Effect of monensin on milk production efficiency and milk composition in lactating dairy cows fed modern diets

Since the US Food and Drug Administration’s approval of monensin in 2004, significant nutritional advances have been made to increase feed efficiency and milk fat production. Recent evidence suggests monensin’s adverse effect on milk fat percentage may be absent when diets are formulated to address known diet-induced milk fat depression risk factors. Thus, study objectives were to evaluate effects of monensin level on dry matter intake (DMI), milk production and composition, and efficiency of high-producing cows fed diets formulated to optimize milk fat. Ninety-six lactating Holstein cows (36 primiparous, 60 multiparous; 106 ± 17 d in milk [DIM]) were balanced by parity, DIM, and milk production and were randomly assigned to 1 of 12 pens with 8 cows per pen. All cows received 11 g/t monensin for 5 wk after which pens received 1 of 4 dietary treatments (n = 3) formulated to provide 0 ( CON ), 11 (R11), 14.5 (R14.5), or 18 (R18) g/t monensin for 9 wk. The basal diet was 54% forage, 27% NDF, 29% starch, and 2.3% rumen unsaturated fatty acid load. Pen was the experimental unit and data were analyzed using the Fit Model Procedure of JMP. Effects of treatment, time, and treatment × time interaction were included as fixed effects and pen as a random effect. Least squares means were determined and linear and quadratic contrasts were tested. Dry matter intake tended to decrease linearly with increasing monensin dose. Milk yield, fat percentage, and protein percentage and yield were unaffected by treatment while fat yield was quadratically increased. Milk de novo and mixed fatty acid (FA) yields (g/d) increased quadratically with monensin whereas preformed FA linearly decreased during the experimental period. Energy-corrected milk (ECM) was quadratically increased by monensin. Milk urea nitrogen concentrations increased linearly with increasing monensin dose. Monensin linearly increased feed efficiency (ECM/DMI, 3.5% fat-corrected milk/ DMI, and solids-corrected milk/DMI). Body weight gain did not differ between treatments. Estimated dietary energy tended to increase linearly with increasing monensin level. These data suggest monensin improves component-corrected milk production efficiency, estimated dietary energy, and does not negatively affect milk fat percentage or FA profile.


INTRODUCTION
Monensin is a carboxylic polyether ionophore which has been used extensively in the dairy industry since its 2004 US Food and Drug Administration (FDA) approval.Mechanistically, monensin functions by selectively inhibiting gram-positive bacteria (via interference of ion transport) and enhancing gram-negative metabolism (McGuffey et al., 2001;Vasquez et al., 2021) resulting in increased propionate production and a corresponding decrease in methane (Russell and Strobel, 1989).Increased propionate formation and subsequent hepatic supply enhances glucose output thereby improving productivity (McCarthy et al., 2015a,b;Markantonatos and Varga, 2017).As stated in the FDA claim, monensin linearly increases milk production efficiency (MPE; defined as marketable solids-corrected milk (SCM) per Mcal NE L intake corrected for changes in BW) from 2% to 4% when fed between 11 to 22 g/t (FDA, 2004).Postapproval research has consistently demonstrated improved MPE, though the driver of the change (i.e., increased milk, decreased DMI, increased BW or a combination of the 3) is known to vary across and even within stage of lactation (Duffield et al., 2008a;Akins et al., 2014;Vasquez et al., 2021).Although beneficial responses to monensin are apparent, adverse effects on milk fat percentage have been documented (Phipps et al., 2000;Benchaar et al., 2006;AlZahal et al., 2008) and this response has been a barrier to supplying the most effective dose for feed efficiency.
Diet-induced milk fat depression (MFD) occurs because of interactions between passage rate, dietary components, rumen microbial activity, and the mammary gland (Rico and Harvatine, 2013;Leskinen et al., Effect of monensin on milk production efficiency and milk composition in lactating dairy cows fed modern diets E. A. Horst, 1 S. K. Kvidera, 1 S. Hagerty, 2 P. D. French, 2 D. B. Carlson, 1 K. Dhuyvetter, 1 and A. W. Holloway 1 * 2019).Interference with the biohydrogenation (BH) pathway results in the formation and release of bioactive fatty acids (FA) which directly affect milk fat synthesis (Baumgard et al., 2000;Dewanckele et al., 2020).Monensin has been shown to inhibit BH capacity in vitro (Fellner et al., 1997) and in vivo reports indicate that milk fat depressing effects of monensin are intensified when combined with high rumen UFA load (Alzahal et al., 2008;He et al., 2012) or increased dietary starch (Van Amburgh et al., 2008).When evaluated across multiple studies, monensin decreased milk fat content 0.13%, but the model displayed moderate heterogeneity and publication bias where smaller studies reporting milk fat reductions were more likely to be published than smaller studies reporting increases (Duffield et al., 2008a).In addition, early research utilizing milk fat depression models including monensin to better understand MFD were included in the analysis and likely contributed to this bias.Although milk fat content decreased, no effect of monensin on milk fat yield (which classically defines MFD), was observed in the meta-analysis (Duffield et al., 2008a).
Through continued investigation it has become evident that MFD is a multietiological syndrome influenced by a myriad of factors (i.e., rumen-unsaturated FA load, feed particle size, NDF level, the rate/extent of starch degradation; McCarthy et al., 2018;Dewanckele et al., 2020) and no single dietary component can explain more than 11% of the variation in milk fat percentage (McCarthy et al., 2018).As a univariable metric, monensin dose (150-400 mg/cow per day) was not associated with bulk tank milk fat percentage in a recent epidemiological study (McCarthy et al., 2018) and this was supported by controlled research feeding diets formulated to avoid MFD (Akins et al., 2014;Hagen et al., 2015;Vasquez et al., 2021).Considering the advancements in understanding of MFD biology and diet formulation capabilities, an updated evaluation of monensin's dose dependent effect on feed efficiency and milk components is warranted.Thus, study objectives were to evaluate effects of monensin level on dry matter intake, milk production and composition, and efficiency of high-producing cows fed diets formulated to optimize milk fat yield.

Animals and Experimental Design
The experimental protocol was approved by the Animal Care and Use Committee of PHD R&D research facility (IACUC# 1256).
A total of 96 lactating Holstein cows (36 primiparous, 60 multiparous; 106 ± 17 DIM) were enrolled in a study conducted at PHD R&D contract research organization (Fort Atkinson, WI) from December 2019 to April 2020.Cows were stratified by parity, DIM, and pre-trial milk production and randomly assigned to 1 of 12 pens with 8 cows per pen (3 primiparous and 5 multiparous cows/pen).Due to logistical constraints (i.e., population of cows available) at the study site, primi-and multiparous cows were comingled to balance parity between treatments, this approach has also been implemented in other recent reports (Akins et al., 2014;Hagen et al., 2015).Cows were housed in a cross-ventilated barn with sand-bedded freestalls (8/pen) and were allowed 7 d to acclimate to housing conditions.Following acclimation, cows were enrolled in a 5-wk adaptation period during which all pens received diets formulated to 11 g/t monensin.Monensin supplementation to all pens in the adaptation period was implemented due to pre-existing monensin use at the study site and to be consistent with other recent reports (Akins et al., 2014;Hagen et al., 2015).Carryover effects of monensin may have attenuated the magnitude of treatment differences observed during the experimental period and this should be considered when interpreting results presented herein.During the 9-wk experimental period pens received 1 of 4 dietary treatments (n = 3) formulated to provide 0 (CON), 11 (R11), 14.5 (R14.5), or 18 (R18) g/t monensin.
Diets were formulated to meet or exceed nutrient requirements for lactating dairy cows producing 45 kg of milk (NRC, 2001; Table 1).In addition, key dietary factors including dietary fat level (target supplemental fat at 1.25% of DM with 75% C16:0 and 11% C18:1), diet fermentability (target ≤22% rumen fermentable starch and 42-45% total fermentable carbohydrate), and physically effective uNDF were accounted for to optimize milk fat synthesis.Pens were fed once daily at 0800 h for a target refusal rate of 2% to 5% and orts were recorded.Samples of forages, concentrate mixes, the TMR, and refusals were obtained weekly and composited at the end of the adaptation period and every 3 wk of the experimental period for nutrient analysis (Cumberland Valley Analytical Services, Waynesboro, PA; Table 2).Particle size of each composite corn silage, alfalfa silage, TMR, and refusal sample was determined using the Penn State Particle Separator (3-sieve+pan model).Treatments were administered via 4 type B concentrate mixes which were manufactured by local mills (Complete Feed Service and CP Feeds) and stored on-farm thereafter.Samples of the mixes were obtained at the time of manufacturing for determination of monensin concentration (Eurofins, Greenfield, IN).Analyzed monensin level in the baseline concentrate mix averaged 33.4 g/t which was equivalent to 15.7 g/t in the TMR.Actual monensin intake (using an average of Throughout the study cows were milked 3 times daily and yield was recorded.Daily averages per pen were determined.Samples for milk composition analysis (true protein, fat, lactose, and MUN) were obtained from each cow on 2 consecutive days (6 consecutive milkings) in the last week of the adaptation period and once per week (3 consecutive milkings) during the experimental period and submitted to AgSource Laboratories (Menomonie, WI).Additional samples obtained at the same time points were submitted to Stearns DHIA Laboratories (Sauk Centre, MN) for milk FA profile analysis (de novo, mixed, and preformed FA) by mid-infrared reflectance.Individual BW were measured following each milking (AfiWeigh, AfiMilk USA, Fitchburg, WI) and reported as a pen weekly average.Average BW in the last 3 d of the adaptation period served as initial BW.Individual BCS was measured by 2 trained individuals in the last week of the adaptation period and wk 3, 6, and 9 of the experimental period.

Calculations and Statistical Analysis
Sample size calculation was determined using JMP software (v14.2,SAS Institute Inc., Cary, NC).The sample size of 3 pens/treatment was determined using variance measures from previous pen-based studies conducted at the research facility with the statistical power to detect a 1-kg difference in DMI was 74%, at an α of 0.05.During the adaptation period, one cow assigned to the R11 treatment died and her data were removed.Outlier analysis was performed on individual cow data using the Robust Fit Outliers procedure (Huber M-Estimation, K = 3) option in JMP.Daily DMI was averaged by wk for each pen and milk yield and BW data were averaged by wk for each cow.Milk component yield was calculated by multiplying milk component concentration by average weekly milk yield as previously reported (Akins et al., 2014).All data were averaged for each pen before statistical analysis.Body weight and BCS change data were calculated by subtracting the adaptation value by the value obtained during wk 9 of the experimental period.Energy-corrected milk, 3.5% FCM, and SCM were determined according to the following equations (NRC, 2001): Data were analyzed as a completely randomized design with covariate adjustment where appropriate.Each specific variable's adaptation period value obtained in the final wk (BW and BCS) or final 2 wk (milk yield and composition, DMI) served as a covariate for analysis of the experimental period.Milk yield and composition, DMI, BW, and BCS data were analyzed using the Fit Model procedure and REML method of JMP.The model included fixed effects of treatment, time, and Body weight change, BCS change, MPE (all 4 metrics), and estimated diet energy content data were analyzed using the Fit Model Procedure of JMP with fixed effects of treatment.
Least squares means were determined and contrasts were constructed to test linear and quadratic effects including CON.To account for unequal spacing, the IML procedure of SAS (SAS Institute Inc., Cary, NC) with the ORPOL function was used to obtain the ap-propriate contrast coefficients based on dose levels.Overall P-values for treatment are not presented.Data were considered significant if P ≤ 0.05 and a tendency if 0.05 < P ≤ 0.10.

RESULTS
Over the experimental period, DMI tended to linearly decrease with increasing monensin dose (P = 0.09) and a tendency for a quadratic response was also observed (P = 0.10; Table 3).Milk yield, fat content, and protein content and yield were similar between treatment groups throughout the experimental period (P ≥ 0.36; Tables 3 and 4).Milk fat yield quadratically increased with monensin supplementation (P = 0.01; Table 4).Lactose percentage was similar between dietary treatments during the first 8-wk of the experimental period but increased at the R14.5 relative to R18 dose at wk 9 (4.70% vs. 4.64% in R14.5 vs. R18, respectively; P = 0.04; data not shown).No treatment differences were detected for lactose yield (Table 3).Monensin linearly increased MUN throughout the experimental period (P < 0.01; Table 3).Regardless of dose, monensin increased MUN concentrations relative to CON in the first 3 wk of the experimental period, effects were most consistent across time in the R11 treatment (P < 0.01; data not shown).
For milk fatty acid content (g/100 g FA), a quadratic increase in de novo FA and decrease in preformed FA (P ≤ 0.05; Table 4) was observed with monensin.Treatment differences in mixed and preformed FA content  were most evident from wk 5 to 9 of the experimental period (treatment × time interaction P ≤ 0.02; data not shown).Milk de novo and mixed FA yields (g/d) increased quadratically (P = 0.01) with monensin whereas preformed FA yield linearly decreased (P = 0.02; Table 4).
A quadratic increase in ECM, 3.5% FCM, and SCM with increasing monensin dose was detected during the experimental period (P ≤ 0.02; Table 3).Effects of monensin supplementation on SCM were most pronounced from wk 5 to 9 of the experimental period (P = 0.10; data not shown).Feed efficiency as determined by actual milk yield/DMI did not differ across treatments (Table 3).However, ECM/DMI, 3.5 FCM/ DMI, and SCM/DMI increased linearly with increasing monensin dose (P ≤ 0.03; Table 3).No treatment differences were detected for BW, BCS, or the change in BW and BCS.Estimated diet energy content tended to increase linearly with increasing monensin dose (P = 0.06; Table 3).

DISCUSSION
Since its approval, studies have continued to demonstrate the benefits of monensin to performance and farm profitability across the lactation cycle (Duffield et al., 2008a,b;McCarthy et al., 2015a,b;Vasquez et al., 2021;Richards et al., 2022).However, decreased milk fat content has been a common concern with feeding higher doses of monensin to capitalize on feed efficiency gains, particularly in late lactation (Duffield et al., 2008a).Although inconsistencies exist (Lean et al., 1994;Duffield et al., 1999;Akins et al., 2014), de-creased milk fat content (Phipps et al., 2000;Duffield et al., 2008a;Dubuc et al., 2009) and reduced rate of recovery from diet-induced MFD (Rico et al., 2014) have been observed with monensin.Additionally, altered milk FA profile including decreased de novo FA (g/100 g of FA) has been documented (Alzahal et al., 2008;Duffield et al., 2008a).Monensin's effect on milk fat is thought to be explained by its inhibitory action on bacteria involved in ruminal biohydrogenation and the corresponding increase in formation and absorption of bioactive FA (Fellner et al., 1997;Alzahal et al., 2008;He et al., 2012).However, research indicates that monensin does not independently depress milk fat content, but it occurs because of interactions with other dietary factors (i.e., increased UFA load, increased starch content/fermentability, and so on; McCarthy et al., 2018).This is supported by the finding that recovery from diet-induced MFD through correction of dietary starch and UFA concentration can be achieved in the presence of monensin (Rico et al., 2014(Rico et al., , 2015)).Additionally, recent evidence indicates that milk fat percentage can be maintained with monensin levels at least as high as 18 g/t (Akins et al., 2014;Hagen et al., 2015;Benoit et al., 2021).Therefore, we hypothesized that increasing monensin level would linearly improve feed efficiency while having no negative effect on milk fat when fed in contemporary diets accounting for known diet-induced MFD risk factors.As discussed above, monensin was supplemented to all pens during the adaptation period.Residual effects of monensin (discussed further below) may have attenuated the magnitude of responses observed and this should be considered when interpreting results.Herein, we observed similar milk fat percentage across treatments and milk fat and de novo FA yields quadratically increased.Our results, as well as those observed by others (Akins et al., 2014;Hagen et al., 2015;Benoit et al., 2021), challenge traditional concerns regarding monensin and decreased milk fat and provide evidence that monensin may positively affect de novo FA synthesis at certain doses.As hypothesized by Benoit et al. (2021), increased de novo FA synthesis with monensin may be explained by (1) propionate serving as a carbon source for formation of acyl chains used in FA elongation, (2) increased reducing equivalents (14NADPH + 14 H + ) for FA biosynthesis (Laliotis et al., 2010), or (3) increased intestinal flow of feed-derived AA which may be linked to de novo lipogenesis via mTOR signaling, as recently described (Li et al., 2016).Although de novo FA yield quadratically increased, preformed FA yield linearly decreased with monensin supplementation.Increased de novo lipogenesis may have resulted in reduced mammary uptake of preformed FA as recently proposed by Matamoros et al. (2022).Further investigation into monensin's effect on de novo FA synthesis when fed in modern diets is warranted.
Increased intestinal supply of diet-derived AA and peptides with monensin supplementation has been attributed to its inhibitory action on hyper-ammonia producing bacteria (Poos et al., 1979;Haïmoud et al., 1996;Ruiz et al., 2001).This inhibitory effect explains why decreased ruminal ammonia concentrations are frequently observed in monensin-fed ruminants (Plaizier et al., 2000;Recktenwald et al., 2014).Despite decreased rumen ammonia, blood and milk urea nitrogen frequently increase in response to monensin (Duffield et al., 2008a,b;Recktenwald et al., 2014;Mc-Carthy et al., 2015a,b) and this agrees with changes in MUN observed herein.This paradoxical relationship is presumably a reflection of increased intestinal AA uptake and subsequent hepatic deamination.It remains unclear how monensin's protein sparing capabilities may contribute to improved performance and more investigation is warranted.We observed no treatment differences in milk lactose and protein percentages or yields during the experimental period.Absence of a treatment difference in milk lactose content agrees with much of the existing literature (Hagen et al., 2015;Benoit et al., 2021;Rezaei Ahvanooei et al., 2023).However, decreased milk protein percentage has been reported repeatedly in monensin-fed cows (Duffield et al., 2008a;Akins et al., 2014;Hagen et al., 2015) and is often attributed to a dilution effect (Ipharraguerre and Clark, 2003).Similar milk protein percentage observed in our study may be explained by the absence of increased milk yield.Overall, our results demonstrate that increasing monensin level has no detrimental effect on milk composition and may benefit milk fat synthesis at certain doses.
In agreement with previous literature (FDA, 2004;Duffield et al., 2008a;Hagen et al., 2015), monensin supplementation induced hypophagia during the experimental period as a tendency for a linear response was detected.According to the hepatic oxidation theory (Allen et al., 2009), increased propionate supply and subsequent hepatic oxidation signals satiety, which likely explains our result, although effects on appetite are known to vary by stage of lactation (Akins et al., 2014;McCarthy et al., 2015b;Vasquez et al., 2020).Although DMI decreased in our study, milk yield did not differ between treatments which agrees with previous work (Alzahal et al., 2008;Hagen et al., 2015;Benoit et al., 2021).Yet, quadratic effects were observed for ECM, 3.5% FCM, and SCM.The combined effects of decreased DMI and increased component-corrected milk yields resulted in linear increases in feed efficiency expressed as ECM/DMI, FCM/DMI, and SCM/DMI.Linear increases in efficiency metrics are consistent with early dose titration work (FDA, 2004) and reports demonstrating increased component-corrected milk feed efficiency at an 18 g/t feeding rate (Akins et al., 2014;Hagen et al., 2015).
Although monensin's effect on feed efficiency (ECM/ DMI, FCM/DMI, and SCM/DMI) was linear, a numerical decrease was detected in the R18 treatment.Accurately assessing energy efficiency requires evaluation of multiple energy sinks including milk yield and composition as well as body tissue gain, and accordingly we observed a numerical increase in BW in R18 pens relative to control.Combined effects of monensin on milk yield and components, BW, and DMI resulted in a linear increase in estimated diet energy content which agrees with others (Akins et al., 2014;Hagen et al., 2015).Energy partitioning toward maintenance, body tissue gain, and lactation varies on an individual cow basis and must be considered when evaluating monensin's efficacy.Even though BW was heavier in the R18 treatment, no differences in BW change, BCS or BCS change were detected which suggests little risk of increasing adiposity with continued monensin supplementation in later lactation.
An important consideration when evaluating results presented herein, as well as the results reported by others (Akins et al., 2014;Benoit et al., 2021), is the period of adaptation following monensin removal.In our experimental design, all treatments received monensin during the adaptation period.The rumen requires time to adapt, establish, and stabilize a new bacterial community and thus, residual effects of monensin supplementation likely occur, which may attenuate the magnitude of treatment differences observed during the experimental period.In a recent pooled analysis compiling data from 4 mid-lactation experiments (Akins et al., 2014;Hagen et al., 2015;Benoit et al., 2021), including the current study, it was demonstrated that effects of monensin take up to 3 wk to dissipate following removal (Dhuyvetter et al., 2023).Aside from its importance in assessing data from controlled research trials, this finding has practical implications when evaluating product efficacy in commercial settings.
Overall, monensin supplementation linearly increased component-corrected MPE and estimated diet energy content while having no detrimental effect on milk fat synthesis.Improvements in feed efficiency shown herein are consistent with the FDA approval trials and numerous postapproval research studies.Regardless of monensin supplementation, significant improvement in industry-wide feed efficiency (expressed as SCM/DMI) has been achieved since the 2004 FDA trials, demonstrating continued progress in the industry.Consistent with previous data, the magnitude of improvement in MPE (~3%) with monensin was sustained even in today's more efficient cows; a response which is crucial given today's high feed costs.reported herein, suggest 18 g/t is the most favorable monensin dose when accounting for changes in milk yield, DMI, and BW (as assessed by the estimated diet energy content equation) in mid-lactation.However, when evaluated simply as component-corrected MPE the adequate dose based on our results would likely be 16 g/t.Lastly, our results challenge past concerns regarding the milk fat depressing effects of monensin and provide evidence that milk fat synthesis may be improved at certain monensin doses in diets formulated to maximize milk fat.
Horst et al.: MONENSIN IN MODERN DAIRY DIETS

Table 2 .
Horst et al.: MONENSIN IN MODERN DAIRY DIETS Analyzed nutrient composition and monensin concentration of the diet × time and the random effect of pen within treatment.Pen was the experimental unit. treatment

Table 3 .
Effects of increasing dietary monensin concentration on DMI and lactation performance

Table 4 .
Horst et al.: MONENSIN IN MODERN DAIRY DIETS Effects of increasing dietary monensin concentration on milk fat content, yield, and fatty acid profile