A field study on the effects of dietary monensin on milk production and milk composition in dairy cows

Abstract

The objectives of this study were to quantify the effect of 16 ppm of dietary monensin on milk production and composition of dairy cows, and to investigate factors having a potential impact on this effect. Data were generated from a total of 3577 Holstein dairy cows (47 herds) in Quebec enrolled in a herd-level, randomized clinical trial investigating the effects of monensin supplementation. Milk production and composition data were collected from monthly dairy herd improvement (DHI) testing. Monensin increased milk production by 0.9 kg/cow/d in cows under 150 days in milk (DIM) (P < 0.05). Monensin decreased milk fat percentage by 0.18 percentage points during the whole lactation (P < 0.05). This decreasing effect was larger for component-fed cows (P < 0.05) and for cows being fed low levels of dietary physically effective particles (P < 0.05) when compared respectively to cows fed total mixed ration and cows fed high levels of dietary physically effective particles. The results of this study suggest that monensin influences milk production and milk composition of dairy cows, and that diet composition and feeding system influence those effects.

Introduction

Monensin is a carboxylic polyether ionophore that is produced by the fermentation of Streptomyces cinnamonensis (1,2). Monensin favors bacterial populations that synthesize propionic acid, a precursor of glucose in cows, which improves the energy metabolism of the cow (1). Dietary monensin was approved in 2004 in Canada for use in dairy cows (Rumensin premix; Elanco Animal Health, Guelph, Ontario). The claims approved are for decreasing milk fat percentage (16 to 24 ppm); for minimizing body condition score loss during lactation (8 to 24 ppm); and for improving efficiency of milk protein production (16 to 24 ppm).

The effects of monensin on milk production and milk composition in dairy cows have been investigated by many researchers in the last 20 y. Results on the effects of monensin on milk production are, however, inconsistent in the literature. Some studies have reported that monensin increased milk production, but others have not (3). Most studies reporting an effect of monensin on milk production have stated an increasing impact between 0.4 and 2.8 kg of milk/cow/d (4). This effect has been attributed to an increased supply of glucogenic precursors resulting from changes in the pattern of rumen fermentation (5). Many factors have been reported to influence the impact of monensin on milk production such as parity and body condition score (6,7). Monensin has been shown to decrease milk fat but this effect was variable among studies (3) and seems to be influenced by many factors such as the dose administered, and the fiber, non-fiber carbohydrate, and polyunsaturated oils in the diet (8–10). Duffield et al (10) suggested that stage of lactation may influence monensin’s effect on milk fat, which could be explained by a reduction of rumen biohydrogenation of long-chain fatty acids (11) resulting in the production specific fatty acid isomers (12) that have been associated with decreased milk fat synthesis in the mammary gland (13). Monensin’s effect on milk protein has been inconsistent in previous studies (4).

A recently published field study evaluated the herd-level impact of monensin on bulk tank milk fat percentage (8). Bulk tank assessment does not allow cow-level data such as parity and stage of lactation to be used; therefore, a cow-level study was performed since the cow-level data generated from this herd-level study were available. The 1st objective of the study herein was to quantify the effects of 16 ppm of monensin on milk production and milk composition. The 2nd objective was to investigate if the stage of lactation and parity of cows have an influence on the effect of monensin on milk production and milk composition. The 3rd objective was to identify dietary factors influencing monensin effects. Dietary factors targeted were diet chemical composition, diet physical characteristics, feeding system, feeding management, and feed additives.

Materials and methods

Data generated from 47 Holstein dairy herds (3577 cows) in Quebec enrolled in a herd-level clinical trial investigating the production effects of monensin on bulk tank milk fat percentage were used for this cow-level study (8). Herds enrolled were required to use a monthly dairy herd improvement (DHI) service for individual cow milk recording, and were visited at least monthly by a veterinarian using the computerized data management system DSA (DS@HR, Saint-Hyacinthe, Québec) for reproduction and health management. The trial was performed between November 2005 and May 2006 to avoid excessive changes to the ration. Monensin treatment period was randomly allocated at the herd level in a crossover design. Monensin (Rumensin premix; Elanco Animal Health) was supplemented to the lactating cow diet for a consecutive 12-wk period for every herd. Twenty-four herds received 16 ppm (dry matter basis) of monensin from November through January (period 1), while the other 23 herds received 0 ppm. Then from March through May (period 2), monensin treatment was reversed. No herd was supplemented during a 4-wk period (February) for elimination of potential residual effects of monensin. No additional monensin (premix or controlled release capsule) was allowed during the trial period other than what was randomly assigned for treatment. Monensin was diluted into a premix of minerals, protein supplement, or concentrates at the feed mill. The accuracy of the dose of monensin fed to each herd was validated using the monensin concentration and the daily intake of supplemented feed. This experiment was approved by the Animal Care Committee of the University of Montreal.

Individual test day milk production and milk composition (fat and protein) data for cows enrolled in the trial were collected from monthly DHI testing (Valacta, Ste-Anne-de-Bellevue, Québec). Cow-level data such as calving date, parity, days in milk (DIM) at test day, reproduction events, and dry-off date were recorded in the computerized data management system (DSA). Diet composition data from multiple farm visits and a questionnaire completed by dairy farm managers regarding the nutritional management of the farms were used to build the herd-level nutritional factors databank. Each herd was visited every 60 d during the trial. During each farm visit, dietary particle size evaluation on forages and total mixed rations (TMR) was performed using the Penn State Particle Separator as described elsewhere (14). Physically effective particle level (PEPL) of the lactating cow ration was considered as the sum of the as-fed proportions retained on the top 2 sieves of the Penn State Particle Separator (8). The PEPL method was adapted from the physically effective fiber measurement method (15). To ensure consistency, the PEPL measurement procedure was performed by the same technician during the entire study. The PEPL values for TMR-fed herds came from the TMR samples and were adjusted if long dry hay was fed but not included in the TMR. For component-fed herds, the diet PEPL values came from theoretical calculations using the as-fed proportions of every forage and concentrate fed and their respective individual PEPL measurements.

Statistical analysis

The experimental unit of this study was the herd. All statistical procedures were done using SAS (version 9.1; SAS Institute, Cary, North Carolina, USA). Descriptive data assessment was performed (PROC MEANS). Milk production, milk fat percentage, milk fat yield, milk protein percentage, and milk protein yield retrieved from DHI were considered as outcome variables. Outcome variables were treated as monthly repeated measures within cow in linear mixed models (PROC MIXED). Herd clustering was treated as random effect in all models. Parity was classified as 1, 2, and ≥ 3. Days in milk (DIM) were classified as < 150 and ≥ 150. Preliminary mixed models were used to screen for significant variables and their interactions with monensin (P ≤ 0.15). Treatment (MON), treatment period, parity, DIM, parity*MON, and DIM*MON were forced covariates in all models. A final model for each outcome variable was built using manual backward elimination procedure using a P-value of ≤ 0.05 for partial F-test. Tendency was declared if 0.05 > P < 0.10. If an interaction between a variable and MON was found significant in the final model, the variable was dichotomized for providing more practical results to be used by nutritionists and veterinarians on farms. A Tukey-Kramer test was used to evaluate the difference between classes (P ≤ 0.05). Residuals from the final models were assessed for normality of distribution and homoscedasticity.

Results and discussion

Most herds were housed in tie-stall barns (n = 42) and were fed a TMR (n = 29). Description of the herds, diet composition, and nutritional management have been published (8). Descriptive statistics of production parameter from cows enrolled in the study are presented in Table 1. Monensin treatment effects on milk production and milk composition parameters from final linear mixed models are summarized in Table 2.

Table 1.

Descriptive statistics of production parameters from 3577 cows (47 herds) enrolled in a field study investigating the effects of supplementing 16 ppm of monensin on milk production and milk composition parameters

Variables	Mean	s	Minimum	Maximum
Days in milk (day)	170.8	110.3	1	427
Parity (year)	2.5	1.7	1	14
Milk production (kg/d)	30.1	9.7	6.4	76.2
Milk fat percentage (%)	3.91	0.67	1.51	6.66
Milk fat yield (kg/d)	1.16	0.37	0.20	3.07
Milk protein percentage (%)	3.35	0.37	2.02	4.52
Milk protein yield (kg/d)	0.99	0.27	0.15	2.81

s — standard deviation.

Table 2.

Least square means from linear mixed models of milk production and milk composition parameters when the cows received 0 or 16 ppm of monensin in a field study performed on 3577 cows (47 herds)

	Monensin (ppm)
	0	16	Effecta	Sx̄	P-valueb
Milk yield (kg/d)c
For cows < 150 DIMd	34.7	35.6	+0.9	0.5	< 0.01
For cows ≥ 150 DIMd	24.5	24.6	+0.1	0.5	0.50
Composition (%)
Fat	3.91	3.73	−0.18	0.04	< 0.01
Protein	3.14	3.11	−0.03	0.01	0.07
Constituent yield (kg/d)
Fat	1.29	1.26	−0.03	0.03	0.02
Protein	1.09	1.10	+0.01	0.17	0.14

S_x̄ — standard error.

^aEffect of monensin treatment.

^bFor monensin effect.

^cMilk yield was stratified by stage of lactation because of its interaction with monensin treatment (P = 0.02).

^dDays in milk.

Milk production

The stage of lactation of cows influenced monensin’s effect on milk production (P = 0.02). Two final models were then built for cows with DIM under and over 150 (Table 2). Supplementation of monensin significantly increased milk production by 0.9 kg per cow per day in cows under 150 DIM. The size of this effect of monensin on milk production is similar to an increase of 0.7 kg per cow per day [95% confidence interval (CI): 0.49, 0.85] reported in a recent meta-analysis (3). This increasing effect of monensin on milk production was likely caused by an increased supply of glucogenic precursors resulting from changes in rumen fermentation pattern (5). No significant effect of monensin on milk production was seen for cows greater or equal to 150 DIM. This effect of stage of lactation on milk production has not been reported before and conflicts with another study that found a positive impact of monensin supplementation on milk production in cows throughout the entire lactation (up to 280 DIM; 5). This finding from our study could be explained by our crossover study design because the cows enrolled were supplemented during a period of only 3 mo. The cows in the study by Phipps et al (5), however, were supplemented during their entire lactation period because if our finding is real, it could be hypothesized that the improved energy metabolism of cows supplemented with monensin provides antiketogenic effect (16) and improves milk production in early lactation, but this benefit in later lactation could be more toward improving feed efficiency. Unfortunately, this parameter was not measured in our study so further investigation is needed. Parity did not influence monensin’s effect on milk production, which is similar to the results of a previous study investigating this effect (5).

Milk fat percentage and yield

Monensin treatment decreased milk fat percentage by 0.18 percentage point and milk fat yield by 0.03 kg/d (Table 2). This decreasing effect of monensin on milk fat percentage is relatively similar to the meta-analysis by Duffield et al (3) which reported a decrease of 0.12 percentage point (95% CI: −0.15, −0.08). However, they found no effect of monensin on milk fat yield, which is different from our results. Previous studies have found a decreasing effect of monensin on milk fat yield similar to our study (17,18). This effect could be explained by the fact that monensin causes a reduction of rumen biohydrogenation of long-chain fatty acids (11) resulting in the production specific fatty acid isomers (12) that inhibit milk fat synthesis in the mammary gland (13). Stage of lactation has been suggested by Duffield et al (10) to influence monensin’s effect on milk fat, but this result was not found in our study. Parity did not influence monensin effect on milk fat. The effect of monensin on milk fat percentage and milk fat yield was influenced by the herd feeding system (Table 3). Cows that were in component-fed herds (forages fed separately than concentrates) had a significant milk fat decrease caused by monensin, whereas cows in TMR-fed herds did not. A meta-analysis found that topdress feeding of monensin had a similar negative effect on milk fat while TMR feeding did not (3). The authors explained this situation by suggesting that bolus feeding of larger amounts of monensin in component-fed herds compared to equally proportioned bites of feed in the TMR feeding system could have different rumen bacterial effects (3). Since most component-fed herds enrolled in this trial were fed monensin as a topdress twice a day (15/18, 83%), this explanation could be plausible. The larger meal size and lower frequency of daily feed intake of concentrates in component-fed herds compared with TMR-fed herds could also be involved (3). The influence of the herd feeding system on the effect of monensin on milk fat has already been reported by Duffield et al (10) but their results were the opposite of ours. They reported only a decreasing effect of monensin on milk fat percentage in TMR-fed herds. This was a cross-sectional study that used a dose varying between 9 and 23 ppm while in our trial the study was prospective and the monensin dose was fixed at 16 ppm. Further investigation is needed to clarify these conflicting results. The decreasing effect of monensin of milk fat was influenced by the PEPL (Table 3). This interaction revealed a more important decreasing effect of monensin on milk fat percentage when fed to cows receiving PEPL ≤ 45% (−0.28%) compared to PEPL level > 45% (−0.16%). However, this monensin effect on milk fat yield was significant only for cows receiving a diet low in PEPL (≤ 45%, Table 3). The cows receiving a diet containing higher PEPL (> 45%) did not experience any significant effect of monensin on milk fat yield. These results suggest that the decrease in milk fat percentage associated with monensin supplementation in the cows fed high PEPL (Table 3) is only a dilution effect. The combination of monensin supplementation and low PEPL in the diet, however, exacerbates the effect of monensin on milk fat yield. These results are similar to a previous study reporting that monensin treatment was associated with a reduction in milk fat percentage when fed in low fiber diets to dairy cows (10,19).

Table 3.

Least square means from linear mixed model of milk fat percentages and milk fat yields when the cows received 0 or 16 ppm of monensin. Analysis was stratified by type of feeding system and physically effective particle level of the ration

	Monensin (ppm)
	0	16	Effecta	Sx̄	P-valueb
Milk Fat (%)
Type of feeding
Component-fed	3.95	3.69	−0.26	0.06	< 0.01
TMR-fedc	3.89	3.77	−0.12	0.04	0.14
Diet PEPLd
≤ 45%	3.95	3.67	−0.28	0.03	< 0.01
> 45%	3.89	3.73	−0.16	0.03	< 0.01
Milk fat yield (kg/d)
Type of feeding
Component-fed	1.20	1.15	−0.05	0.03	< 0.01
TMR-fedc	1.25	1.23	−0.02	0.02	0.11
Diet PEPLd
≤ 45%	1.32	1.26	−0.06	0.03	< 0.01
> 45%	1.27	1.26	−0.01	0.03	0.30

S_x̄ — standard error.

^aEffect of monensin treatment.

^bFor monensin effect.

^cTotal mixed ration.

^dPEPL — Physically effective particle level of ration = sum of as-fed proportions of particles retained on top two sieves of Penn State Particle Separator (> 8 mm). The method was adapted from (15).

Monensin had a tendency to decrease milk protein percentage by a 0.03 percentage point (Table 2). However, since monensin had no significant effect on daily milk protein yield and it was associated with an increased milk yield, it was assumed to be a dilution effect. These results are similar to previous studies which reported absence or very little effect of monensin on milk protein yield (5,20). In our study, monensin’s effect on milk protein was not influenced by parity, stage of lactation, or any of the dietary factors investigated.

Even if several dietary interactions with monensin were tested in this field trial, some could not be investigated such as the source and amount of oils fed, and the polyunsaturated fatty acids (PUFA) level in the diet, which has been shown to influence monensin’s effect on milk production and composition (9,21). Further investigations need to be conducted into the influence of those parameters on monensin’s effects.

In conclusion, monensin fed at a dose of 16 ppm to dairy cows in early lactation (0 to 150 DIM) increased milk production. Milk fat percentage and milk fat yield were decreased throughout the entire lactation when cows were supplemented with monensin. Cows that were component-fed experienced a greater decreasing effect of monensin on milk fat percentage than cows that were TMR-fed. Cows fed a ration with a low physically effective particle level experienced a greater decreasing effect of monensin on milk fat percentage and milk fat yield than cows fed a higher level. Parity had no influence on monensin effect on milk production and milk composition.