A CRITICAL LOOK AT ESTIMATES OF APPARENT DIGESTIBILITY OF PROTEIN AND AMINO ACIDS  

Information of the apparent digestibility coefficient (ADC) of nutrients of different ingredients is increasing every year thanks to sustained research efforts. Estimates of ADC are regularly compiled in the reference literature and increasingly used by feed manufacturers who are now formulating their feeds on a digestible protein and amino acid basis.  This progressive move from formulating on a 'total nutrient' basis to formulating on digestible nutrients is praiseworthy. However, increasing reliance by feed millers on published estimates of ADCs makes it critical to ensure that the information available is relevant and reliable.

ADCs of protein and amino acids are reported for a large number of ingredients fed to different aquaculture species. Significant variability is observable across studies and species. Some of the data suggest significant differences in ADCs of protein among species for the same ingredient. Differences in methodology and intrinsic variability in the composition of different batches of the same ingredient make it difficult to meaningfully compare ADCs across studies, let alone across animal species. A critical review of results from digestibility trials suggest that methodological shortcomings (e.g. marker analysis) and calculation errors (biased equations) result in published estimates of ADC that are not always highly reliable. There is frequently a need to examine the original data from trials to be able to identify sources of errors but this is rarely feasible for published studies.

The University of Guelph Fish Nutrition Research Laboratory (UG-FNRL) has been a pioneer in the assessment of the digestibility of fish feed ingredients. Cho and Slinger (1974) published reliable estimates of ADC of protein and energy to rainbow trout of practical ingredients. The sustained use of the same equipment and methodology by the UG-FNRL has highlighted significant improvements in ADC of protein of some ingredients over the past 40 years. These efforts highlighted significant variability in ADCs of protein and amino acids across different lots or batches of the same type of ingredient.  These efforts also indicated that digestibility is a measure of "disappearance" of nutrients but not a direct measure of bio-availability of nutrients. Consequently, ADC of nutrients should ideally be corroborated by more direct assessment of the bioavailability of nutrients of ingredients. The processing and drying equipment and conditions used in the production of ingredients were observed to have a determinant effect on the digestibility of protein. An in vitro assay examining the susceptibility of protein to degradation by intestinal enzymes suggests the differences in ADC of protein observed for similar ingredients have a rational biochemical basis. However, the chemical determinants of digestibility are not well understood and characterized.  Recent research efforts suggest that the level of protein cross-links present in feed ingredients may be strongly correlated with their digestibility of protein and bio-availability of amino acids. Protein crosslinks are either naturally occurring in proteins or formed as a results of heat or chemical treatment/damage. Protein crosslinks are mostly found as 1) disulfide crosslinks involving thiol-disulphide interchange reactions, and 2) amino acid cross-links derived from reactions between certain amino acids and chemical compounds. A simple processing technique targeting the disruption of disulphide cross-links was highly effective in improving the nutritional quality of relatively poorly digestible protein ingredients. Better understanding of the chemical determinants of protein digestibility should enable the development of rapid methods and tools to more reliably estimate the digestibility and nutritive value of protein ingredients.