Dose–response effects of dietary inclusion of agro‐industrial by‐products on in vitro ruminal fermentation and methane production
Agro‐Industrial by‐products: Their effects on in vitro ruminal fermentation
B Xue et al.
Institute for Global Food Security Queen's University Belfast Belfast UK
Sustainable Livestock Systems Agri‐Food and Biosciences Institute Hillsborough UK
School of Agriculture, Policy and Development University of Reading Reading UK
* Correspondence to: K Theodoridou, Institute for Global Food Security, Queen's University Belfast, Belfast BT9 5DL, UK. E‐mail: k.theodoridou@qub.ac.uk (Theodoridou)Abstract
BACKGROUND
As the agro‐industry produces considerable amounts of by‐products globally, it is acknowledged that there is a need to address the environmental issues related to their disposal and the resource competition between food for humans and feed for animals. The aim of this study was to explore, in vitro, the effects of various by‐products from the agro‐industry on rumen fermentation and methane emission. Samples were collected from various food processing industries, including red and green apple pomace (RAP, GAP), hempseed cake (HC), coffee hulls (CH), coffee grounds (CG), spent mushroom compost (SMC) and distiller's dried grains with solubles (DDGS). In doses of 100, 200 and 300 g kg−1, the tested by‐products were incubated in rumen fluid, where the by‐products replaced equal amounts of substrates.
RESULTS
Gas production (GP) and dry matter digestibility (DMD) decreased linearly for most of the tested by‐products with the growth of doses (P < 0.001), while NH3‐N concentration increased linearly. Linear decreases were observed in CH4 production with increasing doses of all by‐products (P < 0.05). The reduction of CH4 production ranged from 21.4% to 33.6% at doses of 100–300 g kg−1, but reductions were only observed at a dose of 100 g kg−1 when CH4 productions were corrected by digested dry matter (P < 0.05). RAP, GAP and HC were higher than CH, CG and SMC for the comparison of key parameters including DMD, GP and volatile fatty acids. Better methane‐mitigating effects were observed for RAP, GAP and HC than for the control group and CH, CG and SMC.
CONCLUSION
Most of the by‐products tested were found to be a potential option for replacing conventional feed ingredients but should not exceed a dose at 200 g kg−1. © 2025 The Author(s). Journal of the Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.
INTRODUCTION
The total amount of agro‐industrial by‐products in the UK is 2.6 million tons, which makes the UK one of the top producers in Europe, only second to Germany with 3 million tons. 1 Agro‐industrial by‐products are residual materials generated during the processing of agricultural and industrial products. These by‐products often retain substantial nutritional or varieties of bioactive components, which depend on the type of by‐products itself and the processing methods. 2 , 3 Nevertheless, nearly half of those by‐products are disposed of in unsustainable procedures such as landfilling and incineration, causing considerable negative impact on the environment. 4 Characterized by high fibre and protein content, by‐products from the food processing industry are widely used for animal feed, especially ruminants, as fibre can be the primary source of their energy supply while not digestible for monogastric livestock. Industry and academia all over the world are increasingly focusing on sustainable approaches to minimize waste and extract value from these by‐products, in which transforming them to animal feed is of relatively high efficiency. 5
As previous studies reported, the doses of by‐products were found to be of vital importance when included in diets for livestock. 6 Researchers have summarized the safe doses at which fruit by‐products can be added to the diets of different livestock, where the feeding risk has been taken into prime consideration. 7 However, doses need to be more accurate when it comes to large‐scale farms, so it is greatly necessary to conduct in vitro assessments for these by‐products to clarify the dose–response effects, thereby providing feasibility and guidance on in vivo experiments and their application. A study investigated the in vitro effect of citrus pulp inclusion levels at 0, 100, 200 and 300 g kg−1 dry matter (DM) in total mixed ration (TMR) in goats. The digestibility of organic matter and DM increased linearly with the inclusion levels, while short‐chain fatty acids and metabolizable energy reached a maximum at 300 g kg−1. 8 Another study that conducted in vitro trials by replacing dried tomato pomace in diets at doses of 60, 120 and 180 g kg−1 DM found differences in total volatile fatty acids (VFA) and NH3‐N after 8 h incubation among groups, but most of the difference disappeared after 24 h. 9 Furthermore, species‐specific differences in optimal doses are significant in in vivo study, even for the same by‐products. A study that set coffee grounds inclusion levels at 0, 100 and 200 g kg−1 DM in TMR found that increasing levels linearly decreased the digestibility of DM, protein, fibre and retained nitrogen. 10 The DMI decreased to the lowest when feeding at 200 g kg−1 DM dose (96.6 versus 94.8 versus 76.8 g/(body weight)0.75). Intriguingly, two research studies on dairy cows, that also set 100 and 200 g kg−1 DM inclusion levels, respectively, found no significant difference in the VFA, N‐nitrogen and gas production among groups. 11 , 12
In vitro studies related to agro‐industrial by‐products usually focus on rumen fermentation parameters, 13 but ignore the presence of bioactive compounds, such as total phenolics, which have already proved to be of high efficiency in inhibiting rumen methanogenic archaea. 14 , 15 For specific by‐products rich in phenolics, the mitigation of CH4 production in in vitro studies is rarely reported. More importantly. although a great number of potential by‐products and their potentially optimal doses were explored for in vitro incubation, 16 , 17 more accurate information is poorly investigated when it comes to the comparison among the by‐products from different agro‐industries. In the study reported here, samples were collected from various food processing industries, including juice, coffee, whiskey, mushroom and hempseed. The different effects caused by dose and type of by‐product are equally important for us to focus on, allowing selection of the best by‐product candidates for application in animal trials.
Therefore, the aim of the study was to explore, in vitro, the effects of various by‐products from agro‐industry on rumen fermentation and methane emission. The specific objectives were: (1) to explore the nutritive value and total phenolics of the by‐products collected and (2) to assess the dose–response effects of including increasing doses (100, 200, 300 g kg−1 DM) of the seven by‐products on in vitro ruminal fermentation and methane production.
MATERIALS AND METHODS
Preparation of samples
The agro‐industrial by‐products used in this study were red and green apple pomace (RAP, GAP), hempseed cake (HC), coffee hulls (CH), coffee grounds (CG), spent mushroom compost (SMC) and distiller's dried grains with solubles (DDGS). RAP and GAP were obtained from MacNeice Fruit Ltd and Moorstown, Co. Tipperary, Ireland, respectively. HC was collected from UK Hemp Co., UK. CH were collected from a local micro‐roastery based in Belfast, UK. CG were collected from a local coffee shop in Belfast. SMC was from Agri‐Food and Bioscience Institute (AFBI). The rest of the substrates for in vitro incubation, silage and concentrate were collected from AFBI in Hillsborough, Northern Ireland, UK. All samples were dried in a freeze dryer for 72 h, and then ground to pass through a 1 mm sieve.
Experimental design
Experiments were conducted to measure in vitro rumen fermentation characteristics of diets with different inclusion levels of agro‐industrial by‐products, to determine the dose–response effects. Tested by‐products on DM basis were incubated for 24 h in doses of 100, 200 and 300 g kg−1 DM substrates. These amounts replaced equal amounts of the mixed ration (500 mg), which was composed of silage and concentrate (70:30).
Based on a literature review, those by‐products with low protein content, apple pomace, coffee products and mushroom compost, were designed to replace the silage portion, while those with high protein content, HC and DDGS, were designed to replace the concentrate portion, keeping nitrogen balanced in the diets. Details are presented in Table 1. Each of the 21 treatments (7 by‐products × 3 doses) was repeated in two independent in vitro runs over 2 weeks. In addition, each run included triplicate of treatment diets, quadruplicate of control diets (substrate alone without by‐products) and blanks. The total number of experimental units were (21 treatments × 3 + control diets × 4) × 2 runs = 134, which were used in the statistical analysis.
Experimental procedures and sampling
In vitro incubations for this study were performed according to Menke and Steingass. 18 Rumen fluid was collected before the morning feeding from three cannulated non‐lactating Holstein cows fed a ration consisting of 700 g kg−1 grass silage and 300 g kg−1 of a commercial concentrate mix (133 ± 7.8 g kg−1 crude protein (CP), 33 ± 5.8 g kg−1 ether extract (EE), 275 ± 6.1 g kg−1 neutral detergent fibre (NDF) and 99 ± 1.1 g kg−1 ash) twice daily at the abattoir centre for AFBI (1 L from each cow). The rumen contents were transferred into three thermos flasks and immediately transported to the laboratory. Rumen contents were strained through four layers of cheesecloth into an Erlenmeyer flask, followed by mixing with 2× in vitro rumen buffer solution under strictly anaerobic conditions. 19 Subsequently, the buffered rumen fluids from three cattle were homogenized. An amount of 50 mL of rumen fluid/buffer mixture with a ratio of 1:2 was transferred into 125 mL serum bottles containing 0.5 g of experimental diets prepared already under continuous flushing with O2‐free CO2 gas. The bottles were sealed with butyl rubber stoppers and caps made of aluminium and incubated at 39 °C with continuous rotation for 24 h.
At the end of incubation, for each serum bottle, gas production was measured using a pressure transducer and a syringe to collect gas stored for 24 h in a 20 mL gas chromatography vial for methane (CH4) analysis. Once opened, the entire contents of each serum bottle were transferred to a pre‐weighed 50 mL falcon tube, followed by measuring pH immediately. Supernatants were sampled into 2 mL Eppendorf tubes frozen at −20 °C for VFA and NH3‐N analysis. Each serum bottle was washed twice with distilled water to recover all the nondegraded particles that were transferred into the 50 mL falcon tube. Tubes were centrifuged at 3400 rpm for 10 min at 4 °C. Once the supernatant was removed, the residue was obtained, followed by transferring to an oven immediately.
Chemical analysis
For the chemical composition of by‐products, contents of ash (method 942.05) and EE (method 920.39) were analysed as described by AOAC (2000). 20 Concentrations of NDF and acid detergent fibre (ADF) were determined following the procedures of Van Soest et al. 21 and Robertson and Van Soest, 22 respectively, using an ANKOM220 fibre analyser unit (ANKOM Technology Corporation, Fairport, NY, USA). Nitrogen (N) concentrations were determined by the Dumas combustion technique employing a Leco FP258 N analyser (Leco Corporation, St Joseph, MI, USA), and CP concentration (g kg−1 DM) was then calculated as N concentration × 6.25. Concentrations of NH3‐N were determined by the phenol–hypochlorite method. 23 The concentration of VFA was analysed using gas chromatography as described by Huhtanen et al. 24 Non‐fibrous carbohydrate (NFC) was calculated as 1000 – CP − EE – ash − NDF based on NRC (2001). 25 All analyses were performed in triplicate. Gas production (GP) was calculated based on pressure measurements according to the following equation 26 :where V h represents head‐space volume (mL), P a atmospheric pressure (psi) and P t pressure transducer reading (psi). Standard P a value of 14.7 psi was used and V h value of 70 mL.
Quantification of total phenolics in by‐products
Measurement of total phenolics was conducted based on the methods using Folin–Ciocalteu reagent. 27 Total phenolics were extracted from the seven by‐products in triplicate through 70% aqueous acetone (n = 3), followed by a series of different dilutions, ×2, ×5, ×20, mixed with Folin–Ciocalteu reagent and Na2CO3 reagent, and finally absorbance measurements of each sample were recorded at 725 nm using a spectrophotometer.
Measurement of methane production
Gas samples were collected from bottles’ headspace for methane (CH₄) analysis. A Terumo™ Agani™ 18‐gauge, 1.5‐inch needle, coupled with a gas stopper and a 12 mL syringe, was employed to extract 10 mL of gas from each serum bottle, which was then transferred into a 12 mL evacuated Exetainer® vial. Methane concentration was quantified via gas chromatography using an HP 5890 Series II chromatograph with an HP‐Innovax column (25 mm × 0.2 mm × 0.2 μm, Supelco). The carrier gas was nitrogen at 1 mL min−1. The injector and detector temperatures were maintained at 250 and 275 °C, respectively, while the oven temperature was held at 110 °C under isothermal conditions. A 0.1 mL gas sample was injected with a 1 mL sample‐lock syringe. Methane levels were calibrated using a standard curve created through manual injections of six different quantities of pure CH₄ in triplicate, and the final CH₄ concentration was expressed in mL of CH₄ per mL of sample.
Calculation and statistics analysis
Calculation of the in vitro DM digestibility (IVDMD) was conducted as follows:where X = initial weight (g) and Y = dry residue weight (g).
Calculation of total phenolics contents (for use with ×20 dilution) was as follows:where X = phenolics (μg) and Y = dry residue weight (mg). Results were expressed as mg gallic acid equivalent (GAE) per g DM.
All data were analysed using mixed linear model in SPSS (Ver.22.0 for Windows; SPSS, Chicago, IL, USA). Model 1 was used to evaluate the overall effects of by‐products on rumen fermentation, with by‐products, dose and their interaction as fixed effects, run as random effect. Model 2 was used to compare the by‐products' effects within the doses and doses’ effects within the by‐products. Linear and quadratic effects of dose within by‐products were evaluated by orthogonal polynomial contrasts. Post hoc multiple comparisons were performed using the Sidak test. These results are presented as means and standard error of means. Statistical differences were declared significant when P < 0.05 and declared as a tendency when 0.05 ≤ P < 0.10.
RESULTS
Chemical composition of ingredients and experimental diets
The CP content was highest in HC at 376.9 g kg−1 DM, numerically followed by DDGS at 317 g kg−1 DM, while for the others it was below 170 g kg−1 (Table 2). The highest NDF was found in CH and CG, over 630 g kg−1 DM, numerically followed by SMC and HC, whereas RAP and GAP had the lowest at 346 and 318 g kg−1. The ADF content of each by‐product was numerically lower than the NDF content (210–529 g kg−1 DM). Apple by‐products have the highest of NFC at 670 g kg−1 DM, while coffee by‐products and SMC have the lowest. Fat and ash contents range from 29 to 136 g kg−1 DM and from 18 to 230 g kg−1 DM respectively.
| Item | Diet ingredients | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GS | Con | RAP | GAP | HC | CH | CG | SMC | DDGS | |
| DM (g kg−1) | 918 | 942 | 923 | 895 | 899 | 947 | 983 | 945 | 889 |
| GE (MJ kg−1 DM) | 19.2 | 18.1 | 13.4 | 18.4 | 22.0 | 20.5 | 23.0 | 16.2 | 21.6 |
| CP (g kg−1 DM) | 137 | 270 | 53.0 | 65.0 | 327 | 165 | 155 | 145 | 317 |
| EE (g kg−1 DM) | 48.0 | 27.0 | 29.0 | 37.0 | 88.0 | 15.0 | 136 | 31.0 | 60.0 |
| NDF (g kg−1 DM) | 441 | 285 | 346 | 318 | 359 | 530 | 537 | 552 | 370 |
| ADF (g kg−1 DM) | 297 | 182 | 210 | 210 | 309 | 429 | 400 | 426 | 225 |
| Ash (g kg−1 DM) | 11.0 | 76.0 | 18.0 | 20.0 | 77.0 | 70.0 | 21.0 | 230 | 59.0 |
| NFC (g kg−1 DM) | 363 | 342 | 554 | 560 | 150 | 221 | 151 | 43.0 | 194 |
For chemical composition in each dietary treatment (Table 3), CP ranged from 152 to 194 g kg−1 DM, with HC and DDGS diets having the highest. NDF ranged from 366 to 428 g kg−1 DM and ADF ranged 236 to 302 g kg−1 DM, with SMC diets having the highest content. A numerically noticeable difference is the ash content in SMC diets at 52.4–92.2 g kg−1 DM but there was only 30.5 g kg−1 DM in control diets. Fat contents in HC and CG diets at 47.8–60 and 50.5–68.1 g kg−1 DM, respectively, were numerically higher than that of control at 41.7 g kg−1 DM. NFC ranged from 261 to 416 g kg−1 DM, with RAP and GAP diets having the highest.
| Item | Dietary treatments | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CON | RAP 10 | RAP 20 | RAP 30 | GAP 10 | GAP 20 | GAP 30 | HC 10 | HC 20 | HC 30 | CH 10 | CH 20 | CH 30 | CG 10 | CG 20 | CG 30 | SMC 10 | SMC 20 | SMC 30 | DDGs 10 | DDGs 20 | DDGs 30 | |
| Chemical composition (g kg−1) | ||||||||||||||||||||||
| DM | 925 | 926 | 926 | 927 | 923 | 921 | 918 | 926 | 926 | 927 | 928 | 931 | 934 | 932 | 938 | 945 | 928 | 931 | 933 | 920 | 915 | 909 |
| OM | 970 | 969 | 968 | 967 | 969 | 968 | 967 | 971 | 972 | 973 | 964 | 958 | 952 | 969 | 968 | 967 | 948 | 926 | 904 | 971 | 973 | 975 |
| GE | 18.9 | 18.3 | 17.7 | 17.1 | 18.8 | 18.7 | 18.6 | 19.3 | 19.7 | 20.0 | 19 | 19.1 | 19.3 | 19.3 | 19.6 | 20.0 | 18.6 | 18.3 | 18.0 | 19.2 | 19.6 | 19.9 |
| CP | 178 | 169 | 160 | 152 | 170 | 163 | 155 | 183 | 188 | 194 | 180 | 183 | 185 | 179 | 181 | 182 | 178 | 179 | 179 | 182 | 186 | 191 |
| NDF | 394 | 385 | 375 | 366 | 382 | 370 | 357 | 402 | 409 | 416 | 403 | 412 | 422 | 404 | 413 | 423 | 405 | 416 | 428 | 403 | 411 | 420 |
| ADF | 263 | 254 | 245 | 236 | 254 | 245 | 236 | 275 | 288 | 301 | 276 | 289 | 302 | 273 | 283 | 293 | 275 | 288 | 301 | 267 | 271 | 275 |
| EE | 41.7 | 39.8 | 37.9 | 36 | 40.6 | 39.5 | 38.4 | 47.8 | 53.9 | 60 | 38.0 | 35.0 | 32.0 | 50.5 | 59.3 | 68.1 | 40 | 38.3 | 36.6 | 45.0 | 48.3 | 51.6 |
| Ash | 30.5 | 31.2 | 31.9 | 32.6 | 31.4 | 32.3 | 33.2 | 29.5 | 28.5 | 27.5 | 36.4 | 42.3 | 48.2 | 31.5 | 32.5 | 33.5 | 52.4 | 74.3 | 96.2 | 28.8 | 27.1 | 25.4 |
| NFC | 357 | 376 | 395 | 414 | 376 | 396 | 416 | 338 | 318 | 299 | 343 | 330 | 316 | 336 | 314 | 293 | 325 | 293 | 261 | 342 | 327 | 312 |
Total phenolics contents of by‐products
The total phenolics contents contained in the tested by‐products are presented in Fig. 1. The two highest levels of total phenolics contents were found in CG and GAP, at 4.99 and 3.82 mg GAE g−1 DM, followed by those of CH, RAP, HC and DDGS, at 2.39, 1.88, 1.82 and 1.07 mg GAE g−1 DM. SMC had the lowest level of total phenolic content at 0.5 mg GAE g−1 DM.
Fermentation characteristics
Based on model 1 (Table 4), IVDMD and pH were significantly affected by the doses and by‐products (P < 0.001) and significant interactions were observed (P < 0.001). Based on model 2, the pH linearly increased with increasing doses of HC, CH, CG and SMC (P < 0.001). As the dose of by‐products increased, a linear decline in IVDMD was observed for most of the by‐products tested (P < 0.001), except RAP with a quadratic trend. CH, CG and SMC at each dose was lower than other groups (P < 0.05). At doses of 100 g kg−1 DM, lower IVDMD and GP were observed for CH, CG and SMC while for doses of 20 and 300 g kg−1 DM, the differences were more significant (P < 0.05).
| Item | Dose (g kg−1) | CON | RAP | GAP | HC | CH | CG | SMC | DDGs | SE | P | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Type | Dose | T × D | |||||||||||
| IVDMD (%) | 0.63 | <0.001 | <0.001 | <0.001 | |||||||||
| 100 | 65.0a | 66.0a | 64.9a | 62.2ab | 60.1b | 59.2b | 60.5b | 62.5ab | |||||
| 200 | 65.0ab | 68.7a | 63.6abc | 49.4e | 56.5cde | 56.3de | 58.3bcd | 62.3abcd | |||||
| 300 | 65.0a | 64.7a | 57.7abc | 46.2e | 53.1cde | 48.5de | 54.7bcd | 62.0ab | |||||
| Contrast † | L | L | L | L | L | L | L | ||||||
| pH | 0.01 | <0.001 | <0.001 | <0.001 | |||||||||
| 100 | 6.18c | 6.20bc | 6.19c | 6.21abc | 6.24abc | 6.26ab | 6.27a | 6.18c | |||||
| 200 | 6.18bc | 6.22b | 6.15c | 6.29a | 6.28a | 6.28a | 6.31a | 6.16bc | |||||
| 300 | 6.18b | 6.18b | 6.14b | 6.33a | 6.32a | 6.33a | 6.34a | 6.19b | |||||
| Contrast | Q | L | L | L | L | L Q | — | ||||||
Based on model 1 (Figs 2 and 3), NH3‐N and GP were significantly affected by the doses and by‐products (P < 0.001) and significant interactions were observed (P < 0.001). Based on model 2, most of the tested by‐products linearly decreased NH3‐N and GP (P < 0.05).
Production of volatile fatty acids
Based on model 1 (Table 5), by‐product type caused significant effects on the production of VFA (P < 0.001) but doses did not. Based on model 2, inclusion of HC, CH, CG and SMC linearly decreased total VFA (P < 0.05). Similar trends were observed in the production of individual VFA, including acetate, propionate, butyrate and valerate (P < 0.05). With an increase in the dose of by‐products, the acetate‐to‐propionate ratio (A:P) linearly increased for CH, CG and SMC (P < 0.05), while linear decreases were observed for RAP, GAP and DDGS (P < 0.05). At doses of 100, 200 and 300 g kg−1 DM of by‐product inclusion, all individual VFA were lower in CG and SMC than in other groups (P < 0.05). At all doses of by‐product inclusion, A:P of HC, CH, CG and SMC was significantly higher than that of RAP, GAP and DDGS (P < 0.05).
| Item | Dose (g kg−1) | CON | RAP | GAP | HC | CH | CG | SMC | DDGS | SE | P | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Type | Dose | T × D | |||||||||||
| TVFA (mmol L−1) | 1.32 | <0.001 | 0.325 | 0.053 | |||||||||
| 100 | 91.2a | 85.2a | 91.3a | 89.9a | 88.5a | 81.6ab | 64.2b | 90.5a | |||||
| 200 | 91.2a | 84.8a | 87.7a | 84.3a | 88.4a | 62.5b | 79.0ab | 94.7a | |||||
| 300 | 91.2a | 88.7a | 88.8a | 85.2a | 79.6a | 58.3b | 75.5ab | 88.1a | |||||
| Contrast † | — | — | L | L | L | Q | — | ||||||
| Acetate (mmol L−1) | 0.93 | <0.001 | 0.308 | 0.081 | |||||||||
| 100 | 60a | 55.5ab | 60.0a | 59.6a | 59.0a | 51.2ab | 43.7b | 59.8a | |||||
| 200 | 60a | 54.7ab | 56.3a | 55.1a | 59.8a | 42.8b | 54.0ab | 62.2a | |||||
| 300 | 60a | 57.4a | 56.4a | 54.7a | 54.5a | 40.3b | 52.2ab | 57.6a | |||||
| Contrast | — | L | L | — | L | Q | — | ||||||
| Propionate | 0.42 | <0.001 | 0.465 | 0.015 | |||||||||
| (mmol L−1) | 100 | 21.4a | 21.5a | 22.1 | 19.8a | 19.8a | 17.9ab | 13.1b | 21.3ab | ||||
| 200 | 21.4a | 21.6a | 22.0a | 19.5ab | 18.9ab | 12.7c | 16.4bc | 22.8a | |||||
| 300 | 21.4ab | 22.9a | 22.1a | 20.1abc | 16.5bc | 11.2d | 15.1cd | 22.2a | |||||
| Contrast | — | — | — | L | L | L Q | — | ||||||
| Butyrate | 0.20 | <0.001 | 0.331 | 0.07 | |||||||||
| (mmol L−1) | 100 | 7.31ab | 6.13ab | 6.94ab | 7.66a | 7.28ab | 6.66ab | 5.48b | 6.89ab | ||||
| 200 | 7.31a | 6.36ab | 7.09a | 6.85ab | 7.06a | 5.30b | 6.39ab | 7.02a | |||||
| 300 | 7.31ab | 6.39abc | 7.91a | 7.17ab | 6.32abc | 5.03c | 5.98abc | 5.88bc | |||||
| Contrast | — | — | — | L | L | L Q | L | ||||||
| Valerate | 0.02 | <0.001 | 0.312 | 0.016 | |||||||||
| (mmol L−1) | 100 | 0.95ab | 0.82bc | 0.93ab | 1.06a | 0.95ab | 0.85abc | 0.70c | 1.01ab | ||||
| 200 | 0.95ab | 0.80bc | 0.87ab | 1.04a | 0.91ab | 0.65c | 0.77bc | 1.05a | |||||
| 300 | 0.95ab | 0.83bc | 0.96ab | 1.19a | 0.78bcd | 0.59d | 0.70cd | 0.98ab | |||||
| Contrast | L Q | — | L | L | L | L Q | — | ||||||
| A:P | 0.06 | <0.001 | 0.648 | <0.001 | |||||||||
| 100 | 2.91abc | 2.72c | 2.83bc | 3.10ab | 3.09ab | 2.98abc | 3.20a | 2.88bc | |||||
| 200 | 2.91c | 2.57d | 2.70cd | 2.98bc | 3.29a | 3.26ab | 3.33a | 2.78cd | |||||
| 300 | 2.91b | 2.56c | 2.71ab | 2.78ab | 3.41a | 3.40a | 3.51a | 2.66ab | |||||
| Contrast | L | L | Q | L | L | L | L | ||||||
Production of methane
As evident from Table 6, all of the by‐products with 10, 20 and 300 g kg−1 DM doses significantly decreased the methane production compared with the control group (P < 0.05). For methane production/DDM, significant differences were observed only at the dose of 100 g kg−1 DM, where RAP, GAP and HC were significantly lower than the other groups.
| Treatments | P | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Item | Control | 10% | 20% | 30% | SE | Dose | L | Q | 10% | 20% | 30% |
| CH4 (mL) | — | — | — | 0.003 | 0.026 | 0.013 | |||||
| RAP | 25.7a | 19.0b | 19.5b | 18.3b | 1.08 | 0.001 | 0.001 | 0.051 | — | — | — |
| GAP | 25.7a | 17.2b | 18.0b | 20.0b | 1.15 | 0.001 | 0.001 | 0.001 | — | — | — |
| HC | 25.7a | 20.2b | 17.0b | 17.0b | 0.88 | <0.001 | <0.001 | 0.010 | — | — | — |
| CH | 25.7a | 22.7ab | 20.0bc | 17.7c | 0.93 | 0.001 | <0.001 | 0.772 | — | — | — |
| CG | 25.7a | 20.4ab | 20.2ab | 17.9b | 1.15 | 0.055 | 0.014 | 0.47 | — | — | — |
| SMC | 25.7a | 19.6b | 20.2ab | 18.6b | 1.05 | 0.045 | 0.021 | 0.253 | — | — | — |
| DDGS | 25.7a | 24.6a | 20.8ab | 18.3b | 1.16 | 0.050 | 0.002 | 0.699 | — | — | — |
| CH4/dDM (mL kg−1) | — | — | — | 0.005 | 0.103 | 0.263 | |||||
| RAP | 79.3a | 57.6b | 57.6b | 56.4b | 3.47 | 0.001 | 0.001 | 0.027 | — | — | — |
| GAP | 79.3a | 53.4b | 57.2b | 69.7ab | 4.74 | 0.017 | 0.345 | 0.004 | — | — | — |
| HC | 79.3a | 65.5b | 70.4ab | 75.7ab | 2.12 | 0.043 | 0.709 | 0.011 | — | — | — |
| CH | 79.3 | 76.8 | 71.2 | 67.7 | 3.09 | 0.357 | 0.080 | 0.927 | — | — | — |
| CG | 79.3 | 69.9 | 71.4 | 74.9 | 3.94 | 0.808 | 0.736 | 0.410 | — | — | — |
| SMC | 79.3 | 65.6 | 69.2 | 68.3 | 3.27 | 0.459 | 0.332 | 0.352 | — | — | — |
| DDGS | 79.3 | 79.2 | 66.8 | 59.2 | 3.75 | 0.055 | 0.008 | 0.516 | — | — | — |
For the three doses of by‐products, CH4 production linearly decreased by 26.0%, 24.4% and 29.0% for RAP diets (P < 0.05); quadratically decreased by 33.1%, 29.9% and 22.2% for GAP diets (P < 0.05); linearly decreased by 21.4%, 33.6% and 33.6% for HC diets (P < 0.05); and quadratically or linearly decreased when CH, CG, SMC and DDGS partially replaced diets at a certain dose. CH4/DDM linearly decreased by 27.3%, 27.4% and 28.8% for RAP diets (P < 0.05). CH4/DDM linearly decreased by 32.7% and 27.9% for GAP diets (P < 0.05). CH4/DDM linearly decreased by 25.3% for HC dose of 100 g kg−1 DM (P < 0.05).
DISCUSSION
The increasing focus on sustainable livestock production has spurred interest in identifying alternative feed ingredients that can reduce environmental impact while maintaining or improving production efficiency. Agro‐industrial by‐products represent a promising solution due to their abundance, cost‐effectiveness and potential to enhance rumen fermentation and reduce greenhouse gas emissions. This study investigated the potential of seven distinct by‐products to enhance rumen fermentation and reduce methane emissions, offering a novel perspective by comprehensively comparing their dose‐dependent effects. The selected by‐products, including high‐protein and high‐fibre by‐products, were chosen for their nutritional diversity, availability and potential bioactive properties. The commercial applicability lies in the ability of these by‐products to provide cost‐effective, sustainable feed options while repurposing agricultural waste and mitigating methane emissions, thus supporting both economic and environmental goals.
Nutrient profiles of tested agro‐industrial by‐products
As these potential feed ingredients are derived from agro‐industrial industries, the nutrients are basically of high content. HC, with its high crude protein content and rich rumen undegraded protein, 28 showed potential to replace soybean meal in livestock diets. This aligns with other studies highlighting HC's high protein value and its potential to reduce reliance on imported soybean meal. 29 Apple pomace (RAP and GAP) contained high NFC, consistent with other studies which highlight its suitability as a rapid energy source in ruminant diets. 30 , 31 The nutritive value of coffee by‐products was similar to that of grass silage, while the fat content (15 g kg−1 DM) is in accordance with previously reported values. 32 Our study found that the ash and NFC content differed from previous reports, likely because of soil contamination. Earlier research supports this, showing lower ash‐to‐NFC ratios when SMC is processed differently. 33 , 34
Effects of agro‐industrial by‐products on ruminal fermentation
The pH values recorded in our study align with the normal physiological range (5.8–6.5) reported for the rumen. 35 This indicates that the buffered rumen fluid used successfully simulated the in vivo ruminal environment, ensuring that the experimental conditions closely reflected the natural ruminal conditions. It is well documented that NFC, including starch, contributes to the accumulation of VFA and lactic acid. 36 Similar to our results, Carlos et al. reported that wine lees cause an increase in pH at a dose of 180 g kg−1 DM, 37 but not 60 and 120 g kg−1 DM, possibly in relation to the NFC values.
A linear reduction in digestibility was expected with increasing doses of high‐NDF by‐products, as NDF negatively influences digestibility. 38 The higher concentrations of NDF, ADF and lignin in these fibrous by‐products contribute to the observed decrease in DM digestibility. 39 , 40 This linear trend aligns with another study where digestibility of SMC diets decreased 6.2% at 0–140 g kg−1 DM and 15.8% at 0–300 g kg−1 DM, possibly due to higher unfermentable ash or lower inclusion levels. 41 In contrast, the digestibility of apple by‐products was less affected by increasing doses but should not exceed 200 g kg−1 DM, as previously noted. 42 GP was utilized to assess the fermentation of substrates, serving as an indicator of the digestibility of tested feedstuff, 43 , 44 strongly correlated with IVDMD. Its reduction with increasing doses of fibre‐rich by‐products aligns with previous studies. 12
NH3‐N is substantially crucial for ruminal nitrogen metabolism as it is the intermediate product of protein degradation. For HC and DDGS, the high CP content could be responsible for the linear increase in ammonia concentration. 45 In contrast to the results in our study, Antonio et al. reported that three by‐products at a dose of 200 g kg−1 DM decreased the NH3‐N concentrations compared to control. 46 One possible explanation would be the negative energy–nitrogen balance in by‐product diets from their study, where excess nitrogen could not be utilized due to a lack of corresponding metabolizable energy. 47 An exception to the quadratic trends caused by increasing doses of apple by‐products was observed, where doses of 200 and 300 g kg−1 DM resulted in a reduction in ammonia concentration. This aligns with the findings of a previous in vivo study, 48 where dietary inclusion levels of 5, 10 and 200 g kg−1 DM reduced ammonia concentration by up to 32.8%, which might be explained with the low CP content in apple by‐products. 49 Another reason might be the presence of bioactive compounds able to contribute to lower ruminal protein degradation due to the complexation of tannins and protein. 48 On the other hand, low ruminal ammonia concentrations are linked to reduced nitrogen excretion, leading to decreased nitrogen emissions from livestock. This comparative analysis is an indication of the potential contribution of apple by‐products to a sustainable livestock production.
Furthermore, VFA are the primary products of carbohydrate fermentation in the rumen and serve as the main energy source for ruminants. Structural carbohydrates mainly produce acetate, while non‐structural carbohydrates lead to higher propionate production. 50 The total VFA concentrations found in our study were similar among the treatments, at doses of 250 and 500 g kg−1 DM, 51 in accordance with previous studies conducted in cattle and sheep, 52 , 53 except for CG and SMC, that could be partly explained by the fewer unfermentable carbohydrates. 54 , 55 Depending on nutrient profiles of diets (fibre, fat, NFC and starch), A:P reflects the rumen fermentation pattern, which is positively correlated to forage. 56 , 57 Propionate‐type fermentation caused by RAP, GAP and DDGS is likely to implicate a promising shift that improves dietary energy utilization efficiency as propionate is the substrate of gluconeogenesis. 58 , 59 Sanz et al. observed that replacing soybean meal and barley with pea (from 0 to 1000 g kg−1 DM) led to a linear decline in acetate production, an increase in butyrate and no significant changes in propionate concentration and A:P. 60 In contrast, our study found that the acetate‐type fermentation caused by SMC might be attributed to its fat content, which can increase the propionate‐to‐acetate ratio. 56
Incorporating agro‐industrial by‐products into livestock diets can help reduce enteric CH₄ emissions due to their chemical composition and bioactive compounds. The methane reduction observed for the by‐products studied was higher compared to previous similar studies, 61 , 62 which might be due to their higher fat and phenolic content, as every 10 g kg−1 increase in fat content reduces CH4 production by up to 3.8%. 63 In addition, the bioactive compounds (i.e. polyphenols) present in the industrial by‐products function by altering the microbial community and fermentation in the rumen, thereby reducing methanogenesis. 64 , 65 The total phenolic content in the tested by‐products was greater compared to previous studies: 0.56–2.96 GAE g−1 DM for most fruits and 1.02–1.48 GAE g−1 DM for coffee and other spent waste by‐products. 66 , 67 Comparable CH₄ reduction rate was observed for grape pomace at 21.3%, 68 likely attributable to similar levels of polyphenolic compounds present in these by‐products. Methanogenesis in rumen is primarily driven by methanogens, which utilize H₂ and CO₂ as substrates to produce methane. 69 During the ruminal fermentation of carbohydrates, acetate and hydrogen are the products of the same biochemical reaction, so higher A:P is linked to increased methane production. 70 Consistent with the results of A:P, high doses (200, 300 g kg−1 DM) of RAP, GAP, HC and DDGs presented methane‐mitigating effects. CH₄ reductions were not observed with CH, SG and SMC inclusion (mL/digested DM), likely due to lower digestibility, as indicated by the A:P results. This is in line with Mounir et al., 71 who reported that increasing the dose of SCG from 0 to 200 g kg−1 DM did not affect CH4 production per kilogram of digested organic matter, even though the polyphenol content was over 10 GAE kg−1 DM. The methane‐mitigating effects of GAP, RAP and HC remained significant, highlighting their substantial potential for enhancing dietary energy utilization efficiency. 58 Intriguingly, it was previously reported that brewer's waste can reduce CH₄ production linearly by 11.1%, 27.2% and 37.0% at inclusion levels of 20, 40 and 60 g kg−1 DM, 72 respectively. The greater reduction at lower inclusion levels may be due to differences in diet ingredients or bioactive compounds in the by‐products.
This study utilized a single source for each by‐product to represent its typical compositional profile, reflecting its practical use in livestock diets. While this approach facilitated controlled comparisons and dose–response analysis, it inherently limits the generalizability of the findings to other sources of the same by‐products. Variability in composition due to differences in processing, storage or geographical origin may lead to variations in fermentation outcomes. Future research should focus on evaluating multiple sources of each by‐product to account for such variability and enhance the robustness of the conclusions. Nonetheless, the observed dose–response trends provide valuable insights into the potential of these by‐products for inclusion in ruminant diets.
CONCLUSION
Based on key fermentation parameters as critical performance indicators, the similar effects observed across by‐products at a 100 g kg−1 DM inclusion level suggest it as an optimal choice for practical application. In contrast, the significant differences in fermentation parameters seen with CH, CG and SMC highlight the relatively superior performance of RAP, GAP, HC and DDGS. Therefore, apple by‐products, HC and DDGS hold great potential for influencing rumen fermentation and reducing methane emissions in agro‐industry applications. The use of these by‐products not only supports sustainable farming practices but also aligns with circular economy principles, helping to reduce waste and improve resource efficiency, contributing significantly to environmental sustainability.
CONFLICT OF INTEREST
The authors declare that they have no conflicts of interest.
ACKNOWLEDGEMENTS
The provisions of by‐products from suppliers and the assistance from staff in Queen's University Belfast and Agri‐Food Biosciences Institute are gratefully acknowledged. Financial support for this work has been provided by the Biotechnology and Biological Sciences Research Council (BBSRC) (grant number BB/T008776/1) as part of the Doctoral Training Partnership FoodBioSystems: biological processes across the agri‐foodsystem from pre‐farm to post‐fork.
DATA AVAILABILITY STATEMENT
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.