1. Introduction
Intensive culture of the carnivorous freshwater Eurasian perch (
Perca fluviatilis L.) is increasing in recirculating aquaculture systems (RAS) and represents an expanding branch of commercial fish farming in Europe. Nevertheless, as a relatively new aquaculture species, production is low [
1]. It is commonly reared on feed formulated primarily for salmonids or marine fish species [
2]. Diets for carnivorous species contain high levels of protein, which have been obtained from marine fishmeal (FM), considered optimal because of its balanced nutritional composition [
3,
4]. Currently, with FM increased cost and unsustainability [
5], plant protein sources, especially soybean meal, are being used in aquaculture to decrease the dependency on FM and reduce feed costs [
4]. High levels of plant protein in feeds can reduce growth performance or induce fish health issues, chiefly due to imbalance in essential amino acid (EAA) content, low feed acceptance and the presence of anti-nutritional factors [
3,
6,
7]. Processed animal proteins (PAP) such as poultry by-product meal, meat meal and meat and bone meal are valid proteins for aquaculture feeds but their use is limited by legislation. In the EU, PAP from poultry and swine have only recently been reintroduced into aquafeed (EC No. 56/2013) after more than 10 years of ban due to Bovine Spongiform Encephalopathy (EC No 999/2001), while in other parts of the world, its use is common practice [
8].
Recently, interest has turned to PAPs from insects as a component of aquafeeds [
9,
10]. Insect larva meals are rich in proteins and their EAA profile is close to that of FM and considered superior to that of plant proteins [
9]. The use of insect PAP has recently been sanctioned by the European Commission (Brussels, Belgium) (Regulation 2017/893/EC, 2017).
The black soldier fly
Hermetia illucens belongs to the family
Stratiomyidae and is among the most promising insect species for mass-rearing for animal feed [
11]. Commercial
H. illucens meal has an average protein content of 55% dry matter (DM) with lipid content ranging from 5% to 35% DM, depending on the defatting process applied during meal production. Research into its efficacy has thus far been contradictory: Similar or better growth performance to that of fish fed conventional protein sources (mainly FM or soybean meal) using commercial
H. illucens meal at inclusion levels from 2.5% to 40% were obtained for Atlantic salmon (
Salmo salar) [
12,
13,
14], rainbow trout (
Oncorhynchus mykiss) [
15,
16], European sea bass (
Dicentrarchus labrax) [
17], yellow catfish (
Pelteobagrus fulvidraco) [
18] and rice field eel (
Monopterus albus) [
19]. Conversely, other authors reported reduced acceptance and growth [
20,
21], with high levels of inclusion. Divergence in results is likely due to the differences among H.
illucens meals and the level of inclusion in the diet and also suggest species differences in adaptation to insect meals. The use of
H. illucens meal in perch diets has not been investigated.
The goal of this research was to determine the effects of partially defatted H. illucens meal as partial substitute for FM on growth performance, somatic indices, occurrence of splenic lipidosis, hematological parameters and proximate whole-body composition of juvenile P. fluviatilis. The research also aimed to provide new data on the economic and environmental sustainability of this novel protein source.
2. Materials and Methods
An 84-day growth trial was carried out at the Faculty of Fisheries and Water Protection of the University of South Bohemia (České Budějovice, Czech Republic). The trial was designed and carried out in accordance with the Czech and European Communities Directive (2010/63/EU) on the protection of animals used for scientific purposes, protocol number MSMT-6744/2018-2.
2.1. Experimental Diets
Four experimental diets were formulated to be isonitrogenous (crude protein, CP: ~54 g 100 g DM); isolipidic (ether extract, EE: ~13 g 100 g DM); and isoenergetic (gross energy, GE: ~23 MJ kg DM). An FM-based diet was used as control (CON) and three additional diets included FM replacement with 20% (H20), 40% (H40) and 60% (H60) partially defatted H. illucens larva meal obtained with a mechanical process performed using high pressure and without solvents was provided by Hermetia Deutschland GmbH & Co. KG (Baruth/Mark, Germany). In order to ensure that diets were isonitrogenous, isolipidic and isoenergetic, the proportion of wheat meal and fish oil was reduced with increase in H. illucens.
The experimental feeds were prepared at the Department of Agricultural, Forest and Food Sciences experimental facility. Finely ground ingredients and fish oil were thoroughly mixed with water and pelleted using a 2 mm meat grinder and dried at 50 °C for 48 h. Feeds were stored in dark bags at −20 °C until use. The ingredients of the experimental diets are reported in
Table 1. An additional control group (BIO) was fed a commercial extruded diet (BioMar Inicio 2 mm, BioMar A/S, Brande, Denmark) containing fish meal, wheat gluten, wheat, pea protein, soybean concentrate, rapeseed oil, fish oil and yeast extract as main ingredients. Proximate composition (on a wet basis) according to manufacturer’s label was CP 52%, crude lipid 23%, carbohydrates 12%, ash 8.7%, fiber 0.9%, total phosphorus (P) 1.2% and GE 23.5 MJ/kg.
2.2. Fish and Feeding Trial
Eurasian perch juveniles were obtained from pond-reared larvae and intensively reared juveniles in an RAS [
22]. The RAS (4360 L total water volume) included fifteen 75 L rearing tanks, a mechanical drum filter AEM 15 (AEM-Products V.O.F., Lienden, The Netherlands), a 1620 L tank with a series of filtration sections, Bioakvacit PP10 (Jezírka Banát s.r.o., Hněvotín, Czech Republic), a moving bed biofilter (1620 L) with media BT10 (Ratz Aqua & Polymer Technik, Remscheid, Germany), UV treatment AquaForte 55 W (AquaForte, Veghel, The Netherlands) and an Eheim Jäger Thermocontrol 300 flow-through heater (Eheim GmbH & Co KG, Stuttgart, Germany) incorporated directly into the recirculation flow. The flow rate in the tanks was approximately 80 L h
−1 with light aeration. Photoperiod was set at 12:12 h (dark: light) with light intensity of 500–700 Lx at the surface. Oxygen saturation (83.7 ± 6.2%), pH (6.83 ± 0.52) and water temperature (22.5 ± 0.7 °C) (HACH HQ 40, Germany) were measured daily at 08.00 and 16.00. Ammonia, nitrate and nitrite concentrations were analyzed at two-day intervals with kits (HACH, LCK 304, LCK 339, LCK 341), using a HACH DR2800 spectrophotometer. The concentration of nitrite-N, nitrate-N and ammonia-N were 0.62 ± 0.44 mg L
−1, 88.88 ± 57.31 mg L
−1 and 2.07 ± 1.02 mg L
−1, respectively.
A total of 750 juvenile European perch were lightly anaesthetized (0.3 mL L−1 of clove oil), individually weighed (initial body weight (BWi) 21.9 ± 4.2 g) using a digital balance (Pioneer, Ohaus Corporation, Parsippany, NJ, USA, d = 0.01 g) and randomly allocated to one of the fifteen 75 L rectangular plastic tanks at a stocking density of 14.6 kg m−3. The four experimental diet groups and the BIO group were randomly allocated to the fifteen tanks, with each diet tested in triplicate. Fish were fed manually to subjectively-judged satiation five times daily (09:00, 11:00, 13:00, 15:00 and 17:00 h). Care was taken to avoid feed waste and to ensure that all supplied feed was consumed. The feeding trial lasted 84 days.
2.3. Growth Performance
At the end of the trial, all fish were individually weighed and growth performance was calculated using following equations:
where Ni and Nf = initial and final number of fish per tank, Ns = number of sampled fish per tank, BWf = final body weight (g), BWi = initial body weight (g), Nd = number of feeding days, TFS = total feed supplied (g), TPS = total protein supplied (g, DM), SD = standard deviation of subsample BW, BWi = initial mean body weight, BWf = final mean body weight, lnBWf = natural logarithm of final body weight, lnBWi = natural logarithm of initial body weight, DM = dry matter, WG = weight gain.
2.4. Condition Factor, Somatic Indexes and Occurrence of Spleen Lipidosis
To calculate condition factor (K), at the end of the growth trial, fifty fish from each tank were anaesthetized (0.3 mL L−1 of clove oil) and individually weighed and measured for total length (TL, mm) and standard length (SL, mm) within 1 mm using a ruler.
K was calculated as:
where BWf = final body weight (g), TLf = final body length (cm).
At the end of the trial, 30 fish/tank were killed by overdosing of anesthesia with clove oil and wet weight of liver, spleen, viscera and perivisceral fat recorded (±0.01 g) for calculation of somatic indices using the equations:
where Wl = liver weight (g), Ws = spleen weight (g), Wv = viscera weight (g), Wpf perivisceral fat weight (g).
Frequency of occurrence splenic lipidosis [
2], was calculated according to the equation:
where Nt is total number of investigated fish and Nsl is number of fish with spleen lipidosis.
2.5. Chemical Analyses
The H.
illucens larva meal chemical analysis was obtained from Renna et al. [
15]. The proximate composition and energy of diets are reported in
Table 1. Feed samples were finely ground (MLI 204; Bühler AG, Uzwil, Switzerland) and analyzed for DM (AOAC, n. 934.01), CP (AOAC, n. 984.13) and ash (AOAC, n. 942.05) content according to AOAC International [
23]. The EE content (AOAC, n. 2003.05) was analyzed according to AOAC International [
24]. The GE content was determined using an adiabatic bomb calorimeter (C7000; IKA, Staufen, Germany). Chitin content was determined following Finke [
25], by correcting for the amino acid (AA) content of the acid fiber detergent (ADF) fraction and assuming the remainder of the ADF fraction to be chitin. The AA composition of H.
illucens larva meal and FM used in the experimental diets is shown in
Table 2. Amino acid quantification was conducted according to De Marco et al. [
26]. After 22 h hydrolysis in 6N HCl at 112 °C under a nitrogen atmosphere, the AA content in the hydrolysate was assessed by HPLC after post-column derivatization. Performic acid oxidation occurred prior to acid hydrolysis for methionine and cystine. Tryptophan was not determined.
At the end of the trial, whole-body homogenate of six fish from each group was analyzed for DM, CP, EE, organic matter (OM) and ash content according to the procedure used for feed analyses [
23,
24]. The DM content was measured according to AOAC (n. 934.01; [
23]).
Fatty acid profiles were determined both in feed and fish whole-body homogenate (3 fish/tank, 9 fish/group), according to the method of Sampels et al. [
27]. Initially, lipids were extracted by the hexan-isopropanol method according to Hara and Radin [
28]. Fatty acid methyl esters (FAME) were prepared by the BF
3 method according to Appelqvist [
29] and analyzed using FAME C 11:0 as an internal standard by a gas chromatograph (Trace Ultra FID; Thermo Scientific, Milan, Italy) equipped with a flame ionization detector, using a BPX 70 column (length 50 m, i.d. 0.22 mm, film thickness 0.25 μm) (SGE Inc., Austin, TX, USA). The peaks were identified using Thermo Xcalibur 3.0.63 (Thermo Fisher Scientific Inc., Waltham, MA, USA) software and quantification was achieved by comparing sample retention times and peak areas to retention times and peak area in 7 levels (1000 ug/mL–15 ug/mL) of the standard mixture Supelco 37 component FAME mix (Sigma-Aldrich, St. Louis, MO, USA). Fatty acid profiles for feed are shown in
Table 3, analysis were performed in triplicate.
2.6. Haematological Analyses
At the end of experiment, three fish per tank (nine fish from each group) were over-anaesthetized with clove oil and blood samples were taken for hematological analysis. Red blood cell count (RBCC), hematocrit (HCT), hemoglobin concentration (Hb), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH) and mean corpuscular hemoglobin concentration (MCHC) were measured according to Svobodova et al. [
30].
2.7. Economic Analysis and Environmental Sustainability of Feeds
To determine the relative efficacy and benefits of tested diets, economic conversion ratio (ECR) and economic profit index (EPI) for each tested group was calculated by the following equations:
where FCR is feed conversion ratio (kg feed per kg fish); DP is cost per kg feed; WG is weight gain. The per kilogram cost in euros, excluding labor and taxes, of all components from commercial retailers was as follows: FM = € 1.48; H.
illucens larva meal = € 3.5; wheat meal = € 0.61; fish oil = € 1.32; gelatinized starch = € 0.75; mineral mixture = € 0.49; vitamin mixture = € 3.85. This resulted in per kg feed cost of CON = € 1.31; H20 = € 1.75; H40 = 2.18; H60 = € 2.61; and BIO = € 2.53. Eurasian perch sale price (SP) was calculated at € 6.50 kg
−1.
Fish-in fish-out (FIFO) ratio was used as a practical measure of the quantity of live fish from capture fisheries required for each unit of farmed fish produced [
31]. This indicator of environmental sustainability of feeds was calculated as follows:
where LFM is level of fishmeal in the diet; LFO is level of fish oil in the diet; YFMw is yield of fishmeal from wild fish; YFOw is yield of fish oil from wild fish; FCR is feed conversion ratio.
We estimated the impact of FM substitution with
H. illucens larva meal rapported to Metric Tons (MT) on freshwater demand (WD, m
3/MT), land demand (LD, ha/MT), energy use (EU, GJ/MT) and greenhouse gas production (GWP, kg CO
2−eq). Mean WD, LD and EU for FM, wheat, fish oil, starch and mineral and vitamin mixes were obtained from Chatvijitkul et al. [
32]. Data of WD, LD, EU and GWP for
H. illucens larva meal was retrieved from Roffeis et al. [
33]. Finally, GWP for FM was sourced from Thevenot et al. [
34] and GWP for wheat meal from Heusala et al. [
35].
2.8. Statistical Analyses
All data were tested for homogeneity of variance using Cochran, Hartley and Bartlett tests. Normality of data was tested by Shapiro-Wilk test. Perivisceral fat index, splenosomatic index, some minor fatty acids, hemoglobin, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration and economic profit index were analyzed using Kruskal-Wallis non-parametric test as these data does not show normality. All other remaining parameter results were analyzed separately by one-way ANOVA. Differences were considered significant at p ≤ 0.05 (post-hoc test: Tukey test). The data were expressed as mean ± SD and statistical analyses were performed using STATISTICA 12.0 (StatSoft CR, Prague, Czech Republic). As BIO was a completely different diet, not comparable with respect to composition, nutrient and energy contents, it was not included in the statistical analyses.
4. Discussion
Intensive culture of Eurasian perch is still a young industry in Europe with the main producers being Ireland, France, Poland, Belgium and Denmark. Insects have been proposed as an efficient and high-quality alternative protein source for poultry [
37,
38], swine [
11,
39] and carnivorous fish [
13,
15,
16,
17,
20,
40] and interest in use of insect meals in perch diets is high. Insects are a viable source of protein and lipids [
9,
10] and a typical component of Eurasian perch natural diet. Nogales-Mérida [
9], confirmed insects as an excellent source of several vitamins and minerals including iron, potassium, calcium and magnesium. Use of H.
illucens insect meal is consistent with production of perch as an organic product, as insect meal can be produced locally on a variety of substrates [
41,
42].
The present study represents the first reported use of defatted black soldier fly H.
illucens larva meal as an alternative feed ingredient for Eurasian perch reared in intensive culture. Bußler et al. [
43] demonstrated that H.
illucens is an appropriate insect species for insect meal production. It has a well-balanced essential amino acid profile, an average protein content of 55% DM and ~35% fat DM, which may be reduced to 5–9% by defatting, making it more digestible. However, complete FM replacement by insect meal has not been shown feasible. Henry et al. [
44], reported that the maximum dietary replacement of FM by H.
illucens meal ranges from 6 to 25%, depending on fish species, with higher inclusion levels reducing growth performance. Sealey et al. [
45], reported up to 50% H.
illucens inclusion without negative effects on growth of rainbow trout. Our study showed that there is no significant effect up to 400 g/kg of H.
illucens in the perch diet on body weight or specific growth rate. Similar results were demonstrated by Renna et al. [
15], where partially defatted H.
illucens larva meal up to 40% of inclusion level was used in rainbow trout diet without negative effects on survival rate, growth performance, condition factor, somatic indices, physical quality or gut morphology. Magalhaes et al. [
17], replaced 45% of the FM in diet of juvenile European seabass with up to 19.5% H.
illucens meal corresponding to 22.5% protein without adverse effects on growth performance and feed utilization. Kroeckel et al. [
20], reported that inclusion higher than 33% of defatted H.
illucens larvae decreased protein digestibility, feed acceptance and growth performance of juvenile turbot. Lock et al. [
12], showed that drying slightly defatted H.
illucens meal (255 g/kg DM) at low temperature is the most suitable procedure and produced a good alternative feed for Atlantic salmon growth.
Proximate composition of fish is driven by endogenous (size, life cycle stage) as well as exogenous factors (water quality, feed) [
46]. To minimize bias, we reared European perch under similar conditions. We found no significant differences in DM, CP and EE in whole-fish homogenate among tested H.
illucens diets. This is in line with Gasco et al. [
47], who found no significant difference in DM and CP content of European sea bass fed mealworm
Tenebrio molitor at different diet proportions. Contrary results were obtained in rainbow trout fed T.
molitor, in which increasing the proportion of insect meal triggered significant decreases in DM, CP and EE [
48], while increased enriched H.
illucens prepupae content resulted in decline in DM and EE [
45].
Reduction in DM and EE content may result from decreased nutrient availability [
15], depending on insect species [
23,
42] or on its culture substrate [
37,
39]. Culture substrate also substantially affects insect ash content [
49,
50]. Although body ash content has been reported similar among fish consuming various insect meal diets [
20,
51], we found a significant difference among our diet groups, with the highest ash content in H60, while lower ash levels were observed in CON, H20 and H40 groups. This is in contrast to the proximate analysis of tested diets per se, in which the ash content decreased with increasing H.
illucens inclusion. Kirchgessner and Schwarz [
52] and Shearer [
46] reported no effect of crude dietary ash on ash content of fish body, provided sufficient levels of essential elements are present. This suggests that the partially defatted H.
illucens meal used in our study may lack some essential element or elements, although this complex mechanism is largely unexplored and needs further study. The GE content decreased significantly with increased
H. illucens larva meal inclusion, reflecting the non-significant decrease in both CP and EE with higher
H. illucens larva content.
We found total n-3 and n-6 fatty acid in Eurasian perch to decrease significantly with higher levels of
H. illucens larva meal in the diet, reflecting lower fish oil content, with the n-3/n-6 ratio being inversely related to H.
illucens inclusion. This is in agreement with findings of Borgogno et al. [
51] and Renna et al. [
15], who reported significant reduction of n-3/n-6 ratio in rainbow trout fed with
H. illucens larva meal. The opposite effect was observed in Atlantic salmon fed H.
illucens meal [
13]. The differences among studies could be related to diet composition. In the present study, as well as those of Borgogno et al. [
53] and Renna et al. [
15], fish oil was used as a fat source, while Belghit et al. [
54], used large quantities of rapeseed oil, which contain high level of n-6 polyunsaturated fatty acids contributing to maintain constant the n-3/n-6 ratio between insect meal based diets.
These comparisons underscore differences among insect species and culture media. We found increased H.
illucens proportions to be associated with significantly higher SFA content in fish homogenate, reflecting that partially defatted H.
illucens meal is rich in SFAs (lauric acid C12:0, myristic acid C14:0 and palmitic acid C16:0), while T.
molitor larva meal is rich in MUFAs and n-6 PUFAs. A similar trend was observed in studies of Jian carp [
55] and rainbow trout [
13], fed H.
illucens larva meal. The positive effect on HSI observed in the present study could be related to reduction of lipid storage in liver, as was demonstrated in Atlantic salmon [
54].
Hematological parameters, essential tools in evaluation of fish welfare related to stress and immune status [
56,
57,
58], are highly influenced by feeding regime [
59]. Studies of FM substitutes such as cottonseed [
60], soybean [
61,
62], housefly (
Musca domestica) maggot [
63] and cricket (
Gryllus bimaculatus) [
64], showed no significant effect of tested meals on hematological parameters of fish of various species. This reinforces our suggestion that dietary H.
illucens larva meal does not impact welfare of Eurasian perch but further investigations of diet formulations and feeding strategies are needed to collect additional data for this new area of study and to obtain more comprehensive results on fish growth rate.
The fish-in fish-out ratio is a practical indicator of environmental sustainability [
31]. This index uses a global average wet weight (whole fish) to fishmeal yield of 22.5% and wet weight to fish oil yield of 5%. A ratio >1 indicates net removal of fish globally. We found the FIFO ratio to be substantially reduced with increasing proportions of insect meal and that FIFO could be decreased by 49% in perch fed an insect-based diet without affecting growth. This downward trend is in agreement with forecast of Tacon and Marc [
65].
Increasing H. illucens larva meal proportions in commercial fish feeds could lead to higher energy and land use and increased greenhouse gas production. A lower impact was found for freshwater use. Insect meal inclusion level, which does not affect growth parameters, led to a 144% increase in greenhouse gas production, 123% increase in energy demand and 77% increase in land use. Fresh water use was decreased by 38% compared to control. These findings suggest ongoing monitoring of agricultural resources and related socio-economic and environmental impact during the shift in resource demands from the oceans onto the land.
Future studies should be focused on fine-tuning for optimal insect meal inclusion in the range of 40% to 60%, as well as evaluation of diets with a higher contribution of plant-based protein in combination with insect meal. Long-term studies of rearing fish to a higher market size (>200 g), in combination with sensory and texture analysis of the final product, should be carried out to explore full potential of insect-based diets for perch.
When the inclusion level was >60%, growth was significantly reduced compared with the control group, suggesting that incorporation of up to 40% H. illucens larva meal in the feed formulation for perch is feasible and can reduce reliance on marine resources. However, even if presents limitations, such as production cost and increased impact in some environment-related parameters, the partial replacement of fishmeal by insect protein will be more important in the future as getting enough amount of fishmeal will be difficult and culture of insects like a H. illucens using waste food means to convert non-resources to important protein resources is a promising solution to cope this problem.