Net portal appearance used to assess feed evaluation system predictions of the digestive flow and gut metabolism of essential amino acids in dairy cows: A meta-analysis.

Accurate prediction of intestinal digestible flows of EAA (DigFlow) is a crucial step for adequately balancing EAA supply to reduce the amount of CP fed to dairy cows, without compromising either milk production or the health status of the animals. The objective of this meta-analysis was to compare the performance of 3 dairy feed evaluation systems (FES) to predict DigFlow (NRC; National Academies of Sciences, Engineering and Medicine [NASEM]; and Cornell Net Carbohydrate and Protein System version 6.5.5 [CNCPS]) in relation to observed net portal appearance (NPA) of EAA. Although the DigFlow cannot be measured directly and none of the FES were designed to estimate the net portal appearance (NPA) of EAA, NPA could be predicted (NPA) from estimations of DigFlow minus the estimated metabolic fecal protein (MFP). This NPA, however, would not include any oxidation for EAA or synthesis for Arg, a semi-EAA, occurring during the absorption process. Intuitively, the prediction errors should be smallest for the DigFlow best predicted across the 3 FES and for the EAA not oxidized by the portal-drained viscera (PDV). The dataset included 83 NPA treatment means from 25 studies. To avoid type I error, mean and linear biases were considered biologically relevant if statistically significant and representing >5.0% of the observed mean (%). The NPA of branched-chain AA (BCAA) and Thr showed a similar pattern across the 3 FES, with all overpredicted relative to NPA (6% to 27%). The NPA for the other EAA were more variable: (1) underprediction of Arg (9% to 20%) with NRC and NASEM, and Met (8%) with NRC, (2) overprediction of Lys and Phe (5% to 11%) with NASEM, and His, Met and Trp (8% to 14%) with CNCPS, and (3) linear biases for Arg (7%) with NASEM, and for His, Lys, Met, and Phe (5% to 14%) with CNCPS. In our previously reported meta-analysis, the mean and linear biases between observed and predicted EAA postruminal outflow were related to the site of digesta sampling (duodenal vs. omasal) for some EAA, but it was not possible to determine which sampling site was representative of the true supply of EAA to the cows. In an effort to solve this issue, predictions of NPA (except Trp) were recalculated removing the mean and linear biases observed in duodenal and omasal studies of our previous meta-analysis (NPA and NPA). Compared with NPA, the pattern of BCAA and Thr remained overpredicted in NPA and NPA across the 3 FES. This strongly suggests oxidation of Ile, Leu, Val, and Thr by the PDV averaging, respectively, 12% (range 6% to 18%), 16% (range 14% to 17%), 24% (range 18% to 27%), and 19% (range 13% to 23%) of NPA across the 3 FES. The magnitude of PDV oxidation, however, would be related to the ratio of DigFlow to digestible energy intake. The negative mean biases for Arg NPA would suggest synthesis by the PDV. For the BCAA, Thr, and Arg, because of the uncertainty on the biological quantification of the differences between NPA and NPA, the relative performance of each FES could not be assessed. Assuming no oxidation of the other EAA (i.e., His, Lys, Met, and Phe) by the PDV and an adequate estimation of MFP, their NPA indicate that the DigFlow of (1) Met is underpredicted by NRC, (2) Lys is overpredicted by NASEM, (3) His and Met are overpredicted with CNCPS, and (4) His, Lys, Met, and Phe present a linear bias with CNCPS, supporting similar observations on the postruminal EAA outflows from our previous meta-analysis. Based on current NPA studies and our previous meta-analysis on postruminal outflows of EAA, sampling digesta at the duodenum appears to be more representative of the true supply of EAA to the cows than sampling at the omasum.