Nitrogen and amino acid ileal and faecal digestibility of rabbit feeds predicted by an in vitro method

摘要

Predicting the amino acid digestibility of ingredients is one of the main challenges in animal nutrition, due to its effect on animal performance, gut health and the environment. The aim of this work was to predict in vivo ileal and faecal digestibility of nitrogen and amino acids by using an in vitro method. Eleven ingredients (soybean meal 480, toasted full-fat soybean, sunflower meal 280 and 360, peas, wheat, wheat shorts, corn, gluten feed, dehydrated alfalfa and soybean hulls), commonly used in rabbit diets, were simultaneously evaluated in vivo, and subjected to a three step enzymatic in vitro digestion. Dry matter (DM), nitrogen (N) and amino acids (AA) were analysed in the residues in order to determine their in vitro digestibility (NDiv, AAD(iv)). The correlation among in vivo apparent faecal and ileal, and true ileal digestibilities (AFD, AID and TID, respectively) of N and AA of the ingredients and the corresponding in vitro values were analysed using a regression procedure. The average NDiv was 0.875 and ranged from 0.764 of soybean hulls to 0.946 of full-fat soybean, AADiv showed similar ranges of variation and their average values varied with the AA (from GLY and CYS: 0.83 to GLU and ARG 0.92). Values of NDiv were higher than the corresponding in vivo values (0.21, 0.10 and 0.05 points as average for AID, AFD and TID, respectively), but the correlation among them was relatively high (averaged 0.77, 0.76 and 0.65, respectively). An average of 72% of the variation in AA ileal digestibility was explained by the NDiv variation, but the latter only explained 55% of AFD AA variation. The accuracy of the prediction was similar among units, its RSD averaged 0.074. Methionine was the limiting amino acid most precisely predicted, as opposed to CYS and THR, which were predicted with the lowest accuracy. In conclusion, NDiv can help to discriminate among feedstuffs that are more or less digestible at ileal level better than chemical analysis. (C) 2016 Elsevier B.V. All rights reserved.