Effects of cashew nutshell liquid on milk production and methane emission of dairy cows in a farm condition

2.1 Animals, diets, and experimental design

This experiment was carried out according to the guidelines specified by the Animal Care and Use Committee of Hiroshima University. The committee approved the procedures in this study (no. E21-1). The experiment was carried out in a free-stall barn equipped with door feeders for individual cows (Roughage Intake Control (RIC) system, HokoFarm Group, Emmeloord, The Netherlands) and AMS (VMS V300, DeLaval, Tumba, Sweden). Ten lactating Holstein dairy cows (body weight = 660 ± 51.0 kg, parity = 1.9 ± 0.56, days in milk = 111 ± 58.9 days) were divided into the control group (n = 5) and CNSL group (n = 5) based on their parity and milk yield immediately before the start of the experiment. The experiment lasted for 28 days, consisting of a 7-day preliminary period (covariate period) and a subsequent 21-day treatment period. During the treatment period, cows in the CNSL and control groups were supplemented with or without CNSL additives, respectively.

Cows were provided with PMR composed of 26% Italian ryegrass silage, 11% oats hay, 14% alfalfa hay, and 46% formula feed on a dry matter (DM) basis (Table 1). The PMR was delivered three times daily (1000, 1200, and 1500 h) into a feed-weigh trough with a slide gate permitting access for a particular cow assigned to each trough. The total amount of PMR was provided to fulfill the TDN requirement of each cow based on the Japanese feeding standard for dairy cows (NARO, 2017). The percentages of PMR delivered at each time point were 20%, 35%, and 45% for 1000, 1200, and 1500 h, respectively. For CNSL supplementation, a commercial CNSL additive (Ruminap, mash type, SDS Biotech K.K., Japan) was top-dressed on the first delivered PMR at 1000 h to provide the equivalent of 10 g/day of pure CNSL. Additionally, cows are also provided with a commercial concentrate diet (2.0–7.5 kg/day) at each milking in the AMS according to the milk yield of each cow. Water and mineral blocks are freely accessed by the animal throughout the experiment. Cows were milked at the AMS throughout the experiment, with free access to the AMS except during cleaning times. The milking time, duration, and milk yield for each session were automatically recorded for each cow.

2.2 Sampling and measurement

During the preliminary and treatment periods, the amount of daily feed offered and refusals were recorded by the Roughage Intake Control (RIC) system. The samples of the offered PMR were collected every morning and dried at 70°C in an air-forced oven for 72 h to measure DM content. The dried samples were pooled every week and ground in a Wiley mill through a 1-mm sieve for further analysis. Daily PMR intake and visiting time to each feed trough were obtained by summing recorded data of the Roughage Intake Control (RIC)honest significant difference system for the amount of feed consumed and time duration in all the visits made per cow per day. The DMI of PMR was calculated based on the recorded feed intake and DM content of the PMR offered each day. Furthermore, diurnal variation of PMR intake and duration at the allotted door feeder was calculated by summarizing the data across specific 4-h intervals throughout the day: from 10:00 to 14:00 h (encompassing the first to the second delivery of PMR), from 14:00 to 18:00 h (corresponding the time frame after the second delivery to the third delivery of PMR), 18:00 to 22:00 h, 22:00 to 02:00 h, 02:00 to 06:00 h, and 06:00 to 10:00 h. The visiting time duration from entry until leaving the door feeder was used as a proximate of the eating time of PMR. The daily intake of the concentrate diet at the AMS was assumed to be the same amount as the record provided by the AMS. A sample of the concentrate diet was collected once during the experiment, and the DM content was measured similarly to PMR. Daily DMI of PMR (total of each time period), concentrate, and total diet (PMR plus concentrate) were averaged over 7 days during the preliminary period and the third week of the treatment period, respectively. Rumination behavior was monitored by a motion sensor (U-motion, Desamis Co., Ltd., Tokyo, Japan) attached to the neck collar of experimental cows, and the 7-day average of recorded daily rumination time by the sensor system was regarded as time spent rumination.

2.3 Rumen fermentation and in vitro batch culture

On the 4th day of the third week of the treatment period, rumen fluid was collected using a stomach tube 3 h after the first delivery of PMR (13:00 h). Immediately after collection, oxidation and reduction potential (ORP) was measured using a portable meter with a metal electrode (D-210P/9300-10D; Horiba Advanced Techno Co., Ltd., Kyoto, Japan). Ruminal pH was measured using a pH meter equipped with a glass electrode (F-72/9680S; Horiba Advanced Techno Co., Ltd.). Within 30 min after collection, the rumen fluid was filtered through a four-layer gauze, and the filtrate was stored at −30°C for volatile fatty acid (VFA) analysis. Another 1 mL of rumen fluid was mixed with 4 mL of 10% formalin in a 0.9% sodium chloride solution with methyl green and stored in a dark tube for protozoa counting.

An in vitro experiment was also carried out to indirectly assess the ruminal microbial response to the effect of CNSL. Within 30 min after filtration, as described above, 10 mL of filtrated rumen fluid collected from each cow and 20 mL of pre-warmed McDougall's buffer solution were transferred into duplicated 50 mL serum bottles containing a 0.30 g dried sample of the same PMR offered to cows. The empty bottles without the PMR sample were treated in the same manner for the blank incubation of each inoculum. The bottles (a total of 22 bottles) were then flushed with CO₂ gas for 10 s to remove air, sealed with butyl rubber stoppers and aluminum crimps, and placed in a water bath at 39°C with a shaking speed of 60 shakes per minute for 24 h. After the incubation for 24 h, the total volume of gas produced during incubation was measured by puncture using a 100-mL glass syringe with a 21G needle, and a portion of the gas was transferred into evacuated tubes for short-term storage until CH₄ analysis.

2.4 Analytical procedure

Dried and ground samples of PMR and concentrate were analyzed for DM, crude protein (CP), ether extract (EE), crude ash, and neutral detergent fiber (aNDF_OM) (AOAC, 1999; Van Soest et al., 1991). Starch contents in the diets were determined using a total starch analysis kit (Megazyme, Bray Business Park, Bray, Co. Wicklow, Ireland). The concentration of VFA in rumen fluid before and after incubation was analyzed by gas chromatography with a flame-ionized detector (GC2014; Shimadzu, Kyoto, Japan) with a BP-21 column (25 m length × 0.53 mm diameter, 0.5 μm in thickness, Trajan Scientific and Medical, Ringwood, Australia). Protozoa were fixed with a methyl-green formalin solution and manually counted under a light microscope (Shinkai et al., 2012). Methane concentration in gas produced in vitro was analyzed by gas chromatography with a flame ion detector (GC-7A; Shimadzu) and a stainless column (2 m × 2 mm) packed with molecular sieve 5A.

2.5 Statistical analysis

Data were statistically analyzed using fit model analysis of JMP software (version 16.2.0; SAS Institute Inc.). For DMI, milk yield, CH₄ production parameters, and eating and rumination time during the third week of the treatment period, the treatment effects were analyzed using a covariate analysis, with the data of each parameter at the preliminary period as the covariate. Cow within the group was regarded as a random effect, and treatment was a fixed effect. The data of these parameters were presented as covariate-adjusted least square means. For milk components, ECM, CH₄ emission per ECM, and rumen fermentation parameters during the third week of the treatment period, the treatment effects were analyzed using a one-way analysis of variance, as these parameters were only obtained during the third week of the treatment period. Diurnal variation data tabulated every 4 h during the third week of the treatment period were analyzed as repeated measurements to test the effects of CNSL supplementation, time period, and their interaction. When the significant time effect was detected, the differences among time periods were analyzed by Tukey's honest significant difference (HSD) test. The data were presented as least square means. Statistically significant differences were declared at p < 0.05, whereas treatments with 0.05 ≤ p ≤ 0.15 were considered as a trend toward significance.

3.1 Milk production and dry matter intake

The supplementation of CNSL additives to the PMR did not yield any noticeable impact on daily milk production, total DMI, eating time of PMR, and rumination time (Table 2). Although there was a slight upward trend (p = 0.13) in concentrate intake from AMS for the CNSL group compared with the control group, this did not exert a significant influence on the overall feed intake. Furthermore, the supplementation of CNSL did not yield any discernable impact on ECM yield, milk compositions, and milk fatty acid profile, as detailed in Table 3.

The supplementation of CNSL had no noticeable impact on PMR intake at each time period, as detailed in Table 4. Among time periods within the day, the PMR intake was higher (p < 0.05) between 10:00 and 14:00 h and between 14:00 and 18:00 h for cows in both groups, while it was lowest from 06:00 to 10:00 h, the period preceding the next day's feeding. The supplementation of CNSL did not influence the time spent eating either at each time period. Within the day, the eating time for both groups was greater (p < 0.05) between 10:00 and 14:00 h and between 14:00 and 18:00 h, gradually declining from 22:00 h until the subsequent day's feeding schedule.

3.2 Methane emission parameters

Among the parameters concerning CH₄ emissions presented in Table 5, the CH₄/CO₂ ratio in the respiration gas of cows tended to be lower (p = 0.145) in the CNSL group compared with the control group. Furthermore, CH₄ yield (CH₄/DMI, L/kg DMI) and MCF (MJ/100 MJ) exhibited significant reductions (p < 0.01) in the CNSL group compared with the control group; however, daily CH₄ production was not affected by CNSL addition.

To assess potential biases in the distribution of gas measurements within the day and diurnal variation in the CH₄/CO₂ ratios, the group averages of the effective observation numbers and the CH₄/CO₂ ratios of individual cows at each time period in the third week of the treatment period were calculated and are presented in Table 6. No noticeable impact attributed to the supplementation of CNSL was observed in the number of effective observations at each time period, even though there was a trend (p = 0.08) of interaction between treatment and time period. For both treatments, the mean number of effective observations exhibited a trend with higher counts during the hours of 10:00–14:00 h and 18:00–22:00 h, while the lowest number of observations was recorded between 06:00 and 10:00 h (time effect: p = 0.10). For within-day variation in CH₄/CO₂ ratios, the highest CH₄/CO₂ ratio (p < 0.05) was recorded between 10:00 and 14:00 h. In contrast, the lower ratios were observed between 06:00 and 10:00 h and between 22:00 and 02:00 h.

3.3 Rumen fermentation

The supplementation of CNSL had no discernible impact on the pH, ORP, and protozoal population in the rumen of cows during the third week of the treatment period, as outlined in Table 7. Ruminal total VFA concentration tended to be lower (p = 0.11) for cows in the CNSL group compared with the control group. Furthermore, the molar proportion of acetic acid tended to be lower (p = 0.12) for cows in the CNSL group. Conversely, the proportion of n-valeric acid was lower (p < 0.05) for cows in the control group compared with those in the CNSL group, while the molar proportion of other VFA remained unchanged.

Table 8 shows the results of the in vitro incubation of the PMR samples using rumen fluid obtained from cows supplemented with or without CNSL. Total gas (p = 0.05), methane production (p = 0.07), and VFA concentration (p < 0.05) tended to be lower for the incubation with the rumen fluid from cows in the CNSL group compared with those in the control group. Among the VFA proportions, the molar proportion of each VFA remained unchanged from both the control and CNSL groups, except for iso-butyric acid.

4.1 Effect of cashew nutshell liquid on methane emission

In this study, the number of experimental animals was small relative to the size needed to obtain sufficient statistical power to detect the significant treatment effects (approximately 20 cows based on our study conditions, unpublished data). However, we planned a covariate period, and there were no differences in the initial body weight and days in milk of cows between the control and CNSL groups. Therefore, the small sample size is not expected to significantly impact the overall results of this study.

The potential of CNSL to mitigate CH₄ emissions from dairy cows in vivo has been studied previously (Branco et al., 2015; Coutinho et al., 2014; Shinkai et al., 2012). For instance, Shinkai et al. (2012) assessed the efficacy of CNSL supplementation in mitigating CH₄ emissions on non-lactating dry cows subjected to restricted feeding of hay and concentrate diets. They observed a 19% reduction in CH₄ yield by incorporating a pellet containing 22% pure CNSL at a rate of 96–104 g/day, equivalent to 4 g of pure CNSL for every 100 kg of body weight. In contrast, our present study utilized a lower supplementation level of 10 g of pure CNSL per day, representing 0.035% of DMI or approximately 1.5 g of pure CNSL for every 100 kg of body weight. This level of supplementation was based on the previous in vitro report by Watanabe et al. (2010), which demonstrated that even a low CNSL concentration of 50 μg/mL in rumen fluid, equivalent to approximately 0.038% of pure CNSL per kg of dietary DM, could achieve a methane reduction of up to 36%. Despite achieving a reduction in CH₄ yield of 4.3% in our study, the efficacy of CNSL supplementation was confirmed even at a lower dosage under practical feeding conditions. The observed effects could be due to the presence of bioactive compounds such as anacardic acid in the CNSL additives used in this study (Compton et al., 2023), influencing microbial fermentation in the cows' rumen, thereby contributing to the observed decrease in CH₄ yield and MCF. It is essential to note the significant difference in DMI between our study (28 kg/day) and the study by Shinkai et al. (2012) (7.1 kg/day). As a result, our CNSL supplementation level of pure CNSL on a DM basis was approximately one tenth of theirs (0.035% vs. 0.30%).

In this study, the total VFA concentration and molar proportion of acetate in the rumen tended to be lower for cows in the CNSL group. In a previous study by Shinkai et al. (2012), CNSL supplementation in dry cows also led to a decrease in total VFA concentration and molar proportion of acetate, alongside an increase in propionate proportion in the rumen. Additionally, cows supplemented with CNSL showed a decrease in the relative abundance of formate-producing species such as Ruminococcus flavefaciens, Butyrivibrio fibri-solvens, and Treponema bryantii. Conversely, there was an increase in propionate-producing species such as Prevotella ruminicolla, Selenomonas ruminantium, Anaerovibrio lipolytica, and Succinivibrio dextrinosolvens (Shinkai et al., 2012). This reduction in the relative abundance of formate and hydrogen producers in the rumen would result in a shortage of available substrates for methanogenesis. Mitsumori et al. (2014) further supported these findings, indicating that CNSL supplementation effectively decreased the supply of metabolic hydrogen in the rumen. Consequently, the reduction in hydrogen availability with the addition of CNSL might lead to a decrease in CH₄ and acetate production in this study. Generally, the fermentation pattern in the rumen plays a critical role in controlling hydrogen production, which is subsequently converted into CH₄. Higher levels of acetate and, to some extent, butyrate production are associated with increased CH₄ generation. Conversely, a higher proportion of propionate is linked to reduced hydrogen release and lower CH₄ production. Furthermore, Danielsson et al. (2014) observed CH₄ inhibition and a shift in both bacterial and archaeal community structures from the addition of cashew nutshell extract using an in vitro system. The authors suggested that the reduced CH₄ levels were due to a direct effect of the product on methanogenic species. In our in vitro study, we also observed a trend toward lower CH₄ production and a significant decrease in VFA production during incubation with the rumen fluid obtained from cows in the CNSL group. This finding suggests an alteration in the specific activity of rumen bacteria or the bacterial profile influenced by CNSL supplementation, although specific alterations in microbial profiles were not determined in this study.

4.2 Dry matter intake and eating behavior

In this study, a subtle upward trend in concentrate intake from the AMS was observed from the CNSL groups due to the automatic adjustment of the amount of concentrate supply at milking. However, total DMI was not affected by CNSL supplementation, which was consistent with the results of previous studies where CNSL was supplemented at the range from 0.012% (Coutinho et al., 2014) to 0.30% of DMI (Shinkai et al., 2012).

The feeding behavior of dairy cows significantly influences feed efficiency, a critical factor for enhancing livestock profitability while concurrently mitigating environmental impact (Llonch et al., 2018). The temporal analysis revealed that the feeding pattern of PMR did not differ between groups of cows, regardless of whether they were supplemented with CNSL. Both groups showed the highest intake levels of PMR between 10:00 and 14:00 h, coinciding with the timing of the first feeding session. This heightened intake during the specified period suggests that the cows were actively engaged in eating at the time of the first feeding at 1000 h. This observation aligns with previous studies indicating that the daily feeding patterns of group-housed dairy cows are largely influenced by the timing of fresh feed delivery (DeVries & von Keyserlingk, 2005). The second and third deliveries of PMR also elicited heightened eating behavior between 14:00 and 18:00 h, suggesting that cows may exhibit increased eating activity in response to feed delivery. (DeVries et al., 2005). Although the eating pattern of cows under AMS could be affected by a combination of factors, including the cows' natural feeding instincts and circadian rhythms (Fuchs et al., 2022), the adjusted PMR feeding regimen implemented in this study to fulfill the TDN requirements of cows, coupled with the limited frequency of PMR delivery, may have constrained the voluntary feeding behavior of cows during the nighttime.

These feeding patterns affected the diurnal variation of CH₄/CO₂ ratios, even though supplementation of CNSL revealed no discernible impact on these fluctuations. For both groups, the most significant increase in CH₄/CO₂ ratios within the day occurred between pre-feeding (06:00 to 10:00 h) and post-first-feeding (10:00 to 14:00 h), with the higher ratio persisting until 18:00 to 22:00 h. These diurnal changes in CH₄/CO₂ ratios may be attributed to the elevated feed intake and activated rumen fermentation during these time frames. The previous study, which assessed the CH₄/CO₂ ratios of dairy cows by a sniffer method, indicated that the heightened ratios occurred immediately after eating PMR as the result of activated fermentation in the rumen (Suzuki et al., 2021). Furthermore, Brask et al. (2015) observed a proportional increase in the CH₄/CO₂ ratio after morning feeding and throughout the day, possibly reflecting higher feed intake during these periods. The diurnal pattern of CH₄ production is influenced by the timing of feed allocation, typically exhibiting increased emissions following feed distribution (Bell et al., 2018; Brask et al., 2015; Judy et al., 2018).

The feeding behavior of ruminants is typically followed by rumination, a process where feed undergoes microbial fermentation to render nutrients accessible (Llonch et al., 2018). Previous research has explored rumination time as a potential indicator for monitoring CH₄ production (Mikuła et al., 2022). It is widely recognized that the chemical and physical properties of the diet are key factors influencing rumination time. Moreover, an increase in neutral detergent fiber (NDF) from forages has been observed to enhance rumination activity, stimulate saliva production, and, by buffering rumen fluid, elevate acetate production in the rumen, consequently leading to higher CH₄ production (Byskov et al., 2015). However, in this study, rumination time was not different between cows in the control and CNSL groups, probably due to the comparable DMI between the two groups.

4.3 Application of sniffer method

This study used the sniffer method in conjunction with the AMS for estimating CH₄ production under a free-stall barn. This method allowed us to monitor the CH₄ production of cows in a practical barn with unrestricted feeding conditions using a simple measurement system with an accurate gas analyzer. This method is primarily utilized for breeding purposes to assess CH₄-related traits in a large number of cows and to derive phenotypic parameters for the selective breeding of low-CH₄-emitting cows (Uemoto et al., 2024). While this method has faced criticism regarding its accuracy, few studies have employed it to evaluate the dietary effects on CH₄ production (Haque et al., 2014) or the impact of CH₄ inhibitory additives (Bach et al., 2023). The former research group used a similar approach to this study in measuring the CH₄/CO₂ ratio and estimating the CH₄-related parameters using theoretical equations based on the oxygen and CO₂ balance of energy metabolism. On the other hand, the latter research adopted the methodology introduced by Garnsworthy et al. (2012), which estimates CH₄ emission based on the measured CH₄ concentration in spot gas samples and a dilution factor of the respiration gas.

In this study, we calculated daily CH₄ production and MCF by utilizing the CH₄/CO₂ ratio and multiple regression equations derived from CH₄ production data obtained through chamber and head box methods (Suzuki et al., 2021). The validity of this approach was confirmed through comparison with total gas collection data from dairy cows using the head box method, as outlined in a previous study (Suzuki et al., 2021). The accuracy of the estimated CH₄ production relies on the ample number of observations to trace both within-day and between-day variations in CH₄ production. In this study, the CH₄/CO₂ ratio was notably higher between 10:00 and 22:00 h, with lower levels during the early morning hours, as discussed above. Further, in this study, there was a slight disparity in the distribution of observation numbers at each 4-h interval within the day between the control and CNSL groups, which might potentially cause bias when calculating the average CH₄/CO₂ ratio. The observed slight variation in an effective number of observations of the CH₄/CO₂ ratio in this study could be due to the differences in the frequency of voluntary visits by individual cows in the AMS, influenced by motivational factors such as access to concentrate and emptying of the udder, as explained by previous researchers (Broucek & Tongel, 2017; Prescott et al., 1998).

4.4 Effect of cashew nutshell liquid on milk production

The results regarding milk production performance in response to CNSL supplementation indicate no significant effects on milk yield and the contents of key constituents in milk, including fat and protein, as well as milk urea nitrogen and fatty acid profile. Coutinho et al. (2014) found that supplementation of technical-grade CNSL up to 0.036% of the total mixed ration also did not affect daily milk yield or major milk component contents in lactating cows. Similarly, in the study by Branco et al. (2015), technical-grade CNSL supplementation at 30 g per day also did not affect milk yield or major milk component contents, despite the CNSL-modified milk fatty acid profiles (C6:0, C13:1, C16:1, and C18:0) observed in these previous studies.

In conclusion, our study suggests that the inclusion of CNSL at a low level in the diet of dairy cows has no significant impact on milk production, DMI, and eating time. Furthermore, there is a slight but significant reduction in CH₄ yield with CNSL supplementation, indicating a potential environmental benefit without compromising productivity or altering cow behavior. Additionally, our findings suggest the viability of using the sniffer method with AMS in free-stall barn conditions, offering a practical approach for monitoring the inhibitory effects of CH₄-mitigating additives.