Introduction
In recent years, insects have gained increasing attention as a valuable alternative nutritional source in animal feed, thanks to their high protein content, efficient feed conversion, low environmental impact, and ability to valorise organic by-products. Thanks to their favourable nutritional profiles, they represent a potential replacement for soybean meals and fish meals while offering the advantage of reduced competition with human nutrition and a lower ecological cost (Makkar et al., Reference Makkar, Tran, Heuze and Ankers2014; Henry et al., Reference Henry, Gasco, Piccolo and Fountoulaki2015; Gasco et al., Reference Gasco, Henry, Piccolo, Marono, Gai, Renna, Lussiana, Antonopoulou, Mola and Chatzifotis2016; Khan et al., Reference Khan, Khan, Sultan, Khan, Hayat and Shahid2016). Their protein and lipid contents, equal to 30–40% of dry matter (DM), are comparable to those of the main conventional protein sources (Henry et al., Reference Henry, Gasco, Piccolo and Fountoulaki2015; Gasco et al., Reference Gasco, Henry, Piccolo, Marono, Gai, Renna, Lussiana, Antonopoulou, Mola and Chatzifotis2016; Waithaka et al., Reference Waithaka, Osuga, Kabuage, Subramanian, Muriithi, Wachira and Tanga2022). It is important to note, however, that the nutritional composition of insects can vary depending on their developmental stage, rearing conditions, and the type of rearing substrates (Van Huis et al., Reference Van Huis, Van Itterbeeck, Klunder, Mertens, Halloran, Muir and Vantomme2013; Jozefiak and Engberg, Reference Jozefiak and Engberg2015). Among these, the black soldier fly (BSF) (Hermetia illucens Linnaeus, 1758; Stratiomyidae: Diptera) has attracted considerable attention due to its industrial potential. Reared on a variety of substrates, H. illucens larvae exhibit rapid development and a high bioconversion capacity, producing a meal with a stable amino acid profile (Diener et al., Reference Diener, Zurbrügg and Tockner2009; Nguyen et al., Reference Nguyen, Tomberlin and Vanlaerhoven2015; Spranghers et al., Reference Spranghers, Ottoboni, Klootwijk, Ovyn, Deboosere, De Meulenaer, Michiels, Eeckhout, De Clercq and De Smet2017). In recent years, black soldier fly larvae (BSFL) have been incorporated into conventional poultry diets as a fat source in broilers (Schiavone et al., Reference Schiavone, Marco, Martinez, Dabbou, Renna, Madrid, Hernandez, Rotolo, Costa, Gai and Gasco2017, Reference Schiavone, Dabbou, De Marco, Cullere, Biasato, Biasibetti, Capucchio, Bergagna, Dezzutto, Meneguz, Gai, Dalle Zotte and Gasco2018; Addeo et al., Reference Addeo, Randazzo, Olivotto, Messina, Tulli, Vozzo, Attia, Mahayri, Iannaccone, Asiry, Moniello and Bovera2022; Ibrahim et al., Reference Ibrahim, Tan, Abdi, Shah and Aadil Rana2025) and as a protein ingredient in laying hens and broiler quail (Veldkamp et al., Reference Veldkamp and Bosch2015; Cullere et al., Reference Cullere, Tasoniero, Giaccone, Miotti-Scapin, Claeys, De Smet and Dalle Zotte2016; Maurer et al., Reference Maurer, Holinger, Amsler, Früh, Wohlfahrt, Stamer and Leiber2016, Reference Cullere, Tasoniero, Giaccone, Miotti-Scapin, Claeys, De Smet and Dalle Zotte2018), with positive effects on performance and intestinal morphology (Willmore et al., Reference Willmore, Hall and Drewnoski2021). Some reviews indicate that BSFL meal can be included up to approximately 20% (DM basis) in broiler and layer diets without compromising performance (Seyedalmoosavi et al., Reference Seyedalmoosavi, Mielenz, Görs, Wolf, Daş and Metges2022; Chen et al., Reference Chen, Sun, Song, Wang, Ma, Tu, Yang, Li, Wang, Meng, Zhang, Li, Tian, Zhao, Yang, Wang and Li2025). However, this should be considered a tentative threshold, as optimal inclusion levels vary depending on diet composition, species, and age, and further research is needed (Abd El-Hack et al., Reference Abd El-Hack, Shafi, Alghamdi, Abdelnour, Shehata, Noreldin and Ragni2020; de Souza Vilela et al., Reference de Souza Vilela, Kheravii, Sharma, Kolakshyapati, Zimazile, Wu, Andrew and Ruhnke2023). Ongoing studies on BSFL and prepupae (BSFP) also report heterogeneous recommendations (Belluco et al., Reference Belluco, Losasso, Maggioletti, Alonzi, Paoletti and Ricci2013; Van Huis, Reference Van Huis2013; Barragan-Fonseca et al., Reference Barragan-Fonseca, Gort, Dicke and Van Loon2017; da Silva and Hesselberg, Reference da Silva and Hesselberg2020; Salahuddin et al., Reference Salahuddin, Abdel-Wareth, Hiramatsu, Tomberlin, Luza and Lohakare2024). The BSF is a scavenger and polyphagous species capable of developing on a wide range of organic substrates, including agro-industrial by-products, livestock residues, and household waste (Hoc et al., Reference Hoc, Noël, Carpentier, Francis and Caparros Megido2019; Nayak et al., Reference Nayak, Rühl and Klüber2024). Its life cycle includes egg, larval, prepupal, pupal, and adult stages, with the larval stage representing the main period of feeding and growth (Facchini et al., Reference Facchini, Shrestha, van den Boer, Junes, Sader, Peeters and Schmitt2022). The main residue of BSFL rearing, frass, has gained attention as an organic fertiliser rich in nitrogen (N), phosphorus (P), and potassium (K) (Addeo et al., Reference Addeo, Scivicco, Vozzo, Bovera, Asiry, Alqurashi, Cacciola and Severino2024a). Its chemical composition depends on the original substrate, and several studies indicate that frass produced from food waste-based diets can contribute to improved soil fertility (Spranghers et al., Reference Spranghers, Ottoboni, Klootwijk, Ovyn, Deboosere, De Meulenaer, Michiels, Eeckhout, De Clercq and De Smet2017; Setti et al., Reference Setti, Francia, Pulvirenti, Gigliano, Zaccardelli, Pane, Caradonia, Bortolini, Maistrello and Ronga2019; Fischer and Romano, Reference Fischer and Romano2021). Understanding the relationship between substrate composition, larval performance, and frass quality is therefore essential for optimising BSF-based circular systems. Recent evidence indicates that diet chemistry not only affects BSFL growth performance but also shapes gut-associated microbial communities, which in turn may contribute to nutrient utilisation and adaptation to fibre-rich or nutritionally imbalanced substrates (Gurusinga et al., Reference Gurusinga, Hurka, Kreft, Röthig, Vilcinskas and Tegtmeier2026). Accordingly, the aim of this study was to evaluate the effects of five inclusion levels of vegetable and fruit waste (VFW): 0%, 25%, 50%, 75%, and 100% in the substrate of BSFL and BSFP from 7 to 23 days of age. Specifically, we assessed growth traits, body measurements, proximate composition of larvae and prepupae, frass chemical and mineral profiles, and economic parameters. Since frass from BSFL rearing has been proposed as a potential organic fertiliser due to its content of N, P, and K, evaluating its fertilising value also provides insights into the sustainability and circularity of BSF bioconversion systems.
Material and methods
Preparation of organic waste
The VFW consisted of a mixture of broccoli (Brassica oleracea var. italica, 30%), celery (Apium graveolens, 26.25%), cabbage (B. oleracea var. capitata, 18.75%), orange (Citrus sinensis, 12.5%), and apple (Malus domestica, 12.5%), collected from a local market in Jeddah, Saudi Arabia. Organic waste consists of partially spoiled waste unsuitable for human consumption. The VFW was cleaned of foreign materials, mixed, and finely ground using a chopping machine to confirm the homogeneity of the mixture. The samples were collected to determine their chemical composition. Five experimental diets were prepared by mixing laying hen feed (LHF) formulated to meet the nutrient requirements of poultry (NRC, 1994) with VFW at different levels: (0) control – 100% LHF, 25% VFW – 75% LHF, 50% VFW – 50% LHF, 75% VFW – 25% LHF, and 100% VFW (table 1). The substrate mixtures were prepared on a wet-weight basis (kg), and the VFW/LHF proportions refer to as-is mass. The diets were mixed well with dry Saccharomyces cerevisiae (manufactured in France by S.I.L, Rue Gabriel Peri, Marcq-en-Baroeul, France) at 1.5 g/kg, Clostat XCL (Bacillus subtilis PB6) (Kemin Industries Inc, USA) at 0.3 g/kg, and an antitoxin (Biosorb-M, BIOS AGRICorp Ltd®, Espoo, Finland) at 2.5 g/kg. Probiotic doses were chosen based on the levels recommended by the manufacturer. These additives were added at the top of the diets and mixed well for 10 min to be homogenised within the diets to improve the fermentation process, enhance the quality of feed, decrease toxins, and enhance the digestion processes and utilisation of feed by BSFL (Rehman et al., Reference Rehman, Rehman, Somroo, Cai, Zheng, Xiao, Rehman, Rehman, Tomberlin, Ziniu and Jibin2019). Diets were prepared weekly to ensure freshness, microbial stability, and uniform substrate composition throughout the experimental period. All treatments and replicates were conducted simultaneously, not in separate runs. Between preparations, VFW components were stored at 4°C in sealed containers to prevent spoilage and maintain nutritional consistency. The VFW was prepared 1 day prior to use and left at room temperature before feeding. All substrates were standardised to approximately 65% moisture content, following the methodology described by Liu et al. (Reference Liu, Chen, Wang, Yang, Ur Rehman, Li, Cai, Li, Mazza, Zhang, Yu and Zheng2017) to ensure comparable rearing conditions and prevent fungal growth. VFW, which had high intrinsic moisture, was pretreated by moisture reduction according to Khairuddin et al. (Reference Khairuddin, Ghafar and Hassan2022) to achieve the same target level. The amount of substrate used (1.5 kg per replicate) refers to wet weight (as-is basis). From 7 to 23 days of age, 5 different substrate treatments were tested, each replicated 6 times with 200 larvae per replicate. The number of larvae was chosen to minimise competition and prevent anaerobic conditions. Before day 7, larvae were reared on a low-quality heterogeneous feed (LHF), also adjusted to ∼65% moisture. During this phase, larvae were maintained in 14 × 19 × 5 cm plastic containers covered with muslin cloth, within a ventilated chamber set at 27 ± 0.5°C, 70 ± 5% relative humidity, and a 16:8 light:dark photoperiod conditions kept constant throughout the experiment (Addeo et al., Reference Addeo, Scivicco, Vozzo, Bovera, Asiry, Alqurashi, Cacciola and Severino2024a). Feeding the five tested substrates was performed only once at the beginning of the growing period (day 7). The BSFL used in this experiment originated from a long-established laboratory colony. The full amount of 1.5 kg of substrate was provided at that time, and no additional feeding was carried out during the rearing period (7–23 days). The larval density was based on previous studies (Diener et al., Reference Diener, Zurbrügg and Tockner2009; Nguyen et al., Reference Nguyen, Tomberlin and Vanlaerhoven2013) and preliminary trials, aimed at minimising competition, anaerobic conditions, and substrate degradation. The container size provided sufficient surface area for larval development and feeding, facilitating reliable performance measurements. The experiment concluded at 23 days, at which point approximately 30% of the larvae had undergone transformation into prepupae, as indicated by the darkening of the cuticle. This stage was chosen to enable a meaningful comparison between larval and prepupal stages in terms of nutritional and economic traits. Larvae and prepupae were separated from frass on day 23. For chemical composition analysis, selected samples were fasted for 24 h. For growth performance evaluation, 20 larvae per replicate were weighed to the nearest milligram on days 7 and 16. On day 23, both larvae and prepupae were weighed after 6 h of fasting, and their length and width were measured using an electronic digital Vernier caliper. Measurements were recorded in millimeters (mm). Growth rate was calculated to assess the average daily gain in larval body weight during the main feeding phase. The formula used was:
Composition of the experimental diets (substrate) fed during 7–23 days of age
a The substrates were supplemented by 1% of feed additives over the composition with a mixture of S. cerevisiae and B. subtilis (Clostat XCL 300 dry). S. cerevisiae was obtained from a local market and used at 1.5 g/kg feed. Clostat XCL 300 dry, containing B. subtilis (ATCC-PTA 6737, 4 × 108 cfu/g), was used at 300 mg/kg feed concentration by Kemin Industries Inc. (USA). The probiotic doses were chosen based on the levels recommended by the manufacturer. The diets were also supplemented with antitoxins to eliminate dietary toxins in the feeds.
Growth rate (mg/day) = (Body weight at day 16 − body weight at day 7)/9
This reflects the average increase in fresh body weight per larva over the 9-day period from days 7 to 16, when the larvae were actively feeding and growing. The weight values used for this calculation were obtained from the individual measurements of 20 larvae per replicate. No spoilage was observed during the study, and survival rates exceeded 96% across all treatments, confirming the adequacy of the substrate quantity relative to larval density. To assess larval body condition, the term body mass index (BMI) was adapted for insect morphometry. In this study, BMI was evaluated using ratios between body length, width, and weight to reflect the general condition and biomass allocation of individual larvae and prepupae. The length-to-body weight, width-to-body weight, and length-to-width indices were then calculated as follows:
LWI = (L/W) × 100
WWI = (D/W) × 100
LWR = L/D
BWI = W/(L/D)
where LWI = length-to-body weight index, WWI = width-to-body weight index, LWR = length-to-width ratio, BWI = body-weight-to-(length/width) ratio, L = body length (mm), D = body width (mm), and W = body weight (mg).
At 23 days of age, the number of live larvae and prepupae was counted in each replicate, and the survival rate was calculated as follows:
Survival (%) = (Ns/N 0) × 100
where Ns = number of surviving larvae and prepupae at day 23 and N₀ = initial number of larvae at day 7 (initial, n = 200).
The percentage of transformation from larvae to prepupae was calculated as follows:
Transformation (%) = (Np/N 0) × 100
where Np = number of prepupae at day 23 (identified by dark cuticle).
Feed intake (FI) was measured for each replicate by subtracting the frass and feed residuals (together) in each pen from the amount of substrate added, assuming that BSFL and BSFP consumed all the feed because of the difficulty in separating feed from frass. Although other studies have separated larvae, prepupae, and frass using representative subsampling techniques, in our experimental setup, the full separation of these components was not feasible due to the intimate mixing of uneaten feed and frass within the substrate matrix. This limitation only applies to FI measurements. However, for assessing growth parameters, survival rate, and transformation percentage, we manually separated larvae and prepupae based on distinct morphological characteristics (e.g. cuticle colour and rigidity), which allowed for accurate individual counting and weighing. The economic efficiency study was carried out based on input and output analyses, with the input presenting the cost of feeding as the only variable, while other factors were constant among treatments (breeding cost, BSFL cost, and others). The cost of 1 ton of laying hens feed was 1800 Saudi Riyal (SR; 1 USD = 3.756 SR) for the control, 1550 SR for the 25% VFW diet, 1300 SR for the 50% VFW diet, 1050 SR for the 75% VFW diet, and 800 SR for the 100% VFW diet. The price of 1 ton of the black soldier flies = 1000 SR, while a ton of organic waste (VFW) = 800 SR. Economic calculations were based on feed costs (input) and insect yield (output), and computed as follows:
FC = A × C
TR = Wᵢ × P
NR = TR − FC
EE (%) = (NR/FC) × 100
where FC = feeding cost, A = amount of feed per animal, C = cost per ton of feed, TR = total revenue, Wᵢ = weight of produced insect biomass (ton), P = price per ton of insect biomass, NR = net revenue, and EE = economic efficiency.
Larvae and prepupae were randomly chosen from each replicate, euthanised by immersion in hot water at approximately 90–95°C for 1–2 min, dried in a force-ventilated electric oven at 65°C, and stored at room temperature. Then, BSFL and BSFP were ground to a fine powder using an electric dry mill. The powder was then stored in well-tied black plastic bags at room temperature (∼25°C) for further analysis.
Analytical methods
Proximate analysis of experimental materials
On day 23, samples comprising larvae, prepupae, substrate, and frass were collected. Larvae and prepupae were manually removed, while the frass, consisting of larval excreta and unconverted substrate, was collected and analysed as a single material, as its components were visually and physically indistinguishable following larval activity. This composite sample was subjected to proximate chemical composition, mineral content, and chitin analyses, following the methods established by the Association of Official Analytical Chemists (AOAC, 2003). The DM was determined using a force-ventilated oven at 70°C until constant weight. The number of samples was equal to the number of replicates per treatment. Crude protein (CP) was determined using the Kjeldahl method, and a nitrogen-to-protein conversion factor of 4.76 for BSFL and BSFP (Janssen et al., Reference Janssen, Vincken, van den Broek, Fogliano and Lakemond2017) and 6.25 for substrates was applied. Ether extract (EE) was measured using diethyl ether as a solvent with a series of automatic solvent extractors (SER 158/6). The crude fibre (CF) was analysed using a Velp fibre analyser (FIWE 6) according to Van Soest et al. (Reference Van Soest, Robertson and Lewis1991). The ash content was determined via complete combustion in a muffle furnace at 550°C for 24 h. Gross energy (GE) was analysed using a bomb calorimeter (Model ASTM D240): a GDY-1C high-precision coal calorific value meter oxygen bomb calorimeter (ASTM D240). The NFE represents the digestible carbohydrate fraction of the sample, including sugars and starches. All values were expressed on a DM basis. Nitrogen-free extract (NFE) was calculated as:
NFE (%) = 100 − (CP + EE + CF + ash)
Samples were digested in 6 N HCl, and an atomic absorption spectrophotometer single-beam DW-AA4530F (8 Lamp Stand) was used to analyse sodium, K, magnesium, calcium, zinc, and copper (Cu) in BSF frass, in addition to iron, manganese, and selenium in the substrate. The P was determined using flame atomic absorption spectrometry (Kirkbright and Marshall, Reference Kirkbright and Marshall1973). Chitin content was quantified as described by D’Hondt et al. (Reference D’Hondt, Soetemans, Bastiaens, Maesen, Jespers, Van den Bosch, Voorspoels and Elst2020). The non-purified, dried sample was first hydrolysed to release glucosamine and acetate, followed by ultra-performance liquid chromatography–tandem mass spectrometry (AB Sciex QTRAP 5500) analysis to determine the glucosamine concentration and HPLC-RID (Agilent Technologies, T200) for acetate measurements. Chitin content was calculated according to Kaya et al. (Reference Kaya, Sargin and Mujtaba2016) as:
Chitin (%) = (m_c/m_b) × 100
where m_c = mass after demineralisation and deproteinisation (DM) and m_b = initial biomass (DM).
The fertilising value of BSF frass was expressed using total N, total P, and total K concentrations (g/kg, DM basis), as commonly reported for frass characterisation and comparison with organic fertilisers (Gärttling and Schulz, Reference Gärttling and Schulz2022; Lopes et al., Reference Lopes, Jean, Yong and Cecilia Lalander2022). Frass macronutrient concentrations are known to vary with the rearing substrate and processing conditions; therefore, N–P–K values are typically discussed within literature-reported ranges rather than as fixed specifications (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022). For each replicate, nutrient ratios were calculated as: NP = N/P, NK = N/K, and PK = P/K (Lopes et al., Reference Lopes, Jean, Yong and Cecilia Lalander2022). In addition, the NPK ratio was reported descriptively in a normalised fertiliser-like form with N set to 1 as 1:(P/N):(K/N), where P/N and K/N were calculated per replicate and then summarised within each substrate level as mean ± SD (Gärttling and Schulz, Reference Gärttling and Schulz2022). This normalisation is algebraically equivalent to scaling the N:P:K composition by N (i.e. dividing all components by N) to facilitate comparison of relative nutrient balance among treatments (Gärttling and Schulz, Reference Gärttling and Schulz2022).
Statistical analysis
The sample size and normal distribution of the data and their errors were tested using the Shapiro–Wilk test (Reference Shapiro and Wilk1965). The random distribution of the data collected for analyses confirmed the reversibility of the four analyses of variance assumptions (ANOVA): (1) individual observations are mutually independent, (2) the data adhere to an additive effect statistical model comprising fixed effects and random errors, (3) the random error is normally distributed, and (4) the random errors have homogeneous variation. Homoscedasticity (variance homogeneity) was evaluated using Levene’s test (SAS, 2009). The data were statistically analysed as preplanned using one-way ANOVA with SAS (2009), and replicates were used as experimental units for most of the traits using the following statistical model:
Yij = μ + Ti + eij
where Yij = the dependent variable, μ = the overall mean, Ti = the effect of substrate level, and eij = the random error.
For the growth traits and chemical composition of larvae and prepupae at 23 days of age, two factorial analyses of variance were applied, with substrate level and stage of development (larvae vs. prepupae) as main effects, including their interaction.
The statistical model used is as follows:
Yijk = μ + Ti + Sj +(TS)ij + eijk
where Yijk = the dependent variable, μ = the overall mean, Ti = effect of substrate level, Sj = stage of development, (TS)ij = the interaction between substrate level and stage of development, and eijk = the random error.
Variables showing significant variation were tested using Tukey’s test (SAS, 2009). The results were presented based on means ± SD. For frass fertilising traits, one-way ANOVA was applied to N, P, K, NP (N/P), NK (N/K), PK (P/K), P/N, and K/N using replicate as the experimental unit. The normalised NPK ratio reported as 1:(P/N):(K/N) was treated as a descriptive index; therefore, no significance letters were assigned to the NPK column, while statistical inference refers to P/N and K/N separately.
Results
Growth performance and body measurements at 16 days of age
The effects of different substrate concentrations (i.e. substrate levels, defined as the percentage of VFW replacing LHF: 0%, 25%, 50%, 75%, and 100%) on BSFL and BSFP growth and body measurements are presented in table 2. The results indicated that the larval body weight at 7 days of age was similar across all feed mixture groups, indicating no significant initial differences at the start of the experiment. At 16 days of age, the growth rate of larvae between 7 and 16 days of age showed that larvae raised on a substrate containing 75% VFW had a higher live body weight than the other groups, except for larvae that were provided a substrate containing 50% VFW. There were no significant differences between the groups with 25% and 100% VFW substrates. The length, width, body weight ratio, and length/width ratio (BMI) were not affected by the different substrate compositions at 7 and 16 days of age. However, larval length at 16 days of age in the 50% and 75% VFW substrates was significantly longer than that of the control and 100% VFW groups. The larval length in the 25% VFW group was intermediate. Changes in larval length at 16 days of age reflected changes in body weight and growth rate. It is worth mentioning that at 23 days of age, the larvae developed into prepupae, and the conversion percentages (from larval to prepupal) ranged from 32.9% to 50.8% and were not significantly different among the various substrates. These data are shown in table 2. Statistical analysis using one-way ANOVA followed by Tukey’s test (p > 0.05) confirmed that the differences among groups were not significant.
Effect of different levels of VFW (substrate) on body measurements (means ± SD) of BSFL and BSFP
abc Indicates that within a row, different subscripts are significantly different.
Body weight of BSFL and BSFP as affected by substrate level and/or stage of development at 23 days of age
Table 3 shows the influence of different levels of substrate and the developmental stage on the growth rate and body measurements of 23-day-old larvae and prepupae. Growth rate, body length, length/width ratio, and body weight index were not significantly affected by substrate level, except for body width, which was significantly increased in groups that provided 25%, 50%, and 75% substrate compared to the 100% substrate level, whereas the control substrate displayed intermediate values. Most body measurements were significantly affected by the developmental stage (larvae vs. prepupae) at 23 days of age. The results showed that the larvae had a significantly higher body weight than the prepupae, as well as a greater length-to-body weight index (as defined in the Methods section), although they exhibited lower body width. The growth performance of larvae and prepupae at 23 days of age was not significantly affected by the interaction between substrate level and stage of development.
Effect of different levels of VFW (substrate) and/or development stage on growth performance (means ± SD) of BSFL and BSFP at 23 days of age
ab Indicates that within a row, different subscripts are significantly different.
Feed intake and feed conversion
Table 4 represents the effects of various levels of substrate on FI, substrate reduction, feed conversion, and survival rate of BSFL and BSFP at 23 days of age. The highest (p < 0.05) FI was recorded in the 100% VFW group, whereas the lowest was observed in the control group (0% VFW). FI increased by increasing substrate levels in a stepwise manner, showing a VFW dilution effect on nutrient concentrations. The highest reduction in substrate % (p < 0.05) was recorded for the groups reared on substrates containing 75% and 100% VFW, and the smallest was from the control group (0% VFW), again showing a stepwise decline. The best feed conversion (FCR; g feed/g gain) and feed conversion (%) were recorded in the groups reared on the control substrates (0% VFW). Increasing the VFW in the substrate significantly impaired the FCR of the larvae, reaching the highest (i.e. worst) values in the 75% and 100% substrate groups. This impairment in feed conversion efficiency at higher VFW inclusion levels may be related to diet chemistry, particularly the increased fibre content of the substrates, which has been shown to influence BSFL growth performance through diet-driven shifts in gut microbial communities and associated digestive functions (Gurusinga et al., Reference Gurusinga, Hurka, Kreft, Röthig, Vilcinskas and Tegtmeier2026). The survival rates ranged from 96.7% to 99.1% and were similar among the different substrates, showing no significant differences among the different treatment groups.
Effect of different levels of VFW (substrate) on feed intake (FI, g), substrate reduction, feed conversion, and survival rate (means ± SD) of BSFL and BSFP at 23 days of age
abcde Indicates that within a row, different subscripts are significantly different.
Nutrient composition of larval and prepupae
Table 5 shows the proximate chemical composition of 23-day-old BSFL and BSFP as affected by substrate level and developmental stage. BSFL reared on the 75% VFW substrates showed significantly lower DM percentages compared to those reared on 0–50% VFW. The 100% VFW group showed intermediate DM. This may reflect a higher water content or lower accumulation of dry solids in the larvae, although the exact mechanism cannot be determined from the current data. Here, DM content refers to the percentage of total body weight. The CP content was significantly higher in the 25% VFW substrate group than in the other groups, but that of the NFE was lower and confounded by the interaction between the substrate level and stage of development. The CF was significantly the highest in the 50% VFW group and the lowest in the control and 75% VFW groups, respectively. The 25% VFW group had higher CF than the 100% VFW group. Additionally, the NFE of the 50% VFW group was significantly lower than that of the 100% VFW substrate group. The effects of substrate levels on the CF and NFE were confounded by the significant interaction between the dietary substrate level and the development stage. The results showed that CF content was significantly higher in the prepupae groups than larvae corresponding to concatenations of substrates. The difference in NFE was obvious between larvae and prepupae when the control group (0), 50%, and 100% VFW substrates were compared in favour of prepupae. In addition, the number of larvae and prepupae was not equal in each replicate due to natural variation in transformation timing. Therefore, comparisons between larvae and prepupae are presented for descriptive purposes only and are not intended as direct statistical contrasts. In addition, the 100% VFW substrate group showed a lower extract content than the other groups, but the effect was offset by the significant interaction between the dietary substrate level and the development stage. The results showed that larvae had a higher extract content than prepupae when the corresponding levels were compared, which may be due to the physiological cessation of feeding during the prepupal stage. The crude ash and GE values were not affected by the different VFW substrates or the interaction between the dietary substrate composition and the development stage. Chitin content was higher in the control and 25% VFW substrate groups than in the other groups, whereas the 50% VFW substrate group displayed the lowest chitin levels; however, the effect was confounded by the interaction between dietary substrate level and the development stage. The results indicated that larvae had lower chitin than prepupae when the corresponding levels were compared. The results indicated that CP, EE, and GE values were significantly higher in the larval groups than in the prepupae groups, whereas DM, crude ash, CF, chitin, and NFE were lower.
Effect of different levels of VFW (substrate) and/or development stage on nutrient composition (means ± SD) of BSFL and BSFP at 23 days of age
abcde Indicates that within a row, different subscripts are significantly different. GE = gross energy. NFE = 100 − (CP + EE + CF + ash); all values on DM basis, DM = dry matter. CP = crude protein. EE = ether extract. CF = crude fiber. NFE = nitrogen free extract.
Nutrient composition of frass of BSFP and BSFL
Table 6 represents the effects of different substrate levels on the nutrient composition of the frass of 23-day-old BSF. It was found that the DM% was significantly lower than the 25% VFW group, 75% and 100% VFW groups. Meanwhile, 50% of the VFW group was intermediate. The results indicated that the highest (p < 0.05) ash% in frass were from BSF reared on the control substrate, whereas those of the other groups were similar. The percentage of CP increased (p < 0.05) in a stepwise manner as substrate levels increased above 25%. The difference between the control level and the 25% level was insignificant. There was a significant increase in frass EE% with substrates containing different VFW compared to the control group, and the effect was maximised at 100% VFW. The lowest EE was observed in BSF reared on the control substrate. Significantly higher CF content (%) was recorded in the BSF groups reared on control substrates than in the other groups, whereas the lowest CF content was observed in BSF reared on substrates containing 100% VFW, whereas the groups reared on 25%, 50%, and 75% substrate levels were intermediate. There was no significant difference in crude ash in the groups offered substrates containing 25%, 50%, 75%, and 100% VFW, but the value was significantly lower than the control. NFE showed only significant differences between groups, provided the control and 25% VFW diets and was higher in the latter group.
Effect of different levels of VFW (substrate) on nutrient composition (means ± SD) of frass of 23-day-old BSFL and BSFP
abcd Indicates that within a row, different subscripts are significantly different. NFE = 100 − (CP + EE + CF + ash); all values on DM basis, DM = dry matter. CP = crude protein. EE = ether extract. CF = crude fiber. NFE = nitrogen free extract.
Frass mineral composition
Table 7 presents the effects of different substrate levels on the mineral composition of frass of 23-day-old larvae and prepupae. The results indicate a stepwise decrease (p < 0.05) in Zn and Cu (ppm) in frass with increasing substrate levels. The highest Zn and Cu concentrations were observed in the frass collected from BSF reared on the control substrates, whereas the lowest amounts were observed in BSF reared on 100% VFW substrates. The highest Na% was observed in BSF reared on substrates containing 25% VFW and was similar to the content of BSF reared on substrates containing 50% VFW, whereas the lowest content was observed in BSF reared on substrates supplied with 75% VFW. The control and 100% substrate groups were similar and displayed intermediate values. K% showed a stepwise increase (p < 0.05) with increasing substrate levels and was maximised at 100% VFW substrate and minimised in the control. The differences between the 50%, 75%, and 100% substrate levels were insignificant. The highest Mg% was observed in BSF reared on control substrates and was similar to that of BSF reared on substrates containing 100% VFW, but these values were higher than those of the 50% VFW substrate groups. The highest Ca was found in BSF reared on the control and up to 75% VFW substrate concentrations but providing 100% VFW substrate decreased frass Ca contents in BSF frass. The frass P% was similar in the 25–100% substrate groups, but higher than that of the control group.
Effect of different levels of VFW (substrate) on mineral composition (means ± SD) of frass of 23-day-old BSFL and BSFP
abcd Indicates that within a row, different subscripts are significantly different.
Frass fertilising value
Table 8 shows the effect of different substrate levels on the fertilising value of frass derived from 23-day-old BSFL and BSFP. Organic matter was significantly higher in frass from the 50–100% VFW treatments (372.5–383.0 g/kg) compared with the control and 25% VFW groups (361.0–369.1 g/kg) (p < 0.0001). Frass N (g/kg) increased significantly with increasing VFW inclusion above 25%, with no difference between the control and 25% VFW, and the highest N content observed at 100% VFW (16.26 g/kg) (p < 0.0001). P (g/kg) was significantly lower in the control (26.59 g/kg) than in all VFW treatments (27.26–27.38 g/kg) (p < 0.0001). K (g/kg) increased with VFW level, reaching the highest value in the 100% VFW treatment (39.02 g/kg), whereas the control showed the lowest K content (33.21 g/kg) (p < 0.0001). The N:P ratio (NP) increased from 0.514 in the 25% VFW group to 0.596 at 100% VFW (p < 0.0001), while the N:K ratio (NK) was highest in the control and in the 75–100% VFW treatments (0.417–0.430) and lowest in the 25–50% VFW treatments (0.394–0.398) (p = 0.0003). Conversely, the P:K ratio (PK) declined from 0.804 in the control to 0.699 in the 100% VFW group (p < 0.0001). The normalised NPK ratio (reported descriptively as 1:P/N:K/N) ranged from 1:1.95:2.55 in the 25% VFW treatment to 1:1.68:2.40 in the 100% VFW treatment. In line with Tukey’s post hoc test, substrate level significantly affected both P/N (p < 0.0001) and K/N (p = 0.0002): P/N was highest at 25% VFW, intermediate in the control and 50% VFW groups (not different from each other), and progressively lower at 75% and 100% VFW, with the minimum at 100% VFW. For K/N, the 25% VFW group showed the highest values, the 50% VFW treatment was intermediate (not different from either 25% or the lower groups), whereas the control and the 75–100% VFW treatments formed the lower range.
Effect of different levels of VFW (substrate) on the fertilising value (means ± SD) of frass of 23-day-old BSFL and BSFP
Note: The NPK ratio is reported descriptively as 1:(P/N):(K/N) (normalised to N = 1). P/N and K/N are presented as mean ± SD. No significance letters are assigned to the NPK ratio.
abcd Different letters within a column indicate significant differences among treatments (p < 0.05). Organic matter (g/kg) = DM (g/kg) − ash (g/kg).
Economic efficiency
Table 9 represents the effect of different levels of substrate on the economic efficiency traits of BSF at 7–23 days of age. The highest feeding costs were recorded for the 25%, 50%, and 75% VFW groups, whereas the lowest costs were recorded for the control diet and 100% VFW groups. Total and net revenues did not differ significantly among the different substrate groups. However, according to the input and output analyses, the best economic efficiency was recorded for the groups reared on the control diet and those reared on the complete replacement of the layer diet with VFW, compared with the 75% VFW groups. The 25% and 50% VFW groups were similar to those of the other groups.
Effect of different levels of VFW (substrate) on economic efficiency traits (means ± SD) of BSFL and BSFP during 7–23 days of age
ab Indicates that within a row, means with different superscripts differ significantly (p < 0.05). Control diet (layer feed) = 1800 SR/ton. Diets replacing layer feed with VFW were priced as follows: 25% VFW = 1550 SR/ton; 50% VFW = 1300 SR/ton; 75% VFW = 1050 SR/ton; and 100% VFW = 800 SR/ton. Price of BSF biomass = 1000 SR/ton. Exchange rate: 1 USD = 3.756 SR. Feeding cost (FC) = FI per animal × feed price (SR/ton), after converting FI to tons (1 ton = 1000 kg). Total revenue (TR) = insect biomass × 1000 SR/ton (after converting biomass to tons). Net revenue (NR) = TR − FC. Economic efficiency (EE, %) = (NR/FC) × 100.
Discussion
The current results showed that up to 75% of the VFW substrate can be used for rearing BSF from 7 to 23 days of age without negative effects on growth traits (body weight, growth rate, and body length). Although several studies identify days 15–16 as the optimal harvesting window to maximise larval biomass yield and nutrient concentration (Diener et al., Reference Diener, Zurbrügg and Tockner2009; Nguyen et al., Reference Nguyen, Tomberlin and Vanlaerhoven2015; Barragan-Fonseca et al., Reference Barragan-Fonseca, Gort, Dicke and Van Loon2017), the present study was designed to extend the analysis beyond the larval stage, evaluating the nutritional and economic potential of prepupae under delayed harvesting conditions. This approach allows a more comprehensive assessment of production scenarios that may occur in commercial systems, where harvesting is not always synchronised. It is well established that the prepupal stage is characterised by the cessation of feeding and the mobilisation of endogenous energy reserves, processes that are associated with a decline in feed conversion efficiency (Barragan-Fonseca et al., Reference Barragan-Fonseca, Gort, Dicke and Van Loon2021) and may consequently reduce overall bioconversion performance. However, under our rearing conditions (27 ± 0.5°C), approximately 30% of larvae naturally transitioned to the prepupal stage by day 23, providing a realistic snapshot of commercial harvesting scenarios where synchronous development is not always achieved. Despite this, survival rates remained high (>96%), and total biomass and economic efficiency were not adversely affected. Future studies may optimise harvest timing based on targeted production goals (e.g. maximal protein yield vs. bioconversion efficiency). However, harvesting BSF during the mid-larval stage (approximately day 16) appeared to yield better growth performance than harvesting at 23 days, suggesting this developmental window may be more suitable for optimising biomass yield. From an economic point of view, up to 100% VFW did not differ from the control diet, showing that replacing laying hen diets with VFW could be used as a substrate for BSF, while reducing the cost of substrates by 41.7% and increasing economic efficiency compared with 75% VFW. Although the GE content of the 100% VFW substrate was numerically higher compared to the control, the FCR was significantly impaired, indicating reduced feed efficiency. This suggests that energy use by BSF may not correspond directly with gross caloric content. A key factor could be the higher fibre content in the 100% VFW substrate, which is known to reduce nutrient digestibility and absorption. Moreover, the absence of direct measurements of carbohydrate and simple sugar fractions represents a limitation for a comprehensive evaluation of dietary energy balance and nutrient utilisation. Although NFE is commonly used as a proxy for the carbohydrate fraction, in the present study, it was calculated by difference (100 − (CP + EE + CF + ash)). Consequently, this indirect estimation is inherently influenced by the cumulative analytical errors associated with the quantification of the other proximate components, thereby constraining the accuracy of NFE as an indicator of true carbohydrate availability. Previous studies have shown that the protein-to-carbohydrate ratio plays a crucial role in insect feed efficiency and growth (Barragan-Fonseca et al., Reference Barragan-Fonseca, Gort, Dicke and Van Loon2021). The high fibre and unquantified sugar content in the 100% VFW diet may have contributed to lower effective feed quality despite apparent GE values. The higher FCR indicates a decline in feed utilisation efficiency, likely attributable to dietary energy dilution and the associated incremental increase in FI. This stepwise increase in FI was observed between days 7 and 23 across the increasing VFW inclusion levels (from 0% to 100%), as reported in table 4. The decline in efficiency reflects a reduced ability of the larvae to convert feed into biomass under the 100% VFW substrate condition. The current study indicated that different VFW substrates did not negatively affect the survival rate, suggesting that VFW up to 100% as a substrate for rearing BSF had no toxic or harmful effects on viability. In accordance with the current study, several studies have investigated the effects of different FW diets and streams on BSF production (Nguyen et al., Reference Nguyen, Wilcox, Skjaerven and Baird2007; Tinder et al., Reference Tinder, Puckett, Turner, Cammack and Tomberlin2017) and concluded that the substrate and temperature are key aspects of the life cycle of BSFL (e.g. survival, growth rate, maturity, and reproduction). BSFL has been successfully reared on a wide range of substrates such as domestic poultry litter/manure, cattle manure, vegetable residues, dried distillers’ grains with soluble, catering waste, straw, poultry, and fish offal (St-Hilaire et al., Reference St-Hilaire, Cranfill, McGuire, Mosley, Tomberlin, Newton, Sealey, Sheppard and Irving2007; Nguyen et al., Reference Nguyen, Tomberlin and Vanlaerhoven2015; Tinder et al., Reference Tinder, Puckett, Turner, Cammack and Tomberlin2017; Van Huis and Tomberlin, Reference Van Huis and Tomberlin2017; Meneguz et al., Reference Meneguz, Schiavone, Gai, Dama, Lussiana, Renna and Gasco2018; Addeo et al., Reference Addeo, Randazzo, Olivotto, Messina, Tulli, Vozzo, Attia, Mahayri, Iannaccone, Asiry, Moniello and Bovera2022). As observed in the present study, at 16 days of age, rearing substrate influences the weight and length of larvae, and adding a mineral and vitamin mixture affects growth traits (Starcevic et al., Reference Starcevic, Lozica, Gavrilovic, Heruc and Masek2019). However, the developmental parameters documented herein differed according to the rearing medium/substrate used (Sideris and Tsagkarakis, Reference Sideris and Tsagkarakis2017). The latter authors showed that the quantity and source of the substrate substantially affected the immature developmental period. In general, a shorter immature stage was observed in BSFL-fed poultry feed than in catering waste. Furthermore, as previously demonstrated (e.g. Diener et al., Reference Diener, Zurbrügg and Tockner2009; Nguyen et al., Reference Nguyen, Tomberlin and Vanlaerhoven2013), increasing the daily feed amount per larva significantly enhances growth and reduces the development period. In the present study, the substrate treatments that provided higher FI per larva (∼250 mg/day based on batch allocation) were associated with longer larvae and greater weight at 16-day-old, indicating faster development. Conversely, groups with lower FI (∼150 mg/larva/day) exhibited shorter length and reduced growth. These findings suggest that larval length is a good proxy for evaluating nutrient intake and developmental progress. The relationship between dietary provision and larval growth performance is illustrated in table 2, which reports body length and weight, and in table 4, which summarises FI. The feeding range used aligns with batch feeding protocols described by Diener et al. (Reference Diener, Zurbrügg and Tockner2009), who identified 100–250 mg/larva/day as optimal for growth and bioconversion. Based on our observations (table 4), the observed increase in larva-to-prepupal conversion percentage with higher FI suggests that greater nutrient availability may accelerate larval development. This likely enables larvae to reach the physiological threshold for pupation more rapidly, as also reported by Sideris and Tsagkarakis (Reference Sideris and Tsagkarakis2017). This may be due to the higher nutrient availability accelerating larval development and allowing the larvae to reach the physiological threshold necessary for pupation more rapidly. It is important to clarify that the term bioconversion used here refers to the overall transformation of the substrate into insect biomass, representing a yield-based indicator (e.g. biomass produced per substrate; Nugroho et al., Reference Nugroho, Rofiq, Santoso, Yanuar, Hanifa and Nadirah2023). In contrast, feed conversion refers specifically to the FCR, which is an efficiency measure calculated as the amount of feed consumed per unit of biomass gain (g feed/g gain). Although these metrics are related, they are not interchangeable. For example, higher FIs observed in the 100% VFW group – may increase bioconversion in absolute terms but result in a poorer feed conversion ratio due to lower nutrient efficiency. Therefore, the values observed do not conflict but rather reflect different aspects of performance. Lim et al. (Reference Lim, Mohd-Noor, Wong, Lam, Goh, Beniers, Oh, Jumbri and Ghani2019) achieved the highest bioconversion efficiency of 14% in larval biolipids, after accounting for a metabolic loss of 54%. The feed conversion percentage observed herein ranged from 10.5% to 23.5%, which agreed with those reported by Lim et al. (Reference Lim, Mohd-Noor, Wong, Lam, Goh, Beniers, Oh, Jumbri and Ghani2019). In the current study, substrate reduction ranged from 67.2% to 89.4% and was significantly minimised in the control substrate (0% VFW), maximised in 75% and 100% VFW substrates, and parallels the alterations in FI, and in contrast with FCR as g feed/g weight gain or as %. It should also be noted that substrate mass loss may not be solely attributable to larval ingestion. The nutrient profiles on day 23 BSFL and BSFP of age were altered by the VFW level, showing that DM, organic compounds (CP, EE, and NFE), and chitin were significantly altered. The results showed that DM content increased in control, 25%, and 50% VFW substrate groups, but decreased significantly in the 75% VFW, especially in prepupae. This decrease in DM may be related to higher moisture retention or differences in metabolic activity between larval stages. In addition, the CP content was the highest among the groups reared on substrates containing 25% VFW. In accordance with the current results, Fadhillah and Bagastyo (Reference Fadhillah and Bagastyo2020) found that the CP content of BSFL ranged from 48.75% to 63.1% depending on the substrate composition. From the DM and CP standpoints, the 25% VFW substrate resulted in the highest value for animal nutrition. However, the observed reduction in EE in the 100% VFW group should also be interpreted in light of the developmental stage. During the 23 days, a higher proportion of prepupae were present, and prepupae are known to cease feeding and mobilise stored lipids, resulting in naturally lower lipid content compared to larvae. This developmental factor likely contributed to the lower EE values observed, potentially confounding the direct effects of substrate composition. This also explains why differences in EE were not significant across larval groups alone, as shown in table 5. The relationship between substrate composition and BSFL nutrient profile was recently reported by Tschirner and Simon (Reference Tschirner and Simon2015) and Fischer and Romano (Reference Fischer and Romano2021). The current results agree with several reports that the substrate used for rearing insects determines the nutritional composition of the BSFL (Nguyen et al., Reference Nguyen, Tomberlin and Vanlaerhoven2013; Spranghers et al., Reference Spranghers, Ottoboni, Klootwijk, Ovyn, Deboosere, De Meulenaer, Michiels, Eeckhout, De Clercq and De Smet2017). BSFL are strongly influenced by the media in which they breed because if they are rich in protein, larvae will contain high protein, and the same is true in media with high levels of fat or fibre (Addeo et al., Reference Addeo, Vozzo, Secci, Mastellone, Piccolo, Lombardi, Parisi, Asiry, Attia and Bovera2021, Reference Addeo, Randazzo, Olivotto, Messina, Tulli, Vozzo, Attia, Mahayri, Iannaccone, Asiry, Moniello and Bovera2022, Reference Addeo, Tucciarone, Bovera, Vozzo, Secci and Parisi2024b). Meneguz et al. (Reference Meneguz, Schiavone, Gai, Dama, Lussiana, Renna and Gasco2018) reported that the CP of BSFL was 41.9%, 30.8%, 34.4%, and 53.0% using a mixture of vegetables and fruits only, winery, and brewery by-products, as substrates, respectively. Additionally, Danieli et al. (Reference Danieli, Lussiana, Gasco, Amici and Ronchi2019), Moula et al. (Reference Moula, Scippo, Douny, Degand, Dawans, Cabaraux, Hornick, Medigo, Leroy, Francis and Detillick2018b), Oonincx et al. (Reference Oonincx, Van Huis and Van Loon2015), and Oonincx et al. (Reference Oonincx, Laurent, Veenenbos and van Loon2020) documented that substrate affected the CP content of BSF and BSF-derived products. On the other hand, Meneguz et al. (Reference Meneguz, Schiavone, Gai, Dama, Lussiana, Renna and Gasco2018) indicated that not only the level but also the quality of protein (fruit waste vs. brewery by-products) has a crucial impact on its content in larvae. Moreover, Barragan-Fonseca et al. (Reference Barragan-Fonseca, Gort, Dicke and Van Loon2021) estimated that both protein and carbohydrate levels similarly affect protein content in BSF biomass. In addition, Barragan-Fonseca et al. (Reference Barragan-Fonseca, Gort, Dicke and Van Loon2019) and Beniers and Graham (Reference Beniers and Graham2019) reported that the CP content in the BSF biomass was negatively correlated with increased CP concentration in the rearing medium, the EE content varied over a considerable range (approximately 6% to above 30%), and increasing concentrations of substrate CP and carbohydrates resulted in high fat levels. Kim et al. (Reference Kim, Kim, Jeong, Lee, Kim, Lee and Lee2020) and Spranghers et al. (Reference Spranghers, Ottoboni, Klootwijk, Ovyn, Deboosere, De Meulenaer, Michiels, Eeckhout, De Clercq and De Smet2017) observed that the nutritional value of BSFL varies with substrate and age, and its high CP content makes it an excellent source of nutrition for animal feed (Shumo et al., Reference Shumo, Osuga, Khamis, Tanga, Fiaboe, Subramanian, Ekesi, van Huis and Borgemeister2019). Meneguz et al. (Reference Meneguz, Schiavone, Gai, Dama, Lussiana, Renna and Gasco2018) indicated that not only the level but also the quality of protein has a crucial impact on its content in BSFL. Aristi et al. (Reference Aristi, Mudji, Soepranianondo, Haridjani and Aprilia2020) reported that VFW affects the protein content of BSFL, with the highest average protein content found in tomato medium and the highest average fat content found in carrot medium. The larval biomass produced through the bioconversion of the substrate is a high-quality source of proteins and lipids/fat/oil and is increasingly used in a number of industries for animal feeding and the production of biodiesel and biopol (Nyakeri et al., Reference Nyakeri, Ogola, Ayieko and Amimo2017; Kim et al., Reference Kim, Kim, Jeong, Lee, Kim, Lee and Lee2020; Proc et al., Reference Proc, Bulak, Wia and Bieganowski2020). Lim et al. (Reference Lim, Mohd-Noor, Wong, Lam, Goh, Beniers, Oh, Jumbri and Ghani2019) administered a mixed waste of coconut endosperm and soybean crude residue substrates to increase the protein nutritional constituents in BSFL. Aminuddin et al. (Reference Aminuddin, Elias, Moh and Zainalabidin2023) and Ellawidana et al. (Reference Ellawidana, Mutucumarana, Ruwandeepik and Magamage2023) documented that BSF can efficiently convert various VFWs (as livestock manure, soybean curd residue, and food waste) and produce larval biomass rich in protein (Abd El-Hack et al., Reference Abd El-Hack, Shafi, Alghamdi, Abdelnour, Shehata, Noreldin and Ragni2020; Dortmans et al., Reference Dortmans, Egger, Diener and Zurbrügg2021). The frass analysed in this study is a composite matrix of larval excreta mixed with unconverted substrate residues, because complete separation was not feasible; therefore, statistical differences in frass composition should be interpreted as reflecting both larval metabolism and the nutrient profile of the remaining substrate fraction (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). The substrate level had a clear and significant effect on the fertilising traits of the resulting frass. The consistent significance patterns across organic matter and macronutrients indicate that increasing VFW inclusion modifies nutrient retention in the frass matrix, rather than producing random variation. In particular, the fact that total P differed mainly between the control and VFW-fed groups but showed limited separation among VFW levels suggests that P enrichment is achieved once VFW is included, while further increases in VFW do not proportionally alter total P. In contrast, the significant trends observed for N and especially K across substrate levels indicate a stronger substrate-driven shift in these nutrients (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). Although 100% VFW produced statistically higher nutrient concentrations in some traits, this should be interpreted together with the performance data, because reduced feed efficiency under nutritionally unbalanced substrates may limit nutrient utilisation by larvae and thereby increase the fraction remaining in the residual frass matrix. Diet-dependent shifts in nutrient partitioning between larval biomass and frass have been associated with differences in substrate chemistry and digestive processes in BSFL reared on chemically diverse side streams (Gurusinga et al., Reference Gurusinga, Hurka, Kreft, Röthig, Vilcinskas and Tegtmeier2026). Diet macronutrient concentrations and balance are known to affect BSF performance and survival, supporting the plausibility of this trade-off (Barragan-Fonseca et al., Reference Barragan-Fonseca, Gort, Dicke and Van Loon2021). The current results indicate that the frass–residual substrate mixture obtained after BSF rearing contains agronomically relevant macronutrient levels and may therefore represent a potential source of soil amendment and nutrient recycling. Based on the elemental composition (DM basis), total nutrient levels fall within the broad ranges reported for BSF frass in the literature, while remaining strongly dependent on substrate type and processing conditions. (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). These results are consistent with those of numerous studies highlighting the importance of BSFL in managing organic waste and assessing decomposable waste (Pastor et al., Reference Pastor, Velasquez, Gobbi and Rojo2015; Nyakeri et al., Reference Nyakeri, Ogola, Ayieko and Amimo2017; Spranghers et al., Reference Spranghers, Ottoboni, Klootwijk, Ovyn, Deboosere, De Meulenaer, Michiels, Eeckhout, De Clercq and De Smet2017; da Silva and Hesselberg, Reference da Silva and Hesselberg2020; Adebayo et al., Reference Adebayo, Kemabonta, Ogbogu, Elechi and Obe2021; Purkayastha and Sarkar, Reference Purkayastha and Sarkar2021). In addition, the leftover material at the end of the breeding period of BSF can be recycled to yield compost, such as soil amendment and plant fertilisers (Bortolini et al., Reference Bortolini, Macavei, Saadoun, Foca, Ulrici, Bernini, Malferrari, Setti, Ronga and Maistrello2020). Aminuddin et al. (Reference Aminuddin, Elias, Moh and Zainalabidin2023) reported that BSF can renovate various organic wastes, such as livestock manure, soybean crude residue and food waste, and produce larval-rich protein sources, chitin, biofertiliser, and biodiesel. The latter authors revealed that banana peels could be a valuable resource for producing BSFL as a protein supplement, with the resulting frass serving as a nutrient-rich by-product (Acevedo et al., Reference Acevedo, Carrillo, Flórez-López and Grande-Tovar2021; Riyandani et al., Reference Riyandani, Rasyid and Baja2021). P concentrations in organic fertilisers can vary widely depending on the raw materials and processing methods. For example, composts and manures typically contain relatively low P levels, whereas bone-derived amendments may contain substantially higher P contents (Penhallegon, Reference Penhallegon2003; Adekiya et al., Reference Adekiya, Ejue, Olayanju, Dunsin, Aboyeji, Aremu, Adegbite and Akinpelu2020). In BSF systems, published frass analyses also report a broad range of P concentrations, reflecting strong substrate dependence (Setti et al., Reference Setti, Francia, Pulvirenti, Gigliano, Zaccardelli, Pane, Caradonia, Bortolini, Maistrello and Ronga2019; Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). In the present study, total P in frass was higher in all VFW treatments than in the control but showed limited differentiation among VFW inclusion levels, suggesting that once VFW was included, P enrichment was achieved without a proportional increase across higher VFW levels. While the macronutrient balance expressed as an NPK ratio (reported here in normalised form as 1:(P/N):(K/N)) provides an initial indication of the potential fertilising profile of BSF frass, it does not fully capture nutrient bioavailability to plants. Actual nutrient uptake depends on nutrient speciation and release kinetics (e.g. organic vs. mineral N forms and soluble vs. sparingly soluble P), as well as soil pH and interactions with the soil microbiome (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). Although the present study quantified total N, P, and K in the frass matrix, we did not assess nutrient solubility or plant-available fractions through extraction assays or pot/field trials. Therefore, the NPK values should be interpreted as total nutrient content and relative balance, rather than direct fertiliser efficiency (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). Previous research (Setti et al., Reference Setti, Francia, Pulvirenti, Gigliano, Zaccardelli, Pane, Caradonia, Bortolini, Maistrello and Ronga2019; Romano et al., Reference Romano, Sinha, Powell and Fischer2022) has shown that BSFL frass can release a portion of its N and P in plant-available forms, but the extent depends strongly on the rearing substrate and composting conditions. Future work should evaluate mineral availability through pot experiments or extraction assays to better define the agronomic value of BSFL frass as fertiliser. To summarise nutrient balance and compare treatments, the relative proportions of N, P, and K in the frass matrix were expressed as an NPK ratio on an elemental basis. The ratio was reported in a normalised fertiliser-like form with N set to 1, i.e. 1:(P/N):(K/N), derived from analytical concentrations (g/kg DM). This normalisation is a simple scaling of the N:P:K composition that facilitates comparison of relative nutrient balance among treatments, but it should be interpreted as a descriptive indicator rather than a direct fertiliser recommendation (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). In the present study, the normalised NPK ratio ranged from 1:1.95:2.55 in the 25% VFW group to 1:1.67:2.39 in the 100% VFW group, with intermediate values observed in the other treatments. Importantly, statistical inference was applied to the underlying components of this index rather than to the combined NPK string: substrate level significantly affected both P/N and K/N. Specifically, P/N was highest in the 25% VFW treatment and declined progressively toward the 100% VFW group, indicating a relative shift toward greater N enrichment compared with P at higher VFW inclusion. In contrast, K/N showed a different pattern, with higher values in the 25–50% VFW treatments compared with the control and the 75–100% VFW groups, suggesting that moderate VFW inclusion increased the relative contribution of K compared with N, whereas further increases in VFW did not proportionally enhance this balance. Similar fertiliser-oriented profiling and the strong substrate dependence of BSF frass nutrient composition have been highlighted in the literature (Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022; Gärttling and Schulz, Reference Gärttling and Schulz2022). Published studies report wide variability in BSF frass nutrient profiles depending on the substrate and processing conditions; for example, Magee et al. (Reference Magee, Halstead, Small and Young2021) reported frass with an N–P–K grade of 4.9–2.6–1.7 (reported on a fertiliser-grade basis) (Magee et al., Reference Magee, Halstead, Small and Young2021). From an agronomic perspective, the practical fertiliser value of frass depends not only on its N, P, and K content, but also on organic matter input, nutrient release dynamics, soil properties, and crop-specific nutrient requirements across growth stages. Therefore, NPK ratios should be interpreted as an initial indicator of nutrient balance rather than a direct prescription for fertilisation rates (Zewide and Reta, Reference Zewdie and Reta2021; Basri et al., Reference Basri, Azman, Ahmad, Suja, Jalil and Amrul2022). The K content of organic fertilisers ranges from 1.95% to 3.79% in different manure types (Adekiya et al., Reference Adekiya, Ejue, Olayanju, Dunsin, Aboyeji, Aremu, Adegbite and Akinpelu2020); 1.0–2.0% in regular compost, 2.5% in bat guano (Penhallegon, Reference Penhallegon2003), and 4% in wood ash (Mikkelsen, Reference Mikkelsen2008). This is not the only criterion for assessing the need for fertilisers, especially when the soil is deficient in certain essential minerals such as Cu and Zn, which are the most commonly deficient in soils globally (Willmore et al., Reference Willmore, Hall and Drewnoski2021). In the present study, Zn ranged from 95 to 102 ppm and Cu content ranged from 47.5 to 44.7 ppm in the frass of the control substrate and 100% VFW substrate, which are a reflection of the primary rearing substrate levels. Despite 2–3-fold higher Zn and 5–6-fold higher Cu concentrations than the rearing substrates, respectively, these concentrations were within the range found in the livestock manure and reflected the dietary mineral content in the livestock feeds (DeLaune and Moore, Reference DeLaune and Moore2016). The concentrations of Zn and Cu found in the BSFL frass were in agreement with published reports (Gärttling and Schulz, Reference Gärttling and Schulz2022). However, in the present study, Cu was not analysed in either the larvae or the original substrate. Therefore, although the Cu content in the frass appeared to reflect trends in the diet, we cannot confirm whether Cu was enriched in the larval biomass. As previous research suggests that BSFL may bioaccumulate Cu (Addeo et al., Reference Addeo, Scivicco, Vozzo, Bovera, Asiry, Alqurashi, Cacciola and Severino2024a), further studies should directly assess Cu levels in larvae and substrates to confirm such accumulation. It may be suggested that farming BSFL on animal manure and other substrates rich in microelements may enhance the micronutrients of BSFL while reducing manure volume, pathogenic contamination and green gas emission (Beskin et al., Reference Beskin, Holcomb, Cammack, Crippen, Knap, Sweet and Tomberlin2018; Lalander et al., Reference Lalander, Diener, Zurbrügg and Vinnerås2019). Moreover, Ca found in frass ranged from 3.33% to 3.55% within a narrow range; however, Ca is essential for plants to maintain their cellular and structural integrity. The Na content ranged from 1.46% to 1.59% showing 10 times higher than that in the substrate, which competes with Ca uptake, along with Mg, K, and water, resulting in leaf and root burns (Bryson and Barker, Reference Bryson and Barker2002). Increasing Na levels in frass is undesirable because of increasing water and soil salinity. The concentrations found herein were relatively low in the frass, and substantially lower than inorganic fertilisers containing Na, such as sodium nitrate (NaNO3) and sodium borate (borax; Na₂[B₄O₅(OH)₄]8 H₂O). Finally, Mg was higher in frass produced from the 0% VFW control (2.56%) than in the 50% substrate group (2.30%), and these values were higher than those of various manure sources reported in the literature 0.10–0.64% (Adekiya et al., Reference Adekiya, Ejue, Olayanju, Dunsin, Aboyeji, Aremu, Adegbite and Akinpelu2020). Romano et al. (Reference Romano, Sinha, Powell and Fischer2022) concluded that levels of Mg, Ca, K, P, Fe, Ni, Zn, and Mn in frasses generally reflected the levels in the initial substrate; however, different values are quoted in various reports. These authors mentioned that fruit peels are unlikely to be optimal substrates for BSFL growth and development; even though they provide relatively high levels of minerals in the BSFL and frass and therefore could be a beneficial mineral (fertiliser) supplement. In general, organic fertilisers can provide a more equal distribution of nutrients to plants and are less prone to nutrient runoff during heavy rains compared with chemical fertilisers, making them environmentally friendly and potentially more economical (Shaviv, Reference Shaviv2001; Shaji et al., Reference Shaji, Chandran and Mathew2021). Moreover, insect frass contains chitin, which reportedly enhances plant health while reducing pests (Sharp, Reference Sharp2013). Nevertheless, the quality of fertilisers is often quickly gauged by N–P–K values, which are macronutrients required by plants (Magee et al., Reference Magee, Halstead, Small and Young2021). In the current study, N values measured in the frass ranged from 14.0 to 16.3 g/kg; however, it is often reported that BSFL frass exceeds 20 g/kg (Setti et al., Reference Setti, Francia, Pulvirenti, Gigliano, Zaccardelli, Pane, Caradonia, Bortolini, Maistrello and Ronga2019; Gärttling and Schulz, Reference Gärttling and Schulz2022), even when N-low fruits are used as the sole source of substrate (Fischer and Romano, Reference Fischer and Romano2021). Such N values of approximately 20 g/kg are comparable to those of regular compost containing 15–35 g/kg (Penhallegon, Reference Penhallegon2003) or in various manures, which contain 1.01–2.91% (Adekiya et al., Reference Adekiya, Ejue, Olayanju, Dunsin, Aboyeji, Aremu, Adegbite and Akinpelu2020). There were two main differences in mineral composition and levels between BSFL and frass. First, Ca, Cu, Ni, and K contents were higher in the BSFP than in the BSFL. In the case of Ca, this was likely due to the BSFL continually shedding their chitinous exoskeleton, which is rich in calcium carbonate (Barragan-Fonseca et al., Reference Barragan-Fonseca, Gort, Dicke and Van Loon2017). Additionally, while VFW often had the highest mineral content in BSFL, only the K and Mn contents in the frass were significantly higher when frasses were composted with BSFL. According to the input and output analyses, the best economic efficiency was recorded for the groups reared on the control diet (0% VFW) and 100% VFW substrate. These results are compatible with those of Diener et al. (Reference Diener, Zurbrügg and Tockner2009) and Spranghers et al. (Reference Spranghers, Ottoboni, Klootwijk, Ovyn, Deboosere, De Meulenaer, Michiels, Eeckhout, De Clercq and De Smet2017), who indicated that the most important economic advantages of BSF technology are the comparatively low cost of processing organic waste and high revenue from the sale of high-value proteins produced from larvae. Onsongo et al. (Reference Onsongo, Osuga, Gachuiri, Wachira, Miano, Tanga, Ekesi, Nakimbugwe and Fiaboe2018) reported that full-fat BSFL can replace up to 15% of the soybean and fish meal diets in broiler diets, lowering the feed cost by 19%. This suggests that BSFL may already be a more affordable ingredient option with no negative effects on poultry production parameters in some countries. Waithaka et al. (Reference Waithaka, Osuga, Kabuage, Subramanian, Muriithi, Wachira and Tanga2022) demonstrated that BSFL meal might be used as a necessary element of the manufacturing of poultry feed as a meal or live larvae for better indigenous chicken, potentially lowering the cost of total feed while maintaining optimal productivity and lowering the price of meat and egg products. Additionally, the waste stream left behind after collecting fly larvae may be converted into fertiliser, resulting in additional financial gain from the production of insects for food (Roffeis et al., Reference Roffeis, Wakefield, Almeida, Alves Valada, Devic, Koné and Muys2018; Xiao et al., Reference Xiao, Mazza, Yu, Cai, Zheng, Tomberlin and Zhang2018). In addition, the advantages of the BSF system make it fit well in a circular economy, because it is profitable to process organic matter that has decomposed into larval biomass (Bortolini et al., Reference Bortolini, Macavei, Saadoun, Foca, Ulrici, Bernini, Malferrari, Setti, Ronga and Maistrello2020). Because no residue remains after the process, it can be reintroduced into the framework to provide compost-like properties as a soil amendment.
Conclusion
This study demonstrates that VFW can be effectively used as a rearing substrate for BSFL and BSFP from 7 to 23 days of age at inclusion levels up to 75% without compromising growth performance. The 25% VFW substrate yielded the highest CP content in the larvae, while the 100% VFW substrate produced frass with the highest nutrient content, although it negatively affected feed conversion efficiency. From an economic perspective, both the control (0% VFW) and 100% VFW groups exhibited the highest economic efficiency. However, only the 75% VFW substrate provided a favourable balance between insect biomass production and frass fertilising value. Therefore, the use of up to 75% VFW in the substrate is recommended to optimise insect growth performance, nutrient recycling, and frass quality for potential agricultural applications.
Acknowledgement
This project was funded by the Deanship of Scientific Research (DSR) at King Abdulaziz University, Jeddah, Saudi Arabia under grant no. (IPP: 99-155-2025), The Authors, therefore, acknowledge with thanks DSR for technical and financial support.
Competing interests
The authors declare no conflict of interest.
