Fermentation-Based Preservation of Okara and In Vitro Evaluation of Its Application in Dairy Cattle Diets

Abstract

Okara, a protein-rich byproduct of soymilk production, is highly perishable because of its high moisture content. This study evaluated the preservation and nutritional value of okara fermented by lactic acid bacteria for use in dairy cattle diets. Fermentation effectively reduced pH within 2 weeks and maintained quality for up to 6 weeks. However, aerobic exposure increased the concentration of ammonia, indicating a decline in stability. In vitro assessments revealed no significant differences in in vitro true dry matter digestibility, in vitro neutral detergent fiber digestibility, or gas production between fermented and fresh okara, although fermented okara had a higher concentration of ammonia nitrogen. In situ analysis revealed slightly lower dry matter effective degradability (ED) in fermented okara, but similar rumen-degradable and undegradable protein fractions. When fermented okara was used to replace soybean meal in total mixed rations, 25–50% inclusion-maintained digestibility and fermentation characteristics, with 25% replacement yielding the highest ED at a low ruminal passage rate (0.02 h⁻¹). Taken together, these results suggest that fermented okara can be strategically incorporated into dairy rations as a sustainable protein alternative, supporting both rumen function and bypass protein supply.

1. Introduction

Okara, a major byproduct of tofu and soymilk production, is generated in large quantities worldwide, particularly in Asia, where soybean processing is widespread. Globally, approximately 14 million metric tons of okara are produced annually, with China and Japan contributing approximately 2.8 and 0.8 million metric tons, respectively [1]. In terms of dry matter (DM), okara contains a relatively high crude protein (CP) content, ranging from 15.2% to 33.4% [2], indicating its major potential as a protein source for animal feed. However, the extremely high moisture content of okara, which typically ranges from 70% to 80% [3], poses major challenges for storage and transportation, often resulting in rapid microbial spoilage and nutrient degradation [2]. Rahman et al. [1] recommended using fresh okara within 3 days in cases of direct feeding to livestock. Notably, the presence of antinutritional compounds, such as trypsin inhibitors, saponins, phytic acid, hemagglutinins, stachyose, and raffinose, limits the direct inclusion of okara in animal diets [4,5]. Thus, manufacturers often dispose of okara through composting or incineration, which leads to the loss of a valuable protein-rich resource.

Global events such as regional conflicts have increased the prices of imported feed ingredients, including soybean meal. This economic pressure has increased interest in sustainable, locally sourced, and cost-effective feed alternatives [6]. Microbial fermentation offers a promising approach for enhancing the stability and nutritional quality of okara. Fermentation not only extends the shelf life of okara but also mitigates the effects of antinutritional compounds [2]. Lactic acid bacteria (LAB) fermentation is recognized as one of the most effective techniques for preservation. This technique lowers pH through lactic acid production, which inhibits the growth of spoilage microorganisms [7,8]. Additionally, LAB fermentation may enhance the nutritional value of feed by breaking down fibers and proteins, thereby increasing the concentration of free amino acids, and converting macronutrients into smaller, more digestible molecules [3,8,9,10].

To evaluate the nutritional value of okara in ruminant diets, the digestibility of okara should be examined. Although in vivo trials provide direct evidence, they are often expensive, time-consuming, and difficult to conduct for single ingredients [11]. In vitro and in situ techniques offer valuable alternatives for evaluating rumen digestibility with minimal animal use [12,13]. In vitro methods can be used to evaluate DM and neutral detergent fiber (NDF) digestibility, fermentation characteristics, and gas production patterns [14,15]. In situ methods can be used to estimate dynamic DM and CP degradation, including effective degradability (ED) [16,17]. The kinetics of feedstuff degradation can be evaluated through fermentative gas production. Gas measurements also provide valuable information on the digestion kinetics of both the soluble and insoluble fractions of feedstuffs [18]. A well-constructed model may link gas production data to animal performance by estimating the extent of ruminal degradation [19]. Combining the results of in vitro digestibility and fermentation products, such as volatile fatty acids (VFAs) and ammonia, with gas kinetics may offer a practical approach for examining how ruminant diets are utilized [20].

This introduces a two-phase approach addressing the challenges of okara preservation and protein source diversification in ruminant nutrition. In summary, this study aimed to develop an optimal fermentation process for okara preservation by evaluating its effects on chemical composition, fermentation quality, and aerobic stability. Concurrently, it sought to determine the feeding value of the optimally fermented okara, including its degradability and potential as a soybean meal replacement in dairy cattle total mixed rations (TMRs), through comprehensive in vitro and in situ digestibility trials.

2. Materials and Methods

2.1. Okara Source

Fresh okara was obtained from Kuang Chuan Ranch in Taoyuan, Taiwan. It was immediately collected after soymilk processing, which involved soaking, grinding, and heating soybeans, followed by filtration. Upon sample arrival, the internal temperature of the okara samples ranged from 60 to 70 °C, with an average moisture content of approximately 80%. All samples were analyzed to determine their chemical composition, including DM, CP, ether extract (EE), NDF, acid detergent fiber (ADF), acid detergent lignin (ADL), ash, and water-soluble carbohydrates (WSCs).

2.2. Experimental Design

This study was conducted in two phases: (1) laboratory-scale fermentation of fresh okara and (2) large-scale evaluation of fermented okara as a replacement for soybean meal in dairy cattle diets through in vitro and in situ trials. The first phase focused on evaluating a small-scale fermentation process to examine its effects on the chemical composition, fermentation quality, and aerobic stability of okara. The second phase involved determining the feeding value of fermented okara through in vitro and in situ trials, particularly focusing on its degradability and potential to replace soybean meal in dairy cattle total mixed rations (TMRs). These two complementary phases were designed to provide a comprehensive understanding of the storage stability and practical use of fermented okara in ruminant feeding.

2.2.1. Phase 1: Laboratory-Scale Fermentation Trial

In the first phase, a fermentation strategy was developed and evaluated to improve okara preservation by determining its effects on chemical composition, fermentation quality, and aerobic stability. Each replicate sample in a treatment group consisted of 200 g of fresh okara.

Fermentation conditions were tested using nine different combinations: moisture content at 60%, 70% or 80%, and inoculum concentration at 10⁶, 10⁷, or 10⁸ CFU/g. Changes in pH and ammonia concentration were monitored on days 3, 6, 12, and 24 after inoculation. These parameters were analyzed to identify the optimal conditions for future fermentation applications.

A commercial fermenting starter (Synferm Feed, Shenghe Biotech, Kaohsiung City, Taiwan) containing Lactobacillus plantarum, Lactobacillus fermentum, Lactobacillus paracasei, Lactobacillus rhamnosus, and Saccharomyces cerevisiae was used. The concentrations of LAB and yeast were 2.2 × 10¹⁰ and 5.2 × 10⁷ CFU/g, respectively. The inoculum was diluted in sterile water and applied at 0.2% (v/w) okara. The mixture was then vacuum-sealed in plastic bags and fermented at room temperature. A control group without an inoculum or moisture adjustment was also included. After screening, fermentation was conducted at a moisture content (unadjusted) of 80% and an inoculation level of 10⁶ CFU/g for 24 days for evaluation. Finally, the fermented samples were analyzed to determine their chemical composition, fermentation quality, and aerobic stability after 48 h of air exposure.

Fermentation quality was determined by opening the fermentation bags at 1, 2, 4, 6, and 8 weeks. Samples from the top, middle, and bottom layers were mixed and analyzed for pH, microbial count, VFAs, ammonia nitrogen (NH₃-N), and chemical composition. Aerobic stability was evaluated by exposing 4 week fermented samples to the air for 48 h, with microbial counting conducted every 24 h [8].

2.2.2. Phase 2: Feed Utilization Trial

In this phase, the nutrimental potential of fermented okara was evaluated by examining its digestibility, fermentation characteristics, and gas production patterns both in vitro and in situ. On the basis of the fermentation results, fermented okara was incorporated into TMRs at varying replacement levels for soybean meal, and the digestibility of each formulation was evaluated. These replacement levels were set at 0%, 25%, 50%, 75%, and 100%, resulting in five experimental groups. All diets were formulated in accordance with National Research Council [21] guidelines to maintain comparable levels of CP, rumen degradable protein (RDP), rumen undegradable protein (RUP), and net energy for lactation. These formulations were designed for a multiparous lactating dairy cow weighing 680 kg and producing 35 kg/day of milk containing 3.0% true milk protein and 3.5% milk fat. Differences in mineral and vitamin content between the experimental diets were adjusted using a premixture to meet standard dairy cattle requirements. Ingredient proportions and nutrient compositions are presented in

Table 1. TMR formulas and nutrient compositions used in the okara replacement experiment.

Ingredient %DM 2		Diets (Soybean Meal Replacement %) 1
		Fermented Okara
	Control	25	50	75	100
Alfalfa hay	25.4	25.4	25.4	25.4	25.4
Oat hay	25.4	25.4	25.4	25.4	25.4
Soybean meal	6.4	4.8	3.2	1.6	0.0
Fermented okara	0.0	1.6	3.2	4.8	6.4
Corn meal	23.7	19.7	15.8	11.5	7.6
Wheat bran	10.6	10.6	11.4	13.1	13.6
Corn gluten meal	4.7	3.4	2.5	1.7	0.6
DDGS 2	2.1	7.4	11.2	14.7	19.3
NaHCO3	0.8	0.8	0.8	0.8	0.8
Salt	0.4	0.4	0.4	0.4	0.4
Vitamin premix	0.5	0.5	0.5	0.5	0.5
Nutrient Composition (%DM) 3
CP	16.4	16.4	16.4	16.4	16.4
RDP	10.4	10.4	10.4	10.4	10.4
RUP	6.0	6.0	6.0	6.0	6.0
NEl	1.5	1.5	1.5	1.5	1.5
NDF	36.8	38.8	40.5	42.5	44.4
ADF	22.6	23.4	24.1	24.9	25.7
Ca	0.7	0.7	0.7	0.7	0.7
P	0.5	0.5	0.5	0.5	0.5

¹ Nutrient requirements recommended by the National Research Council [21] for dairy cows: CP = 16.4%, RDP = 10.4%, RUP = 6%, NE_l = 1.52 Mcal/kg, Ca = 0.61%, and P = 0.35%. ² DM = dry matter; DDGS = distilled dried grains with soluble components. ³ CP = crude protein; RDP = rumen degradable protein; RUP = rumen undegradable protein; NDF = neutral detergent fiber; ADF = acid detergent fiber.

. Each formulation was evaluated using in vitro and in situ methods.

2.3. Chemical Composition Analysis of Fresh Okara and Fermented Products

2.3.1. Sample Preparation

10 g of fresh or fermented okara was homogenized with 90 mL of chilled sterile water (4 °C) by using a high-speed blender (Oster, Ontario, CA, USA) for 20 s and then filtered through four layers of cheesecloth. The filtrate was used for pH measurement and microbial counting. The remaining liquid was centrifuged twice at 12,000× g at 4 °C for 15 min to remove impurities. The supernatants were then analyzed for microbial CP (MCP), VFA, and NH₃-N.

Approximately 100 g of fresh and fermented okara was dried at 65 °C for 2 days, ground using a mill (DM-6, 25,000 rpm; XiangTai, Taishan District, New Taipei City), and sieved through a 20 mesh screen for use in a chemical composition analysis and in vitro and in situ digestion trials.

2.3.2. pH, Microbial Count, Microbial Crude Protein, Volatile Fatty Acids, and NH₃-N

After fermentation, the pH of the fluid was measured using a pH meter (F-71; Horiba Scientific, Kyoto, Japan). The serum bottle was then recapped and placed in an ultrasonicator containing cold water (approximately 4 °C) for 5 min and sonicated under cold water for 15 min to dislodge the microbes attached to the residual substrates. After the sample filtrate was diluted with phosphate-buffered saline, LAB and mold or yeast counting was conducted using a 3M Petrifilm Lactic Acid Bacteria Count Plate and a YM-Media Pad, respectively. The sonicated fermented fluid was then centrifuged at 400× g at 4 °C for 5 min. At this point, half of each supernatant fraction was collected and lyophilized for MCP analysis by using the purine content method described by Zinn and Owens [22]. MCP values from the blank fermentation group (containing rumen inoculum but no substrate) were subtracted during MCP synthesis calculations. The remaining half of the supernatant fraction was centrifuged at 13,500× g at 4 °C for 15 min to analyze VFAs, lactic acid, and NH₃-N. The concentration of NH3-N was immediately determined using a colorimetric method [23]. The samples used for VFA and lactic acid analysis were acidified with 25% metaphosphoric acid (4:1, w/v) and filtered through a 0.22 µm membrane filter (Millipore, Bedford, MA, USA). All VFA samples were stored at −20 °C for analysis. The VFA concentration was determined using high-performance liquid chromatography (LC-4000; Jasco, Tokyo, Japan) with a Rezex ROA-Organic Acid H⁺ (8%) column (300 mm × 7.8 mm; Phenomenex, Torrance, CA, USA) and a UV detector as an integrated part of the LC-4000 system [20].

2.3.3. Chemical Composition Analysis

The organic matter (AOAC 942.05), DM (AOAC 934.01), EE (AOAC 920.39), CP (AOAC 990.03), ADL (AOAC 973.18), and ash (AOAC 942.05) contents of fresh and fermented okara were determined using standard Association of Official Analytical Chemists (AOAC) [24] methods. NDF and ADF were analyzed in accordance with the procedure described by Van Soest et al. [25]. WSCs were determined using the methods reported by Dreywood [26] and Morris [27]. MCP synthesis efficiency is expressed as MCP (milligrams) per gram of digested organic matter.

2.4. In Vitro Digestibility and Gas Kinetics Assay

Rumen fluid was collected from two rumen-fistulated dry Holstein cows (approximately 800 kg body weight) housed in a livestock farm at the College of Agriculture, National Taiwan University. These cows were fed a TMR with a forage-to-concentrate ratio of 60:40, consisting of 180 g/kg Bermuda hay, 180 g/kg alfalfa hay, and 240 g/kg commercial concentrate (16% CP, DM basis). Each cow received 16 kg of food per day and had ad libitum access to fresh water. The two cows were housed in individual pens and fed the same diet for 14 days before rumen fluid collection. On the sampling day, rumen fluid was collected through a cannula from each cow 2 h after the morning feeding session. This fluid was then strained through four layers of cheesecloth into a prewarmed serum bottle (40 °C) flushed with carbon dioxide before use. Freeze-dried okara samples were used in the in vitro assay. A portion of the fermented okara was frozen at −80 °C for 24 h and then freeze-dried at −50 °C for 96 h. Freeze-dried samples were used for in vitro tests to minimize the impact of high-temperature drying on digestibility.

In vitro degradability was determined using an Ankom Daisy II Incubator (Ankom Technology, Macedon, NY, USA) over a 48 h incubation period. Digestion and sample collection were conducted as described by Spanghero et al. [28]. Briefly, F57 filter bags (25 μm pore size; Ankom Technology) were prewashed with acetone, soaked for 10 min, and dried overnight at 65 °C. After cooling, the bags were weighed to determine their initial weights. A total of 0.5 g of freeze-dried sample was placed in each filter bag, and all bags were sealed twice using a heat sealer (four replicates per treatment). Blank bags without samples were prepared in the same manner and served as controls. Weight loss was determined and used to calculate in vitro true DM digestibility (IVTDMD). Residual NDF was used to determine in vitro NDF digestibility (IVNDFD).

To conduct a gas production assay, 400 mg of each experimental diet was weighed and added to 100 mL serum bottles, with four replicates per treatment. Rumen fluid from each cow was then combined with artificial saliva solution, as described by Menke and Steingass [18], at a ratio of 1:4 (v/v). This mixture was maintained at 39 °C in the presence of continuous flushing with carbon dioxide. Subsequently, 80 mL of the rumen inoculum mixture was added to each serum bottle in the presence of carbon dioxide flushing. After inoculation, the bottles were connected to an Ankom pressure sensor module, and gas production was monitored every 2 min by using a wireless Ankom RF gas production system (Ankom Technology) [29].

2.5. In Situ Degradability Assay

2 g of freeze-dried sample was sealed in a 5 cm × 10 cm nylon bag (50 μm pore size) and incubated in the rumen of cannulated cows, with three replicates per treatment. The cows were fed the same diet described in Section 2.4. The digestion bags were removed at 2, 4, 8, 12, 24, 48, and 72 h; washed; dried at 60 °C for 48 h; and analyzed for DM, CP, and NDF. Degradation parameters (a, b, c, and t₀) were modeled following the method described by McDonald [30]. RDP, RUP, and ED were calculated using the models proposed by Cone et al. [15] and Ørskov and McDonald [16].

2.6. Statistical Analysis

The chemical composition of fresh and fermented okara, the fermentation characteristics of fermented okara, and the digestibility and fermentation parameters of the in vitro trials were analyzed using analysis of variance (ANOVA) in SAS software version 9.4 (SAS Institute, Cary, NC, USA). Significant differences between treatment means were evaluated using Tukey’s honestly significant difference test. Gas production kinetics and in situ degradability data were modeled using the one-phase nonlinear association model in GraphPad Prism 8 version 2019 (GraphPad Software, San Diego, CA, USA). A nonlinear regression analysis was conducted using the NLIN procedure in SAS software, and gas production parameters were estimated using the Marquardt iterative method. The following model, based on McDonald [30], was used to calculate degradability values:

P1 = a (t ≤ t₀)

P2 = a + b (1 − e^−ct) (t > t₀)

where P is degradation at time t a is the soluble fraction, b is the potentially degradable fraction, c is the degradation rate of the b fraction, and t₀ is the lag time before degradation begins. A p value of <0.05 was considered statistically significant, and values between >0.05 and ≤0.10 were interpreted as indicative of a statistical trend.

3. Results

3.1. Laboratory-Scale Fermentation

3.1.1. Fermentation Condition Selection

According to the results of the moisture content and inoculum concentration combination tests (

Table 2. Temporal changes in pH and ammonia nitrogen (NH₃-N) content during fermentation at different inoculation levels and moisture conditions on days 3, 6, 12, and 24.

Item	Days	Control	Inoculation Level (CFU/g)									SEM	p-Value
			106			107			108
			Mositure %			Mositure %			Mositure %
			80	70	60	80	70	60	80	70	60
pH	3	4.61 a	4.32 b	4.26 c	4.54 b	4.25 c	4.37 b	4.61 a	4.39 b	4.45 a	4.54 a	0.03	<0.01
	6	4.47 a	4.28 c	4.27 c	4.38 b	4.23 c	4.34 c	4.60 a	4.36 b	4.43 b	4.57 a	0.02	<0.01
	12	4.36 a	4.15 c	4.26 b	4.30 b	4.28 b	4.30 c	4.42 a	4.32 b	4.33 a	4.41 a	0.02	<0.01
	24	4.37 b	4.23 c	4.38 b	4.34 b	4.39 b	4.41 b	4.52 a	4.47 a	4.50 a	4.55 a	0.01	<0.01
NH3-N (g/kg DM)	3	7.90 a	4.61 b	2.38 d	1.15 e	3.79 c	2.77 d	2.03 d	3.41 c	2.66 d	2.45 d	0.09	<0.01
	6	6.34 a	4.68 b	1.56 d	0.69 e	3.41 c	2.06 d	1.38 de	3.03 c	2.40 d	1.61 d	0.11	<0.01
	12	2.54 a	1.58 b	1.21 b	0.85 c	1.70 b	1.37 b	0.96 c	1.73 b	1.31 b	1.06 bc	0.07	<0.01
	24	2.23 a	1.41 b	1.32 b	0.93 b	1.65 b	1.27 b	1.04 b	2.40 b	1.53 b	0.78 b	0.08	<0.01

Different superscript letters indicate significant differences between treatments. CFU = colony-forming unit; DM = dry matter; SEM = standard error of the mean.

), fermentation with a moisture content of 80% and an inoculum level of 10⁶ CFU/g resulted in the lowest pH value (approximately 4.1) on day 12 after inoculation and maintained a relatively low ammonia concentration through day 24. These findings suggest that fresh okara can be directly fermented using a low inoculum concentration without requiring additional moisture adjustment. Based on these screening results (Table 2), the optimal condition, characterized by 80% moisture content and an inoculum concentration of 10⁶ CFU/g, achieved the lowest pH on day 12 and maintained low ammonia through day 24, which effectively inhibited spoilage microorganisms and promoted a stable microbial population. Consequently, this condition was chosen for comprehensive evaluation of chemical composition, fermentation quality, and aerobic stability.

3.1.2. Chemical Composition of Fresh and Fermented Okara

Fresh okara contained 21.70% DM, 22.40% CP, and 7.72% EE on a DM basis (

Table 3. Chemical composition of fresh okara.

Item	%DM Basis
Dry matter (%)	21.70 ± 0.09
Crude protein (%DM)	22.40 ± 0.13
Ether extract (%DM)	7.72 ± 0.07
Neutral detergent fiber (%DM)	34.29 ± 0.75
Acid detergent fiber (%DM)	21.02 ± 0.31
Acid detergent lignin (%DM)	13.52 ± 1.79
Ash (%DM)	3.42 ± 0.46
Water soluble carbohydrate extract (%DM)	2.07 ± 0.03

Data are expressed as mean ± standard deviation. DM = dry matter.

). Its fiber content included 34.29% NDF, 21.02% ADF, and 13.52% ADL. After 8 weeks of fermentation, DM ranged from 18.06% to 19.06% and CP ranged from 17.45% to 21.04% (

Table 4. Chemical composition of fermented okara at constant moisture (80%) and inoculation concentration (10⁶ CFU/g).

	Week After Inoculation
Item	0	1	2	4	6	8	SEM	p-value
Chemical composition (%DM)
DM (%)	18.72	19.05	18.80	19.06	18.85	18.06	0.22	0.89
Crude protein (%DM)	21.04	20.41	20.28	20.27	20.06	17.45	0.52	0.33
Neutral detergent fiber (%DM)	37.31	38.76	37.64	37.89	39.84	34.31	0.87	0.20
Acid detergent fiber (%DM)	23.64	25.92	25.08	25.42	26.45	25.28	0.53	0.47
Ash (%DM)	2.61	2.33	2.39	2.34	2.24	0.23	0.04	0.87

DM = dry matter.

). Fiber fractions fluctuated during fermentation, with NDF ranging from 34.31% to 39.84% and ADF ranging from 23.64% to 26.45%. Ash content ranged from 2.24% to 2.61%. All p values were >0.05, indicating that the chemical composition of okara remained stable throughout the fermentation process.

3.1.3. Quality of Fermented Okara

Table 5. Fermentation quality of fermented okara at constant moisture (80%) and inoculation concentration (10⁶ CFU/g).

	Week After Inoculation
Item	1	2	4	6	8	SEM	p-value
pH	5.29 a	4.65 b	4.51 b	5.18 a	5.58 a	0.15	<0.01
NH3-N 1 (g/kg DM)	6.09 b	13.20 b	15.95 a	42.81 a	51.38 a	5.59	<0.01
Lactate (g/kg DM)	27.54 b	44.89 a	44.74 a	6.96 c	4.21 c	5.58	0.01
Acetate (g/kg DM)	7.47	6.23	7.58	17.90	44.76	2.16	0.08
Lactic acid bacteria (log CFU/g FM 2)	9.39 bc	9.98 ab	9.88 a	10.16 a	8.73 c	0.19	0.02
Yeast (log CFU/g FM)	5.24 a	5.42 a	-	3.78 b	-	0.22	0.01

¹ NH₃-N = ammonia nitrogen. ² FM = fresh matter. Different superscript letters indicate significant differences between weekly treatments. - Indicates not analyzed. CFU = colony-forming unit; DM = dry matter; SEM = standard error of the mean.

presents the fermentation quality of fermented okara. Fermentation had a significant effect on pH, NH₃-N, and lactate concentration (p < 0.01). For example, pH dropped from 5.29 in week 1 to a minimum of 4.51 by week 4 and then increased to 5.58 by week 8. In week 4, the concentration of NH₃-N significantly increased from 6.09 to 51.38 g/kg DM. In week 2, the concentration of lactate peaked at 44.89 g/kg DM and remained high through week 4 (44.74 g/kg DM) before substantially decreasing in subsequent weeks. LAB counts peaked in week 6 (10.16 log CFU/g), whereas yeast populations significantly declined after week 4.

3.1.4. Aerobic Stability of Fermented Okara

After 48 h of aerobic exposure, major changes were observed in the chemical and microbial composition of fermented okara, particularly in NH₃-N and acetate concentrations and microbial proliferation (

Table 6. Changes in the chemical composition and microbial count of fermented okara after 48 h of aerobic exposure.

	Time After Open (h)
Item	12	24	36	48	SEM	p-value
Temperature above RT (°C)	1.50	0.60	1.10	1.20	0.31	0.32
pH	4.91	4.56	4.64	4.86	0.12	0.16
NH3-N 1 (g/kg DM)	12.21 b	15.66 a	14.79 a	16.27 a	0.63	0.01
Lactate (g/kg DM)	32.92	27.20	23.24	22.88	6.28	0.23
Acetate (g/kg DM)	8.81 b	12.65 a	14.67 a	14.53 a	0.80	<0.01
Lactic acid bacteria (log CFU/g FM 2)	-	9.24	-	9.27	0.05	0.72
Yeast (log CFU/g FM)	-	5.02 b	-	6.79 a	0.12	0.01
Aerobic bacteria (log CFU/g FM)	-	9.20 b	-	10.05 a	0.29	0.03

¹ NH₃-N = ammonia nitrogen. ² FM = fresh matter. Different superscript letters indicate significant differences between weekly treatments. - Indicates not analyzed. CFU = colony-forming unit; DM = dry matter; SEM = standard error of the mean.

). The temperature of fermented okara remained stable and did not significantly differ over time (p = 0.32), indicating minimal heat generation during the aerobic exposure period. The pH values ranged from 4.56 to 4.91 and did not significantly differ (p = 0.16).

The concentration of NH₃-N significantly increased over time (p = 0.01), rising from 12.21 g/kg DM at 12 h to 16.27 g/kg DM at 48 h. The concentrations observed at 24, 36, and 48 h were significantly higher than that observed at 12 h. The concentration of acetate also significantly increased from 8.81 g/kg DM at 12 h to approximately 14.5 g/kg DM at 48 h (p < 0.01). By contrast, the concentration of lactate remained stable over time (p = 0.23).

Microbial counts demonstrated signs of aerobic deterioration. Yeast populations significantly increased from 5.02 to 6.79 log CFU/g FM (p = 0.01), and aerobic bacterial counts increased from 9.20 to 10.05 log CFU/g FM (p = 0.03), indicating active microbial growth under aerobic conditions. By contrast, LAB counts remained relatively stable between 24 and 48 h (9.24 vs. 9.27 log CFU/g FM, p = 0.72), suggesting the persistence of LAB during aerobic exposure.

3.1.5. In Vitro Fermentation and Digestibility of Fermented Okara

After 48 h of in vitro fermentation, differences in fermentation end-products and gas kinetics were observed among the three treatment groups: control, fresh okara (lyophilized), and fermented okara (

Table 7. Fermentation products, gas kinetic parameters, and in vitro digestibility of fresh and fermented okara after 48 h of in vitro fermentation.

Item 1	Control	Fresh	FO	SEM	p-Value
	Fermentation Parameters
Gas production (mL/g DM)	183.6	195.1	219.2	11.8	0.13
pH	5.65	5.65	5.62	0.01	0.26
NH3-N 2 (mg/kg)	4.26 b	4.06 b	6.71 a	0.37	<0.01
Acetate (mM)	52.11	48.62	62.25	6.32	0.57
Propionate (mM)	24.54	24.04	27.18	2.45	0.37
Butyrate (mM)	1.21	1.10	1.78	0.15	0.84
MCP 3 (mg/g DM)	28.97	27.86	21.07	4.46	0.91
	Gas Kinetics Parameter
Soluble fraction a (%)	1.61	0.92	1.75	0.62	>0.05
Degradable fraction b (%)	98.40	99.08	98.25	0.62	>0.05
Degradable rate of b (h−1)	0.10	0.10	0.12	0.01	>0.05
	In Vitro Digestibility
IVTDMD 4 (%)	85.76	88.42	86.83	2.44	0.70
IVNDFD 5 (%)	58.62	65.09	57.26	8.44	0.84

¹ Control = ensiled okara without incubation; Fresh = lyophilized fresh okara; FO = lyophilized fermented okara (80% moisture, 10⁶ CFU/g inoculation concentration, 4 week fermentation). ² NH₃-N = ammonia nitrogen. ³ MCP = microbial crude protein; DM = dry matter; SEM = standard error of the mean. ⁴ IVTDMD = in vitro true dry matter digestibility. ⁵ IVNDFD = in vitro neutral detergent fiber digestibility. Different superscript letters indicate significant differences between groups.

). No significant between-group difference was observed in total gas production (p = 0.13). The pH values remained constant across the three groups, averaging approximately 5.6, indicating stable fermentation conditions (p = 0.26). The fermented okara group had a significantly higher NH₃-N concentration (6.71 mg/kg) than did the control (4.26 mg/kg) and fresh okara (4.06 mg/kg) groups (p < 0.01), suggesting increased protein deamination during fermentation. No significant differences were observed in the profiles of VFAs (acetate, propionate, and butyrate) or MCP production (p > 0.05).

Regarding gas kinetics, no significant between-group differences were observed in the soluble fraction (a) or the degradable fraction (b) (p > 0.05 for both). However, no significant between-group difference was observed in the values of in vitro digestibility. In addition, no significant between-group difference was observed in the degradation rate constant (c).

In terms of digestibility, IVTDMD and IVNDFD did not significantly differ between the groups (p > 0.05). The IVTDMD values ranged from 85.8% to 88.4%, and the IVNDFD values ranged from 57.3% to 65.1%, indicating that fermentation did not negatively affect the DM or fiber digestibility of okara.

3.1.6. In Situ Degradation of Fresh and Fermented Okara

Table 8. Gas kinetic parameters, and in vitro digestibility of fresh and fermented okara after 48 h of in vitro fermentation.

Item 1	Control	Fresh	FO	SEM	p-Value
	DM Degradation Parameters
Rapidly degradable fraction a (%)	20.69 c	30.69 a	21.42 b	0.96	<0.01
Slowly degradable fraction b (%)	79.22 b	65.73 c	78.52 b	2.89	<0.01
Rate of degradation (Kd, %/h)	0.08	0.08	0.07	0.01	0.12
Effective degradability 2 (%)	60.96 a	64.29 a	58.78 b	1.44	<0.01
	Protein Degradation Parameters
Rapidly degradable fraction a (%)	2.98	5.38	4.75	3.63	0.54
Slowly degradable fraction b (%)	20.52	30.07	38.26	5.72	0.98
Indigestible fraction c (%)	76.50	64.55	56.99	6.77	0.23
Rate of degradation (Kd, %/h)	0.09	0.05	0.03	0.25	0.76
RDP 3 (% CP)	13.83	17.48	15.53	4.11	0.98
RUP 4 (% CP)	86.17	82.52	84.47	4.11	0.98

¹ Control = ensiled okara without incubation; Fresh = lyophilized fresh okara; FO = lyophilized fermented okara (80% moisture, 10⁶ CFU/g inoculation concentration, 4 week fermentation). ² Effective degradability = a + (b × K_d)/(K_d + K_p), where Kp is the assumed rumen particle passage rate (0.08/h). ³ RDP = rumen degradable protein; CP = crude protein; DM = dry matter; SEM = standard error of the mean. ⁴ RUP = rumen undegradable protein. Different superscript letters indicate significant differences between groups.

presents the in situ rumen degradation kinetics of DM and CP for fresh and fermented okara. Significant between-group differences were observed in the DM degradation characteristics. The rapidly degradable DM fraction (a) was highest in the fresh okara group (30.69%), followed by the fermented okara (21.42%) and control (20.69%) groups (p < 0.01). By contrast, the slowly degradable DM fraction (b) was significantly lower in the fresh okara group (65.73%) than in the fermented okara (78.52%) and control (79.22%) groups (p < 0.01). No significant between-group differences were observed in the degradation rate constant (K_d) (p = 0.12). The ED of DM, calculated at a rumen passage rate of 0.08/h, was significantly higher in the fresh okara group (64.29%) than in the fermented okara group (58.78%, p < 0.01), whereas the ED of DM in the control group (60.96%) did not significantly differ from that of the fresh okara group, but was significantly higher than that of the fermented okara group (58.78%, p < 0.01).

No significant between-group differences were observed in protein degradability. However, the fermented okara group had a numerically higher slowly degradable protein fraction (p > 0.05) compared with the other groups. The RDP and RUP contents also did not significantly differ among the three groups.

3.2. Feed Utilization Trial

3.2.1. Gas Production and Fermentation Characteristics of Total Mixed Rations with Okara Replacement

The inclusion of fermented okara at various replacement levels (0–100%) in TMRs significantly affected several in vitro fermentation parameters at both 24 and 48 h incubation periods (

Table 9. Gas production and fermentation products of TMR groups with different fermented okara replacement ratios.

Item	Replacement (%)
	0	25	50	75	100	SEM	p-value
24 h 1
Gas production (mL/g DM)	126.3 b	175.5 a	159.0 a	173.2a	163.0 b	6.1	<0.01
pH	5.85	5.81	5.80	5.87	5.80	0.02	0.12
NH3-N 2 (g/kg DM)	0.53 b	1.08 a	1.09 a	1.17 a	1.06 a	0.07	<0.01
Acetate (mM)	32.34 a	24.92 b	44.86 a	49.78 a	46.80 a	2.49	0.02
Propionate (mM)	12.15 b	9.74 c	15.72 b	18.09 a	18.84 a	0.96	<0.01
Butyrate (mM)	3.03	2.61	3.24	4.02	4.14	0.36	0.72
MCP 2 (mg/g DM)	19.40	20.13	23.36	19.60	13.91	2.50	0.50
MCP (mg/g OM digestibility)	26.79	28.31	33.42	38.66	20.82	4.28	0.41
48 h 1
Gas production (mL/g DM)	220.5 a	228.7 a	224.0 a	183.5 c	206.1 a	5.9	<0.01
pH	5.61 c	5.64 c	5.66 c	5.70 b	5.79 a	0.01	<0.01
NH3-N (g/kg DM)	2.24 b	2.57 a	2.56 a	1.83 b	2.06 b	0.08	<0.01
Acetate (mM)	56.14	58.93	56.37	44.49	52.55	2.75	0.08
Propionate (mM)	18.91	19.47	18.55	15.55	17.48	1.60	0.17
Butyrate (mM)	4.98 a	5.46 a	4.94 a	3.91 b	4.20 b	0.25	0.02
MCP (mg/g DM)	43.78	49.22	56.40	46.56	44.38	2.96	0.79
MCP (mg/g OM digestibility) 3	55.49	63.78	73.73	61.06	58.98	3.85	0.26

¹ 24 h = incubated for 24 h; 48 h = incubated for 48 h. ² NH₃-N = ammonia nitrogen; MCP = microbial crude protein; DM = dry matter; OM = organic matter; SEM = standard error of the mean. Different superscript letters indicate significant differences between groups.

). At 24 h, gas production significantly increased with 25–75% okara replacement, reaching 175.5 mL/g DM at 25% replacement (p < 0.01). The concentration of NH₃-N was significantly higher in all fermented okara replacement groups than in the control group (p < 0.01), indicating enhanced protein degradation or nitrogen release. The concentrations of acetate and propionate were lowest at the 25% replacement level (p < 0.05), whereas acetate peaked at 75% (49.78 mM) and propionate peaked at 100% (18.84 mM). No significant differences were observed in pH, butyrate concentration, or MCP production across treatments at a 25% replacement level.

At 48 h, total gas production remained high in the 0–50% replacement groups but significantly declined at 75% replacement (p < 0.01), suggesting a threshold for optimal fermentability. The pH values linearly increased with higher replacement levels (p < 0.01), with the highest pH observed at 100% replacement (5.79), indicating decreased acid production or buffering effects. The concentration of NH₃-N was significantly higher in the 25% and 50% replacement groups than in both the control and the 75% replacement group (p < 0.01). The concentration of butyrate was significantly lower in the 75% and 100% replacement groups than in the other groups (p = 0.02). No significant differences were observed in acetate, propionate, or MCP production among treatments at 48 h (p > 0.05).

3.2.2. In Vitro Gas Kinetics and Digestibility of Total Mixed Rations

Replacing soybean meal with fermented okara in TMRs significantly affected gas production kinetics (see Section 2.6), particularly the soluble and degradable fractions (

Table 10. In vitro gas production kinetics of TMR groups with different soybean meal replacement ratios.

Item	Replacement (%)
	0	25	50	75	100	SEM	p-value
Soluble fraction a (%)	3.90 bc	5.17 ab	5.91 a	2.94 c	4.59 ab	0.52	0.03
Degradable fraction b (%)	82.95 a	81.81 a	79.80 a	68.32 c	72.88 b	1.32	<0.01
Degradable rate (c, h−1)	0.124 a	0.122 a	0.122 a	0.116 b	0.123 a	0.019	0.03
Half time (h)	5.57 b	5.52 b	5.68 b	6.09 a	5.66 b	0.11	0.02
IVTDMD 1 (%)	78.90	77.17	76.52	76.40	75.11	1.28	0.15
IVNDFD 2 (%)	42.60	41.11	42.11	44.53	43.99	3.09	0.69

¹ IVTDMD = in vitro true dry matter digestibility. ² IVNDFD = in vitro neutral detergent fiber digestibility. SEM = standard error of the mean. Different superscript letters indicate significant differences between groups.

). The soluble gas fraction (a) significantly differed among treatments (p = 0.03), with the highest value observed at 50% replacement (5.91%), indicating the increased availability of rapidly fermentable components. The degradable fraction (b) significantly decreased from 82.95% in the control group to 68.32% with increasing replacement levels of okara (75%, p < 0.01). The degradation rate of the degradable fraction (c) was significantly lower at 75% replacement in the fermented okara group (0.116 h⁻¹) than in the other groups (0.122–0.124 h⁻¹, p = 0.03), suggesting a slower fermentation process at higher replacement levels. Consistently, the half-time for gas production was significantly longer at 75% replacement for fermented okara (6.09 h, p = 0.02), further indicating delayed substrate utilization. No significant differences were observed in IVTDMD or IVNDFD among the treatments.

3.2.3. In Situ Degradation of Total Mixed Rations with Okara Replacement

Table 11. In situ rumen DM degradation kinetics of TMR groups with different fermented okara replacement ratios.

	Replacement (%)
	0	25	75	SEM	p-Value
Rapidly degradable fraction a (%)	43.31	44.43	44.85	0.60	0.09
Slowly degradable fraction b (%)	47.06 a	45.59 b	44.61 b	0.34	0.15
Rate of degradation (Kd, %/h)	0.083	0.087	0.077	0.007	0.59
Effective degradability0.02 1	80.92 ab	81.47 a	80.19 b	0.24	0.04
Effective degradability0.06	70.23	71.38	69.84	0.59	0.17
Effective degradability0.08	66.93	68.15	66.66	0.57	0.19

¹ Effective degradability = a + (b × K_d)/(K_d + K_p), where K_p is the assumed rumen passage rate (0.02, 0.06, or 0.08/h). SEM = standard error of the mean. Different superscript letters indicate significant differences between groups.

presents the in situ rumen degradation characteristics of TMRs containing varying levels of fermented okara as a replacement for soybean meal. DM degradability exhibited minimal changes at 25% and 75% replacement levels. ED, calculated at a rumen passage rate of 0.02/h, significantly differed among the treatments (p = 0.04). The 25% replacement group had the highest ED value (81.47%), followed by the 0% (80.92%) and 75% (80.19%) groups. However, no significant differences were observed at higher assumed passage rates of 0.06/h and 0.08/h (p = 0.17 and 0.19, respectively), where ED values ranged from 66.66% to 71.38%.

4. Discussion

4.1. Fermentation Efficacy and Preservation

4.1.1. Fermentation Characteristics

In the laboratory-scale fermentation trials, the pH of all treatments decreased below 4.5 by day 12, reaching a minimum at that time point (Table 2). The group treated with 10⁶ CFU/g at 80% moisture had the lowest pH, which dropped to 4.15 (p < 0.01) and remained low through day 24 (4.23, p < 0.01). This acidification process was primarily attributable to lactic acid production by Lactobacillus spp. [31], which creates unfavorable conditions (pH < 4.5) for undesirable microorganisms such as Escherichia coli and Staphylococcus aureus [32]. Excessive moisture can reduce porosity and increase viscosity, whereas insufficient moisture can limit substrate solubility [33]. These findings are consistent with those of a previous study on solid-state fermentation [34]. NH₃-N reflects protein and amino acid degradation [35], any reduction in its concentration over time indicates active nitrogen utilization. Unlike LAB, which rarely assimilate inorganic nitrogen for amino acid synthesis [36], S. cerevisiae likely contributes to NH₃-N reduction by utilizing yeast-assimilable nitrogen compounds, such as glutamine and asparagine, during alcoholic fermentation [37]. In the present study, by day 24, treatment with 10⁶ CFU/g at 80% moisture resulted in both the lowest pH and a relatively high concentration of NH₃-N, making this the most favorable condition for scaling up okara fermentation.

During storage, the chemical composition (DM, CP, NDF, ADF) remained stable over the 8-week period (Table 4), highlighting the role of optimal fermentation in maintaining nutritional quality. However, preservation stability declined after week 4, as evidenced by rising pH, declining lactate, and increasing NH₃-N concentrations (Table 5). The increased lactic acid level at this point reflects the active proliferation of LAB and rapid acidification, which are both essential for suppressing spoilage microorganisms and maintaining silage quality [7]. Lactic acid is a primary catabolic product of pyruvate metabolism by Lactobacillus spp. during fermentation [31]. Its accumulation reduces environmental pH, which in turn inhibits the growth and reproduction of undesirable microbes such as E. coli and S. aureus [32]. This study confirmed that LAB fermentation effectively reduced the pH of okara, thereby enhancing its preservation potential. The pronounced pH decline observed after 4 weeks of fermentation was primarily attributable to lactic acid accumulation, consistent with findings reported by Takenaka and Echo [38]. These results support previous studies indicating that LAB fermentation can stabilize high-moisture byproducts, such as okara, when the fermentation conditions are optimized [2,8].

Overall, LAB fermentation effectively stabilized okara during the early weeks of storage, but preservation beyond 4 weeks showed signs of instability, suggesting a practical feeding window of 2–4 weeks for maintaining quality. These changes may be attributable to the depletion of available carbon sources, decreased LAB activity, or microbial contamination [39]. By weeks 6 and 8, the concentration of NH₃-N exceeded 40 g/kg DM, a threshold commonly associated with low silage quality and decreased palatability for ruminants [40]. Simultaneously, a marked increase was observed in acetate concentration and LAB counts, followed by a resurgence of yeast populations during the early stages and a subsequent decline by week 6, further supporting the occurrence of microbial succession and aerobic instability.

4.1.2. Aerobic Stability

During the aerobic stability test (week 4), NH₃-N concentrations increased significantly with prolonged air exposure, peaking at 48 h, while acetate also rose markedly from 12 h onward (Table 6). Yeast and aerobic bacterial counts increased after 48 h, confirming a progressive decline in aerobic stability. Aerobic stability of silage is defined as the period after aerobic exposure during which the internal temperature of silage remains within 2 °C of the ambient temperature. Once aerobic microorganisms begin oxidizing organic acids and WSCs, carbon dioxide and water are produced, typically leading to an increase in internal silage temperature [35]. In this study, although no significant increase was observed in temperature, likely because of the high ambient temperature, the concentrations of NH₃-N and acetate significantly increased after 24 h of aerobic exposure. Yeast and aerobic bacterial counts also markedly increased after 48 h, indicating a gradual decline in silage quality during aerobic exposure. This pattern is characteristic of aerobic deterioration, in which microbial activity accelerates nutrient breakdown. The increase observed in the concentration of acetate suggests a shift from anaerobic lactic acid fermentation to aerobic metabolism, likely driven by spoilage yeast, such as S. cerevisiae [37]. From a practical perspective, fermented okara can be stored anaerobically for up to 4 weeks with acceptable quality, but once opened, its aerobic stability declines rapidly within 24–48 h. To maintain nutritional value and palatability, fermented okara should be fed within 12–24 h of opening, emphasizing the need for airtight storage and timely incorporation into TMRs.

4.2. Enhanced Nutritional Value of Fermented Okara

4.2.1. Fermented Okara Characteristics and Digestibility for Feed Use

The comparison among control, fermented okara, and fresh okara treatments showed no significant differences in gas production, in vitro kinetics, IVTDMD, or IVNDFD, indicating that LAB fermentation did not negatively affect the digestibility of okara. However, the fermented okara group exhibited significantly higher NH₃-N concentrations, consistent with reports that fermentation increases protein solubility and non-protein nitrogen release [2,9]. Despite these positive findings, high NH₃-N concentrations may also reflect suboptimal microbial protein synthesis, particularly when not accompanied by adequate fermentable carbohydrate availability [13]. This inefficiency may be linked to the acetate-to-propionate (A/P) ratio, which serves as an indicator of energy utilization efficiency in the rumen. A lower A/P ratio generally corresponds to higher energetic efficiency [41]. A decrease in this ratio is associated with a reduction in NH₃-N accumulation as a result of enhanced microbial nitrogen capture [42]. Chen et al. [20] reported a negative correlation between the A/P ratio and MCP synthesis efficiency. In the present study, the A/P ratios were 2.1, 2.0, and 2.29 for the control, fresh, and fermented okara groups, respectively. The high A/P ratio observed in the fermented okara group suggests low-efficiency fermentation and nitrogen utilization. According to the literature, NH₃-N accumulation may result from an imbalance between protein degradation and carbohydrate availability, thereby limiting microbial protein synthesis [43].

Gas production is strongly associated with microbial activity and fermentation dynamics in the rumen [44]. Yang [45] reported a gas yield of 207 mL/g DM after 48 h of ensiled okara fermentation in the rumen. This value is comparable to those observed in the present study, along with similar acetate and propionate concentrations. VFAs, including acetate, propionate, and butyrate, are key end-products of the microbial fermentation of organic matter in the rumen. They serve as essential energy sources and substrates for gluconeogenesis [46]. Therefore, VFA profiles not only reflect fermentation characteristics but also have direct implications for ruminant growth and milk production. Overall, although the fermented okara group exhibited a high NH₃-N concentration, gas production and MCP synthesis did not significantly differ among treatments, suggesting stable overall fermentation efficiency and microbial growth potential. This supported the feasibility of fermented okara inclusion in dairy cow diets. Provided that adequate fermentable carbohydrates are supplied, fermented okara can serve as a stable protein source in TMRs without impairing digestibility.

4.2.2. In Situ Rumen Degradability in Feed Applications

In situ rumen degradation kinetics revealed that fermentation significantly affected the DM degradability of okara, although its CP degradation dynamics remained unchanged (Table 8). Specifically, fermented okara exhibited a lower ED of DM compared with fresh okara (58.78% vs. 64.29%, p < 0.01), primarily because of a reduction in the rapidly degradable fraction (a). By contrast, the slowly degradable fraction (b) was significantly higher in both the control and fermented okara groups than in the fresh okara group, suggesting that fermentation induces structural modifications into the okara matrix. Given that dietary fibers constitute approximately 40–60% of okara’s dry weight, which is predominantly composed of insoluble components such as cellulose and hemicellulose [2], this increase in fraction b can be attributed to enzymatic action, particularly β-glucanase production by S. cerevisiae, which can degrade glucans and alter fiber structure [9,31]. In addition, the sum of a + b remained high in both the control (99.9%) and fermented okara (99.6%) groups, exceeding that of fresh okara (96.4%), indicating an overall increase in the proportion of rumen-accessible substrates after fermentation. Although a shift from rapid to slowly degradable fractions may reduce the immediate availability of fermentable substrates, it may also promote more sustained nutrient release, thereby enhancing rumen stability, particularly in diets rich in rapidly fermentable carbohydrates. In the present study, the ED of fermented okara (58.78%) was comparable to that of commonly used forage ingredients, such as oat hay (59.97%) and alfalfa hay (58.89%) [47], indicating its potential as a functional fiber and protein source in dairy cow diets formulated for gradual nutrient delivery and enhanced rumen function.

In terms of protein utilization, no significant between-group difference was observed in the concentrations of RDP and RUP. This finding indicated that LAB-based fermentation preserved the protein fraction without inducing excessive proteolysis or altering the ruminal degradability profile. In situ test results revealed that both fresh and fermented okara had a higher proportion of RUP (approximately 82–86%) compared with the higher proportion of RDP (approximately 65%) found in soybean meal [48]. Notably, the high degradability of soybean meal may not be suitable for high-producing dairy cows because it increases urinary nitrogen loss [49]. This high RUP content of okara positions it as a promising source of bypass proteins, which are capable of delivering a constant amino acid supply to the small intestine of ruminants. This characteristic may help reduce feed cost and mitigate reliance on expensive animal-derived protein supplements in ruminant diets.

4.3. Application of Okara in Ruminant Diets

Effects of Fermented Okara Replacement Levels on TMRs

Analysis of the in vitro fermentation trial results revealed that partial replacement of soybean meal with fermented okara in TMR formulations can positively affect ruminal fermentation dynamics and nutrient utilization (Table 9). Moderate inclusion of fermented okara (25–50% replacement of soybean meal) exhibited significantly increased gas production at 24 h, indicating enhanced fermentability and microbial activity [44]. This finding suggests that moderate fermented okara inclusion provides a fermentable substrate profile favorable for ruminal microbial metabolism, potentially because of the presence of lactic acid, pre-fermented carbohydrates, and reduced fiber complexity [2]. This interpretation is further supported by the VFA profiles. Although the concentrations of acetate and propionate increased with fermented okara inclusion, the highest values were observed at 50–100% replacement levels, indicating a shift toward more efficient fermentation pathways. Despite these findings, the decline observed in total gas production and VFA yield at 75% replacement at 48 h indicates that excessive replacement may impair fermentation efficiency, likely as a result of fiber recalcitrance, an imbalanced energy-to-protein ratio, or a decline in ruminal buffering capacity. In addition, the increased NH₃-N concentration observed at 24 h in all fermented-okara-containing diets indicates increased protein solubility and deamination. Although this phenomenon may increase nitrogen availability for microbial use, it also underscores the importance of balancing fermented okara inclusion with adequate fermentable carbohydrate sources to optimize microbial protein synthesis [13].

An increase in the soluble-gas-producing fraction (a) at 25–50% fermented okara replacement levels indicates that moderate inclusion enhances the availability of rapidly fermentable substrates (Table 10), likely because of the increased levels of WSCs and small peptides generated during fermentation [7]. These components may contribute to rapid initial microbial fermentation and improved energy availability for ruminal microbes during early digestion. By contrast, the reduction in the degradable fraction (b) and degradation rate observed at 75% fermented okara replacement suggests that excessive inclusion impairs fermentability. This reduction may be attributable to an increase in structural fiber content or the accumulation of residual fermentation products that inhibit microbial access to fermentable substrates [2,9]. This interpretation is further supported by the prolonged half-time for gas production at high fermented okara levels, indicating delayed microbial activity and decreased substrate utilization efficiency [15]. Although IVTDMD and IVNDFD remained unaffected, consistent with earlier findings that okara inclusion does not impair digestibility when properly balanced [4,50]. Taken together, these results suggest that moderate fermented okara inclusion supports or enhances fermentation dynamics, whereas excessively high replacement levels (≥75%) may reduce gas production efficiency without compromising final digestibility outcomes.

In this study, the in situ DM degradation kinetics of TMRs formulated with varying levels of fermented okara (Table 11) revealed that neither the rapidly degradable fraction (a) nor the slowly degradable fraction (b) was significantly affected by fermented okara replacement. This finding indicates that fermented okara inclusion does not alter the inherent degradability characteristics of basal TMR ingredients. Research has indicated that ruminal passage rates and digestive metabolism significantly differ between high-producing dairy cows and animals with lower productivity or intake, such as dry cows or those exposed to heat stress [13]. In this study, ED at a low ruminal passage rate (0.02/h), a condition representative of slow rumen turnover, was found to be significantly higher at 25% fermented okara inclusion than at 75% fermented okara inclusion. This finding indicates that moderate fermented okara inclusion (i.e., 25%) may improve nutrient availability under conditions of decreased DM intake, supporting a degradability profile comparable to that of soybean meal–based rations.

5. Conclusions

LAB fermentation stabilizes okara by enhancing its aerobic stability and reducing pH, particularly under conditions of approximately 80% moisture and an inoculation level of 10⁶ CFU/g. Although fermentation improves certain nutritional and preservation characteristics, its integration into ruminant diets must be carefully balanced with fermentable energy and nitrogen sources to prevent excessive nitrogen excretion, as indicated by increased NH₃-N concentrations. Notably, the relatively high RUP content of fermented okara highlights its potential as a valuable bypass protein source, capable of delivering a constant amino acid supply to the small intestine. From a ration formulation perspective, fermented okara can be strategically incorporated at a 25–50% replacement level for soybean meal (on a CP basis). Overall, fermented okara is a feasible alternative protein source for on-farm feeding and stably contributes to the bypass protein fraction in TMRs.