Total volatile fatty acid concentrations are unreliable estimates of treatment effects on in vivo ruminal fermentation

Monday, July 21, 2014: 10:00 AM
2103B (Kansas City Convention Center)
Mary B. Hall , U. S. Dairy Forage Research Center, USDA-ARS, Madison, WI
Tamilee D. Nennich , Purdue University, West Lafayette, IN
Perry H. Doane , ADM Research, Decatur, IL
Abstract Text: The sum of ruminal acetate, propionate, butyrate, and valerate concentrations ([VFA], mM) are used to assess impact of dietary treatments on pattern of carbohydrate fermentation, typically with intent to indirectly relate microbial products to nutrient supply to the cow. However, discrepancies in statistical results of treatment effects between [VFA] and VFA pool size (VFAmol) within one study suggested there were problems with this approach. We investigated relationships among [VFA], VFAmol, and rumen digesta liquid kg (LIQ) measured 2 h post-feeding using individual lactating cow data (175 observations) measured in 7 separate feeding trials. Regression analyses were performed using mixed models with “trial” as a discrete random variable; dependent (Y) and independent (X) variables are in Table 1. Correlations (r) are the mean of individual trial values. Across studies, [VFA] had a numerically smaller within-study coefficient of variation (13%) than did VFAmol (23%). Rumen liquid increased with VFAmol. Change in LIQ was likely a function of water flux in the rumen based on the osmotic gradient of rumen liquid and blood; VFAmol represents a large ruminal pool of solute. Rumen liquid and [VFA] were poorly correlated. Concentration of VFA increased with VFAmol. The ratio of [VFA]/VFAmol ranged from 9.0 to 24.1 and changed as a function of 1000/LIQ. The equation reflects the inherent relationship among the 3 variables, and that [VFA] at differing LIQ differ in their relationship to VFAmol.  Mean within-trial LIQ was 73 kg (standard deviation=11.2), with an average within-study range of 43 kg. Assumption of equivalent ruminal liquid volumes is incorrect. Occupying variable LIQ, [VFA] are not on the equivalent basis needed for comparison of treatments. Alternate approaches must be developed to appropriately use [VFA] to assess treatment effects.

Table 1. Relationships between rumen measures.1

Y X Intercept (SE) Slope (SE) p-value of X Average r
[VFA] VFAmol 83.9 (5.1) 5.83 (0.38) <0.01 0.74
LIQ VFAmol 32.8 (2.15) 3.68 (0.18) <0.01 0.83
[VFA] LIQ 113 (10.9) 0.473 (0.124) <0.01 0.26

1 Y = dependent variable, X = independent variable, SE = standard error, r = correlation coefficient, [VFA] = VFA concentration, mM, VFAmol = moles of VFA, LIQ = rumen liquid, kg.

Keywords: rumen, VFA, liquid