The chemical composition of the basal diet (Table 2) was calculated using the manufacturer’s nutrient profiles of the ingredients (Yazd Mokamel Co., Yazd, Iran) and standard output values from the National Research Council (NRC) 2007 standards [10]. Before sample collection, the foals underwent a 14-day adaptation period to a controlled maintenance diet formulated according to the recommendations of the NRC [10]. The ration (basal diet) consisted of a mixture of 1.8 kg of mash-formulated concentrate and 7 kg of lucerne hay (Medicago sativa). Meals were provided three times daily at 06:00, 14:00, and 22:00.

A basal substrate was prepared from a mixture of 7 kg lucerne hay and 1.8 kg mash-formulated concentrate to simulate the donor diet in the in vitro system. Both ingredients were ground separately to pass through a 2-mm mesh screen and subsequently thoroughly mixed by hand in a large plastic vessel to ensure homogeneity. To determine DM content, a 1 kg aliquot of the homogenized basal diet mixture was spread thinly in a pre-weighed Teflon container and dried at 65 °C for 24 hours. The container was reweighed after cooling in a desiccator. Portions of this dried basal diet (300 mg) were weighed into sterile 50 mL serum bottles to serve as the in vitro assay substrate.

Four healthy Arabian gelding foals (average age, 14 months; body weight 285 ± 9.65 kg) were selected as donor animals. Before the trial began, horses were checked for parasites and overall health by a certified veterinarian. Growing horses were chosen as the biological model for this study because they possess a developing hindgut microbiome that is particularly susceptible to digestive disturbances and oxidative stress. This heightened sensitivity makes this demographic clinically superior to adult horses for evaluating the efficacy of phytogenic interventions that modulate fermentation kinetics and gut health [11]. The equine hindgut microbiome undergoes significant successional changes and does not reach complete stability until adulthood [12]. Consequently, evaluating phytogenics during this transitional phase is physiologically critical because the developing microbiome is more responsive to dietary modulation and more susceptible to dysbiosis than that of mature horses [13].

The foals were housed in individual, fully ventilated stables equipped with ceiling fans and automated stainless-steel waterers to ensure ad libitum access to fresh water. Stalls were bedded with wood shavings and were cleaned daily. The environmental conditions during the trial (mid-spring 2020) were controlled to maintain a mean ambient temperature of 28°C and a relative humidity of 28%.

On day 14 of the adaptation period, fecal samples were collected via rectal stimulation 2 h pre-prandial to minimize post-feeding microbial variations. Approximately 500 g of feces per horse was immediately collected into pre-warmed (39°C), CO₂-flushed (for 10 min) vacuum flasks to ensure anaerobic conditions. The sealed flasks were transported to the laboratory within 30 min of collection.

Upon arrival, equal-weight aliquots from the four donors were pooled under continuous CO₂ for 10 mins to create a composite inoculum, a step used to minimize individual host variability. The composite sample was homogenized for 30 s under continuous CO₂ flushing to maintain anaerobiosis. Subsequently, the homogenate was diluted 1:4 (w/v) with McDougall’s [14] buffer (a bicarbonate-phosphate artificial saliva, pH 6.9), which had been pre-warmed to 39°C and saturated with CO₂ for 10 mins. The resulting slurry was gently mixed for 1 min and filtered through a double-layered cheesecloth to remove large particulates while retaining the microbial fraction.

The in vitro gas production procedure followed the protocol described by Nobari et al. [15]. This method utilizes gas accumulation as an indicator of digestion kinetics, reflecting the production of volatile fatty acids (VFAs), CO₂ methane, and trace hydrogen during anaerobic fermentation. Incubation was initiated by adding the buffered inoculum to the prepared serum bottles (containing substrate and treatments) under anaerobic conditions. Inoculation was performed by dispensing 20 mL of the filtered fecal-buffer mixture into pre-warmed 50 mL serum bottles containing 300 mg of the basal diet as the substrate. This volume results in a headspace of approximately 30 mL per bottle. Six replicates (n = 6) were prepared for each treatment to ensure uniformity and statistical power.

Immediately following inoculation, the headspace was flushed with CO₂ to establish anaerobic conditions for 60 s, and the bottles were hermetically sealed with butyl rubber stoppers and aluminum crimp caps. The bottles were sealed and maintained at 39°C in an incubator-shaker (Model KM11, Fan Azma Gostar, Iran). Gas volume was measured at 2, 4, 6, 8, 12, 16, 24, 36, 48, 72, 96, and 120 h of incubation using Fedorah and Hrudey’s [16] water displacement method, in which the evolved gas displaces water in graduated tubes connected to the fermentation vessels. To account for non-substrate gas production, six “blank” bottles containing only buffered inoculum were included as negative controls. The net fermentation gas production for each treatment was calculated by subtracting the mean gas volume produced by the blank bottles from the volume recorded for each treatment bottle at each time point.

The study employed a completely randomized design with a 2 × 2 factorial arrangement, yielding four distinct treatment groups (n = 6 replicates per treatment). Two stock solutions were prepared using distilled water as follows:

Stock Solution A (Asafoetida): 6 g of F. asafoetida alcoholic extract diluted to 1 L (6 mg/mL; 0.6% w/v).

The specific inclusion levels (30 mg for F. asafoetida and 50 mg for RPS) were selected based on a preliminary pilot study in our laboratory, which established these concentrations as the optimal range for maximizing fermentation kinetics while maintaining pH stability. The experimental treatments were prepared by adding these solutions to serum bottles containing 300 mg of basal diet as follows:

Control (C): A basal diet only (no additives).

The experiment was conducted as a completely randomized design with a 2 × 2 factorial arrangement of treatments. The factors included Asafoetida extract (0 and 30 mg/mL) and RP extract (0 and 50 mg/mL), resulting in four treatment combinations with six replications per treatment (n = 6).

Since gas production is a curve (starts slowly, accelerates, and then decelerates), cumulative gas production data recorded over time were fitted to the nonlinear Gompertz equation using the PROC NLIN procedure in SAS 9.4 (SAS Institute, Cary, NC) [17]. The Gompertz equation is the standard biological model for S-shaped growth or fermentation curves. PROC NLIN uses an iterative algorithm (i.e., guessing the parameters, checking the error, adjusting the guess, and repeating) until it finds the “best fit” curve [17]. The Marquardt iterative method was employed to minimize the residual sum of squares and estimate the following kinetic parameters:

Where:

V = Cumulative gas production (mL) at time t.

The goodness of fit for the nonlinear models was assessed by calculating the Pseudo-R² and ensuring convergence of the iteration limit.

Derived parameters (b, c, L), fermentation pH, NH₃-N, and digestibility coefficients of nutrients were subjected to analysis of variance using PROC GLM. Before the analysis, the data were screened for normality using the Shapiro-Wilk test (W > 0.90) and for homogeneity of variance using Levene’s test. The statistical model used was as follow:

Where:

Y_ijk is the dependent variable.

In significant F-tests were observed (p < 0.05), treatment means were separated using the Tukey–Kramer test to adjust for multiple comparisons and control Type I error. Trends are discussed at 0.05 < p < 0.10.

To determine apparent digestibility, blank correction was performed gravimetrically after drying. The mean residue weight of blank bottles (n = 8 containing only buffered inoculum) was subtracted from the treatment residues. This step corrects for contributing buffer salts and inoculum-derived microbial biomass.

The chemical composition of risin residues has been analyzed, highlighting their potential as nutrient- and fiber-rich ingredients in animal diets. The findings align with those of prior studies, demonstrating the value of raisin pomace in supporting equine nutrition through its tannin compounds (Table 3). Kolláthov et al. [22] reported that raisin pomace supplementation can potentially improve the antioxidant status of the Slovak Warmblood breed.

The chemical profile of Asafoetida comprises approximately 40%–64% resin, 25% gum, and 10%–17% essential oil. The resin fraction is rich in sulfur-containing compounds, such as butyl propenyl disulfide, which contribute to its characteristic aroma and therapeutic efficacy [23]. To the best of our knowledge, this analysis constitutes the first comprehensive GC-MS biochemical profile of F. asafoetida specifically established for equine nutritional applications. By identifying key bioactive constituents, such as ϒ-Gurjunenepoxide-(2) (Table 4), specific chemical metabolites can be linked to the observed shifts in hindgut fermentation kinetics.

CO₂ extracts from the plant’s underground parts are particularly high in unsaturated fatty acids, such as oleic (46.1%) and linoleic acids (43.0%), along with essential amino acids [24]. GC-MS analysis of the F. asafoetida extract, based on the NITS mass spectral libraries, revealed a diverse array of bioactive compounds, including coumarins, phenolics, terpenoids, organic acids, and other metabolites with known pharmacological activities (Figure 1, Table 2).

Notably, over 60% of these compounds exhibit anti-inflammatory and antiparasitic properties, underscoring the extract’s potential to modulate gut health (25.92%), reflecting its antioxidant, antimicrobial, and anti-inflammatory properties. Ferulic acid, a prominent constituent, along with polysulfides and coumarins, further supports the pharmacological relevance of the extract [25]. Traditionally used across Iran, Iraq, and Egypt, Asafoetida is used to treat gastrointestinal disorders, including bloating, dyspepsia, and intestinal spasms. Its antispasmodic and antimicrobial effects are well documented [26, 27].

The addition of Asafoetida extract (A), condensed tannins (G), and their combination (A × G) significantly influenced the fermentation kinetics (Table 4). The cumulative gas production parameter (V) was highest in treatments G (358.5 mL at 120 h) and A × G (355.6 mL), showing an increase of approximately 53% over the control treatment (228.8 mL), which was significant (p < 0.01). The maximum fermentation rate (b) also increased from 55.4 mL in the control to 150.2 mL in G and 187.1 mL in A × G, with corresponding reductions in lag times. The results indicate a marked improvement in gas production and fermentation rates with the addition of F. asafoetida extract (A) and condensed tannins (G), either individually or in combination. Treatment G produced the highest cumulative gas volume and maximum fermentation rate, which is consistent with the results of previous studies demonstrating the fermentative efficiency of plant-derived tannins [28]. The A × G combination yielded gas production comparable to that of G alone but showed a higher fermentation rate constant (c).

The observed improvements in fermentation parameters may be attributed to a potential synergistic interaction between the sulfur-containing compounds in Asafoetida and the tannins. Sulfur compounds, such as butyl propenyl disulfide, have been previously documented to possess antimicrobial properties [29]. It is hypothesized that these compounds might modulate the composition of the gut microbiota, potentially creating an environment more conducive to efficient fermentation. However, specific shifts in microbial populations need to be confirmed via genomic analysis. Tannins are widely recognized for their ability to modulate microbial populations, often by selectively inhibiting specific bacterial groups while preserving others [30]. The combination of Asafoetida and tannins may influence digestion through complementary mechanisms [31]. Previous studies on plant-derived bioactive compounds suggest that such combinations can alter the profiles of fermentation by-products, including NH3-N and VFA profiles [32].

The slightly prolonged lag time (L) in the A×G treatment compared to G suggests a period of microbial adaptation to the complex mixture, a phenomenon consistent with observations in other herbivorous species [33].

Although individual VFAs were not quantified by chromatography in this study, GP provides an estimate of aggregate short-chain fatty acid (SCFA) synthesis. In a bicarbonate-buffered system, gas production is stoichiometrically proportional to SCFA generation due to the release of CO₂ from buffering reactions [34]. The significant increase in the cumulative gas volume observed with the Asafoetida and Tannin treatments suggests a net increase in the overall fermentative activity. Furthermore, the c parameter implies that these additives may have facilitated more rapid fermentation once the initial lag phase was overcome, aligning with the findings of Lowman et al. [35] in equine models. We hypothesize that the observed synergy results from a novel dual-action mechanism. The organosulfur compounds in F. asafoetida (e.g., disulfides) likely modulate microbial activity through redox interactions, whereas condensed tannins regulate nitrogen availability via reversible protein complexation. This complementary action optimizes the fermentation environment more effectively than either additive alone [36, 37].

The pH of the fermentation medium remained relatively stable across all treatments during the 120-h incubation period, within the physiological range conducive to fibrolytic activity (Table 6). Although the combined treatment (A × G) recorded the lowest terminal pH (6.33), this value did not fall below the threshold for subclinical acidosis, suggesting that the system’s buffering capacity was not overwhelmed by the accelerated fermentation rates.

Regarding nitrogen dynamics, tannin inclusion significantly influenced proteolytic activity. NH₃-N concentrations were elevated in the G (26.0 mg/dL) and A × G (25.5 mg/dL) treatments compared with the control (24.5 mg/dL; p < 0.01). While condensed tannins are often associated with protein-binding that reduces ammonia accumulation, the increased NH3-N levels observed here suggest active microbial protein degradation and turnover. This implies that the specific tannins in RP, at the concentrations used, did not inhibit proteolytic bacteria but instead supported a robust microbial population capable of accessing dietary nitrogen. Alternatively, the sulfur-containing compounds in Asafoetida may have modulated redox potential, further facilitating amino acid deamination [37].

Both DMD and CPD were significantly improved in the grape pomace (G) and combined (A × G) treatments compared with the control. This enhancement suggests a favorable modulation of the fermentative environment. Although specific microbial populations were not quantified, previous studies have indicated that tannins can selectively modulate the microbiome, potentially favoring fibrolytic activity and VFA production [38]. The A × G treatment demonstrated comparable benefits, although the values were slightly lower than those with G alone.

The sulfur-rich compounds in F. asafoetida may interact with tannins via nucleophilic addition to quinones, possibly moderating their binding capacity or enzymatic inhibition [39, 40]. Consequently, the observed increase in substrate degradation implies a greater potential for VFAs for energy yield. It is important to acknowledge that condensed tannins have a dual nature. While moderate doses are beneficial, high concentrations can exert antinutritional effects, including chelation of essential minerals and formation of irreversible irreversible tannin-protein complexes that reduce nitrogen availability [41]. Therefore, in vivo trials are essential to determine the optimal inclusion rates that maximize digestibility without compromising nutrient bioavailability or negatively impacting animal health and performance, while these in vitro results indicate improved nutrient accessibility.

In addition, excessive phenolic loads have been linked to reduced feed palatability or even animal health issues in in vivo settings [42]. However, the significant improvement in CPD in the current study suggests that the 50 mg inclusion level remained below the threshold for these harmful effects, likely modulating proteolysis without compromising nutrient accessibility.