1622 Evaluation of in vitro and in situ starch digestibility assays

Abstract

The objective of this experiment was to evaluate commercial in vitro and in situ starch digestibility assays to estimate ruminal starch digestibility (RSD) and rate of ruminal starch degradation (kd). Twelve commercial dairy herds located in Wisconsin, Pennsylvania, and Missouri were used (4 free-stall and 8 tie-stall housing). Fecal samples were collected from 5 high- and 5 low-producing cows from each herd, and samples of total mixed rations (TMR) were collected after feed delivery from each pen (n = 8) or directly in front of each cow (n = 80). Corn silage (CS; n = 13) and corn (dry, n = 9, and high-moisture, n = 7) samples were also collected from each farm. Feed samples were thoroughly mixed and split for analysis. Fecal and TMR samples were sent to Rock River lab for starch analysis. Total tract starch digestibility (TTSD) was calculated from TMR and fecal starch content. The following equation was used to estimate RSD from TTSD: y = 82.224 + (0.185 × ruminal), in which y = TTSD (Ferraretto et al., 2013). Samples of TMR, corn, and CS were sent to Rock River for 7-h in situ starch digestibility and to Dairyland lab for analysis of nutrient composition and 7-h in vitro starch digestibility. All herds tested milk through Dairy Herd Improvement Association and individual milk yield, milk fat, and milk protein content data were collected from test day closest to day of sample collection. The REG procedure of SAS was used to determine the relationship between in situ, in vitro, and in vivo RSD. In situ starch digestibility was related to in vivo RSD (R2 = 0.19, P < 0.002). There was no relationship between in vitro and in vivo RSD (R2 = 0.01, P = 0.45). The Cornell Net Carbohydrate and Protein System (CNCPS) uses ruminal starch kd to predict microbial protein production. Actual milk yield was related to model-predicted milk yield using model default kd for RSD (R2 = 0.69, P < 0.0001). Ruminal starch kd was calculated for CS and corn using in situ data and entered into the CNCPS model (version 6.5; AMTS, LLC). This relationship was improved when measured in situ starch kd was entered for corn and CS (R2 = 0.76, P < 0.0001). In conclusion, in situ starch digestibility at 7 h is a good approach for estimating RSD in vivo and using kd from this method improves milk yield predictions from the CNCPS model.