The measurement of dietary amino acid digestibility in pigs, rats and chickens: a comparison of methodologies

Abstract In a collaborative study between three centres several methods of estimating ileal digestibility of protein and amino acids were compared. There were five trials with pigs, each with three to four animals: one with T-cannulas (TC), one with postvalve ileocolic reentrant cannulas (IPVC) and three using end-to-end ileorectal anastomosis (IRA). There was one trial with 18 rats with end-to-end ileorectal anastomosis and two trials with 12 intact and 12 caecectomised cockerels. Three isonitrogenous diets were used: one based on barley (B), one based on dried milk (M) and one which was a mixture of the other two (BM). Diet B was reground for pigs with IPVC to avoid cannula blockage and also for rats because of feed refusals. The reground diet was also tested with IRA pigs. True digestibility was calculated by correction for endogenous losses using values obtained with protein-free diets. Overall mean digestibility of nitrogen and amino acids, both for apparent and true digestibility, was higher for IPVC (apparent 0.83; true 0.92) than for the other pig trials, which gave similar mean values (apparent 0.71–0.74; true 0.83–0.85). Mean values with rats were similar to these (0.75 and 0.83 respectively). Mean apparent digestibility was lowest with cockerels (0.66), but true digestibility values (0.86) were similar to those of the pigs (except IPVC) and rats. However, trials giving the same or similar mean values ranked individual diets differently. Analysis of the correlations between trials showed that the pattern of variation between trials was similar for most amino acids. In this respect, the trials with cockerels were distinctly different from those with pigs and rats, giving lower digestibility of threonine, serine and cystine. In all trials, the digestibility, both apparent and true, of Diet BM was essentially equal to that expected from the estimates with Diets B and M.