1. Introduction
The use of fishmeal (FM) and fish oil (FO) in salmon feeds has declined markedly over the last two decades and this trend is predicted to continue [
1]. The reduction in FM/FO has been facilitated, to a certain extent, by an improved understanding of salmon nutrition, genetic selection, advances in feed technologies, and innovation in feed delivery methods [
2,
3,
4,
5]. Other attributes driving the reduction in the use of FM/FO have been the rising price of both commodities in the international marketplace [
6] and the emergence of a more refined and opinionated consumer. Prices for FM/FO have risen as competition for both resources has increased while availability has generally weakened. Reduced catches of forage fish, which provide the raw material for the production of FM/FO, have emerged due to over-fishing, natural and climate-induced effects (e.g., El Niño–Southern Oscillation, the Madden–Julian Oscillation), and the increased consumption of forage fish resources by society [
7,
8,
9,
10]. Another concern that has forced the aquafeed sector to reduce its reliance on FM/FO is public awareness of the negative direct and indirect effects of forage fisheries on marine and coastal ecosystems [
11]. Consumers are now more mindful of issues surrounding sustainability, including losses to biodiversity such as the consequences of intense fishing operations on populations of marine mammals, seabirds, and reptiles [
12,
13,
14]. Put plainly, nowadays consumers insist that food production systems are safe, ethically, and environmentally reactive and follow the principles of the circular economy with zero discard, combined with more rigorous criteria for animal health and welfare.
In response to the above and other issues, aquafeed manufacturers and researchers have continued to examine the prospect of further reducing or eliminating FM/FO from Atlantic salmon feeds. While this has met with varying degrees of success, the most satisfactory responses have been gained using animal protein-based diets and those comprising blends of plant protein concentrates, and their combinations cf. [
15,
16,
17,
18,
19,
20,
21,
22,
23]. Indeed, some longer-term studies support the concept that, with judicious selection and blending of various alternative proteins and oils, it may prove possible to completely replace FM/FO in salmon feeds [
24,
25,
26,
27,
28,
29,
30]. Due to the potential savings inherent in employing FM/FO alternatives, even given poorer Feed Conversion Ratios (FCRs) and reduced growth rates, profit margins can increase [
31,
32,
33] while simultaneously addressing some consumer concerns. In fact, previous studies reveal that shoppers will pay higher prices for salmon produced under sustainable conditions, provided account is taken of fish welfare, and that food safety is guaranteed [
34,
35,
36]. For example, consumers are prepared to pay a 25.3% premium for organic salmon in the UK [
37], 20% in Denmark [
38], 15% in Norway [
39], 11–20% in Croatia [
40], and an 11% premium in France [
41].
Comparatively few studies have evaluated the quality attributes of Atlantic salmon fed on completely marine resource-free feeds. Most reports have evaluated fish fed diets that substitute either FM or FO rather than excluding both ingredients simultaneously [
24,
42,
43,
44,
45,
46,
47]. Since refined consumers will not accept sub-standard foodstuffs, there is a need to determine whether concurrent FM/FO replacement negatively influences the overall quality of farmed Atlantic salmon. Accordingly, we examined the impact of feeding post-smolt Atlantic salmon with various diets over a 90-day period and assessed the impact of each on the mechanical and organoleptic/palatability characteristics of resultant fillets. Fillet color characteristics were examined as a quality indicator and fin and eye condition indices were assessed to determine whether the marine resource-free diets influenced overall animal welfare. In addition, we explored the value of uncovering carbon and nitrogen isotope ratios as a means of confirming the marine resource-free nature of feeds.
3. Results
There were no differences in weight, length, or
K between dietary groups and this extended to gutted weights, headed-gutted weights, fillet weights and yields, and body surface area (
Table 2). Percentage yield did however vary between groups, ranging from 79.88 to 83.00% with lower (
p < 0.001) values observed for fish fed experimental diets void of KM and salmon fed the P2 formulation (
Table 2). Variances (
p < 0.03) between dietary groups were apparent for pigmentation in each of the three tested fillet regions and Student’s paired
t-tests revealed that, overall, belly color was lighter than either the NQC or back portions (
p < 0.001).
Diet had a significant impact on SL, with fish fed the P1 and A3 diets being shorter than those fed the FM/FO-based diet (
Table 3). The relative widths (RFW) and heights (RFH) of the dorsal fins did not differ between groups, varying from 42.44 to 45.69 and from 29.81 to 33.88, respectively, with the relative fin index falling (RFI) in the range of 13.4 to 13.8 (
Table 3). The number of frays (#F) and fray widths (FW) of the dorsal fin frays also did not differ between treatments (
Table 3). No cataracts were discerned in any of the fish, irrespective of dietary treatment.
Mean fillet stable isotope differences (
p < 0.05,
Table 4,
Figure 2) were apparent between salmon fed on the control, FM/FO-based feeds, and all other samples, with δ
13C and δ
15N concentrations being lower and higher, respectively. Δ
13C levels also varied with diet when considering the experimental feeds, with fish fed the diets containing 5% KM and the 2.5% animal protein formulation returning elevated concentrations when compared to P1, P2, and A1 feeds. There were no differences in readings, however, for measured fillet δ
15N levels between experimental diets. Nevertheless, the significant variation observed in fillet δ
13C had no overall impact on fillet δ
13C: δ
15N ratios from experimental feeds, which, as illustrated by
Figure 2, were all lower (
p < 0.05) than the control fillets.
Table 5 and
Table 6 summarize measurements for hardness, cohesiveness, springiness, and resilience of the anterior and posterior halves of the fillets derived from the different dietary treatments, respectively. There were no differences discerned in any measured parameter when comparing the anterior fillets. However, evaluation of springiness and resilience for the posterior half of the fillet revealed a higher value for springiness (
p < 0.005) in A1-fed salmon fillets when compared against all others, which returned comparable results (
Table 6). Additionally, A1 fillets expressed higher resilience readings than those observed in other experimental diets (
p < 0.007). The fillets from the FM/FO-fed fish (C1), however, did not differ in resilience to any of the experimental feeds (
Table 6).
The average age of the untrained taste testers was 43.1 ± 16.3 years. Declared preferences for specific fillets were in the following order: P3 > P2 > no preference > A2 = A3 (
p > 0.05;
Figure 3). Testers of Asian, African, and Latino descent expressed preference only for fish fed plant-based, KM-containing feeds (P2 and P3), whereas Caucasians were apparently less discerning, selecting fillets derived from each dietary group except for the control (C1). Interestingly, taster testers who selected no preference were generally older (66.4 ± 12.1 years) than those who selected the P3 fillets (42.2 ± 4.6 years).
4. Discussion
Seafood consumers generally make purchasing decisions based on safety and quality. Willingness to pay (WTP) for products by more sophisticated buyers, however, may include consideration of a variety of other factors. These include issues that are predominantly aligned to commercial fisheries (e.g., by-catch and at-sea discards, ghost fishing, toxic fisheries subsidies, carbon emissions from fleets, human rights infringements) as well as those that are more affiliated with aquaculture (e.g., animal welfare, environmental degradation, genetic pollution, carbon emissions from feed manufacturers, human rights infringements in production and supply chain sectors) [
34,
54,
55,
56,
57,
58,
59,
60]. Irrespective of whether a salmon is fished or farmed, however, the first impression a consumer obtains is generally visual. How the salmon presents, therefore, represents a significant element in a purchaser’s decision-making process and WTP for the product.
During cultivation, Atlantic salmon are exposed to various stressors that can influence their overall appearance. For example, eyes can become variably opaque due to the presence of cataracts [
61]. This condition, which is generally considered to be irreversible, may be caused by nutritional, environmental, chemical, and infectious insults [
62]. During the culture of Atlantic salmon, cataracts and the attending poorer vision can lead to abnormal behavior, reduced feed intake, inferior performance, and lowered survival, which also influences profitability. Sissener et al. [
63] observed a 6–14% occurrence of cataracts with post-smolts 3 months following transfer to net pens, while Tröße [
64], reported mild cataracts and smaller-sized lenses in Atlantic salmon fed on a substantially plant-based diet for 12 months. She suggested that increased risks for cataracts existed due to the low histidine and N-acetylhistidine content of the plant diet. In the present trial, however, there were no signs of eye opacity or hemorrhaging between treatment groups. Thus, over the duration of the trial, the rearing environment and nutrition were acceptable, an observation identical to that of Sissener et al. [
63]. Nonetheless, the possibility of cataract development should be thoroughly examined, especially if FM/FO-free feeds are to be deployed over a full production cycle since it has been suggested that dietary levels of histidine should be 14.4 g kg
−1 feed [
65], which is higher than used herein.
More obvious than cataracts are external injuries to the skin and fins. High-density holding of Atlantic salmon has been reported to elevate fin erosion [
66]. However, in the current study stocking densities were maintained below that which is considered stressful [
66,
67]. Moreover, there were no signs of skin or opercular damage at trial end, indicating a low level of aggressive interaction within tanks and further evidence that the rearing environment used was appropriate. However, cultured salmonids often suffer from fin erosion—damage to the epidermis, dermis, and fin rays—resulting in fraying, splitting, and changes in histology and fin size [
68]. Fin erosion is considered an important welfare issue in fish since caudal, dorsal, and pectoral fins are nociceptive. The causes of fin erosion, which can achieve a prevalence of 60–90%
+ [
63,
69,
70,
71], appear diverse [
68,
72,
73,
74,
75] but, unlike fin rot, it is not due to bacterial infection. Nor is fin erosion transmissible, but it may, nonetheless, open fish to secondary bacterial infections. In salmonids, damaged fins can result in downgrading by processors on aesthetic grounds and reduced WTP by consumers [
76]. In the present study, fin erosion was noted in approximately 45% of the fish examined with no differences between dietary treatments. Moreover, the average fray width of affected fish was 2.12 mm, with the majority only exhibiting modest levels of erosion. A lower prevalence of fin erosion has been observed by others [
77,
78], although the latter studies used fish maintained in net pens at stocking densities one-quarter to one-tenth that used during the present trial—differences that may provide partial explanations for the detected variations. Nevertheless, further examination of the causes and mitigation of fin damage is warranted.
Flesh color is another important visual quality characteristic, and, in the marketplace, the pink-red color of salmon not only separates them from other species but represents an indicator of freshness and flavor [
79]. Most salmon are sold as fillets, and therefore, color is more apparent than eye or fin attributes. The color of salmon also contributes to a consumers’ general enjoyment and gastronomic delight [
80]; a detail that increases a shopper’s WTP for redder fillets [
79]. Most studies with Atlantic salmon evaluate production characteristics using much larger fish (4+ kg) than employed here, with the consequence that comparison of morphological relationships and sensory characteristics is problematic. For example, several authors have commented on the relationship between the development of flesh color and fish size, which itself is linked to the duration and quantity of pigmented feed fed, dietary lipid concentration, pigment intensity, type, digestibility, deposition, and retention [
81,
82,
83]. In the current study, the belly regions of examined fillets were lighter than the back or Norwegian Quality Cut (NQC), which corroborates the report of Young et al. [
84] and others who examined larger fish. However, herein, the overall intensity of color attained for fish of the size range evaluated was in accord with previous observations [
85]. Interestingly, there was a tendency toward a more vivid hue in salmon fed the alternative protein diets when compared to the control group, which contrasts with the observations with salmon and trout, where increased fillet lightness has been reported [
86,
87,
88]. Because alternate lipid sources appear to have no influence on pigmentation [
30,
31,
89], the color differences encountered in the current study are difficult to resolve since the darkest hues, as measured by the
SalmoFan™, were from fish fed the plant-based diet void of krill meal. The likeliest explanation for the more pronounced coloration of the alternative protein diets, therefore, was the incorporation of
Schizochytrium sp. oil at 4.5% of the diet. The oil from these heterotrophic microalgae, belonging to the order Thraustochytriales, not only provides a source of docosahexaenoic acid (DHA)/eicosapentaenoic acid (EPA) but is also known to contain astaxanthin and other pigments [
90].
In the present study, irrespective of the animal’s dietary background, instrumental measurements failed to reveal any differences in texture for anterior fillets. However, some subtle differences were observed in springiness and resilience for one of the posterior fillets. Indeed, herein, overall comparisons of anterior and posterior fillets failed to reveal significant anteroposterior differences within treatments. In contrast, under normal conditions of rearing, using FM/FO-based feeds, Casas et al. [
91] reported differences in fillet textural measurements taken at anterior and posterior positions of samples from 3
+ kg animals, with posterior sites registering greater hardness but lower cohesiveness and springiness. These anterior-to-posterior differences in muscle characters were recorded earlier by [
92,
93] and others. The noted divergence between the present and other studies may simply reflect disparities in the age/size of fish examined, the dietary formulations employed, or even the season and method of slaughter used [
94]. Indeed, it is difficult to compare the findings from the present study with those of others since most have examined the effect of supplementary ingredients or only partial replacements of FM with alternative proteins, and often with larger fish. Nevertheless, both [
95,
96] reported increased firmness in salmon flesh fed supplementary glutamate, while [
44], using an FM-free diet and similarly sized post-smolts to those used here, observed no effect of diet on shear force. Augmented muscle hardness, gumminess, and chewiness have also been reported for other species in which diets were supplemented with high levels of plant and animal proteins and distiller’s grains [
89,
97,
98] although contrary effects have also been encountered [
99,
100]. KM has been demonstrated to increase fillet firmness in Atlantic salmon and the impact of poultry and plant-based meals on the mechanical and organoleptic characteristics of fillets has received attention elsewhere [
2,
44,
101].
Quality and sensorial analyses of Atlantic salmon have generally employed the so-called NQC as being a typical sample [
46,
101]. However, as pointed out previously [
102,
103,
104], the distribution of fat in salmon is not homogenous and this can affect organoleptic qualities and distribution of omega-3 fatty acids along the fillet. Accordingly, Nøstbakken and colleagues [
105] suggest that the whole fillet rather than just the NQC be used to obtain an impartial compositional analysis, and here we extended this concept to organoleptic evaluation. Additionally, because a small, trained taste panel could never represent the varied perceptions of a naïve target market [
106,
107], we elected to use regular consumers of salmon (naïve assessors) for the sensory evaluation. Indeed, many studies have determined a null effect of training over consumer assessments for various foods [
108,
109,
110]. Quality elements generally considered relevant to the judgment of seafood include taste, texture, and appearance, and, for salmonids, [
80] indicates that taste and texture are especially significant in revealing consumer preferences for salmonid flesh. It is the triumvirate of taste, odor, and color, however, that has true physiologic importance since these attributes stimulate appetite and gratification. An individual’s perceived quality determinants do, however, differ with age, gender, income, educational attainment, racial origin, culinary heritage, and other factors [
111,
112,
113,
114]. The latter may explain differences in preference reported here since the naïve testers were of varied age, gender, ethnicity, and educational attainment, while also being of differing culinary heritage (Caucasian, Afro-American, Hispanic, Asian) and geographic origin.
The methods of cooking employed, including baking, as used in the present study, do not appear to interfere with salmon fillet lipid quality or omega-3 fatty acid dynamics [
115,
116,
117] and, by way of confirmation, Barrows et al. [
6] reported that a 75 g serving of fish from the present study, irrespective of dietary treatment, was sufficient to exceed daily intake recommendations for EPA+DHA. The use of algal oil as a replacement for FO was suitable given the study findings. The association of fish size to body fat levels and the relationship between fillet fattiness and sensory quality in Atlantic salmon is well documented [
118,
119,
120,
121], but size (this study) and body fat did not vary between treatments [
6], implying that neither issue impacted sensorial evaluations. Interestingly, Atlantic salmon fed with non-FM/-containing feed, but with 5% KM (P3), proved preferable to all other treatments, with 42% of the respondents having a positive view. Of the remainder, 21% of the panel expressed a liking of P2-fed fish, 15% had no preference, while fillets derived from A2 and A3 feeds polled 6% and 9%, respectively. Noteworthy was that the ‘no preference’ group was of the greatest mean age whereas the P3 group was younger. This is perhaps not surprising given that a higher prevalence of taste disorder has been described in older people from various geographic regions [
122,
123,
124,
125], and also reflects the age skew of the taste testers.
Contemporary consumers demand safer and higher-quality food and, when paying premium prices, insist on increased transparency relating to product identity. For salmon, whole, filleted, and cutlet products are generally easy to recognize. Less straightforward, however, is the determination of provenance and production processes for specific seafoods, such as marine resource-free or organic variations. Indeed, according to the FAO, fish products are among the food products most susceptible to fraud [
126]. For this reason, a variety of methods have been examined to authenticate, for example, wild and farmed salmon, adulteration (particularly of processed foodstuffs), organic products, and to determine pigment and element levels in salmon muscle [
126,
127,
128,
129,
130,
131,
132,
133]. In this regard, stable isotope analyses have proven especially useful [
131], and we have previously demonstrated its utility in verifying the marine resource-free status of farmed largemouth bass [
133]. In the current trial, however, evaluation of muscle δ
13C: δ
15N ratios at the trial end did not provide the degree of discrimination previously encountered, for example, with largemouth bass. While there was a clear separation between FM/FO and the alternative plant and animal protein diets, the method was unable to separate fish fed diets supplemented with KM. The reason for this remains obscure but further studies are required to substantiate the suitability of the method as a technique for verifying the fidelity of marine resource-free farmed salmon.