 0	100AO	75AO:25FO	50AO:50FO	25AO:75FO
^a Lipid corrected Toxic Equivalence. Values (n = 3) are mean ± SD. Values in the same line with different superscripts are significantly different (P < 0.05).
Total PCBs (ng g^−1)	0.17	16.59	20.75 ± 1.47^a	18.33 ± 1.74^ab	16.54 ± ^bc	13.54 ± ^c
Total PCDDs (pg g^−1)	0.22	0.22	0.89 ± 0.47^a	0.31 ± 0.15^ab	0.24 ± 0.13^b	0.14 ± 0.12^b
Total PCDFs (pg g^−1)	0.55	0.33	0.74 ± 0.34	0.58 ± 0.09	0.47 ± 0.02	0.46 ± 0.02
PCB TEQ (pg g^−1)	0.003	0.38	0.59 ± 0.04^a	0.54 ± 0.10^ab	0.49 ± 0.01^ab	0.41 ± 0.01^b
PCDD TEQ (fg g^−1)	1.0	1.0	1.9 ± 1.4	0.6 ± 0.7	0.2 ± 0.1	0.1 ± 0.1
PCDF TEQ (pg g^−1)	0.04	0.04	0.09 ± 0.06	0.07 ± 0.01	0.06 ± 0.00	0.05 ± 0.01
Total TEQ (pg g^−1)	0.04	0.42	0.68 ± 0.09^a	0.61 ± 0.09^ab	0.55 ± 0.01^ab	0.46 ± 0.01^b
Total TEQ (pg g lipid^−1)a		2.03	3.08 ± 0.29^a	2.67 ± 0.11^b	2.44 ± 0.21^c	2.14 ± 0.02^d

Individual sablefish in this study were highly variable in flesh lipid content (values ranged from 15.1% to 27.8%). Using ANCOVA, lipid content was shown to influence flesh TEQ levels (P < 0.0001) with every percentage point increase in lipid content raising flesh total TEQ by 0.024 pg g⁻¹ WHO_SUM-TEQ. Dietary FO concentration had no effect on the accumulation efficiencies of the different contaminants under investigation. However, there were differences in AE between PCBs and PCDD/Fs. PCBs were deposited at rates similar to the deposition of dietary lipids into sablefish flesh (Table 5), whereas AE values for PCDD/Fs were lower.

Table 5. Accumulation efficiencies % (AE) of organohalogens.

	AEWB (%)	AEFillet (%)
Values are mean (n = 12) ± SD in brackets. AEWB, Percentage of consumed contaminants deposited in the sablefish; AEFillet, Percentage of consumed contaminants that were deposited in the edible portion of the sablefish.
Lipid	85 (3)	33 (1)
PCBs	76 (7)	31 (3)
Dioxins	13 (14)	5 (6)
Furans	52 (13)	21 (5)
Total TEQ	78 (4)	31 (2)
Tetra PCBs	84 (11)	34 (4)
Penta PCBs	81 (7)	32 (3)
Hexa PCBs	71 (3)	28 (1)
Hepta PCBs	66 (3)	27 (1)
Octa PCBs	63 (6)	25 (2)
Nona PCBs	61 (9)	24 (4)
Deca PCB	52 (8)	21 (3)

Influence of diet treatment on terminal fillet fatty acid compositions

The terminal fatty acid compositions of the fillets generally reflected the fatty acid profiles of the diet treatments (Table 6). In this regard, examination of the eight fatty acids that were found to be present at the highest concentrations in the sablefish muscle was positively correlated with their respective dietary concentrations (Fig. 1). The Pearson R² values (except for one fatty acid viz., 18:1n-9) were found to be greater than 0.98. The slopes for each of the fatty acids shown in Fig. 1 were less than one. This indicates that reduced retention efficiency of each fatty acid in the fillet decreased as its dietary concentration (intake) was increased.

Figure 1. Relationship between fatty acids in the diet and in the sablefish fillet. A diagonal line in the centre of each graph represents equality between the concentration of the fatty acid of interest in the fillet and the diet treatment.

Table 6. Sablefish muscle fatty acid composition (% of total fatty acids)

Fatty acid	Diet
	Day 0	100AO	75AO:25FO	50AO:50FO	25AO:75FO
Values (n = 3) are mean ± SD. Values in the same line with different superscripts are significantly different (P < 0.05).
12:0	0.09	0.09 ± 0.01 a	0.08 ± 0.01 ab	0.06 ± 0.00 b	0.05 ± 0.00 c
14:0	4.04	4.96 ± 0.68 a	4.31 ± 0.47 ab	3.54 ± 0.33 bc	2.71 ± 0.48 c
15:0	0.33	0.40 ± 0.05 a	0.35 ± 0.03 ab	0.30 ± 0.03 bc	0.24 ± 0.03 c
16:0	16.87	17.56 ± 1.13 a	16.25 ± 0.80 ab	15.07 ± 0.86 bc	13.66 ± 1.13 c
17:0	0.61	0.35 ± 0.01 a	0.29 ± 0.04 ab	0.27 ± 0.02 bc	0.23 ± 0.00 c
18:0	3.91	3.60 ± 0.08	3.62 ± 0.10	3.64 ± 0.11	3.70 ± 0.06
20:0	0.18	0.21 ± 0.03 a	0.20 ± 0.02 ab	0.17 ± 0.02 ab	0.17 ± 0.01 b
Σ Saturated	26.03	27.18 ± 1.63 a	25.11 ± 1.24 ab	23.05 ± 1.18 bc	20.75 ± 1.60 c
16:1n-7	7.31	8.07 ± 0.80 a	6.91 ± 0.42 b	5.79 ± 0.34 c	4.61 ± 0.52 d
18:1n-9	31.28	26.76 ± 1.80	26.63 ± 0.77	27.85 ± 0.94	28.95 ± 0.77
18:1n-7	2.79	4.06 ± 0.40 a	3.83 ± 0.57 a	3.42 ± 0.57 ab	2.76 ± 0.73 b
20:1n-9	1.61	1.02 ± 0.06	0.97 ± 0.06	0.99 ± 0.03	0.93 ± 0.03
22:1	0.82	0.55 ± 0.01	0.57 ± 0.05	0.55 ± 0.09	0.61 ± 0.03
24:1n-9	0.02	0.38 ± 0.04 a	0.37 ± 0.04 a	0.31 ± 0.05 b	0.26 ± 0.02 c
Σ mono-unsaturated	43.82	40.84 ± 0.63 a	39.28 ± 0.19 ab	38.93 ± 1.32 ab	38.12 ± 0.39 b
18:2n-6	8.45	5.27 ± 0.18 d	6.36 ± 0.40 c	7.37 ± 0.06 b	8.48 ± 0.08 a
18:3n-6	1.67	0.27 ± 0.05	0.26 ± 0.03	0.25 ± 0.07	0.23 ± 0.07
18:3n-3	0.21	1.36 ± 0.07 d	6.52 ± 0.47 c	12.1 ± 0.36 b	17.6 ± 0.55 a
18:4n-3	1.24	1.69 ± 0.04 a	1.45 ± 0.01 b	1.21 ± 0.08 c	0.93 ± 0.05 d
20:2n-6	0.40	0.28 ± 0.03	0.26 ± 0.06	0.23 ± 0.02	0.20 ± 0.01
20:3n-6	0.20	0.23 ± 0.09	0.18 ± 0.07	0.19 ± 0.11	0.16 ± 0.06
20:4n-6	0.62	1.05 ± 0.09 a	0.97 ± 0.08 a	0.75 ± 0.06 b	0.60 ± 0.10 c
20:4n-3	0.85	0.90 ± 0.06 a	0.82 ± 0.05 ab	0.71 ± 0.08 b	0.58 ± 0.06 c
20:5n-3	7.44	9.01 ± 0.31 a	7.95 ± 0.37 b	6.27 ± 0.26 c	4.74 ± 0.25 c
22:4n-6	0.32	0.46 ± 0.03	0.37 ± 0.02	0.31 ± 0.09	0.30 ± 0.12
22:5n-3	1.89	2.14 ± 0.11 a	1.92 ± 0.15 ab	1.72 ± 0.25 b	1.27 ± 0.16 c
22:6n-3	6.85	9.33 ± 0.52 a	8.53 ± 0.63 a	6.89 ± 0.87 b	6.03 ± 0.75 b
Σ n-3	18.48	24.43 ± 1.04 c	27.20 ± 1.52 bc	28.92 ± 1.75 ab	31.16 ± 1.77 a
Σ n-6	11.67	7.56 ± 0.38 d	8.40 ± 0.35 c	9.10 ± 0.40 b	9.97 ± 0.23 a
n-3 n-6^-1	1.58	3.23 ± 0.04	3.24 ± 0.29	3.18 ± 0.07	3.12 ± 0.11
Σ polyunsaturated	30.16	31.98 ± 1.41 c	35.60 ± 1.29 bc	38.02 ± 2.13 ab	41.13 ± 1.99 a
Σ n-3 HUFAS	14.30	18.34 ± 0.79 a	16.49 ± 0.99 a	13.16 ± 1.12 b	10.78 ± 0.99 c

No evidence of bioconversion of 18:3n-3 to 20:5n-3 was observed for sablefish in this study, as the fillet concentrations of 18:4n-3 across treatments were lower than respective values observed for the diets. Also, the other metabolic derivatives of 18:3n-3 were each found to be inversely related to the dietary FO concentration. Therefore, a per cent decline was observed for flesh concentrations of n-3 HUFAs when fish fed the diet with highest FO concentration were compared with those fed diet 100AO. In contrast, the mean overall flesh concentrations of n-3 fatty acids and polyunsaturated fatty acids in the sablefish were found to bear a direct relationship to the dietary concentration of FO, mostly because of progressive increases in 18:3n-3, as more AO was replaced by FO in the supplemental dietary lipid. Interestingly, the ratios of n-3 to n-6 fatty acids in the flesh lipids were uninfluenced by diet treatment, and ranged from 3.12 to 3.24.

Influence of diet treatment on fish health

The replacement of marine fish oil with FO did not result in any adverse effects on the health of the sablefish reared under the experimental conditions described here. Table 7 presents the data for the various health assays used to assess fish health after the 105-day feeding trial was completed. There were no significant differences in the various health indicators between any of the diet treatments, and results for fish fed the experimental diets were similar to those detected for fish fed the control diet.

Table 7. Health parameters measured in juvenile sablefish fed the test diets

Parameter	Diet
	100AO	75AO:25FO	50AO:50FO	25AO:75FO
Values (n = 3) are mean ± SD. No significant diet effects were found (P > 0.05) for any of the health parameters.
Total no. leucocytes (× 10^4 mL^−1)	3.79 ± 0.57	3.29 ± 0.66	3.17 ± 0.27	3.63 ± 0.92
Plasma lysozyme activity (HEWL U mL^−1)	15.1 ± 1.3	15.8 ± 3.1	15.8 ± 0.26	14.6 ± 2.1
Total no. erythrocytes (× 10^6 mL^−1)	1.58 ± 0.16	1.59 ± 0.12	1.72 ± 0.22	1.63 ± 0.076
Mean erythrocyte volume (nm^3)	234 ± 24	237 ± 24	228 ± 39	244 ± 26
Haematocrit (% packed cell volume)	36.6 ± 1.38	36.6 ± 2.48	37.5 ± 1.46	39.0 ± 3.45
Respiratory burst activity (% NBT positive)	0.21 ± 0.072	0.33 ± 0.19	0.21 ± 0.26	0.33 ± 0.072

Discussion

Influence of diet treatment on fish performance

The results of this study demonstrate that the growth performance (i.e., growth, feed intake, feed and protein utilization), survival, condition factor (weight length⁻¹ relationship) and yield of edible tissue of juvenile sablefish were not compromised, when up to 75% of the supplemental AO was replaced with FO. Few studies have been conducted on sablefish of size similar to those used in the present study. However, the present results agree closely with those of Clarke et al. (1999). Both studies found that sablefish grew from a weight of 150 g to 420 g in about 100 days. Juvenile sablefish (<10 g in weight) have specific growth rates that exceed 10% of body weight day⁻¹ (Sogard & Olla 2001). These high growth rates appear to be short lived, as the fish used in this study had a maximum growth rate of 1.2% body weight day⁻¹.

While no other studies have examined the use of alternative lipids in juvenile sablefish, vegetable oils have been incorporated successfully into the feeds of other marine species such as red seabream (Glencross, Hawkins & Curnow 2003), gilthead seabream (Izquierdo et al. 2005), European sea bass (Mourente, Dick, Bell & Tocher 2005b), Atlantic cod (Bell et al. 2006) and Atlantic Halibut (Martins, Valente & Lall 2009). Depending upon the lipid contributed by the dietary protein sources (e.g., fish meal), complete replacement of the supplemental fish oil with vegetable oil has not been successful, as marine fish species have essential dietary requirements for n-3 HUFAs (Kanazawa et al. 1979; Higgs & Dong 2000). Substitution of vegetable oil for either 80% or 100% of the supplemental dietary fish oil decreased fish growth and feed efficiency in gilthead seabream (Izquierdo et al. 2005) and sea bass (Yildiz & Sener 2004). In the present study on sablefish, the dietary n-3 HUFA concentrations ranged from about 2.7 (Diet 25AO:75FO) to 5.7 (diet 100AO)% of dry weight or they varied between 13.7% and 29% of the dietary lipid concentration without any differences in fish performance. These results suggest that the preceding range in dietary levels of n-3 HUFAs is adequate for growth of sablefish. However, until more work and long-term feeding trials are conducted to define the essential fatty acid needs of sablefish, this conclusion must remain tentative. For instance, other studies on the essential fatty acid needs of marine finfish species have found evidence for optimal dietary levels of 20:5n-3 and 22:6n-3, with some requirement for 20:4n-6 (reviewed by Higgs & Dong 2000).

Influence of diet treatment on proximate analysis and digestibility

Dietary treatment was not found to significantly influence whole body or fillet proximate compositions. There were no differences among test diets in dietary protein and energy contents or ratios of digestible protein to lipid or in feed intake among the groups, given the different diet treatments. These variables are known to influence the proximate compositions of other fish species (Higgs, Macdonald, Levings & Dosanjh 1995). Fillet lipid concentrations were similar to those in the test diets (~ 19%), and such high lipid levels in the flesh were not unexpected as this species is also commonly known as butterfish. Lipid values of 18.7% (Nakayama, Kawai, Mori, Matsuoka & Akehashi 1978) and 15.1% (Stansby 1976) have been reported in wild sablefish. In addition, the results did not suggest a propensity of sablefish to deposit dietary lipid into the liver, as liver weights expressed as a percentage of body weight remained less than 3.2% across all diet treatments. Minkoff and Clarke (2003) also indicated that juvenile sablefish are able to utilize dietary lipid up to 22% of dry matter without enlargement of the liver. The lack of a dietary effect on liver size also suggests that the sablefish were not limiting in essential fatty acids, because increases in liver lipid deposition can be associated with essential fatty acid deficiency in marine fish (Izquierdo et al. 2005; Mourente, Dick et al. 2005b).

The estimated apparent digestibility coefficients for protein, energy and organic matter in the test diets were not influenced by diet treatment. However, the values for each of the foregoing digestibility coefficients were found to be relatively low. Probably, this occurred because of the faeces collection procedure that was used. Dissection of faeces can underestimate digestibility coefficients, especially for protein, due to contamination of samples with intestinal fluids and mucus (Hajen, Beames, Higgs & Dosanjh 1993). This undoubtedly occurred in this study, but intestinal dissection had to be employed for faecal collection, because the sablefish had soft bellies that precluded collection of faeces by stripping. In addition, it was not possible to obtain a sufficient number of juvenile sablefish to conduct a separate digestibility trial using the modified ‘Guelph system’ of faecal collection as described by Hajen et al. (1993).

Influence of diet treatment on flesh contaminant concentrations

Flesh contaminant concentrations in the flesh were found to reflect the dietary treatment, as the diet is the largest source of PCBs in the flesh of fish, and there is negligible uptake of these compounds through the gills (Morrison, Gobas, Lazar, Whittle & Haffner 1997). Similar results have been found in other studies of finfish species when they have been fed diets based on alternative lipids of plant origin. In Atlantic salmon, replacement of 100% of the supplemental fish oil with either a 1:1 blend of rapeseed oil and linseed oil (Bell, McGhee, Dick & Tocher 2005) or a blend of 55% rapeseed oil, 30% palm oil and 15% linseed oil (Berntssen, Lundebye & Torstensen 2005) or 100% rapeseed oil (Bethune, Seierstad, Seljeflot, Johansen, Arnesen, Meltzer, Rosenlund, Froyland & Lundebye 2006) significantly lowered total flesh TEQ levels by 66%, 93% and 64% respectively. Other studies with trout (Isosaari et al. 2002; Karl et al. 2003; Carline et al. 2004) and Atlantic salmon (Lundebye et al. 2004) have examined aquaculture feeds with different concentrations of PCDD/Fs and PCBs, and similarly found flesh POP concentrations to be decreased when the fish were fed less contaminated feed.

The higher accumulation efficiencies noted for PCBs over PCDD/Fs in sablefish is consistent with results that have been reported for Atlantic salmon (Berntssen et al. 2005; Isosaari, Kiviranta, Lie, Lundebye, Ritchie & Vartiainen 2004; Berntssen, Giskegjerde, Rosenlund, Torstensen & Lundebye 2007; Berntssen, Maage, Julshamn, Oeye & Lundebye 2011) and rainbow trout (Isosaari et al. 2002). The observed PCB AE_WB of 76.4% in this study was similar to that (> 70%) determined in Atlantic salmon (Lundebye et al. 2004; Berntssen et al. 2005, 2007). Furthermore, in this study, it was determined that the retention of consumed PCBs decreased as their level of chlorination increased Table 5. In terms of human intake assessments, examination of the AE_fillets values indicated that sablefish deposited only 31% of the consumed PCBs, 5% of the PCDDs and 21% of the total PCDFs into the edible flesh tissue.

Influence of diet treatment on fillet fatty acid compositions

The fatty acid compositions of sablefish fillets were generally found to mirror their respective dietary fatty acid compositions. These results are consistent with those of other studies on marine fish that have substituted FO for fish oil (Izquierdo et al. 2005; Montero et al. 2005). Some fatty acids, especially the saturated fatty acids, 18-1n9 and 18:3n-3, may have been preferentially utilized by the sablefish as sources of non-protein energy, as range in concentration of these fatty acids in the fillets was much smaller than the range in the dietary treatments. The elevated concentration of 18:1n-9 in the flesh lipids versus dietary lipids of all groups may reflect the high retention of this exogenous fatty acid in the triglyceride fraction of the sablefish muscle over time.

Progressively lower levels of EPA and DHA were found in the flesh of sablefish when they were fed diets that contained increased amounts of FO by replacement of supplemental AO. This partially occurred because FO does not contain n-3 HUFAs. In addition, marine fish have been found to have insufficient elongase and desaturase activity, and consequently, they do not produce significant amounts of either n-3 HUFAs or arachidonic acid from C18 fatty acids (Rodriquez et al. 2002; Mourente & Dick 2002; Bell et al. 2006). In this study on juvenile sablefish, there was likewise no indication of bioconversion of dietary 18:3n-3 to n-3 HUFAS in the fillet lipids, whereas there may have been low level, but certainly not extensive bioconversion of 18:2n-6 to 20:4n-6.

Influence of diet treatment on human health

Adequate consumption of EPA and DHA has been found to have numerous health benefits such as prevention of cardiovascular disease, reduced likelihood of some cancers and inflammatory conditions and promotion of neural and ocular development and cognitive function (see 'Introduction'). The absolute amounts of EPA and DHA in the flesh of the sablefish ranged from 1900 mg of EPA and DHA in a 100 g raw serving for fish fed diet 75FO to 3200 mg of EPA and DHA per 100 g serving for fish fed diet 100AO. According to the USDA nutritional data bank (U.S. Department of Agriculture and Agricultural Research Service 2006), the only other seafood products that provide comparable amounts of EPA and DHA in a 100 g serving size are: American shad (2400 mg), Atlantic mackerel (2300 mg), wild chinook salmon (2000 mg) and farmed Atlantic salmon (1900 mg). It is recommended by the American Dietetic Association and Dietitions of Canada that humans consume 500 mg of EPA and DHA per day (Kris-Etherton & Innis 2007). The human intake of these n-3 HUFAs, however, remains below 100–200 mg d⁻¹ (Kris-Etherton, Taylor, Yu-Poth, Huth, Moriarty, Fishell, Hargrove, Zhao & Etherton 2000). Consumption of 16 g d⁻¹ (108 g week⁻¹) of flesh from sablefish fed the 100AO diet or 26 g d⁻¹ (184 g week⁻¹) of flesh from fish fed diet 75FO would provide consumers with the recommended daily intake of 500 mg of EPA and DHA. It should also be mentioned parenthetically that 726 g week⁻¹ of 100AO sablefish or 1046 g week-¹ of 75FO sablefish would have to be consumed by a 70 kg person before exceeding the lower end of the World Health Organization recommended maximum daily intake of 1–4 TEQ pg g⁻¹ kg⁻¹ body weight (Van Leeuwen, Feeley, Schrenk, Larsen, Farland & Younes 2000). Consumption of other popular seafood products would require larger daily serving sizes to obtain the same levels of EPA and DHA as those measured in the fillets of sablefish i.e., 136 g d⁻¹ (951 g week⁻¹) Atlantic cod; and 281 g d⁻¹ (1966 g week⁻¹) of tilapia (U.S. Department of Agriculture and Agricultural Research Service 2006). In some cases, depending on factors such as the source and age of the species, the consumption of seafood products to obtain n-3 HUFA requirements may not be possible due to concurrently low levels of EPA and DHA and high contaminant concentrations. Take, for example, a study examining both contaminants and fatty acids in commonly consumed fish purchased on the retail market (Domingo, Bocio, Falco & Llobet 2007). For a 70 kg person to obtain 250 mg day⁻¹ of EPA and DHA from consumption of tuna, red mullet, sardines or anchovies, they would exceed the lower end of the World Health Organization recommended maximum daily intake of 1–4 TEQ pg g⁻¹ kg⁻¹ body weight by 2.6, 2.3, 1.4 and 1.2 times respectively.

Influence of diet treatment on fish health

There did not appear to be any adverse health effects of feeding the juvenile sablefish, diets containing FO in place of marine fish oil. Despite replacement of 75% of the supplemental AO with FO, alterations in haematology or innate immunity of sablefish were not observed after 105 days of feeding the test diets. The sablefish were healthy throughout this study, and the lack of adverse health impacts at the end of the trial was supported by the observation of no fish mortality during the study. The inability to detect differences in fish health suggests that the fish were able to meet their essential fatty acid requirements, even at the highest level of marine fish oil replacement, over a 105-day period. However, it remains unknown whether these fish would have been able to maintain their excellent state of health over a much longer period of time.

Conclusions

Flaxseed oil was found to be a suitable alternative to AO in sablefish feed. Seventy-five per cent of the supplemental dietary AO was replaced without any negative effects on fish growth, fish health, feed intake, feed efficiency, protein utilization, condition factor, survival or the apparent digestibility coefficients for dietary protein, organic matter and energy. The fillet fatty acid compositions were similar to those of the dietary lipid compositions. While the muscle EPA and DHA concentrations decreased as the dietary FO concentration was increased, there were attendant elevations in the concentrations of linolenic acid and total polyunsaturated fatty acid without marked elevations in the linoleic acid concentration. These changes may prove to have positive benefits for human health in subsequent studies. The use of FO was also found to beneficial in reducing flesh levels of POPs.

Acknowledgements

The authors gratefully acknowledge the financial assistance provided by AquaNet, the Department of Fisheries and Oceans Canada (DFO)/Aquaculture Collaborative Research and Development Program, Skretting Canada, Vancouver, BC and Ewos Canada, Surrey, BC. Also, the authors express their appreciation for the laboratory help given by N. Ghani, M. Rowshandeli, J. Oakes, J. McDonald, R. Dominelli, J. Gordon, D. MacQuarrie, J. Sutton and N. Rowshandeli. We also thank Ewos Canada, Microtek International, Saanichton, BC and Taplow Feeds, North Vancouver, BC as well as NRC/IRAP for providing financial assistance to Dr S.K. Balfry.