Feed Ingredients: Assessment and Enhancement of Nutritional Value

The nutritional value of feed ingredients comprises all nutrients needed for maintenance and production of animals. This paper mainly focuses on the energy value of feed ingredients, which depends on the nutrient composition (starch, fat and protein), the efficiency of nutrient digestion and absorption in the intestinal tract, and the presence of anti-nutritional factors (ANFs). Although the composition of feed ingredients can be analysed by quick methods like NIR accurately, it does not take differences in the digestibility of nutrients in the intestinal tract into account, which can vary among different batches of the same feedstuff. Garnsworthy et al. (2000) indicated that NIR methods predicted the chemical and agronomic characteristics of wheat with high accuracy, whereas prediction of the nutritional value was much less accurate due to animal variation. Recently, Black et al. (2014) indicated NIR calibrations for nutritional value were improved substantially.

Processing of feedstuffs, like heat treatment to eliminate ANFs or for drying purposes and milling/pelleting, affects nutrient digestibility values. Good quality parameters for eg. soybean meal are still topic of debate, because of a questionable reliability of current parameters like urease activity, protein dispersibility index, KOH protein solubility and trypsin inhibitor. These techniques have low consistency and sensibility among laboratories or are very tedious (De Coca et al., 2008).

Feedstuff tables generally comprise one batch of single feed ingredients that can be used in feed formulation. In case of known variable nutritional values, more batches might be included, often still accepting a fixed ratio between digestible amino acids per unit of crude protein, and/or similar nutrient digestibility values for different qualities of a feedstuff and types of poultry. Of course tabulating nutritional values for several batches of a single feedstuff is only relevant if different qualities could be identified via quick, cheap and reproducible methods. It should be realized that only using variable nutrient contents, like eg. the different crude protein contents of soybean meal, are not be the best discriminating factors. Maize batches with a similar nutrient composition were shown to have distinctly different nutritional values (Gehring et al., 2012). Prediction of the quality of feed ingredients is therefore very important to be able to formulate diets the meet the nutrient requirements of the birds. 

Feedstuff databases are traditionally built on digestibility studies with adult roosters. These birds enable high through-put studies to evaluate the feeding value of single feedstuffs and to indicate variation in nutritional quality among different batches of the same feedstuff. Therefore, also many NIR calibrations are based on in vivo data with adult roosters. However, it is shown that feeding values vary between adult roosters and broilers (eg. Rodrigues et al. 2001, 2002), which justifies developing separate feeding tables for young broilers and adult birds. Apparent AMEn values for different types of poultry are summarized in Table 1, as modified from Cozannet et al. (2010). 

Table 1 - The MEn¹ value in feeds for roosters, layers, broilers and turkeys (adapted from Cozannet et al., 2010)

It is indicated in Table 1 that the difference between roosters and broilers can be quite variable, eg. for wheat. Svihus and Gullord (2002) concluded from their study with twenty batches of wheat that the low-AME wheat phenomenon was observed in broilers but not in roosters, and was related to a high feed intake and low starch digestibility. This indicates that low quality feedstuffs might be evident in young broilers, but not in roosters.

For the purpose of building a more extensive dataset on AMEn values for laying hens, Schothorst Feed Research recently conducted two experiments evaluating sixteen feedstuffs in both broilers (at 21-24 days of age) and laying hens (at 24-26 weeks of age; minimum laying rate 90%). These feedstuffs were added to a basal diet (which varied in composition for broilers and laying hens) to obtain the experimental diets. Diets were pelleted (pellet temperature < 65°C) and fed for ad libitum intake during ten to twelve days, followed by a three-day excreta collection period (collection once daily. Samples were immediately stored after collection at -20°C (assay similar to Bourdillon et al., 1990, except for using TiO₂ as an inert marker and omitting a feed withdrawal period)). The AMEn value of test feedstuffs were calculated from results of experimental and basal diets. All measurements were done with six replicates using twelve broilers or eight layers per replicate cage. Results are given in Figure 1. 

Figure 1 - The AMEn value in feedstuffs for laying hens and broilers (data Schothorst Feed Research). Line X=Y is indicated.

Based on Table 1 it was concluded that in general AMEn values in broilers are lower compared to roosters, whereas Figure 1 indicates that feedstuff AMEn values in laying hens can be both lower (min -20%) and higher (max +35%) than in broilers. Therefore, the validity of rooster data is not only questionable for broilers, but also for producing laying hens. This will have consequences for the type of bird that should be used for calibration of methods to predict the nutritional values of feed ingredients for poultry, as was also concluded by Yegani et al. (2013). 

The nutritional value of feed ingredients is dependent on the contents of energy delivering nutrients (starch, fat, protein) and their corresponding digestibility values. Starch is the main source of metabolizable energy in poultry diets. Although its digestibility is assumed to be almost complete at the lower ileum, potential variation has been reported as reviewed recently by Svihus (2011). Gutierrez del Alamo et al. (2009) showed that starch digestion rate in 30-d old broilers varied with wheat origin and cultivar, whereas Svihus (2011) indicated the following potential factors affecting starch digestibility: Cereal characteristics like starch granule size, interactions with the gluten matrix in wheat or sorghum, starch entrapment in the cell wall material, and the amylose/amylopectin ratio and high feed intakes of modern broilers when fed pelleted diets. Amylopectin is due to its branched structure more easily degraded by amylase compared to the more linearly arranged amylose. Collins et al. (2003) reported an AMEn value of 2860 kcal/kg for waxy maize (3% of starch as amylose) and 2790 kcal/kg for normal maize (27% of starch as amylose) in 7-9 day old broilers, whereas such a difference was not evident for roosters (average AMEn of 3250 kJ/kg). Also Ertl and Dale (1997) did not observe differences between waxy and normal corn in roosters, despite a difference between maize varieties with a similar amylose/amylopectin ratio of 140 kcal/kg DM. Ravindran et al. (2007) observed lower AMEn values in waxy hull-less barley compared to normal hull-less barley (being 2410 and 3015 kcal/kg DM, respectively), which was in their study related to the higher soluble β-glucan content in waxy barley cultivars.

Protein-starch interactions were not only reported in wheat (Svihus, 2011), but also in maize (Gehring et al., 2012) and sorghum (Selle, 2010). Gehring et al. (2012) showed that salt-soluble protein in maize (ranging from 25 to 50%), indicating differences in the starch accessibility related to the starch-protein interface, was correlated with dietary AMEn (ranging from 3262 to 3342 kcal/kg) in 30-d old broilers. However, the AMEn value was not correlated with starch digestibility, and therefore needs further validation.

In vitro techniques have been developed to estimate the potential digestibility of nutrients enabling a more accurate calculation of the energy value of feed ingredients. Weurding et al. (2001) showed a good correlation between in vitro starch digestion during 2h and 4h incubation and the rate of in vivo starch digestion in the small intestine of broilers for twelve feed ingredients, whereas Yegani et al. (2013) showed a high correlation between in vitro digestibility of gross energy and in vivo AMEn values in 14-day old broilers using six batches of wheat and two batches of triticale. However, as this in vitro method that was optimized for wheat and triticale samples, it needs to be validated for other feed ingredients. In general, techniques that estimate the rate of nutrient digestion predict the actual feeding value more accurately than end point titrations, as was summarized by Van der Klis and Kwakernaak (2013). Black et al. (2014) highlight the value of NIR as rapid assessment of feed ingredient quality in their Australian Poultry Science Symposium paper and will therefore not be discussed here. 

Tabulated data on nutritional values of feed ingredients are generally based fecal apparent nutrient digestibility values, i.e. not corrected for endogenous losses and for fermentation. Therefore, ANFs that normally induce endogenous losses will directly impact such nutrient digestibility values (especially of amino acids) and to a lesser extend AMEn values. Moreover, nutritional values of feed ingredients are supposed to be additive, without interactions. Absence of interactions between feed ingredients, however, is unlikely. Nutritionists try to limit those via constraints in diet formulation with respect to e.g. the ratio between unsaturated and saturated fatty acids, a minimum starch inclusion level and/or maximizing inclusion levels of ANF containing feed ingredients. As said interactions between feed ingredients do exists, like effects of low AMEn wheat varieties becoming more pronounced with increasing inclusion levels of animal blended fat, resulting in decreasing AMEn values of wheat (Van der Klis et al., 1995).

This example indicates that the nutritional value of feed ingredients should not be regarded separately from the intestinal physiology of the animal. For example, feed retention time in different intestinal segments, viscosity, reflux of chyme and microbial activity in the proximal parts of the small intestine clearly affect the efficiency of nutrient digestion and absorption. Increasing gastro-intestinal development and reflux by structural components in poultry diets (like adding inert fiber and/or increasing feed particle size) improves AMEn conversion (MJ/kg BWG) in case of suboptimal conditions in the intestinal tract (Figure 2). Highly methylated citrus pectin increases chyme viscosity and microbial growth in the small intestine and thereby reduces nutrient absorption and production performance.

Coarse grinding of cereals stimulates the development of the “proventriculus plus gizzard” in broilers, potentially improving the digestibility of the feed. Indeed coarse grinding was shown repeatedly to stimulate production performance (Amerah et al., 2007). However, milling for small particle sizes will increase the surface area of the substrate for enzymes. Fine grinding of maize (milled to pass a 2 mm sieve) was shown to have a higher ileal digestibility of crude protein and energy than more coarse grinding (milled to pass a 4 mm sieve): 0.82 vs 0.79 and 0.76 and 0.73 respectively), whereas starch digestibility was not affected (0.96 on average) (Bhuiyan et al, 2013). Péron et al. (2005) indicated that finely ground hard wheat improved the digestibility of starch (fine was milled to pass a 2 mm sieve (GMD 380 μm) and coarse to pass a 6 mm sieve (GMD 955 μm). This might indicate that starch granules are entrapped in cell walls and/or the protein matrix in cereals, limiting digestibility (Svihus, 2011). 

Figure 2 - The AMEn conversion (MJ/kg BWG) in 21-day old broilers fed a corn/soy diet with and without 3% high methylated citrus (HMC) pectin. Pelleted diets were fed with finely ground corn (milled to pass a 3 mm sieve), coarsely ground corn (milled to pass a 8 mm sieve) with 5% added oat hulls on isocaloric basis (1) or added on top (2). (LSD 0.432; P< 0.01), van der Klis (2012).

Nutritional values of feed ingredients can also be improved by dietary supplementation of exogenous enzymes. The use of endoxylanases and β-glucanases in cereal-based poultry diets containing high levels of soluble non starch polysaccharides has become common practice, as it improves the nutritional value of the diet. In contrast, effects of carbohydrases, amylases and proteases to corn/soy diets can be inconsistent, being small to negligible. Meng et al. (2005) screened combinations of carbohydrases (cellulase, pectinase, xylanase, glucanase, and mannanase) in vitro on wheat, soybean meal, canola meal and peas, showing positive effects of these enzymes on all feed ingredients, achieving more pronounced effects when these carbohydrases were used in concert. Effective enzyme combinations were evaluated in vivo using 5 to 18 day old broilers fed mash diets, which improved dietary AMEn and ileal digestibility values of starch and protein (only the full combination of enzymes is shown in Table 2). They concluded that these carbohydrases were effective in degrading cell wall polysaccharides and nutrient utilization, which probably has eliminated nutrient encapsulating affects of cell wall polysaccharides. 

Table 2 - The effect of a combination of carbohydrases in the AMEn value and ileal digestibility of protein and starch in 18 day old broilers.

Yegani and Korver (2013) discussed potential reasons for inconsistent response of broilers when supplementing corn/soy diets with exogenous enzymes, being differences in the types and activities of the enzymes, using single enzymes or mixtures of enzyme activities, the nutritional quality of the dietary ingredients, form of the diet and age of the birds. They showed the largest effects in the grower phase from 12 to 28 days of age, but only in one out of the three corn batches tested, despite similar analyzed nutrient contents. Unfortunately, details on nutrient composition of diets and characteristics of feed ingredients in such experiments is too limited to start a proper evaluation of the reason for these inconsistent responses. 

Amerah AM, Ravindran V, Lentle RG & Thomas DG (2007) World’s Poultry Science Journal 63: 439-454.

Askbrant US (1988) British Poultry Science 29: 445-455.

Bourdillon AB, Carré B, Conan L, Francesch M, Fuentes M, Huyghebaert G, JanssenWMMA, Leclercq B, Lessire M, McNab J, Rigoni M & Wiseman J (1990) British Poultry Science. 29: 567-576.

Barrier-Guillot B & Métayer JP (2001) Journal of Rech. Avicol 4: 131-134.

Bhuiyan MM, Islam F, Cowieson AJ & Iji PA (2013) Asian Journal of Poultry Science 7: 1-16.

Black JL, Hughes RJ, Diffey S, Tredrea AM, Flinn PC, Spragg JC & Kim JC (2014) Proceedings of the Australian Poultry Science Symposium 25: (this issue).

Collins NE, Moran ET & Stilborn HL (2003) Journal of Applied Poultry Research 12: 196-206.

Cozannet P, Lessire M, Gady C, Métayer JP, Primot Y, Skiba F & Noblet J (2010) Poultry Science 89: 2230-2241.

de Coca-Sinova A, Valencia DG, Jiménez-Moreno E, Lazaro R & Mateos GG (2008) Poultry Science 87: 2613-2623.

Ertl D & Dale N (1997) Journal of Applied Poultry Research 6: 432-435.

Garnsworthy P, Wiseman J & Fegeros K (2000) Journal of Agricultural Science 135: 409-417.

Gehring CK, Bedford MR, Cowieson AJ & Dozier WA (2012) Poultry Science 91: 1908-1914.

Gutierrez del Alamo A, Verstegen MWA, den Hartog LA, Perez de Ayala P & Villamide MJ (2009) Poultry Science 88: 1666-1675.

Lessire M, Revol N, Rudeaux F & Hallouis JM (1995) Inra Productions Animales 8: 189-195.

Macleod MG, ValentineJ, Cowan A, Wade A, McNeill L & Bernard K (2008) British Poultry Science 49: 368-377.

Meng X, Slominski BA, Nyachoti CM, Campbell LD & Guenter W (2005) Poultry Science 84: 37-47.

Parsons CM, Potter LM & Bliss BA (1982) Poultry Science 61: 2241-2246.

Péron A, Bastianellib D, Ouryc FX, Gomez J & Carré B (1995) British Poultry Science 46: 223-230.

Ravindran V, Tilman ZV, Morel PCH, Ravindran G & Holes GD (2007) Animal Feed Science and Technology 134: 45-55.

Rodrigues PB, RostagnoHS, Albino LFT, Gomes PC, Barboza WA & Santana RT (2001) Revista Brasileira de Zootecnia 30: 1767-1778.

Rodrigues PB, Rostagno HS, Albino LFT, Gomes PC, Nunes RV & Toledo RS (2002). Revista Brasileira de Zootecnia 31: 1771-1782.

Selle PH, Cadogan DJ, Li X & Bryden WL (2010) Animal Feed Science and Technology 156: 57-74. Svihus B & Gullord M (2002) Animal Feed Science and Technology 102: 71-92.

van der Klis JD, Scheele C & Kwakernaak C (1995) Proceedings of the 10th European Symposium on Poultry Nutrition, Antalya, Turkey, pp. 160-168.

van der Klis JD (2012) Proceedings of the 28th FEDNA Symposium on Advances in Animal Nutrition, Madrid, Spain, pp. 123-141.

van der Klis JD & Kwakernaak C (2013) Poultry Science Symposium, San Diego. Weurding RE, Veldman A, Veen WAG, van der Aar PJ & Verstegen MWA (2001) Journal of Nutrition 131: 2336-2342.

Yegani M & Korver DR (2013) Poultry Science 92: 1208-1220.

Yegani M, Swift ML, Zijlstra RT & Korver DR (2013) Animal Feed Science and Technology 183: 40-50. Aust. Poult. Sci. Symp. 2014.....25 37