JL Ellis, J. Dijkstra, A. Bannink, E. Kebreab, SE Hook

Quantifying the effect of monensin dose on the rumen volatile fatty acid profile in high-grain-fed beef cattle J. L. Ellis, J. Dijkstra, A. Bannink, E. Kebreab, S. E. Hook, S. Archibeque and J. France J ANIM SCI 2012, 90:2717-2726. doi: 10.2527/jas.2011-3966

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://www.journalofanimalscience.org/content/90/8/2717

www.asas.org

Downloaded from www.journalofanimalscience.org at Wageningen UR Library on September 4, 2012

Quantifying the effect of monensin dose on the rumen volatile fatty acid profile in high-grain-fed beef cattle1 J. L. Ellis,*†2, J. Dijkstra,* A. Bannink,‡ E. Kebreab,§ S. E. Hook,† S. Archibeque,# and J. France† *Animal Nutrition Group, Wageningen University, Wageningen 6708 WD, the Netherlands; †Centre for Nutrition Modelling, Department of Animal and Poultry Science, University of Guelph, Guelph, ON N1G 2W1, Canada; ‡Wageningen UR Livestock Research, Wageningen University Research Centre, Lelystad 8200 AB, the Netherlands; §Department of Animal Science, University of California, Davis 95616; and #Animal Sciences, Colorado State University, Fort Collins 80523

ABSTRACT: Monensin is a common feed additive used in various countries, where 1 of the associated benefits for use in beef cattle is improved efficiency of energy metabolism by the rumen bacteria, the animal, or both. Modeling fermentation-altering supplements is of interest, and thus, it is the purpose of this paper to quantify the change in VFA profile caused by monensin dose in high-grain-fed beef cattle. The developmental database used for meta-analysis included 58 treatment means from 16 studies from the published literature, and the proportional change in molar acetate, propionate, and butyrate (mol/100 mol) as well as total VFA (mM) with monensin feeding dose (mg/kg DM, concentration in the feed) was evaluated using the MIXED procedure (SAS Inst. Inc., Cary, NC) with the study treated as a random effect. The mean monensin dose in the literature database was 30.9 ± 3.70 mg/kg DM and ranged from 0.0 to 88.0 mg/kg DM. Mean DMI was 7.8

± 0.26 kg DM/d, mean concentrate proportion of the diet was 0.87 ± 0.01, and mean treatment period was 42 ± 5.6 d. Results produced the following equations: proportional change in acetate (mol/100 mol) = −0.0634 (± 0.0323) × monensin (mg/kg DM)/100 (P = 0.068), proportional change in propionate (mol/100 mol) = 0.260 (± 0.0735) × monensin (mg/ kg DM)/100 (P = 0.003), and proportional change in butyrate (mol/100 mol) = −0.335 (± 0.0916) × monensin (mg/kg DM)/100 (P = 0.002). The change in total VFA was not significantly related to monensin dose (P = 0.93). The results presented here indicate that the shift in VFA profile may be dose dependent, with increasing propionate and decreasing acetate and butyrate proportions (mol/100 mol). These equations could be applied within mechanistic models of rumen fermentation to represent the effect of monensin dose on the VFA profile in high-grain-fed beef cattle.

Key words: methane, modeling, monensin, volatile fatty acids © 2012 American Society of Animal Science. All rights reserved. INTRODUCTION Monensin (Elanco Animal Health, Ontario, Canada) is a common feed additive used in North America, Australia, and New Zealand, where the associated benefits for use in cattle include improved efficiency of energy and N metabolism by the rumen bacteria, the animal, or both, and retardation of digestive disorders resulting from abnormal rumen fermentation 1The authors thank the Canada Research Chair and NSERC strategic programs for funding. 2Corresponding author: [email protected] Received February 14, 2011. Accepted February 27, 2012.

J. Anim. Sci. 2012.90:2717–2726 doi:10.2527/jas2011-3966

(Bergen and Bates, 1984). Positive effects on energetic efficiency are the result of increased propionate production in the rumen as a consequence of a general resistance to monensin of gram-negative bacteria that reduce succinate to propionate, whereas a reduction in population size and activity occurs in gram-positive bacteria groups (McGuffey et al., 2001). Protozoa and fungi tend to be inhibited (Bergen and Bates, 1984), and Gyulai and Baran (1988) showed that some strains are more sensitive to monensin than others, although the data for fungi are more sparse and contradictory than for rumen bacteria (McGuffey et al., 2001). An increase in propionate production also results in a decrease in available substrate for methanogens,

2717 Downloaded from www.journalofanimalscience.org at Wageningen UR Library on September 4, 2012

2718

Ellis et al.

mainly hydrogen and formate, and reduces the amount of energy lost as methane (CH4; reviewed by Ellis et al., 2008), and there has been some interest in the potential use of monensin as a CH4 mitigation strategy (Beauchemin et al., 2009), although in vivo results are inconclusive. Dynamic mechanistic models have been used increasingly to explore mitigation and feeding strategies and to aid in the development of future experiments (Benchaar et al., 2001; Bannink et al., 2010; Ellis et al., 2011, 2012), but to date, none have accounted for the use of non-nutritional supplements such as monensin. Because most high-grain-fed beef cattle trials currently being conducted at North American institutions include monensin in the diet, one either has to exclude these data from exercises of model development and challenge or has to discount the potential effect monensin will have on the VFA stoichiometry. It is therefore the purpose of this study to quantify the change in VFA profile with monensin dose for high-grain-fed beef cattle. MATERIALS AND METHODS Animal Care and Use Committee approval was not obtained for this study because the data were obtained from existing literature. Literature Database The database used for this evaluation was constructed from a systematic literature search and, after application of various criteria for selection, comprised 58 treatment means from 16 studies spanning from 1976 to 2009. It included the studies of Ralston and Davidson (1976), Raun et al. (1976), Richardson et al. (1976), Utley et al. (1977), Heinemann et al. (1978), van Maanen et al. (1978), Thorton and Owens (1981), Zinn (1987), Morris et al. (1990), Clary et al. (1993), Zinn and Borques (1993), Zinn et al. (1994), Surber and Bowman (1998), Ives et al. (2002), Fandiño et al. (2008), and Meyer et al. (2009). The criteria for study selection were that the experiment involved beef cattle, that monensin was a treatment comparable against the appropriate control diet, that the basal diet was at least 80% grain, and that information was available on monensin dose (mg/d), DMI (kg/d), and ruminal acetate, propionate and butyrate expressed as a fraction of total VFA,or as mM. There was a preference for studies that also reported total VFA (tVFA, mM) so that the effect of monensin and tylosin (Elanco Animal Health) on tVFA could be evaluated, but Ralston and Davidson (1976), van Maanen et al. (1978), Zinn (1987), and Zinn and Borques (1993) did not report this information. Some studies included tylosin as an additive with monensin,

while others included it with both monensin and control treatments. The mean monensin dose in the literature database was 30.9 ± 3.70 mg/kg DM and ranged from 0.0 to 88.0 mg/kg DM, mean DMI was 7.8 ± 0.26 kg DM/d, mean concentrate proportion of the diet was 0.87 ± 0.01, and mean treatment period was 42 ± 5.6 d. A summary of the database is presented in Table 1, and a summary of the VFA profile for each study is given in Table 2. Statistics The change in tVFA (mM) and the proportional change in acetate, propionate, and butyrate (in mol/100 mol tVFA) relative to the control diet were examined against the main effects of monensin dose (mg/kg DM) and tylosin dose (mg/kg DM) within the MIXED procedure (SAS Inst. Inc., Cary, NC). Mixed model analysis was chosen because the data were compiled from multiple studies, thereby making it necessary to consider analyzing not only the fixed effects of the dependent variables but also the random effect of study. This accounts for differences such as physiological status of the animals, experimental design, measurement methods, techniques, and varying laboratories (St-Pierre, 2001). In several studies in the developmental database, tylosin was included either in all treatment diets (Heinemann et al., 1978; Surber and Bowman, 1998) or only in the monensin diet and excluded from control diets (Zinn, 1987; Morris et al., 1990; Clary et al., 1993; Ives et al., 2002). Tylosin is an antibiotic commonly included in high-grain diets, often with monensin, to reduce the occurrence of liver abscesses (e.g., Brown et al., 1973; Heinemann et al., 1978). In vitro work by Baldwin et al. (1982) showed that tylosin decreased tVFA concentration, and Nagaraja et al. (1987) reported that tylosin decreased the molar concentration of propionate. However, in vivo work done by Heinemann et al. (1978) showed no effect of tylosin on the VFA profile at similar doses to those evaluated here (11 mg/kg DM), and no interaction with monensin dose was evident. Nagaraja et al. (1999) also found, in vivo, that tylosin did not alter the concentration of tVFA in the rumen. In attempts to develop multiple regression equations including both monensin and tylosin dose, tylosin was nonsignificantly related to the change in acetate, propionate, butyrate and tVFA (P > 0.70) and was therefore excluded from the analysis. Multivariate analysis, through introduction of the correlation of dependent variables (acetate, propionate, and butyrate) via the random effects statement as well as via the residual variances (repeated statement; Strathe et al., 2010), were tested but resulted in over-


2719

Monensin dose effect on rumen volatile fatty acids

Table 1. Description of the database Data source

Monensin doses, mg/kg DM

DMI1, kg/d

Concentrate in diet, %

BW, kg

Tylosin, mg/kg Treatment DM period, d

Ralston and Davidson (1976) Raun et al. (1976)

0.0, 5.3, 11.4, 23.9, 36.2 0.0, 2.7, 5.5, 11, 22, 33, 44, 88

8.86 ± 0.254

100

386

0.0

8.84 ± 0.284

97

363

0.0

Richardson et al. (1976)

0.0, 14.5, 79

6.70 ± 0.185

95

415 ± 4.0

0.0

Utley et al. (1977)

0.0, 85

6.63 ± 0.735

100

266 ± 4.3

0.0

Heinemann et al. (1978) van Maanen et al. (1978) Thorton and Owens (1981)

0.0, 5.5, 11, 33

9.76 ± 0.207

82

446 ± 1.7

5.53

0.0, 39

3.85 ± 0.000

80

155 to 253

0.0

0.0, 55

4.10 ± 0.000

94

375

0.0

Zinn (1987)

0.0, 33

6.57 ± 0.427

80

295 ± 7.3

0.0, 114

Morris et al. (1990)

0.0, 29

9.31 ± 0.052

95

512 ± 0.2

0.0, 113

Zinn and Borques (1993)

0.0, 33

3.99 ± 0.000

88

180

0.0

Clary et al. (1993)

0.0, 25

10.9 ± 0.09

90

430

0.0, 103

Clary et al. (1993)

0.0, 25

10.6 ± 0.02

90

430

0.0, 103

Zinn et al. (1994)

0.0, 28

4.82 ± 0.014

90

234

0.0

14

Zinn et al. (1994)

0.0, 28

4.83 ± 0.001

80

234

0.0

14

112

83.0% steam rolled barley, 10.0% beet pulp, 5.0% molasses, 1.0% protein supplement, and 1.0% oyster shell flour 28 Diet 1 (CA-40): 61.2% corn, 20.0% corn cobs, 3.1% alfalfa meal, 8.7% soybean meal, 5.0% molasses, and 2.0% vitamins and minerals Diet 2 (CA-90): 70.0% corn, 10.0% corn cobs, 5.0% alfalfa meal, 8.0% soybean meal, 5.0% molasses, and 2.0% vitamins and minerals 10 to 30 CA-90 diet: 70.0% corn, 10.0% corn cobs, 5.0% alfalfa meal, 8.0% soybean meal, 5.0% molasses, and 2% vitamins and minerals 84 Diet 1: 74.0% ground shelled corn, 20.0% peanut hulls, 6.0% supplement Diet 2: 75.8% acid-treated rolled corn, 18.6% peanut hulls, 5.6% supplement2 85 66.4% ground corn, 15.0% dehydrated beet pulp with molasses, 18.0% sun-cured alfalfa hay, and 0.6% vitamin/mineral 35 64.0% cracked corn, 16.0% protein supplement, and 20.0% chopped alfalfa hay 15 62.8% rolled corn grain, 10.0% soybean meal, 6.0% alfalfa meal, 5.0% molasses, 14.0% cottonseed hulls, and 2.2% vitamins and minerals 14 50.6% corn, 1.0% barley, 20.0% whole cottonseed, 6.0% molasses 15.0% sudangrass, 5.0% alfalfa hay, and 2.4% vitamins and minerals 18 76.9% steam-flaked milo, 5.0% ground alfalfa cubes, 5.0% cottonseed hulls, 1.5% yellow grease, 7.0% molasses, 1.0% premix, 1.0% urea, and 2.6% vitamins and minerals5 14 74.9% steam-flaked corn, 6.0% alfalfa hay, 6.0% sudangrass hay, 4.0% yellow grease, 6.0% molasses, 1.1% urea, and 2.0% vitamins and minerals 19 78.0% rolled corn, 10.0% prairie hay, and 8.0% supplement Diet 1: 0.0% tallow and 4.0% molasses Diet 2: 4.0% tallow and 0.0% molasses 19

21

Surber and Bowman (1998)

0.0, 41.6

6.50 ± 0.011

80 to 87

430

Ives et al. (2002)

0.0, 32

7.99 ± 0.230

90

345

0.0, 12.93

21

0.0, 32

8.47 ± 0.100

95

345

0.0, 11.73

21

Fandiño et al. (2008)

0.0, 31.3

7.55 ± 0.050

90

NA

0.0

24

Meyer et al. (2009)

0.0, 26.4

8.15 ± 0.450

93

512

0.0

28

Ives et al. (2002)

1

Basal diet, % DM

152

Diet 1: 79.8% steam-flaked corn, 2.5% alfalfa hay, 7.5% sudangrass hay, 3.0% yellow grease, 4.0% molasses, 1.1% urea, and 2.1% vitamins and minerals Diet 2: 69.8% steam-flaked corn, 5.0% alfalfa hay, 15.0% sudangrass hay, 3.0% yellow grease, 4.0% molasses, 1.1% urea, and 2.1% vitamins and minerals Diet 1: 70.0% cracked corn, 20.5% grass hay, 3.8% soybean meal, 1.2% dry molasses, 1.0% urea, 0.2% canola oil, and 3.3% vitamins and minerals Diet 2: 80.0% cracked barley, 12.7% grass hay, 2.1% soybean meal, 1.3% dry molasses, 1.0% urea, 0.2% canola oil, and 2.7% vitamins and minerals1 Diet 1: 72.1% rolled corn, 12.0% soybean meal, 10.0% alfalfa, 4.0% molasses, and 1.9% vitamins and minerals Diet 2: 62.9% rolled corn, 30.0% wet corn gluten feed, 5.0% alfalfa, and 2.1% vitamins and minerals 27.9% ground barley grain, 25.6% ground corn grain, 10.9% soybean meal, 6.7% soybean hulls, 5.7% corn gluten feed, 5.7% tapioca, 7.2% sunflower meal, 10.0% barley straw, and 0.3% vitamins and minerals 66.0% high-moisture corn, 16.5% dry-rolled corn, 7.5% alfalfa hay, 5.0% molasses, and 5.0% supplement

Mean ± SEM. fatty acid results are the mean over these 2 diets.

2Volatile 3Tylosin

(Elanco Animal Health, Ontario, Canada) was included in both the control and monensin (Elanco Animal Health) diets. tylosin was included in the control diet, but tylosin was included in the monensin diet. 5As-fed basis. 4Zero


2720

Ellis et al.

Table 2. Mean VFA profile, by study, for the database1 tVFA2, mM

Data source Ralston and Davidson (1976)

—

Acetate, Propionate, Butyrate, mol/100 mol mol/100 mol mol/100 mol tVFA tVFA tVFA 46.1 ± 0.76

48.4 ± 0.99

5.5 ± 0.32

Raun et al. (1976)

80.1 ± 0.91

48.4 ± 0.37

40.2 ± 0.44

7.4 ± 0.35

Richardson et al. (1976)

83.2 ± 2.57

51.0 ± 2.52

38.8 ± 3.52

5.7 ± 0.70

Utley et al. (1977)

101.0 ± 2.40

53.1 ± 0.15

33.4 ± 2.70

10.1 ± 2.00

Heinemann et al. (1978)

64.0 ± 1.81

60.9 ± 1.16

32.3 ± 1.70

6.8 ± 0.60

van Maanen et al. (1978)

—

64.7 ± 1.80

25.2 ± 2.75

10.2 ± 0.90

132.9 ± 24.4

70.3 ± 4.00

17.3 ± 2.80

6.4 ± 0.15 11.0 ± 0.25

Thorton and Owens (1981) Zinn (1987)

—

65.1 ± 0.15

24.0 ± 0.10

129.0 ± 2.15

46.5 ± 1.95

39.6 ± 0.35

9.9 ± 2.05

—

49.3 ± 1.35

40.6 ± 3.05

10.2 ± 1.75

Clary et al. (1993)

111.4 ± 2.26

49.5 ± 2.14

37.4 ± 2.83

8.7 ± 0.13

Zinn et al. (1994)

123.0 ± 5.93

49.6 ± 1.86

39.0 ± 1.49

8.4 ± 0.57

Surber and Bowman (1998)

51.4 ± 0.69

57.5 ± 0.89

20.1 ± 1.24

15.8 ± 0.59

Ives et al. (2002)

Morris et al. (1990) Zinn and Borques (1993)

101.7 ± 3.89

52.3 ± 1.39

26.7 ± 1.09

15.7 ± 0.38

Fandiño et al. (2008) 110.7 ± 1.90

54.9 ± 0.40

26.4 ± 1.15

12.6 ± 0.40

Meyer et al. (2009)

50.5 ± 0.40

30.7 ± 2.15

12.1 ± 0.85

1Mean

107.0 ± 2.25

± SEM. total VFA.

2tVFA =

parameterization of the model and a lack of convergence or a non-positive R matrix. Acetate, propionate, and butyrate regressions were thus fitted simultaneously using a univariate approach. The statistical model was

Yij = B1Xij + biXij + eij ,

[1]

where Yij is the dependent variable, B1 is the overall fixed effect regression coefficient of Y on X, Xij is the value of the continuous predictor variable, bi is the random effect of study on the regression coefficient of Y on X, and eij is the residual error. In this analysis, monensin and tylosin dose (independent variables) were centered around the mean, where, for example, monensin_new = monensin – monensin_mean. Centering the monensin or tylosin dose places the intercept at the mean of all the doses and therefore removes the correlation between slope(s) and intercept(s) in the regression model (van Landeghem et al., 2006). Without centering, both the mean value and the variation around that mean are involved in selecting model factors. Because this study is only interested in the change (slope) in VFA profile, this was an appropriate adjustment. The joint distribution of random effects was assumed to be multivariate normal, and the dual quasi-Newton technique was used for optimization with an adaptive Gaussian quadrature as the integration method. Analysis was performed with the assumption

that variance distribution for the estimates followed a multivariate normal distribution. Recently, Sauvant et al. (2008) highlighted the importance of considering additional sources of systematic variation in meta-analysis studies, and so this database was selected to minimize these sources. For example, the database was limited to beef cattle fed a diet of >80% concentrate, which severely cuts down on the number of available studies in the literature and, subsequently, the variation in diets. Monensin was expressed as milligrams per kilogram DM to exclude the interfering factor of DMI. This study was specifically interested in the relative change in VFA profile as monensin adjustment equations will be integrated with model-determined VFA yields, which would account for the effect of diet. Therefore, the effect of the control or basal value was removed, and equations were based on the proportional change in values relative to the control. This essentially removes any inter-study bias effect and examines only the slope. Goodness of fit of the statistical model (inclusion or exclusion of random effects, variance or covariance structure selection) was evaluated using the Bayesian information criterion (BIC) fit statistic and the restricted log-likelihood function (−2LL; SAS Inst. Inc.), where smaller values indicate better fit, and significance of the fixed effect model parameters were tested against a P value of 0.05. RESULTS AND DISCUSSION The relative amounts of VFA produced in the rumen are of particular interest in ruminant nutrition because the major VFA feed into important metabolic pathways in other organs. Propionate is a substrate for gluconeogenesis and is the main source of glucose for the ruminant, whereas non-glucogenic acetate and butyrate are precursors for long-chain fatty acid synthesis. The VFA profile, particularly the nonglucogenic to glucogenic VFA ratio, is associated with effects on end-product composition and energy balance in ruminants (Thomas and Martin, 1988; van Knegsel et al., 2007). The type of VFA formed in the rumen is also essential in mechanistic models that predict enteric methanogens (Alemu et al., 2011; Bannink et al., 2011; Mills et al., 2001) because propionate is a hydrogen sink whereas acetate and butyrate are hydrogen sources, and hydrogen is the major substrate for CH4 formation (Wolin, 1960). Methane represents an energy loss to the animal, where observations can range from 2% to 12% of GE intake (Johnson and Johnson, 1995), and it is also a greenhouse gas. Monensin, through its impact on the rumen VFA profile, has been questioned as a potential CH4 mitigation strategy. This paper focused on



high-grain-fed feedlot cattle as within the beef industry it is the most likely sector targeted for CH4 mitigation. Feedlots contain large numbers of animals that are intensively managed and are typically fed ionophores. The mean propionate molar proportion was relatively high (31.3 ± 1.87 mol/100 mol tVFA) and that of acetate was low (54.5 ± 1.70 mol/100 mol tVFA; means for control treatments). This is expected as, in general, the low-fiber and high-starch content of highconcentrate diets, from stoichiometric principles, is known to result in a high-propionate-producing rumen profile (Bannink et al., 2008). Relative changes in molar proportions of acetate, propionate, and butyrate (mol/100 mol tVFA) were related to monensin dose in the diet. The relationships developed in the present study modify model-determined proportions of individual VFA (mol/100 mol tVFA) as follows: Acetate (mol/100 mol tVFA) = Default acetate (mol/100 mol tVFA) × [1.0 − 0.0634 (± 0.0323) × monensin (mg/kg DM)/100],

[2]

Propionate (mol/100 mol tVFA) = Default propionate (mol/100 mol tVFA) × [1.0 + 0.260 (± 0.0735) × monensin (mg/kg DM)/100],

[3]

Butyrate (mol/100 mol tVFA) = Default butyrate (mol/100 mol tVFA) × [1.0 − 0.335 (± 0.0916) × monensin (mg/kg DM)/100],

[4]

where the default VFA proportion (acetate, propionate, or butyrate; mol/100 mol tVFA) is the proportion of VFA determined as a result of the VFA stoichiometry represented in a given model and the new calculated VFA proportion (mol/100 mol tVFA) is related to the monensin dose via a proportional change. Proportional changes were significant for propionate (P = 0.003) and butyrate (P = 0.002) but not for acetate (P = 0.068). The joint BIC value was 1291 and −2LL was 1277. The root-mean-square prediction error (RMSPE) values (percent of observed mean; Bibby and Toutenburg, 1977) for the 3 regressions tested back against the developmental database were 5.2%, 9.4%, and 17.0% for acetate, propionate, and butyrate, respectively, and the error was mostly random (93.9%, 91.8%, and 96.5%, respectively). Concordance correlation coefficient (CCC; Lin, 1989) values were 0.063, 0.327, and 0.357, respectively, with R values (indicator of precision) of 0.122, 0.412, and 0.488 and Cb values (indicator of accuracy) of 0.512, 0.792, and 0.732, respectively. A low R value for the non-significant acetate equation causes

2721

the CCC value to be poor compared to analysis using RMSPE, which does not pick up on this error (Ellis et al., 2010). Regressions of the proportional changes in Eq. [2] through [4] are illustrated in Figure 1. It is evident from Figure 1 that the majority of trials tested monensin doses in the range of 0.0 to 40.0 mg/kg DM. As a result, some caution should be used in applying these equations to greater doses, although the slope of the best fit regression appears similar to the slope of the trials that tested doses greater than 40.0 mg/kg DM (Figure 1).

Figure 1. Proportional change in observed rumen molar (top) acetate, (middle) propionate, and (bottom) butyrate, measured in moles/100 moles of total VFA, vs. monensin dose (mg/kg DM), where the dotted lines represent individual study (Table 1) regression lines and the solid lines represent the overall best fit regression (Eq. [2], [3], and [4]).


2722

Ellis et al.

Across studies, the molar proportion of propionate consistently increased with monensin dose, whereas the most variation in regression slopes occurred with the molar proportion of butyrate and the least variation occurred with the change in molar proportion of acetate (Figure 1). The reason for high variation in the butyrate slope between studies is unknown but might be due to it being the smallest fraction of the three VFA examined, resulting in proportionally more error in measurement. Henderson et al. (1981) examined the in vitro effect of monensin on bacteria and found differences in the amount of inhibition between 2 strains of Butyrivibrio bacteria in the presence of monensin. Differences in the Butyrivibrio rumen bacteria communities between studies may also contribute to the increased variation in the change in butyrate results seen here. Furthermore, protozoa are known to quickly metabolize sugars (e.g., see Dijkstra, 1994) and produce relatively large amounts of butyrate. Studies have suggested that the sensitivity of protozoa to monensin might contribute to the initial decrease in CH4 as protozoa and methanogens interact closely in the rumen (Richardson et al., 1978; Hino, 1981; Habib and Leng, 1987). However, this effect may subside due to the adaptation of the protozoal population

to monensin over time (Johnson and Johnson, 1995; Tedeschi et al., 2003; Guan et al., 2006). The net result is that a decrease in CH4 production might be of greater magnitude initially (see the review by McGuffey et al., 2001). Therefore, variation between studies in the length of the treatment period (varied from 14 to 152 d) could lead to variation in the phase of the response of protozoa and therefore heightened variation in the response of butyrate to monensin dose. While this study took a linear approach to the analysis of monensin dose on VFA profile, it is possible that the slope may be non-linear, particularly at high doses (Figure 1). However, without more high-dose studies, it is not possible to develop a reliable nonlinear equation to describe the data. Analysis of the residuals (predicted minus observed values) in Figure 2 shows that error in prediction is not systematically related to the observed value, DMI, or the acetate, propionate, and butyrate molar proportion. For the residual vs. butyrate (mol/100 mol) plot, there is a seemingly negative slope, but this is largely influenced by a single datum point. Removal of this datum point reduces the slope from −2.07 to −0.02. In this paper it was assumed that the change in VFA molar proportion due to monensin addition was related

Figure 2. Residual (predicted minus observed) of the proportional change in rumen molar proportion of (left) acetate, (middle) propionate, and (right) butyrate, measured in moles/100 moles of total VFA, vs. (top) equation-predicted proportional change (Eq. [2] through [4]), (middle) DMI (kg/d), and (bottom) molar proportion of acetate, propionate, or butyrate (mol/100 mol of total VFA).



to the observed VFA molar proportion. In this approach, the magnitude of the change is assumed to depend on the actual VFA molar proportion (relatively small change if VFA molar proportion is small and vice versa), rather than fixed and independent of the VFA molar proportion. If in real life the correction should not depend on the basal VFA molar proportion, the residuals would have shown a consistent pattern of over- and under-prediction at low and high VFA molar proportions. The absence of consistent patterns indicates that a correction on a relative basis is valid. The residuals are also not systematically related to the treatment period length (data not shown). Because Eq. [2] through [4] were developed independently and yet are related when expressed as moles/100 moles, the post-adjustment sum of moles/100 moles acetate, propionate, and butyrate is not necessarily identical to the pre-adjustment total. For example, a diet that produces a VFA profile that is 55 mol acetate/100 mol tVFA, 35 mol propionate/100 mol tVFA, and 8 mol butyrate/100 mol tVFA (sums to 98 mol/100 mol tVFA), with a 40 mg/kg DM monensin dose, becomes 53.6 mol acetate/100 mol tVFA, 38.6 mol propionate/100 mol tVFA, and 6.93 mol butyrate/100 mol tVFA (sums to 99.2 mol/100 mol tVFA). This has to be corrected by scaling total acetate, propionate, and butyrate (mol/100 mol tVFA) back to 98 mol/100 mol tVFA when it is also presumed that the fraction of other VFA remains the same. Other minor VFA (e.g., valerate, isovalerate, and isobutyrate) were not consistently reported in the database studies, and the studies reporting these VFA were not numerous enough to allow meta-analysis. Total VFA was not reported in studies by Ralston and Davidson (1976), van Maanen et al. (1978), Zinn (1987), and Zinn and Borques (1993). Analysis of the effect of monensin dose on the change in tVFA for the remaining studies (using the same MIXED model analysis described above but replacing individual VFA proportions with tVFA) was not significant: The change in tVFA (mM), which equals (0.0007 ± 0.06041) × monensin dose (mg/kg DM), where P = 0.99, is illustrated in Figure 3. Aside from 1 study that observed a 31% decrease in tVFA (mM) with a 55 mg/kg DM monensin treatment (Thorton and Owens, 1981), this is in agreement with other publications reporting no change in tVFA with monensin feeding (Guan et al., 2006). The relationship remained non-significant whether the Thorton and Owens (1981) study was left in or removed, as the equation became the change in tVFA (mM) = 0.0034 (± 0.0368) × monensin dose (mg/kg DM), where P = 0.93 when it was excluded. Nagaraja et al. (1997) examined the average change in acetate, propionate, and butyrate concentration (in mM, mol/100 mol tVFA and mol/d) for the 70% and 50% roughage diets originally reported by Prange et al. (1978) and Rogers and Davis (1982). In that summary Nagaraja

2723

Figure 3. Proportional change in rumen total VFA (tVFA, measured in mM) vs. monensin dose (mg/kg DM), where the dotted lines represent individual studies and the solid line represents the overall best fit regression (Table 1).

et al. (1997) illustrated a 6% decrease in acetate, a 29% increase in propionate, and a 14% decrease in butyrate (mol/100 mol tVFA) on the 70% roughage diet and a 4% decrease in acetate, a 15% increase in propionate, and a 25% decrease in butyrate (mol/100 mol tVFA) on the 50% roughage diet, with monensin doses of 33 g/kg DM. When a 33 mg/kg DM monensin dose is assumed here, with the equations developed in this study, based on diets with an average of 13% roughage, equations predict a 2.1% decrease in acetate, an 8.6% increase in propionate, and an 11.1% decrease in butyrate. Aside from the large change in butyrate reported for the 50% forage diet, these results suggest a greater impact on the VFA profile for diets at greater forage % diets. However, the studies of Prange et al. (1978) and Rogers and Davis (1982) also had treatment periods of 21 d compared with the slightly greater 28-d median value for this study, which may have some minor impact on the magnitude of the results (e.g., because of the longevity of protozoa effects). Also, results here are based on a higher number of studies and may therefore be more reliable. However, this observation also agrees with the results of Thorton and Owens (1981), where both the magnitude of change in the VFA profile and CH4 emissions were amplified for high-forage diets at the same monensin dose. In agreement with the dose-dependent results presented here, Beauchemin et al. (2008), in a summary of the literature, found that the effect of monensin on CH4 may be dose dependent. In that review, doses of less than 15 mg/kg DM monensin had no effect on CH4 production (g/d or g/kg DMI) in dairy cows, doses of less than 20 mg/kg DM either had no effect or reduced total CH4 (g/d but not g/kg DMI) in dairy cows, and greater doses (24 to 35 mg/kg DM) reduced CH4 production (g/d by 4% to 10% and g/kg DMI by 3% to 8%) in both beef cattle and dairy cows, with short-term decreases of up to 30% for both high- and low-forage diets. The equations


2724

Ellis et al.

developed here may represent part of the longer-term impact of monensin on the rumen fermentation pattern, as opposed to transient modifications in the rumen (e.g., effects on protozoa). Changes in the VFA profile appear to remain constant and persist over time (Guan et al., 2006). In terms of other ionophores used in beef cattle production, monensin and lasalocid (Elanco Animal Health) are licensed in the United States and Canada, laidlomycin propionate is licensed only in the United States, and salinomycin (Elanco Animal Health) is licensed only in Canada (NRC, 2000). Martineau et al. (2007) found no significant differences in acetate, propionate, or butyrate (mol/100 mol tVFA) between lasalocid and monensin treatments, and neither did Clary et al. (1993). However, although Morris et al. (1990) found no differences between lasalocid and monensin in terms of acetate and propionate, butyrate and valerate were significantly decreased with monensin treatment (mol/100 mol). Domescik and Martin (1999) concluded that laidlomycin propionate altered fermentation by mixed ruminal microorganisms in a manner similar to monensin but that monensin seemed to be a more potent inhibitor. Galyean et al. (1992) also found no difference in the VFA profile (mol/100 mol tVFA) between laidlomycin propionate and monensin treatments. However, the VFA profile for the ionophore treatments compared to the control treatment were also nonsignificantly different. Although no direct comparison with monensin was made, Merchen and Berger (1985) found the VFA profile shifted in a linear manner toward more propionate and less acetate and butyrate (mol/100 mol tVFA) with an increasing dose of salinomycin. The NRC (2000) groups all ionophores together in terms of their effects, and so it is possible that the equations developed here for monensin could also apply to other ionophores or that similar equations could be developed if the dose-response amounts differ between ionophores (Domescik and Martin, 1999). In summary, this paper provides equations to quantify the change in VFA profile with monensin feeding in high-grain-fed beef cattle. The results of this meta-analysis indicate that the changes in VFA profile are dose dependent and that monensin should not be ignored when considering beef cattle data in modeling exercises on rumen fermentation profiles. No change in tVFA was evident. The equations produced can be used to adjust VFA stoichiometry within a model of rumen fermentation that may provide more accurate estimates of VFA production for beef cattle being fed monensin.

LITERATURE CITED Alemu, A. W., J. Dijkstra, A. Bannink, J. France, and E. Kebreab. 2011. Rumen stoichiometric models and their contribution and challenges in predicting enteric methane production. Anim. Feed Sci. Technol. 166–167:761–778. Baldwin, K. A., J. Bitman, and M. J. Thompson. 1982. Comparison of N,N-dimethyldodecanamine with antibiotics on in vitro cellulose digestion and volatile fatty acid production by ruminal microorganisms. J. Anim. Sci. 55:673–679. Bannink, A., J. France, S. Lopez, W. J. J. Gerrits, E. Kebreab, S. Tamminga, and J. Dijkstra. 2008. Modeling the implications of feeding strategy on rumen fermentation and functioning of the rumen wall. Anim. Feed Sci. Technol. 143:3–26. Bannink, A., M. C. J. Smits, E. Kebreab, J. A. N. Mills, J. L. Ellis, A. Klop, J. France, and J. Dijkstra. 2010. Simulating the effects of grassland management and grass ensiling on methane emission from lactating cows. J. Agric. Sci. 148:55–72. Bannink, A., M. W. van Schijndel, and J. Dijkstra. 2011. A model of enteric fermentation in dairy cows to estimate methane emission for the Dutch National Inventory Report using the IPCC Tier 3 approach. Anim. Feed Sci. Technol. 166–167:603–618. Beauchemin, K. A., M. Kreuzer, F. O’Mara, and T. A. McAllister. 2008. Nutritional management for enteric methane abatement: A review. Aust. J. Exp. Agric. 48:21–27. Beauchemin K. A., T. A. McAllister, and S. M. McGinn. 2009. Dietary mitigation of enteric methane from cattle. CAB Rev. Perspect. Agric. Vet. Sci. Nutr. Nat. Resour. 4(35):1–18. Benchaar, C., C. Pomar, and J. C. Hiquitte. 2001. Evaluation of dietary strategies to reduce methane production in ruminants: A modeling approach. Can. J. Anim. Sci. 81:563–574. Bibby, J., and T. Toutenburg. 1977. Prediction and Improved Estimation in Linear Models. John Wiley, Chichester, UK. Bergen, W. G., and D. B. Bates. 1984. Ionophores: Their effect on production efficiency and mode of action. J. Anim. Sci. 58:1465–1483. Brown, H., N. G. Elliston, J. W. McAskill, O. A. Meunster, and L. V. Tonkinson. 1973. Tylosin phosphate (TP) and tylosin urea adduct (TUA) for the prevention of liver abscesses, improved weight gains and feed efficiency in feedlot cattle. J. Anim. Sci. 37:1085–1091. Clary, E. M., R. T. Brandit Jr., D. L. Harmon, and T. G. Nagaraja. 1993. Supplemental fat and ionophores in finishing diets: Feedlot performance and ruminal digesta kinetics in steers. J. Anim. Sci. 71:3115–3123. Dijkstra, J. 1994. Simulation of the dynamics of protozoa in the rumen. Br. J. Nutr. 72:679–699. Domescik, E. J., and S. A. Martin. 1999. Effects of laidlomycin propionate and monensin on the in vitro mixed ruminal microorganism fermentation. J. Anim. Sci. 77:2305–2312. Ellis, J. L., A. Bannink, J. France, E. Kebreab, and J. Dijkstra. 2010. Evaluation of enteric methane prediction equations for dairy cows used in whole farm models. Glob. Change Biol. 16:3246–3256. Ellis, J. L., J. Dijkstra, A. Bannink, A. J. Parsons, S. Rasmussen, G. R. Edwards, E. Kebreab, and J. France. 2011. The effect of high-sugar grass on predicted nitrogen excretion and milk yield simulated using a dynamic model. J. Dairy Sci. 94:3105–3118. Ellis, J. L., J. Dijkstra, J. France, A. J. Parsons, G. R. Edwards, S. Rasmussen, E. Kebreab, and A. Bannink. 2012. Effect of highsugar grasses on methane emissions simulated using a dynamic model. J. Dairy Sci. 95:272–285.


Monensin dose effect on rumen volatile fatty acids Ellis, J. L., J. Dijkstra, E. Kebreab, A. Bannink, N. E. Odongo, B. W. McBride, and J. France. 2008. Aspects of rumen microbiology central to mechanistic modeling of methane production in cattle. J. Agric. Sci. 146:213–233. Fandiño, I., S. Calsamiglia, A. Ferret, and M. Blanch. 2008. Anise and capsicum as alternatives to monensin to modify rumen fermentation in beef heifers fed a high concentrate diet. Anim. Feed Sci. Technol. 145:409–417. Galyean, M. L., K. J. Malcolm, and G. C. Duff. 1992. Performance of feedlot steers fed diets containing laidlomycin propionate or monensin plus tylosin, and effects of laidlomycin propionate concentration on intake patterns and ruminal fermentation in beef steers during adaptation to a high-concentrate diet. J. Anim. Sci. 70:2950–2958. Guan, H., K. M. Wittenberg, K. H. Ominski, and D. O. Krause. 2006. Efficacy of ionophores in cattle diets for mitigation of enteric methane. J. Anim. Sci. 84:1896–1906. Gyulai, F., and M. Baran. 1988. Effect of monensin on rumen ciliate protozoa in sheep. Arch. Tierernahr. 38:153–157. Habib, G., and R. A. Leng. 1987. The effects of monensin rumen fungi population in sheep given diets based on oaten chaff or wheat straw. Page 1A in Recent Advances in Animal Nutrition in Australia. University of New England, Armidale, Australia. Heinemann, W. W., E. M. Hanks, and D. C. Young. 1978. Monensin and tylosin in a high energy diet for finishing steers. J. Anim. Sci. 47:34–40. Henderson, C., C. S. Stewart, and F. V. Nekrep. 1981. The effect of monensin on pure and mixed cultures of rumen bacteria. J. Appl. Bacteriol. 51:159–169 Hino, T. 1981. Action of monensin on rumen protozoa. Japan J. Zootech. Sci. 52:171–179. Ives, S. E., E. C. Titgemeyer, T. G. Nagaraja, A. del Barrio, D. J. Bindel, and L. C. Hollis. 2002. Effects of virginiamycin and monensin plus tylosin on ruminal protein metabolism in steers fed corn-based finishing diets with or without wet corn gluten feed. J. Anim. Sci. 80:3005–3015. Johnson, K. A., and D. E. Johnson. 1995. Methane emissions from cattle. J. Anim. Sci. 73:2483–2492. Lin, L. I. K. 1989. A concordance correlation coefficient to evaluate reproducibility. Biometrics 45:255–268. Martineau, R., C. Benchaar, H. V. Petit, H. Lapierre, D. R. Ouellet, D. Pellerin, and R. Berthiaume. 2007. Effects of lasalocid or monensin supplementation on digestion, ruminal fermentation, blood metabolites, and milk production of lactating dairy cows. J. Dairy Sci. 90:5714–5725. McGuffey, R. K., L. F. Richardson, and J. I. D. Wilkinson. 2001. Ionophores for dairy cattle: Current status and future outlook. J. Dairy Sci. 84:E194–E203. Merchen, N. R., and L. L. Berger. 1985. Effect of salinomycin level on nutrient digestibility and ruminal characteristics of sheep and feedlot performance of cattle. J. Anim. Sci. 60:1338–1346. Meyer, N. F., G. E. Erickson, T. J. Klopfenstein, M. A. Greenquist, M. K. Luebbe, P. Williams, and M. A. Engstrom. 2009. Effect of essential oils, tylosin, and monensin on finishing steer performance, carcass characteristics, liver abscesses, ruminal fermentation, and digestibility. J. Anim. Sci. 87:2346–2354. Mills, J. A. N., J. Dijkstra, A. Bannink, S. B. Cammell, E. Kebreab, and J. France. 2001. A mechanistic model of whole-tract digestion and methanogenesis in the lactating dairy cow: Model development, evaluation and application. J. Anim. Sci. 79:1584–1597.

2725

Morris, F. E., M. E. Branine, M. L. Gaylean, M. E. Hubbert, A. S. Freeman, and G. P. Lofgreen. 1990. Effect of rotating monensin plus tylosin and lasalocid on performance, ruminal fermentation, and site and extent of digestion in feedlot cattle. J. Anim. Sci. 68:3069–3078. Nagaraja, T. J., C. J. Newbold, C. J. Van Nevel, and D. I. Demeyer. 1997. Manipulation of ruminal fermentation. Pages 523–632 in Rumen Microbial Ecosystem. P. N. Hobson and C. S. Stewart, ed. Chapman and Hall, London, UK. Nagaraja, T. G., Y. Sun, N. Wallace, K. E. Kemp, and C. J. Parrot. 1999. Effects of tylosin on concentrations of fusobacterium necrophorum and fermentation products in the rumen of cattle fed a high-concentrate diet. Am. J. Vet. Res. 60:1061–1065. Nagaraja, T. G., M. B. Taylor, D. L. Harmon, and J. E. Boyer. 1987. In vitro lactic acid inhibition and alteration in volatile fatty acid production by antimicrobial feed additives. J. Anim. Sci. 65:1064–1076. NRC. 2000. Nutrient Requirements of Beef Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. Prange, R. W., C. L. Davis, and J. H. Clark. 1978. Propionate production in the rumen of Holstein steers fed either a control or monensin supplemented diet. J. Anim. Sci. 46:1120–1124. Raun, A. P., C. O. Cooley, E. L. Potter, R. P. Rathmacher, and L. F. Richardson. 1976. Effect of monensin on feed efficiency of feedlot cattle. J. Anim. Sci. 43:670–677. Ralston, A. T., and T. P. Davidson. 1976. The effect of varying levels of monensin in finishing rations for beef cattle. Spec. Rep. 452. Agric. Exp. Stn., Oregon State Univ., Corvallis. Richardson, L. F., E. L. Potter, and C. O. Cooley. 1978. Effect of monensin on ruminal protozoa and volatile fatty acids. J. Anim. Sci. 47(Suppl. 1):45. (Abstr.) Richardson, L. F., A. P. Raun, E. L. Potter, C. O. Cooley, and R. P. Rathmacher. 1976. Effect of monensin on rumen fermentation in vitro and in vivo. J. Anim. Sci. 43:657–664. Rogers, J. A., and C. L. Davis. 1982. Rumen volatile fatty acid production and nutrient utilization in steers fed a diet supplemented with sodium bicarbonate and monensin. J. Dairy Sci. 65:944–952. Sauvant, D., P. Schmidely, J. J. Daudin, and N. R. St-Pierre. 2008. Meta-analysis of experimental data in animal nutrition. Animal 2:1203–1214. Strathe, A. B., A. Danfaer, A. Chwalibog, H. Sorensen, and E. Kebreab. 2010. A multivariate nonlinear mixed effects method for analyzing energy partitioning in growing pigs. J. Anim. Sci. 88:2361–2372. St-Pierre, N. R. 2001. Invited review: Integrating quantitative findings from multiple studies using mixed model methodology. J. Dairy Sci. 84:741–755. Surber, L. M., and J. G. Bowman. 1998. Monensin effects on digestion of corn or barley high-concentrate diets. J. Anim. Sci. 76:1945–1954. Tedeschi, L. O., D. G. Fox, and T. P. Tylutki. 2003. Potential environmental benefits of ionophores in ruminant diets. J. Environ. Qual. 32:1591–1602. Thomas, P. C., and P. A. Martin. 1988. The influence of nutrient balance on milk yield and composition. Pages 97–118 in Nutrition and Lactation in the Dairy Cow. P. C. Garnsworthy, ed. Butterworths, London, UK. Thorton, J. H., and F. N. Owens. 1981. Monensin supplementation and in vivo methane production by steers. J. Anim. Sci. 52:628–634. Utley, P. R., G. L. Newton, D. M. Wilson, and W. C. McCormick. 1977. Dry and propionic acid treated-high moisture corn fed with and without monensin to feedlot heifers. J. Anim. Sci. 45:154–159.


2726

Ellis et al.

van Knegsel, A. T. M., H. Van Den Brand, J. Dijkstra, W. M. Van Straalen, R. Jorritsma, S. Tamminga, and B. Kemp. 2007. Effect of glucogenic vs. lipogenic diets on energy balance, blood metabolites, and reproduction in primiparous and multiparous dairy cows in early lactation. J. Dairy Sci. 90:3397–3409. van Landeghem, G., P. Onghena, and J. Van Damme. 2006. Separability in the fixed part of multilevel models. Comput. Stat. Data Anal. 50:2902–2919. van Maanen, R. W., J. H. Herbein, A. D. McGilliard, and J. W. Young. 1978. Effects of monensin on in vivo rumen propionate production and blood glucose kinetics in cattle. J. Nutr. 108:1002–1007.

Wolin, M. J. 1960. A theoretical rumen fermentation balance. J. Dairy Sci. 43:1452–1459. Zinn, R. A. 1987. Influence of lasalocid and monensin plus tylosin on comparative feeding value of steam-flaked versus dry-rolled corn in diets for feedlot cattle. J. Anim. Sci. 65:256–266. Zinn, R. A., and J. L. Borques. 1993. Influence of sodium bicarbonate and monensin on utilization of a fat-supplemented, high-energy growing-finishing diet by feedlot steers. J. Anim. Sci. 71:18–25. Zinn, R. A., A. Plascencia, and R. Barajas. 1994. Interaction of forage level and monensin in diets for feedlot cattle on growth performance and digestive function. J. Anim. Sci. 72:2209–2215.


References

This article cites 54 articles, 26 of which you can access for free at: http://www.journalofanimalscience.org/content/90/8/2717#BIBL
