Amino acid bioavailability and the digestibility assay

As outlined by Fuller⁽Reference Fuller⁴⁾, the bioavailability of a dietary AA can be defined as the proportion of the AA intake that is absorbed in a chemical form that can be utilized for metabolic processes such as protein synthesis⁽Reference Stein, Seve and Fuller⁵⁾. Accurate estimates of the bioavailable AA content of feeds and foods allow for greater precision when matching AA intake with AA requirements. However, currently no direct measures of bioavailability are available. Measures of AA digestibility are used widely to estimate AA bioavailability in feed and food ingredients for monogastrics, with the terms digestibility and bioavailability sometimes being used interchangeably and incorrectly. Slope-ratio assays are still considered the benchmark for determining (relative) bioavailability of AA in feeds and foods. This method involves supplying graded dietary levels of an individual AA from a particular protein source and measuring a response variable, such as whole body protein retention or AA oxidation⁽Reference Fuller and Tome^1, Reference Stein, Seve and Fuller⁵⁾. Generally, the response to the test ingredient is compared with the response to a reference protein in which the AA bioavailability is known or is assumed to be 100 %, which yields relative estimates of bioavailability⁽Reference Stein, Seve and Fuller⁵⁾. The slope-ratio assay is a time consuming, expensive, and laborious method that only provides a measure of bioavailability for one AA at a time. Moreover, estimates of bioavailability obtained with the slope-ratio assay may be confounded with endogenous AA losses, dietary nutrient imbalances, leading to incorrect estimates of bioavailability⁽Reference Langer and Fuller⁶⁾ and may lack precision and/or repeatability⁽Reference Sauer and Ozimek⁷⁾. It should be noted that when absorbed AA are all available for metabolism, bioavailability estimates obtained with this assay more closely represent apparent digestibility than true digestibility, as it reflects the net uptake of AA from the digestive tract. To evaluate bioavailability, as defined earlier, measures of true ileal digestibility combined with chemical availability are required. In principle, the digestibility assay is simpler than in vivo assessments of AA bioavailability and allows for simultaneous estimation of the bioavailability of all AA in an experimental diet or an individual diet ingredient. However, measures of digestibility are based on AA disappearance from the digestive tract and do not reflect the form in which AA and AA metabolites are absorbed⁽Reference Stein, Seve and Fuller⁵⁾.

Apparent faecal digestibility coefficients

Apparent faecal, or total tract, digestibility (AFD) is the most basic in vivo method for estimating AA bioavailability and reflects disappearance of AA along the entire gastrointestinal tract. The faecal digestibility method has long been considered an inappropriate approach to determining AA bioavailability for monogastric animals, largely because of enteric fermentation in the hindgut⁽Reference Sauer and Ozimek⁷⁾.

Undigested protein that enters the hindgut is susceptible to microbial degradation via deamination with carbon skeletons used as an energy source⁽Reference Vince, Dawson and Park⁸⁾. Studies have shown that protein, both of dietary and endogenous origin, is the major substrate for intestinal ammonia production⁽Reference Wrong, Vince and Waterlow^9–Reference Mouille, Robert and Blachier¹²⁾. Using an infusion of isotopic urea, it was determined that the contribution of urea hydrolysis to faecal ammonia in humans is less than 10 % with the remainder derived from fermentation of undigested dietary and endogenous protein⁽Reference Wrong, Vince and Waterlow^9, Reference Wrong and Vince¹⁰⁾. It has also been shown that an increase in protein flow into the hindgut results in an increase in digesta ammonia concentration⁽Reference Mouille, Robert and Blachier¹²⁾ providing further evidence that protein is the major substrate for gut ammonia production. Ammonia produced in the hindgut is readily absorbed by the animal, transported to the liver for ureagenesis, and excreted largely in urine. Therefore, the majority of nitrogen disappearance from the hindgut is likely due to fermentative losses and is of limited nutritional value to the animal.

Some studies, however, suggest the hindgut is not entirely incapable of AA absorption. Heine et al. ⁽Reference Heine, Wutzke and Richter¹³⁾ demonstrated that when labelled yeast was infused into the colon of human infants, the majority of absorbed label (i.e. not excreted in faeces) was retained in body protein (i.e. not excreted in urine). The authors used these observations to conclude that the colon is capable of absorption of intact AA that can be utilized for metabolic processes. However, the ¹⁵N-protein infused in this study may have been absorbed as non-protein nitrogen (i.e. ammonia) and utilized for endogenous production of dispensable AA or utilized in the upper gut for indispensable AA synthesis. The presence of AA transporters, such as the ATBo+ and B^o+ transport systems, have been demonstrated in the colonic mucosa⁽Reference Blachier, Mariotti and Huneau¹⁴⁾ indicating the capacity for AA absorption. In addition, uptake of AA through the colonic mucosa has been demonstrated⁽Reference James and Smith^15, Reference Smith and James¹⁶⁾. However, it should be noted that measurement of AA uptake in these studies was done in isolated colonocytes grown in culture or on mounted sections of tissue and does not necessarily indicate what would occur in vivo. In addition, these studies were performed in newborn pigs in which it has been demonstrated that gut structure and function differs from older pigs. Indeed, it was shown that the ability of colonic tissue to transport AA was reduced in 4-day-old pigs and virtually eliminated in 10-day-old pigs⁽Reference James and Smith¹⁵⁾. This would suggest that the ability of the colon to absorb AA is lost quickly and would not be a significant factor in the growing pig and humans, except perhaps for young infants.

The inadequacy of using AFD coefficients to estimate AA bioavailability is further supported by a large number of studies in which protein or individual AA have been infused into the hindgut of pigs with little or no retention of the infused AA⁽Reference Zebrowska^17–Reference Wunsche, Hennig and Meinl²²⁾. In several of these studies statistically insignificant improvements in whole body nitrogen retention were observed, suggesting that some of the experimental observations may not have been sufficiently sensitive to identify retention of AA that were infused into the hindgut⁽Reference Moughan²³⁾. However, in carefully controlled studies, and with AA that were limiting in the diet, infusion of AA into the hindgut of pigs did not improve whole body AA retention. For example, Darragh et al. ⁽Reference Darragh, Cranwell and Moughan²⁴⁾ infused lysine or methionine into the proximal colon of young pigs fed either lysine or methionine deficient diets with the assumption that nitrogen balance would be improved if the dietary indispensable AA were absorbed intact thus correcting the diet deficiency. However, no improvement in nitrogen balance was observed; in fact nitrogen balance was somewhat reduced when compared to pigs receiving additional lysine or methionine orally. Wünsche et al. ⁽Reference Wunsche, Hennig and Meinl²²⁾ performed a similar study examining lysine and isoleucine infusion into the colon of animals receiving lysine or isoleucine deficient diets. Again it was found that while faecal digestibility was not affected by the infusion, nitrogen balance was decreased and urinary ammonia increased in these animals⁽Reference Wunsche, Hennig and Meinl²²⁾. In another study, nitrogen digestibility and retention was compared in pigs in which a casein supplement was given orally or infused into the terminal ileum. Total tract nitrogen digestibility from the hindgut was decreased only slightly when casein was administered ileally, suggesting that nitrogen was being absorbed from the hindgut, but there was a significant reduction in nitrogen retention in these pigs [Zebrowska et al. (1978) as referenced in Buraczewski⁽Reference Buraczewski²⁵⁾], which indicates that nitrogen absorption from the hindgut was of no nutritional significance to the animal. In a review of studies in which casein was infused into the caecum of pigs, Fuller & Reeds⁽Reference Fuller and Reeds²⁶⁾ found that even though faecal digestibility was not different with casein infusion, indicating that nitrogen was being absorbed from the hindgut, there was little or no improvement in whole body nitrogen balance. This is consistent with the observation that nitrogen that is absorbed from the hindgut contributes primarily to urinary nitrogen excretion. Indeed, while it was shown that approximately 18 % of total nitrogen absorption occurred in the colon of pigs⁽Reference Krawielitzki, Zebrowska and Kreienbring^27, Reference Krawielitzki, Zebrowska and Schadereit²⁸⁾, the absorption of labelled nitrogen from the colon was entirely accounted for by label excretion in urine⁽Reference Krawielitzki, Zebrowska and Schadereit²⁸⁾.

The AA profile of faecal protein, and therefore AA digestibility, is extensively modified by enteric microbes, which can be attributed to both fermentative AA catabolism and de novo synthesis of microbial protein. As a result the AA profile of faecal protein bears little resemblance to the dietary AA profile and even to the AA present in ileal digesta. When comparing ileal and faecal digestibility values for lysine in different feed ingredients, Rademacher et al. ⁽Reference Rademacher, Mosenthin and Sauer²⁹⁾ demonstrated that there can be significant differences between the two measures and that faecal digestibility is not always the greater value. This demonstrates the combined effects of protein fermentation and microbial synthesis of AA in the hindgut of pigs and their effects on estimates of AA digestibility. The faecal digestibility method is therefore considered inappropriate with respect to determining AA bioavailability.

Only in two scenarios can ammonia absorbed from the hindgut provide a benefit in supplying nitrogen to the host⁽Reference Heger^30, Reference Reeds³¹⁾. Firstly, when the amount of absorbed AA is insufficient to satisfy the host's demand for dispensable AA, ammonia may be used in transamination reactions yielding dispensible AA⁽Reference Heger³⁰⁾. Secondly, urea that is generated from absorbed ammonia may be secreted, or recycled, into the gut and be used by microbes in the upper gut for the microbial production of indispensable and dispensable AA. Like dietary AA, microbially produced AA may be absorbed from the upper gut and can contribute to the supply of AA to the host. The latter has been shown to occur in pigs⁽Reference Torrallardona, Harris and Fuller^32–Reference Torrallardona, Harris and Milne³⁴⁾, humans⁽Reference Tanaka, Kubo and Shiraki^35–Reference Metges³⁸⁾, and rats⁽Reference Torrallardona, Harris and Coates³⁹⁾. It has been estimated that about 30 % of non-protein nitrogen that is generated in the pig's body is recycled into the gut and, therefore that only 70 % of the whole body non-protein nitrogen flux is excreted in urine⁽Reference Fuller and Reeds²⁶⁾. It can thus be speculated that recycling of non-protein nitrogen is an important contributor to nitrogen economy and impacts the host's utilization of indispensable AA and total nitrogen. It follows then that nitrogen disappearance from the hindgut due to protein fermentation, and thus AFD, may in some scenarios be important for meeting the total nitrogen needs of the animal. However, for both of these scenarios the efficiency of capturing ammonia that is generated in the hindgut, and therefore the net contribution of protein fermentation in the hindgut to the host's supply of dispensable and indispensable AA, remains to be quantified⁽Reference Fuller and Reeds²⁶⁾.

Apparent ileal digestibility coefficients

Amino acid disappearance from the gut proximal to the terminal ileum, reflected in apparent ileal digestibility (AID), is considered a better estimate of AA bioavailability than AFD⁽Reference Sauer and Ozimek^7, Reference Moughan²³⁾. The AID of AA largely represents enzymatic protein digestion and absorption of intact AA with some influence from microbial fermentation in the upper gut⁽Reference Stein, Fuller and Moughan⁴⁰⁾. In order to determine AID coefficients digesta samples must be obtained from the distal ileum. The latter requires rather invasive sampling methods, involving the use of surgically modified subjects and digestibility markers. Practical aspects of sampling ileal digesta from animals and humans are addressed by Fuller⁽Reference Fuller⁴⁾. Caution must be used when using protein or nitrogen digestibility as an indicator of AA digestibility since there can be considerable differences between the digestibility values for individual AA⁽Reference Moughan and Smith^41, Reference Sauer, Just and Jorgensen⁴²⁾.

In pigs several studies have shown that AID values provide more accurate estimates of AA bioavailability than AFD⁽Reference Rademacher, Mosenthin and Sauer^29, Reference Moughan and Smith^41, Reference Tanksley and Knabe^43–Reference Just, Jorgensen and Fernandez⁴⁵⁾ and that AID is a more sensitive assay than AFD for identifying differences in nutritional value among protein sources⁽Reference Sauer, Stothes and Parker^46–Reference Van Weerden, Huisman and van Leeuwen⁴⁸⁾. For example, Low et al. ⁽Reference Low, Partridge and Keal⁴⁴⁾ and Just et al. ⁽Reference Just, Jorgensen and Fernandez⁴⁵⁾ showed that growth performance and body protein gain of pigs is more highly correlated with ileal rather than faecal protein digestibility. Just et al. ⁽Reference Just, Jorgensen and Fernandez⁴⁵⁾ determined AID and AFD of AA and protein in 24 diets and found that nitrogen retention was slightly more correlated with ileal rather than faecal digestible daily crude protein intake (0·81 vs. 0·75). However, AID and AFD intake of AA and nitrogen are highly confounded across diets in this study and the dietary AA balances vary among the diets and, therefore, these findings should be interpreted with caution. In contrast, Dierick et al. ⁽Reference Dierick, Vervaeke and Decuypere⁴⁹⁾ compared feed efficiency in 8 diets formulated to contain similar levels of faecal digestible protein, total dietary lysine, and ratios among key indispensable AA but to differ in ileal digestible nitrogen content. It was determined that there was a higher correlation between feed efficiency (0·87 vs. 0·65) or growth rate (0·76 vs. 0·34) and ileal nitrogen digestibility vs. faecal nitrogen digestibility. Moreover, it was shown that pig performance suffered as ileal digestibility of protein decreased⁽Reference Dierick, Vervaeke and Decuypere⁴⁹⁾. Moughan & Smith⁽Reference Moughan and Smith⁴¹⁾ and Rademacher et al. ⁽Reference Rademacher, Mosenthin and Sauer²⁹⁾ showed that AID values accurately predicted AA bioavailability. Rademacher et al. ⁽Reference Rademacher, Mosenthin and Sauer²⁹⁾ formulated diets with decreasing amounts of AID lysine by replacing soyabean meal with canola meal, meat meal, alfalfa meal, beet pulp, and wheat bran in which the AID of lysine was previously determined. It was shown that as the amount of AID lysine was decreased in the diets the growth performance of the pigs decreased confirming the validity of using AID values for estimating lysine bioavailability⁽Reference Rademacher, Mosenthin and Sauer²⁹⁾. Likewise, Moughan & Smith⁽Reference Moughan and Smith⁴¹⁾ determined the AID of lysine in barley, meat and bone meal, fish meal, and pea meal and then compared the growth performance and carcass measures in pigs fed a diet based on all the protein sources to a casein based diet formulated to have the same level of AID lysine as the barley diet. Mough & Smith⁽Reference Moughan and Smith⁴¹⁾ found no difference between any of the measured responses between the two diets and concluded that the AID lysine in each feedstuff could be used to accurately formulate and predict the AID lysine content of a complete diet.

The benefit of using AID coefficients is also demonstrated when comparing the differences in protein quality among and within foods, feeds, and dietary protein sources. For example, Sauer et al. ⁽Reference Sauer, Kennelly and Aherne⁴⁷⁾ showed that unlike AFD, AID coefficients differed among four sources of barley. In a study by Holmes et al. ⁽Reference Holmes, Bayley and Leadbeater⁵⁰⁾ comparing AID and AFD of nitrogen and AA between soyabean meal and rapeseed meal in pigs, it was found that while AID values for soyabean meal correlated well with AFD values, this was not the case for rapeseed meal. Moreover, the difference between nitrogen digestibilities at the faecal level for these two protein sources was much less than observed at the ileal level. On average, Austic⁽Reference Austic⁵¹⁾ reported that AFD is 6·5 percentage units higher than AID and that this discrepancy becomes greater (ileal digestibility lower and/or faecal digestibility higher) for low digestibility protein sources. This is likely a combination of decreased digestibility and increased excretion of microbial protein when high-fibre feedstuffs are included in diets. Sauer et al. ⁽Reference Sauer, Kennelly and Aherne⁴⁷⁾ showed that measures of ileal digestibility are more sensitive than faecal digestibility and can be used to identify the impact of diet processing on the nutritional value of protein sources.

It should be noted that in some instances AID coefficients overestimate AA bioavailability, especially for lysine in heat-processed protein sources, for threonine in diets containing large amounts of fermentable fibre, and when large endogenous gut protein losses are incurred. This will be addressed in further detail in a subsequent section of this review.

A critical issue with AID coefficients is the curvilinear increase in AID of AA with increasing dietary AA content, which occurs even when the dietary AA content is altered by varying the inclusion level of a particular protein source in the diet (Fig. 1)⁽Reference Fan, Sauer and Hardin⁵²⁾. This variability in AID coefficients for individual protein sources raises concerns about the impact of protein levels used in experimental diets to establish AID coefficients and about the lack of additivity of AID coefficients among food or feed ingredients. By considering the impact of endogenous gut protein losses on AID coefficients these concerns can largely be overcome.

Fig. 1

Impact of dietary methionine level on observed apparent ileal digestibility (AID) and estimated standardized ileal digestibility (SID) of methionine in growing pigs [adjusted from Fan et al. ⁽Reference Fan, Sauer and Hardin⁵²⁾; variable dietary methionine levels are the result of varying the level of soyabean meal across diets; SID coefficients were established by subtracting basal endogenous ileal methionine loss from total ileal methionine flows for each measurement of AID of methionine (Fig. 2)].

Endogenous gut amino acid losses and ileal digestibility coefficients

Measures of AID of AA are confounded with endogenous gut AA losses (EAAL). The EAAL represent AA and synthesized proteins that are secreted into the intestine and are not reabsorbed prior to the terminal ileum, and originate from saliva, gastric secretions, bile, pancreatic juice, digestive enzymes, sloughed epithelial cells, and mucins⁽Reference Nyachoti, de Lange and Schulze^53–Reference Tamminga, Schulze and Van Bruchem⁵⁵⁾. The amount of protein from endogenous origin may exceed the amount of ingested dietary protein and is, just like dietary protein, partly digested and absorbed. Since endogenous secretions cannot be distinguished from dietary proteins in conventional digestibility studies, ileal digestibility coefficients that are not corrected for EAAL should be referred to as AID coefficients⁽Reference Stein, Seve and Fuller⁵⁾.

The curvilinear increase in AID of AA when increasing the dietary AA content (Fig. 1) can largely be attributed to the relatively large contribution of EAAL to total ileal AA flows at low dietary AA levels⁽Reference Fan, Sauer and Hardin⁵²⁾. This is further illustrated in Fig. 2, in which the AID coefficients for methionine that are provided in Fig. 1 are used to calculate total ileal methionine flow per unit of feed DM intake. Figure 2 shows a highly linear relationship between total ileal methionine flow and the dietary methionine level. Similar highly linear relationships between ileal AA flow and dietary AA level have been shown for other indispensable AA and in studies in which five or more dietary protein levels have been evaluated⁽Reference Fan, Sauer and Hardin^52, Reference Eklund, Mosenthin and Piepho⁵⁶⁾. The impact of dietary AA level on AID of AA (Fig. 1) can thus be attributed fully to the intercept of the relationship between total ileal AA flow and dietary AA level (Fig. 2). Furthermore, the slope of the linear relationship between total ileal AA flow and dietary AA level represents an important measure of digestibility. Historically, it was assumed that the increase in total ileal AA flow with increasing dietary AA level could be attributed entirely to an increased flow of undigested dietary AA. Hence this slope was used to derive estimates of ‘true’ ileal AA digestibility. However, by using more advanced techniques, which allow for separate identification of endogenous and dietary AA in ileal digesta, it has been shown that increases in total ileal AA flows can generally be attributed to increase in flow of both EAAL and dietary AA⁽Reference Nyachoti, de Lange and McBride⁵⁴⁾. Therefore, at least two components of EAAL can be identified: (1) the basal EAAL, represented by the intercept in Fig. 2, and (2) ingredient specific EAAL. Basal EAAL may be related to DM intake, which is the approach that has been adopted widely in animal nutrition⁽Reference Jansman, Smink and van Leeuwen⁵⁷⁾, or to the host's metabolic state (e.g. some measure of metabolic body weight), including the host's (gut) health status. The latter may be more appropriate for mature individuals or when levels of food intake are low. This is the case in humans where food intake can be considered low relative to the levels of feed intake observed in highly productive farm animals. The specific EAAL are induced when increasing amounts of different protein sources, fibre, or other EAAL inducing compounds (e.g. anti-nutritional factors such as tannins, lectins, trypsin inhibitors) are included in the diet⁽Reference Nyachoti, de Lange and McBride^54, Reference Moughan⁵⁸⁾.

Fig. 2

Impact of dietary methionine level on total ileal flow of methionine in growing pigs [adjusted from Fan et al. ⁽Reference Fan, Sauer and Hardin⁵²⁾; variable dietary methionine levels are the result of varying the level of soyabean meal across diets; the basal endogenous ileal methionine loss (0·1289 g/kg DM intake) was estimated based on linear regression analyses and extrapolation to feeding a methionine-free diet; the slope of the line of best fit (0·066) yields an estimate of SID of methionine, which is equivalent to the horizontal line representing estimated SID of methionine in Fig. 1: SID = 100 × (1 − slope)].

To differentiate between ileal digestibility coefficients that are corrected for total EAAL (true ileal digestibility; TID; sometimes also referred to as real ileal AA digestibility) and ileal digestibility coefficients that are corrected for basal EAAL only, a new term was introduced: standardized ileal AA digestibility (SID)⁽Reference Stein, Seve and Fuller^5, Reference Nyachoti, de Lange and McBride⁵⁴⁾. The concept of SID coefficients has been adopted around the world for the formulation of pig diets and extensive databases summarizing SID coefficients for AA in a wide range of feed ingredients have been published^(59–Reference Sauvant, Perez and Tran⁶¹⁾. The main advantages of using SID coefficients in feed formulation is that they are independent of dietary AA level, reduce variability in measures of ileal digestibility within food and feed ingredients, are more additive than AID coefficients in mixtures of feed or food ingredients, and can be derived experimentally without too much difficulty by correcting AID for basal EAAL⁽Reference Stein, Seve and Fuller^5, Reference Nyachoti, de Lange and McBride^54, Reference Furuya and Kaji⁶²⁾. Many studies have been conducted to quantify basal EAAL in pigs. In a review of the literature, Jansman et al. ⁽Reference Jansman, Smink and van Leeuwen⁵⁷⁾, suggested that the mean estimate of basal endogenous ileal lysine losses observed in studies where pigs were fed protein-free diets (0·36 g per kg DM intake; SD 0·10; 16 studies) was similar to estimates obtained based on regression analyses, as outlined in Fig. 2 (0·38; SD 0·09; 3 studies) and only slightly lower than estimates obtained from pigs fed purified proteins with an assumed TID of 100 % (0·44; SD 0·17; 11 studies). Relative differences were similar for other indispensable AA. It should be noted that estimates of basal EAAL vary between laboratories, even when using identical methodology, and that it is appropriate to obtain local estimates of EAAL when determining AID coefficients for calculating TID coefficients for AA⁽Reference Stein, Seve and Fuller^5, Reference Seve, Tran and Jondreville⁶³⁾. Until a methodology is in place to routinely measure total EAAL, both basal and diet specific, and these losses have been quantified for a wide range of feed ingredients, SID coefficients will provide the basis for estimated bioavailability of AA in feeds and foods.

When advanced methods have been used to estimate total EAAL, and thus specific EAAL, it has been shown that the specific EAAL varies among protein sources and feed ingredients. For example, when de Lange et al. ⁽Reference de Lange, Souffrant and Sauer⁶⁴⁾ used an ¹⁵N isotope dilution technique, they estimated total ileal endogenous lysine losses to be 1·02, 1·22, 1·10 and 1·11 g per kg feed DM intake (SE 0·10) in growing pigs fed diets that contained soyabean meal, canola meal, wheat, or barley as the only protein source, respectively. These values were all higher than values obtained in pigs that were managed under similar conditions and that were fed either a protein-free diet (0·63) or protein-free diets with a simultaneous intravenous infusion of a mixture of indispensable and dispensable AA (0·56 g per kg DM intake; SE 0·07)⁽Reference de Lange, Sauer and Souffrant⁶⁵⁾. These observations also illustrate that endogenous gut AA are a major contributor to dietary AA requirements and that care should be taken in how EAAL are considered when estimating AA requirements. When diets are formulated based on SID coefficients, the specific EAAL are already accounted for in the nutritional value of the diet and only the basal EAAL, not the specific EAAL, should be considered explicitly when characterizing host or animal factors that contribute to AA requirements. When foods or feeds are evaluated based on TID coefficients, it is critical that both the basal and specific EAAL are considered as part of the animal or host factors contributing to AA requirements.

Refinements to ileal digestibility coefficients for estimating amino acid bioavailability

As mentioned above, in some instances SID, or AID, coefficients do not yield proper estimates of AA bioavailability and the confounding of EAAL and true digestible AA intake should be considered carefully when interpreting estimates of AA bioavailability, especially when these are obtained with the slope-ratio assay. As we move towards a better understanding of the discrepancy between ileal digestibility and bioavailability, refinements may be made to SID coefficients to routinely estimate AA bioavailability in feeds and foods and to better predict the host's response to varying AA intake form different protein sources. Here three specific examples of such refinements are provided.

Conclusions and implications

There is a large body of evidence to suggest that in many instances measures of ileal amino acid (AA) digestibility yield reasonable estimates of bioavailability of AA in feeds and foods. A key limitation of using apparent ileal digestibility (AID) coefficients for diet evaluation and estimating AA requirements is that they are confounded with dietary AA levels. The latter can be attributed to the varying contribution of endogenous gut AA losses (EAAL) to total ileal AA flows. The EAAL can be separated into two main components, basal and ingredient specific EAAL. When correcting AID coefficients for basal EAAL only, standardized ileal digestibility (SID) coefficients are derived that are independent of diet AA level, less variable for a particular dietary ingredient, and more additive in mixtures in feed or food ingredients. The concept of SID is used extensively when formulating diets for monogastric farm animals. When using SID coefficients in diet evaluation, only the basal EAAL should be considered part of the animal or host effects on AA requirements. In some instances SID (as well as AID and TID) coefficients will provide inaccurate estimates of AA bioavailability, in particular bioavailability of lysine in heat treated ingredients, bioavailability of threonine in diets that contain large amounts of fermentable fibre, and that of AA in diets that induce large amounts of specific EAAL. The concept of SID coefficients may be refined for specific feeds and foods to account for instances where SID coefficients do not yield proper estimates of AA bioavailability.