Technical note: Methodological and feed factors affecting measurement of protein A, B, and C fractions, degradation rate, and intestinal digestibility of rumen-undegraded protein.

When formulating dairy cow rations, characterization of protein in feeds requires estimation of protein degradation in the rumen and digestion in the intestine. The objective of this work was to evaluate experimental and feed-related factors that affect characterization using in situ, in vitro, or mobile bag techniques, of 0-h washout (A), potentially degradable (B), and undegradable (C) protein fractions, protein degradation rate (Kd ), and digestibility of rumen undegradable protein (dRUP). Data sets of 136 studies on A, B, C, and Kd and 113 studies on dRUP were amassed from the literature. Mixed-effect linear models were used to relate these variables to methodological and feed factors while accounting for random differences among studies. Predictions of A, B, and C protein fractions were significantly influenced by crude protein and neutral detergent fiber interactions with bag pore size, incubation time, bag area, and sample-to-bag area ratio. For example, a 20.0% decrease in crude protein of a theoretical legume silage sample would increase A fraction prediction by 20.1%, but 34.7% with bag incubation time -1 standard deviation below the mean. Similarly, reported Kd values were significantly influenced by crude protein interactions with bag area and sample-to-bag area ratio and by neutral detergent fiber interaction with pore size. Feed variables and measurement variables influencing protein digestibility measures suggest that these analytical factors are likely associated with variance among differing methodologies and within unique samples of the same feed. When predicting dRUP, the use of mobile bag method produced significantly different estimates compared with the in vitro 3-step method. The use of mobile bag resulted in an 8.9% (±3.8%) higher estimate of dRUP compared with the in situ technique. In 618 and 977 samples, sample variation to sample mean ratio for acid detergent fiber and pepsin-acid incubation time was 63.0 and 58.0%, respectively. Variation in feedstuff content and lack of standardization of methods used to measure protein disappearance led to a lack of robustness in the measurements commonly employed.