All procedures involving the collection and use of rumen fluid from fistulated cattle were reviewed and approved by the Animal Ethics Committee, School of Veterinary Medicine and Biomedical Sciences, IPB University (Approval No. 395/KEH/SKE/XI/2025).

Coconut and palm fatty acid distillates used in this experiment were sourced from the Indonesian refining industry. The fatty acid profiles of CFAD and PFAD differ markedly in chain length distribution (Table 1). CFAD was dominated by medium-chain fatty acids, particularly lauric acid (C12:0; 50.42%) and myristic acid (C14:0; 19.82%), followed by capric acid (C10:0; 5.94%) and palmitic acid (C16:0; 9.29%), with relatively low levels of unsaturated fatty acids such as oleic acid (C18:1 cis; 5.63%) and linoleic acid (C18:2 cis; 1.55%). In contrast, PFAD was primarily composed of long-chain fatty acids, mainly palmitic acid (C16:0; 40.43%) and oleic acid (C18:1 cis; 33.77%), followed by linoleic acid (C18:2 cis; 19.30%) and stearic acid (C18:0; 3.20%), indicating a higher proportion of long-chain and unsaturated fatty acids compared to CFAD. Saponification of CFAD and PFAD with CaCO₃ was carried out using a double-composition approach (Zahera et al., 2024). Fatty acid distillate was heated until fully melted and saponified with NaOH. Subsequently, CaCO₃ dispersed in distilled water was added under continuous heating (70–90 °C) and stirring to facilitate the exchange of Na⁺ with Ca²⁺, forming calcium soap. The mixture was heated until most of the water evaporated, then dried, ground, and stored for further analysis. A molar ratio of 2:1 between CaCO₃ and Free Fatty Acid (FFA) was applied to ensure the final product reached an acid value below 0.75 mg KOH/g. The acid value was used as an indicator of calcium soap formation, with values below 0.75 mg KOH/g considered indicative of successful saponification.

The dairy ration was prepared using a 40:60 forage-to-concentrate balance, as recommended by (Anzhany et al., 2022) to enhance CLA-enriched milk production. The formulated diet consisted of 65.53% DM, 11.60% ash, 3.70% ether extract, 14.30% crude protein, 21.41% crude fiber, and 65.24% total digestible nutrients. Rumen fluid was collected from three rumen-fistulated Friesian Holstein cows housed in a closed dairy nutrition facility at the Faculty of Animal Science, IPB University, prior to morning feeding.

In vitro fermentability and digestibility were evaluated using a modified two-stage (Tilley & Terry, 1963) procedure, comprising rumen fermentation followed by pepsin–HCl digestion. Each 0.5 g sample was incubated anaerobically at 39 °C with 10 mL of rumen fluid and 40 mL of McDougall’s buffer, using five serial tubes per experimental unit. One tube was fermented for 4 h to analyze total VFA, partial VFA, ammonia, protozoa and bacterial counts, and microbial protein synthesis, as this incubation time represents the peak of microbial fermentation activity for most feedstuffs. Two tubes (0 h and 48 h) were designated for biohydrogenation assessment, while the remaining tubes were incubated for 48 h and then digested for 48 h in pepsin–HCl to obtain residues for determining dry matter digestibility (DMD), organic matter digestibility (OMD), neutral detergent fiber digestibility (NDFD), and acid detergent fiber digestibility (ADFD). Rumen pH was measured with a digital pH meter (Hanna HI98191), ammonia concentration was quantified using Conway microdiffusion, and total VFA concentration was determined via steam distillation following a similar method used by (Anzhany et al., 2024). Total protozoa and bacterial populations were measured using the method of (Ogimoto & Imai, 1981). Microbial protein synthesis (MPS) was determined using a modified Lowry method as described by (Rosmalia et al., 2022).

Partial VFA profiles were determined using gas chromatography (Bruker Scion 436-GC equipped with a Bruker-1ms capillary column). Samples were prepared according to standard VFA analytical procedures, and individual VFA peaks were identified and quantified using a certified reference standard (Supelco Volatile Free Acid Mix). Methane production was estimated from the molar proportions of acetate (C2), propionate (C3), and butyrate (C4) using the equation of Moss et al. (2000): CH₄ (mM) = 0.45C2 – 0.275 C3 + 0.40 C4.

Rumen C18 fatty acid profiles were analyzed using gas chromatography (GC-7820A, Agilent Technologies) following the procedure of (Anzhany et al., 2024). Biohydrogenation indices were calculated from changes in fatty acid concentrations between 0 and 48 h. Apparent biohydrogenation of individual and total C18 PUFA was based on proportional loss during incubation, following (Vlaeminck et al., 2008) with minor modifications. Indices describing incomplete biohydrogenation and CLA yield (intermediate accumulation and CLA formed relative to C18:2 cis disappearance) were adapted from (Kalscheur et al., 1997) and (Cruz-Hernandez et al., 2007). Saturation (SFA : total FA ratio) and unsaturation indices (weighted average number of double bonds per FA) were calculated as in (Jalc et al., 2007). Trans/cis isomerase indices for C18:1 and C18:2 followed the trans/cis ratios used by (Kramer et al., 2004) and (Toral et al., 2024). Total C18 recovery was calculated as the ratio of total C18 FA at 48 h to 0 h, similar to (Harvatine & Allen, 2006).

Hydrogen production, consumption, balance, and recovery were calculated from VFA stoichiometry following the approach of (Moss et al., 2000). Net H₂ from VFA, H₂ consumed in methane, and H₂ balance were derived using standard hydrogen balance equations, and H₂ recovery was expressed as the proportion of metabolizable hydrogen accounted for by methane formation.

The study employed a 2 × 5 factorial randomized block design with 12 replications. Rumen fluid was collected from three fistulated cows, pooled, and used as the inoculum. Different collection times across incubation runs were treated as blocks. Factor A was the type of protected fatty acid distillate (coconut or palm), and Factor B was the supplementation level (0%, 1%, 2%, 3%, and 4%) in the dairy cattle diet. The measured parameters included rumen fermentability (pH, ammonia, and total and individual VFAs), nutrient digestibility (dry matter, organic matter, NDF, and ADF digestibility), estimated methane production, microbial populations (protozoa and bacteria), and biohydrogenation indices. Data were analyzed using ANOVA, and significant treatment effects were further assessed using Tukey’s test. Polynomial contrasts were applied to describe response curves and identify the optimum supplementation levels. All statistical analyses were performed using SPSS version 27.

Table 2 summarizes the effects of CFAD and PFAD supplementation on rumen fermentation parameters, including pH, ammonia concentration, VFA profiles, and methane production. Rumen pH remained within the normal physiological range across all treatments and was not significantly affected by fat source or supplementation level. A significant FAD × level interaction was observed for NH₃ concentration. In the CFAD treatments, NH₃ decreased from 0% to 2% supplementation before rising sharply at 3%, whereas PFAD showed a more gradual decline up to 3%, followed by a slight increase at 4%. The contrast between fat sources was most pronounced at 2% and 3%, where CFAD resulted in significantly lower (2%) and higher (3%) NH₃ concentrations than PFAD.

Total VFA increased markedly with supplementation (p<0.05), with the highest values recorded at 2%–3% inclusion for both fat sources; CFAD consistently produced greater total VFA at these levels. Acetate (C2) was not significantly affected by treatment, although a numerical increase was apparent at 1%–3% inclusion; however, the high variation among replicates prevented this trend from reaching statistical significance. Propionate (C3) decreased at 1% and partially recovered at higher levels, resulting in a significantly elevated acetate-to-propionate ratio (C2:C3) at 1% and 4% (p<0.05).

The branched-chain VFA iC4 exhibited a significant interaction, with concentrations declining after 1% CFAD inclusion but remaining relatively high across PFAD levels. No treatment effects were detected for iC5. Straight-chain C4 and C5 were minimally affected, although PFAD yielded a higher mean butyrate (C4) than CFAD (p<0.05). Methane production increased from 0% to 1% supplementation (p<0.05), remained higher than the control at 3%, and rose again at 4%, with similar patterns for both fat sources.

Table 3 presents the responses of total bacterial and protozoal populations, along with microbial protein synthesis (MPS), to different levels of CFAD and PFAD supplementation. Total bacterial and protozoal counts decreased progressively with increasing supplementation but remained comparable to the control up to 2%, with both fat sources showing similar patterns. Neither microbial population was significantly affected by the type of fat. In contrast, a significant FAD × supplementation level interaction was observed for MPS. Under CFAD, MPS declined from 0% to 2% inclusion and then increased at higher levels, reaching its maximum at 4%. For PFAD, the decline in MPS was more gradual, reaching the lowest value at 3% before increasing again at 4%.

Table 4 shows the effects of CFAD and PFAD supplementation on nutrient digestibility, including dry matter, organic matter, NDF, and ADF digestibility. Dry matter digestibility (DMD) and organic matter digestibility (OMD) responded differently to CFAD and PFAD supplementation. With CFAD, both DMD and OMD began to decline at 2%–3% inclusion, whereas PFAD produced no significant changes across levels, except for a slight reduction at 1% relative to the control. NDF digestibility was not significantly affected by either fat source or supplementation level, while ADF digestibility showed a modest but consistent decrease as supplementation increased.

Table 5 presents the hydrogen metabolism indices derived from VFA stoichiometry and methane production, including H₂ produced, H₂ consumed, net H₂ from fermentation, H₂ used in methane formation, H₂ balance, and H₂ recovery. H₂ produced was not significantly affected by fat source or supplementation level. In contrast, H₂ consumed, net H₂ from VFA, H₂ consumed in methane, and H₂ balance were all influenced by supplementation level (p<0.05), with the highest values generally occurring at 1% and 4% inclusion.

A significant fat source × level interaction was detected for H₂ recovery (%). Both fat sources increased H₂ recovery relative to the control, but PFAD showed a more pronounced rise, particularly at 2%–4% inclusion, whereas CFAD displayed a more modest and variable increase across levels. Overall, supplementation altered multiple hydrogen-related indices, with PFAD produ-cing a steeper enhancement in H₂ recovery than CFAD.

Changes in fatty acid transformation, isomerization, and biohydrogenation of dairy cattle rations supplemented with CFAD and PFAD at various levels are presented in Table 6. The biohydrogenation of C18:2 cis and C18:3 n-6, the saturation index, and the C18:1 isomerase index were not significantly affected by either the type of FAD or the supplementation level. Incomplete biohydrogenation, CLA yield, and total C18 recovery were influenced only by FAD source, whereas total PUFA biohydrogenation, the C18:2 isomerase index, and the 48-h unsaturation index were affected solely by supplementation level. Interaction effects between FAD type and supplementation level were observed for the biohydrogenation of C18:3 n-3 and the 0-h unsaturation index.

Correlation analysis revealed several significant associations between hydrogen metabolism variables and C18 fatty acid transformation indices (Table 7). Biohydrogenation of C18:2 cis showed positive correlations with H₂ produced (r = 0.192, p<0.05) and H₂ consumed via propionate (r = 0.220, p<0.05), and a negative correlation with H₂ recovery (r = –0.191, p<0.05). Biohydrogenation of C18:3 n-6 was negatively correlated with net H₂ from VFA (r = –0.189, p<0.05), H₂ consumed in methane (r = –0.191, p<0.05), and H₂ balance (r = –0.186, p<0.05), while positively associated with H₂ consumed via propionate (r = 0.187, p<0.05).

The saturation index displayed a positive correlation with H₂ recovery (r = 0.315, p<0.01). CLA yield was also positively correlated with H₂ recovery (r = 0.170, p<0.05). The C18:1 isomerization index showed a negative correlation with H₂ consumed via propionate (r = –0.386, p<0.01) and a positive correlation with H₂ recovery (r = 0.307, p<0.01). Meanwhile, the C18:2 isomerization index exhibited a weak but significant positive correlation with H₂ recovery (r = 0.174, p<0.05).

To determine the optimal supplementation level of CFAD and PFAD, polynomial response curves were evaluated for all parameters showing significant effects (p<0.05), as presented in Table 8. The table shows that protozoa population, BH-total PUFA, CFAD-DMD, CFAD-OMD, and PFAD-CH₄ followed linear response patterns, whereas the C18:2 isomerase index, CFAD-MPS, and PFAD-iC4 exhibited quadratic responses. Cubic trends were observed for iC5, the acetate-to-propionate ratio (C2/C3), and CH₄. Meanwhile, total VFA, ADFD, CFAD-NH₃, PFAD-MPS, and the 0-h unsaturation index displayed quartic response curves. The R² values were relatively low (<0.50), likely due to the large data pool (n = 120) combined across all blocks in a randomized block design, which inherently increases variability among replicates. Nevertheless, all regression models were statistically significant (p<0.05), indicating that the fitted curves reliably captured the overall response patterns.

The significant interaction between fatty acid source and supplementation level for NH₃ concentration indicates that rumen nitrogen metabolism responded differently to CFAD and PFAD. Although CFAD and PFAD were supplied as calcium soaps, rumen protection is not absolute. Calcium soaps may partially dissociate under ruminal conditions, allowing a fraction of fatty acids to remain available for interaction with rumen microorganisms and ruminal lipid metabolism. Fat supplementation modified rumen fermentation in a source- and level-dependent manner. The increase in total VFA at moderate inclusion suggests enhanced fermentative efficiency, potentially linked to reduced protozoal grazing pressure.

The greater reduction in NH₃ observed with CFAD than with PFAD may be related to differences in fatty acid composition between the two supplements. CFAD was dominated by medium-chain fatty acids (MCFA), whereas PFAD contained predominantly long-chain fatty acids (LCFA). Accordingly, the observed responses are more likely associated with differences in fatty acid chain length and profile than with saturation alone. Medium-chain fatty acids (MCFA), which dominate CFAD, are widely recognized for their strong antiprotozoal activity. Numerous in vivo and in vitro studies have demonstrated that MCFA markedly suppress Entodinium-dominated protozoal communities (Shi et al., 2020); (Luan et al., 2023). Although protozoal species were not directly quantified in the present experiment, the decline in NH₃ concentrations at 1%–2% CFAD is consistent with reduced protozoal turnover of bacterial protein, a pattern also observed in coconut-oil–based MCFA supplementation studies (Shi et al., 2020); (Joch et al., 2023).

The more pronounced reduction in NH₃ under CFAD supplementation suggests a stronger inhibitory effect on proteolytic and deaminating bacteria, which are known to be sensitive to MCFA. Previous studies have demonstrated that key rumen proteolytic taxa, such as Prevotella ruminicola, Butyrivibrio fibrisolvens, and Clostridium aminophilum, exhibit reduced abundance or metabolic activity when exposed to MCFA (Patra, 2013); (Shi et al., 2020). (Zahera et al., 2024) reported a marked decline in Bacteroides spp. with increasing levels of unprotected CFAD. This suppression of proteolytic microbial populations likely reduced deamination activity, resulting in lower ammonia release from dietary protein (Shen et al., 2018); (Diether & Willing, 2019); (Mitchell et al., 2023).

Compared with CFAD, PFAD exerted a milder effect on nitrogen metabolism, consistent with its greater proportion of LCFA and lower concentration of MCFA. Numerous studies have shown that medium-chain fatty acids, particularly lauric acid (C12:0) and myristic acid (C14:0), exert strong antimicrobial and antiprotozoal effects in the rumen, whereas the long-chain fatty acids that predominate in PFAD generally produce milder microbial responses (Shi et al., 2020); (Luan et al., 2023); (Joch et al., 2023). This is reflected in the more stable NH₃ concentrations observed under PFAD supplementation and is consistent with previous reports showing minimal changes in rumen ammonia dynamics with LCFA feeding (Mavrommatis et al., 2021); (Riestanti et al., 2021). Overall, these patterns indicate that CFAD alters fermentation primarily through MCFA-driven microbial suppression, whereas PFAD exerts more modest and stable effects due to its LCFA profile.

The progressive decline in microbial populations with increasing supplementation level indicates that both CFAD and PFAD exerted inhibitory effects on rumen microorganisms, although the magnitude of suppression tended to differ according to fatty acid composition. Total bacterial and protozoal populations declined with increasing supplementation level, but the magnitude of suppression differed between fat sources. CFAD resulted in stronger microbial reductions than PFAD, indicating that the microbial response differed according to fatty acid composition. This finding is consistent with the stronger antimicrobial activity generally reported for MCFA compared with LCFA (Zahera et al., 2024). MCFA disrupt cell membranes through increased permeability and leakage of intracellular components, particularly in gram-positive and anaerobic rumen microbes(Vadroňová et al., 2023); (Obukhova & Murzina, 2024). The stronger protozoal suppression under CFAD aligns with earlier findings showing that MCFA from coconut oil rapidly depresses Entodinium-dominated protozoal communities (Burdick et al., 2022); (Zahera et al., 2024).

Although individual taxa were not quantified in this study, previous research consistently reports that key fibrolytic bacteria such as Ruminococcus albus, Ruminococcus flavefaciens, and Fibrobacter succinogenes are highly sensitive to free fatty acids due to their strictly anaerobic metabolism and membrane composition (Carreño et al., 2019); (Zhang et al., 2020); (Lock et al., 2025). This supports the observed decline in ADF digestibility at higher supplementation levels, as fibrolytic pathways depend heavily on these taxa (Riestanti et al., 2021).

However, NDF digestibility and acetate concentrations remained stable across treatments, indicating that the core fibrolytic capacity of the rumen was largely preserved. (Lock et al., 2025) also reported improved NDF digestibility with fatty acid supplementation, supporting the present findings. Although B. fibrisolvens was not quantified in this study, this species is well known to contribute to hemicellulose and pectin degradation (Palevich et al., 2019); (Sengupta et al., 2022). While B. fibrisolvens is not a major cellulolytic bacterium, its presence has been shown to enhance overall cellulose digestion(Palevich et al., 2019); (Dewanckele et al., 2020), and it is generally less sensitive to moderate levels of fat supplementation compared with strictly cellulolytic species. This interpretation is further supported by (Zahera et al., 2024), who observed no significant changes in B. fibrisolvens abundance with CFAD supplementation up to 3%. Taken together, these results suggest that although certain microbial groups may have been inhibited, the functional components responsible for structural carbohydrate degradation were likely maintained, thereby sustaining NDF digestibility. A similar pattern of selective suppression, where ADF digestibility declines but NDF digestibility remains relatively stable, has also been observed with MCFA supplementation. This likely occurs because MCFA inhibit certain cellulolytic microbes, while others are able to persist and compensate through alternative enzymatic pathways (Vadroňová et al., 2023); (Obukhova & Murzina, 2024).

The interaction patterns observed in dry matter digestibility (DMD), organic matter digestibility (OMD), and microbial protein synthesis (MPS) indicate that CFAD induced stronger but more variable microbial metabolic shifts(Luan et al., 2023)(Lock et al., 2025). This variability mirrors known MCFA effects, where microbial inhibition is both concentration-dependent and species-specific (Patra, 2013); (Luan et al., 2023). In contrast to CFAD, PFAD produced smoother digestibility and MPS responses, suggesting that the LCFA-rich profile of PFAD imposed less microbial stress than the MCFA-rich profile of CFAD (Riestanti et al., 2021); (Sun et al., 2022). Collectively, these findings demonstrate that CFAD-driven MCFA imposes greater microbial pressure, whereas PFAD leads to moderated changes more compatible with stable rumen function.

No significant source × supplementation level interactions were detected for acetate, propionate, C2:C3 ratio, or methane production, indicating that both CFAD and PFAD produced similar response patterns across supplementation levels. The decrease in propionate, accompanied by an increase in the C2:C3 ratio at certain supplementation levels, suggests a shift in rumen fermentation pathways irrespective of fatty acid source. This response is consistent with the sensitivity of amylolytic bacteria, such as Streptococcus bovis and Selenomonas ruminantium, to dietary fatty acids, as reported in both in vitro and in vivo studies (Patra, 2013); (Vadroňová et al., 2023). In addition, the reduction in branched-chain VFA suggests suppression of amino acid-fermenting bacteria, including Peptostreptococcus and Clostridium species, which contribute to the formation of isobutyrate and isovalerate(Patra, 2013); (Mitchell et al., 2023) ;(Hackmann, 2024).

Interestingly, methane production increased at inclusion levels of 1% and 4% despite a decline in protozoal populations. This pattern likely reflects the concurrent reduction in propionate production, which diminished one of the primary hydrogen sinks and redirected more reducing equivalents toward methanogenesis (Moss et al., 2000); (Ungerfeld, 2020). Under these conditions, surviving hydrogenotrophic methanogens, particularly Methano-brevibacter-related groups, may have remained sufficiently active to utilize the increased availability of H₂. These taxa are known for their adaptability and ability to maintain methanogenic activity even when protozoal populations are suppressed (Danielsson et al., 2017); (Ungerfeld, 2020); (Khairunisa et al., 2023).

The significant interactions detected for C18:3 n-3 biohydrogenation and the unsaturation index indicate differential responses of CFAD and PFAD during ruminal fatty acid transformation. Most terminal biohydrogenation steps, including the disappearance of C18:2 cis and C18:3 n-6, were not affected, indicating that key terminal hydrogenating groups, such as Butyrivibrio proteoclasticus, remained functional at the tested fat levels (Jeyanathan et al., 2016); (Toral et al., 2024). The interaction observed for biohydrogenation of C18:3 n-3 likely reflects differences in the sensitivity of early hydrogenating species, particularly B. fibrisolvens, to the distinct fatty acid profiles of CFAD and PFAD. Although the fatty acid sources were supplied as calcium soaps, rumen protection is not intended to completely prevent biohydrogenation. Rather, calcium soap protection may reduce the immediate availability of unsaturated fatty acids and thereby limit the extent of biohydrogenation, allowing partial transformation and accumulation of intermediates such as CLA and vaccenic acid(Dewanckele et al., 2020); (Toral et al., 2024).

Total PUFA biohydrogenation declined at higher supplementation, consistent with established findings that fatty acids becoming available following partial dissociation of calcium soaps impair microbial attachment and inhibit biohydrogenation enzyme systems (Palmquist & Jenkins, 2017);(Toral et al., 2024). This decline suggests that excessive fat supplementation may suppress microbial populations involved in biohydrogenation, thereby reducing the formation of CLA-related intermediates at higher inclusion levels. Compared with CFAD, PFAD generated greater incomplete biohydrogenation and higher CLA-related intermediates, suggesting that the two fatty acid sources differed in their effects on ruminal lipid metabolism (Ventto et al., 2017); (Toral et al., 2024). In contrast to PFAD, the stronger microbial suppression associated with CFAD may have reduced the activity of early CLA-forming bacteria such as B. fibrisolvens, leading to lower CLA yields (Hussain et al., 2016); (Dewanckele et al., 2020); (Vadroňová et al., 2023).

The significant interaction observed for H₂ recovery indicates that the effects of supplementation level on hydrogen utilization differed between CFAD and PFAD. Hydrogen metabolism indices revealed distinct partitioning patterns. Although total H₂ production remained stable, supplementation markedly shifted the pathways through which H₂ was utilized. The calculated hydrogen partitioning in this study aligns with established stoichiometric estimates in the literature, where methanogenesis typically accounts for 40%–50% of reducing-equivalent use, propionate for 20%–30%, and biohydrogenation for roughly 20%–30% (Moss et al., 2000); (Ungerfeld, 2020). The increase in residual H₂ at higher inclusion levels is consistent with the reduced efficiency of hydrogen sinks described by(Ungerfeld, 2020).

Under CFAD, the strong suppression of protozoa and amylolytic bacteria, which are key contributors to H₂ production and propionate formation (Maia et al., 2010); (Choudhury et al., 2022); (Vadroňová et al., 2023), reduced both H₂ generation and the propionate sink, leading to H₂ accumulation. At the same time, CFAD inhibited CLA-forming bacteria and disrupted terminal biohydrogenation pathways (Dewanckele et al., 2020); (Vadroňová et al., 2023), limiting the use of H₂ for fatty acid hydrogenation and ultimately reducing total H₂ recovery. In contrast, PFAD exerted a weaker antimicrobial effect, allowing the propionate pathway to remain active, promoting incomplete biohydrogenation and the accumulation of CLA precursors (Sun et al., 2022); (Choudhury et al., 2022); (Toral et al., 2024), and increasing methane production as more H₂ became available. These combined effects resulted in greater total H₂ recovery under PFAD. Overall, the type of fat directs hydrogen flow either away from (CFAD) or toward (PFAD) hydrogen-utilizing pathways, thereby shaping biohydrogenation outcomes.

Polynomial regressions revealed clear dose-dependent responses for the main variables. CFAD exhibited nonlinear, quadratic-type patterns, consistent with the threshold-based microbial inhibition typically observed for MCFA-rich fats (Patra, 2013); (Vadroňová et al., 2023). In contrast, PFAD generated more linear responses, reflecting the milder and more proportional microbial effects of LCFA (Palmquist & Jenkins, 2017); (Toral et al., 2024). Most fermentation and fatty-acid traits reached optimal values at 2%–3% CFAD, a range previously associated with maximal VFA production, improved NH₃ utilization, protozoal suppression, and stable unsaturation indices (Shi et al., 2020); (Luan et al., 2023). PFAD showed optimal responses at 1.5%–2%, particularly for CLA yield, incomplete biohydrogenation, and fermentation stability (Toral et al., 2024); (Lock et al., 2025). Methane increased linearly with PFAD inclusion, in agreement with prior studies demonstrating proportional methane increases with rising saturated-fat supplementation (Patra, 2013); (Tan et al., 2024).

The combined fermentation, biohydrogenation, and hydrogen-flux responses demonstrate that CFAD and PFAD modulate rumen metabolism through fundamentally different biochemical mechanisms. CFAD, enriched in medium-chain fatty acids, exerted strong suppression of protozoa and several bacterial groups, reducing NH₃ turnover and enhancing fermentation efficiency at moderate inclusion levels(Maia et al., 2010); (Vadroňová et al., 2023). However, this same antimicrobial pressure inhibited CLA-forming species, reduced trans-intermediate formation, redirected hydrogen away from biohydrogenation, and decreased overall H₂ recovery(Ventto et al., 2017); (Ungerfeld, 2020); (Vadroňová et al., 2023). These patterns support supplementing CFAD at 2.5%–3% DM, a range shown to optimize VFA production, ammonia utilization, and fermentation efficiency while avoiding excessive inhibition of fibrolytic pathways (Luan et al., 2023); (Vesga et al., 2024).

In contrast, PFAD, which is characterized by a predominance of long-chain fatty acids, exerted a milder inhibitory effect on rumen microbes, allowing fibrolytic activity to be maintained and promoting greater incomplete biohydrogenation with increased accumulation of CLA precursors (Palmquist & Jenkins, 2017); (Toral et al., 2024). Hydrogen partitioning under PFAD favored both methanogenesis and biohydrogenation, increasing H₂ recovery but also elevating methane at higher inclusion levels (Patra, 2013); (Ungerfeld, 2020); (Kjeldsen et al., 2024). These responses indicate that PFAD is optimally supplemented at 1.5%–2% DM, where fermentation remains stable and CLA-related intermediates are enhanced without excessive methane formation.

The dose–response patterns highlight the importance of aligning fat type with the intended nutritional outcomes. CFAD is more suitable for improving nitrogen efficiency and overall fermentation performance, whereas PFAD is better suited for enhancing CLA precursor formation and maintaining lipid transformation pathways. This approach is particularly valuable for precision lipid supplementation strategies in tropical dairy systems.