Partial Substitution of Cereal Grains With Sugar-Rich Castella By-Product on Nutrient Digestibility and Lactation Performance of Dairy Cows
ABSTRACT
This study investigated the effect of partially replacing corn with castella by-product, a sugar-rich feedstuff, on the lactation performance and nutrient digestibility of dairy cows. Eleven late-lactating Holstein cows were used in the experiment employing a crossover design with two periods of 21 days, including 18 days of adaptation and 3 days of data collection, under group-feeding. Castella by-product was incorporated into the diet at 7% on a dry matter (DM) basis, which replaced 30% DM of the starch-rich grains used in the diet. The results showed that the inclusion of the castella increased the digestibilities of DM, organic matter, and neutral detergent fiber (p < 0.05) despite its high sugar content, while decreasing starch digestibility (p < 0.05), without affecting fecal pH, fecal DM, or microbial protein synthesis in the rumen (p > 0.05). The castella diet did not differ significantly from the corn diet for milk yield, milk composition, milk fatty acid profile, or microbial protein synthesis in the rumen (p > 0.05). In summary, the partial replacement of corn with castella by-product in the diet yielded comparable lactation performance outcomes to those of the control diet in dairy cows.
1 Introduction
Cereal grains are the primary energy source for dairy cows, but their increasing global demand raises concerns about future food security. Thus, utilizing cost-effective alternative energy feeds is necessary. The increasing availability of by-products from the food industry can be considered a potential energy source for ruminant animals. Bakery by-products have been used as cereal grains substitutes due to the large amounts of waste generated from unsold and unused bakery products (Kaltenegger et al. 2020). Previous research reported that substituting grains with bakery by-products can enhance milk production and shift the nutrient profile from a glucogenic to a lipogenic diet (Kaltenegger et al. 2020).
Sugar serves as an energy source due to its rapid fermentation in the rumen and has the potential to enhance microbial protein synthesis (MPS) (Oba 2011). Castella by-product, leftover pieces of castella after the baking process at a food factory, has been shown to contain high sugar content and significantly increase short-chain fatty acids production in the rumen (Nayohan et al. 2024). Despite its potential benefits, the high sugar content in castella (59% dry matter [DM]; Nayohan et al. 2024) may present certain risks for rumen fermentation. Feeds with rapidly fermentable carbohydrates can cause a significant decrease in ruminal pH (Zebeli et al. 2012). This pH reduction may inhibit the activity of fibrolytic bacteria in the rumen (Nagaraja and Titgemeyer 2007), potentially leading to decreased fiber digestibility. Additionally, ruminal acidosis has been associated with reduced milk fat content in dairy cows (Zebeli et al. 2012). Therefore, it is necessary to evaluate the effects of partially replacing conventional cereal grain feeds such as steam-flaked corn with castella by-product on milk production performance, nutrient digestibility, and rumen fermentation in dairy cows. We hypothesized that the higher sugar content in castella by-product might enhance MPS, improve nutrient utilization, and maintain or increase milk production when used as a partial replacement for corn. This study aims to determine the viability of castella by-product as an alternative energy source in dairy cattle diets by analyzing its effects on lactation performance, milk composition, and nutrient digestibility.
2 Materials and Methods
Animal experiments were performed in accordance with the guidelines for Animal Care and Use of Ministry of Agriculture, Forestry, and Fisheries of Japan (18Noukai #307). All animal care and management protocols received approval from Aichi Prefecture Agricultural College (January 23, 2024).
2.1 Experiment Design
The study was conducted at Aichi Prefecture Agricultural College using 11 late-lactation dairy cows (265 ± 42.4 d in milk, lactation number 1.73 ± 0.19). The cows were housed in free stalls under group-feeding conditions and were separated into two groups, comprising 5 and 6 cows, respectively. The experiment was performed using a crossover design over two periods. Each period lasted 21 days, including an 18-day adaptation period and a 3-day data collection period. The treatments included a corn diet and a castella diet. Castella by-product, mainly produced from cutting the edge of castella after the baking process, was provided from a local food company. The castella by-product contained 8.5% CP, 1.5% aNDFom, 2.9% ether extract (EE), 59.0% sugar, 26.2% starch on a DM basis. Castella by-product was included in the total mixed ration (TMR) at 7% on a DM basis by replacing part of the corn. This substitution was equivalent to replacing approximately 30% of the starch-rich grains (corn and barley) with castella by-product. The diet ingredients were collected three times in different days: once during each of the two data collection periods (first and second period) and once during the adaptation period. Details of the diets and their nutrient composition are presented in Table 1. The diets were provided twice daily, in the morning and afternoon, and refusals were collected before each feeding time. Salt and water were offered to the cows ad libitum.
| Ingredients(% DM) | Corn diet | Castella diet | p |
|---|---|---|---|
| Sudangrass hay | 19.8 | 19.8 | |
| Timothy hay | 13.2 | 13.2 | |
| Alfalfa hay | 8.8 | 8.8 | |
| Steam-flaked corn | 13.4 | 6.4 | |
| Castella by-product | 0.0 | 7.0 | |
| Steam-flaked barley | 11.0 | 11.0 | |
| Soybean meal | 4.4 | 4.4 | |
| Wheat bran | 3.3 | 3.3 | |
| Cotton seed | 4.4 | 4.4 | |
| Compound feed | 8.8 | 8.8 | |
| Beet pulp | 8.8 | 8.8 | |
| Amino acid supplement | 1.8 | 1.8 | |
| Mineral supplement | 2.2 | 2.2 | |
| Computated quantity | 100.0 | 100.0 | |
| Nutrient composition (% DM unless stated) | |||
| Dry matter(% of fresh matter) | 40.1 ± 0.12 | 40.1 ± 0.13 | 0.958 |
| Organic matter | 94.7 ± 0.09 | 94.6 ± 0.34 | 0.654 |
| Crude protein | 14.3 ± 0.19 | 14.0 ± 0.38 | 0.469 |
| Starch | 19.2 ± 0.46 | 16.7 ± 0.32 | 0.011 |
| Sugars | 6.5 ± 0.09 | 9.9 ± 0.14 | 0.002 |
| aNDFom | 39.8 ± 0.75 | 39.7 ± 0.05 | 0.950 |
| Ether extract | 3.1 ± 0.15 | 3.1 ± 0.16 | 0.920 |
- Note: Data are shown as mean ± standard error.
- Abbreviation: aNDFom, neutral detergent fiber.
2.2 Urine and Feces Collection
Urine and feces samples were manually collected twice daily, in the morning and afternoon, for 3 consecutive days from eight cows during the data collection period. Feces were collected directly from the rectum. Urine was collected by massaging near the urinary opening to stimulate urination. The pH of the feces was measured using a pH meter (206-2, Testo SE & Co. KGaA, Germany). Urine and fecal samples were stored at −28°C in a freezer for subsequent analysis. Urinary purine derivatives (PD, allantoin, and uric acid) and creatinine concentrations were determined by the following the method described by George et al. (2006). Total urine volume was estimated from the creatinine concentration, assuming that daily creatinine excretion was constant at 21 mg/kg BW (Tamura et al. 2007). Microbial N synthesis in the rumen was estimated using 0.116 as the purine:N ratio, according to the purine derivative excretion method described by Chen and Gomes (1992). The amount of MPS in the rumen was estimated by multiplying N amount by 6.25. Apparent total-tract digestibility of dietary nutrients was calculated using acid-insoluble ash (AIA) as an internal marker in TMR and feces, following the method outlined by Van Keulen and Young (1977), using the formula digestibility (%) = 100 − [100 × (AIA in feed × nutrient content in feces)/(AIA in feces × nutrient content in feed)].
2.3 Milk Collection
A total of 11 dairy cows were milked twice daily for 3 consecutive days to measure milk yield and determine milk composition. Milk composition was measured using an FTIR milk analyzer (CombiFoss 7 DC, FOSS, Hillerød, Denmark) at the milk inspection station of Tokai Dairying Federation of Cooperatives (Okazaki, Japan). The milk components included milk fat, protein, lactose, solids-not-fat (SNF), urea nitrogen (MUN), and fatty acid (FA) profiles (de novo, mixed, and preformed FA) were determined.
2.4 Nutrient Analysis
For feeds and feces, DM, organic matter (OM), CP, and EE were measured according to the official methods of the Association of Official Analytical Chemists International (AOAC International 2000). NDF was evaluated using a heat-stable α-amylase enzyme and corrected for ash content (aNDFom) as described by Van Soest et al. (1991). Starch concentration was measured by a commercial kit (Total Starch Assay Kit, Megazyme, Ireland), whereas sugar levels were analyzed using high-performance chromatography as described by Kondo et al. (2021).
2.5 Statistical Analysis
An independent t-test was performed to compare the nutrient content between the corn and castella diets. A linear mixed-effects model (LMM) was then employed to evaluate the effects of diets on milk yield and composition, milk fatty acids, fecal conditions, digestibility, and MPS using the PROC MIXED procedure in SAS 9.4 (SAS Institute Inc.). Diets and experimental runs were treated as fixed effects, whereas animals were considered random effects. Statistical significance was determined at p < 0.05 and tendency at 0.05 < p < 0.1 for all analyses.
3 Results and Discussion
The nutrient composition of experimental diets is presented in Table 1. Castella by-product was incorporated into the diet at 7% on a DM basis. This substitution resulted in significant changes in carbohydrate fractions while maintaining similar levels of other nutrients. Although there were no significant differences (p > 0.05) between the corn and castella diets in terms of DM, OM, CP, aNDFom, and EE, the starch content was significantly higher in the corn diet compared to the castella diet (p < 0.05). Conversely, the sugar content was significantly higher in the castella diet compared to the corn diet (p < 0.01). These differences were expected, as castella contains high sugar levels, which elevated the sugar proportion in the dietary formulation when incorporated as a partial replacement for corn.
The effects of the dietary treatments on lactation performance are shown in Table 2. Milk yield did not differ significantly between treatments (p = 0.26). Similarly, milk components including fat, protein, lactose, and MUN showed no significant differences (p > 0.05) between the corn and castella diets. In this study, a tendency for difference in SNF (p = 0.07) was observed between treatments; however, no significant differences were detected in its major components, milk protein and lactose. The physiological significance of this slight statistical tendency appears limited. The milk fatty acid profile, including de novo, mixed, and preformed FA (expressed both as g/100 g milk and g/100 g FA), also showed no significant differences between treatments (p > 0.05). Average dry matter intake (DMI) per cow was similar between treatments, with 23.0 kg/day for the corn diet and 23.8 kg/day for the castella diet. This similar DMI between treatments suggests good palatability and acceptability of castella as a feed ingredient for dairy cows, which is important when considering alternative feed ingredients. Due to the group-feeding arrangement of the experiment, individual feed intake data could not be determined.
| Parameter | Treatment | p | |
|---|---|---|---|
| Corn diet | Castella diet | ||
| Milk yield (kg/day) | 26.30 ± 2.04 | 25.68 ± 2.04 | 0.263 |
| Milk components (g/100 g) | |||
| Fat | 4.51 ± 0.17 | 4.62 ± 0.18 | 0.193 |
| Protein | 3.86 ± 0.10 | 3.85 ± 0.10 | 0.357 |
| Lactose | 4.44 ± 0.04 | 4.43 ± 0.04 | 0.280 |
| SNF | 9.31 ± 0.10 | 9.27 ± 0.09 | 0.072 |
| MUN (mg/dL) | 8.60 ± 0.57 | 8.31 ± 0.49 | 0.170 |
| De novo FA | |||
| g/100 g milk | 1.21 ± 0.05 | 1.22 ± 0.04 | 0.507 |
| g/100 g FA | 28.1 ± 0.30 | 27.9 ± 0.38 | 0.476 |
| Mixed FA | |||
| g/100 g milk | 1.59 ± 0.07 | 1.61 ± 0.07 | 0.476 |
| g/100 g FA | 37.1 ± 0.40 | 36.7 ± 0.39 | 0.126 |
| Preformed FA | |||
| g/100 g milk | 1.47 ± 0.05 | 1.54 ± 0.07 | 0.126 |
| g/100 g FA | 34.5 ± 0.60 | 35.2 ± 0.66 | 0.162 |
- Note: Data are shown as mean ± standard error.
- Abbreviations: FA, fatty acid; MUN, milk urea nitrogen; SNF, solid non-fat.
These results indicate that despite the differences in carbohydrate composition, the castella diet maintained comparable milk production and composition to the control corn diet. These findings align with previous research by Kaltenegger et al. (2020), which demonstrated that including bakery by-products containing sugar in dairy cow diets can enhance milk production and reduce MUN levels. However, in our study, the higher sugar content in the castella diet did not result in improved milk production or altered milk composition.
Table 3 presents data on fecal conditions, apparent total-tract digestibility, and microbial protein synthesis in the rumen. Fecal pH and fecal DM did not differ significantly between treatments (p > 0.05). Feeds with rapidly fermentable carbohydrates can cause a significant decrease in ruminal pH (Zebeli et al. 2012), which may inhibit fibrolytic bacteria activity (Nagaraja and Titgemeyer 2007) and potentially lead to decreased fiber digestibility. Additionally, ruminal acidosis has been associated with reduced milk fat content (Zebeli et al. 2012).
| Parameter | Treatment | p | |
|---|---|---|---|
| Corn diet | Castella diet | ||
| Fecal condition | |||
| Fecal pH | 6.28 ± 0.03 | 6.29 ± 0.04 | 0.691 |
| Fecal DM (% DM) | 13.54 ± 0.29 | 13.75 ± 0.18 | 0.410 |
| Digestibility (%) | |||
| DM | 62.12 ± 0.90 | 65.47 ± 0.85 | 0.043 |
| OM | 64.25 ± 0.87 | 67.38 ± 0.83 | 0.049 |
| CP | 58.15 ± 1.00 | 61.87 ± 1.28 | 0.074 |
| Starch | 98.34 ± 0.18 | 98.06 ± 0.22 | 0.011 |
| EE | 70.47 ± 1.56 | 72.92 ± 0.96 | 0.219 |
| aNDFom | 45.30 ± 1.72 | 50.32 ± 1.55 | 0.045 |
| MPS (gCP/day) | 2793 ± 149 | 2922 ± 149 | 0.259 |
- Note: Data are shown as mean ± standard error.
- Abbreviations: CP, crude protein; DM, dry matter; EE, ether extract; MPS, microbial protein synthesis in the rumen; aNDFom, neutral detergent fiber; OM, organic matter.
The similar fecal pH values between treatments suggest that the increased sugar content in the castella diet might not have significantly affected rumen pH, though this can only be inferred indirectly from our measurements. Recent research by Khorrami et al. (2022) has demonstrated the relationship between ruminal pH and fecal pH, showing significant positive correlations between fecal pH and both minimum ruminal pH (r = 0.64) and daily mean ruminal pH (r = 0.54). Their study established that changes in dietary starch content significantly affected fecal pH patterns, with high-starch diets resulting in lower fecal pH and a distinct daily pattern compared to low-starch diets. In our study, the maintenance of normal fecal pH levels may suggest that rumen fermentation likely remained stable. This is further supported by the maintenance of milk fat content, as reduced ruminal pH would typically manifest as depressed milk fat.
The nutrient digestibility was significantly affected by dietary treatments. The castella diet resulted in higher digestibility of DM, OM, and aNDFom compared to the corn diet (p < 0.05). This enhanced digestibility can be attributed to the rapid fermentation of sugars in the castella diet, which may have promoted overall nutrient utilization. Additionally, there was a tendency for higher CP digestibility in the castella diet (p = 0.07). Conversely, starch digestibility appears to be slightly higher in the corn diet compared to the castella diet (98.3% vs. 98.1%), though the biological significance of this small difference should be interpreted with caution despite statistical significance (p < 0.05). The improved aNDFom digestibility in the castella diet is particularly noteworthy, especially considering the concerns about potential negative effects of high sugar content on fiber digestion. This improvement could be explained by several factors. First, despite the higher sugar content, the castella diet might have maintained ruminal pH within a range conducive to fibrolytic bacterial activity, as suggested by the similar fecal pH values. Second, the inclusion of sugar sources in the ruminant diet, increasing the dietary sugar content to approximately 5%–8%, was associated with improved fiber digestibility, likely through enhanced ruminal fermentation (Broderick and Radloff 2004). These findings suggest that sugars may promote the growth of fiber-digesting bacteria and thereby increase fiber digestibility.
Despite the differences in nutrient digestibility between treatments, MPS did not differ significantly between the corn and castella diets (p = 0.26). Contrary to our initial hypothesis that higher sugar content in castella by-product might enhance microbial protein synthesis, no such enhancement was observed. This unexpected result warrants examination, especially considering the higher digestibility of OM and tendency for higher CP digestibility in the castella diet, which theoretically could have provided more energy and nitrogen to support microbial growth. The discrepancy between improved digestibility and unchanged MPS can be attributed to the nature of sugar metabolism in the rumen. Sugar provides fewer carbon units than starch due to a lower monomer yield during hydrolysis (Hall and Herejk 2001), likely reducing its efficiency for MPS despite improved digestibility. It is also possible that the inclusion level of castella (7% DM) was insufficient to significantly alter ruminal microbial dynamics to the extent needed for enhanced MPS, despite its positive effects on overall nutrient digestibility.
The increase in bakery by-product in the diet can modify fermentation patterns in lactating cows (Kaltenegger et al. 2020). Sucrose fermentation typically yields both propionate and butyrate in the rumen (Oba 2011), whereas castella has been shown to promote propionate production in the in vitro rumen (Nayohan et al. 2024). These altered fermentation patterns would theoretically support changes in milk production parameters, either through glucogenic (propionate) or lipogenic (butyrate) pathways. However, our results indicate that despite the improved digestibility of key nutrients, the castella diet did not significantly affect milk yield, composition, or FA profile compared to the corn diet.
In summary, incorporating castella by-product at 7% of diet DM improved nutrient digestibility, particularly aNDFom digestibility, without negative effect on milk production despite its high sugar content. The castella diet maintained stable fecal pH and enhanced fiber digestion, contrary to typical concerns with high-sugar feeds. Although microbial protein synthesis was not enhanced, milk production and composition remained comparable to the corn diet. As this study used late-lactation cows, future research should examine early-lactation cows with higher intake and metabolic demands, along with higher inclusion levels and long-term effects on microbial populations. Our findings suggest that castella by-product can serve as an effective partial replacement for cereal grains in dairy cattle diets while potentially improving nutrient utilization.
Acknowledgments
We extend our sincere appreciation to IMURAYA STARTUP PLANNING CORPORATION (Mie, Japan) for supplying the castella by-products and to KANKYO TECHSYS Co. Ltd. (Aichi, Japan) for their technical support in material preparation.
Conflicts of Interest
The authors declare no conflicts of interest.
