Abstract #T235

# T235
Replacing cereal grains starch with non-forage fiber in diets of dairy cows: A meta-analysis.
Juan I. Sanchez-Duarte*1,2, Kenneth F. Kalscheur3, 1Dairy Science Department, South Dakota State University, Brookings, SD, 2INIFAP, Matamoros, Cohuila, Mexico, 3USDA-ARS, U.S. Dairy Forage Research Center, Madison, WI.

A meta-analysis was conducted to evaluate the effect of reducing cereal grains starch with non-forage fiber sources (NFFS) on the productive response of lactating dairy cows. Thirty-nine studies were selected that used NFFS to replace a portion of starch from cereal grains. Studies reported cow performance, rumen fermentation, or total-tract nutrient digestibility. Data were analyzed through stepwise regression analysis by the mixed effect models of R using the study as a random effect. The variance explained by the models was evaluated calculating marginal R2(m) and conditional R2(c). As dietary starch intake (kg/d) increased, DMI (Y = −0.08x2+1.11x+19.15) and milk fat yield (Y = −0.005x2+0.05x+1.07) responded quadratically; milk yield (Y = 0.34x+31.87), milk protein concentration (Y = 0.02x+3.01), and milk lactose yield (Y = 0.02x+1.46) increased linearly; and milk fat concentration decreased linearly (Y = −0.05x+3.94). DMI was explained by the quadratic effect of DM digestibility (P = 0.04) as starch intake increased. Increases of milk yield and milk protein concentration was explained by the linear increase of rumen VFA concentrations (propionate, acetate to propionate ratio, isobutyrate, isovalerate, and valerate; P < 0.05) and CP digestibility (P = 0.01) in response to the increased starch intake. Decreased milk fat concentration is an effect of the reduction in the rumen acetate concentration (P < 0.001) and NDF digestibility (P < 0.001). Values of R2(m) and R2(c) indicated a better goodness-of-fit for significant models. Residuals and Q-Q plots of the models were symmetrical and their errors were normally distributed. Additionally, intake of DM, CP and NDF, as well as DIM contributed to the variation of the models. It is important to highlight that diets formulated with NFFS (25.9 ± 11.5% on a DM basis) had lower DMI, milk yield, and milk protein concentration than cows fed diets high in cereal grains (27.5 ± 11.5% as DM basis). Cows fed diets higher in NFFS may result in greater milk fat percentage. Therefore, all those factors should be taken into account when NFFS are used to reduce starch from cereal grains in lactating dairy cow diets.

Key Words: meta-analysis, starch, milk