Dietary, processing and animal factors affecting energy digestibility in swine 1

Dietary, processing and animal factors affecting energy digestibility in swine 1 Jean Noblet INRA, F-35590 Saint-Gilles (France) jean.noblet@rennes.inra.fr Introduction The cost of feed is the most important cost of pig meat production (#60-70%) and the energy component represents the greatest proportion. Therefore, it is important to estimate precisely the energy value of feeds, either for least-cost formulation purposes or for adapting feed supply to energy requirements of animals. In addition, energy supply has an important impact on performance of animals. Evaluation of energy content of pig feeds is firstly and most commonly based on their digestible (DE) or metabolizable (ME) energy contents. However, the closest estimate of the "true" energy value of a feed corresponds to its net energy (NE) content which takes into account differences in metabolic utilization of ME between digested nutrients. The objectives of this review paper are to present the effects of some dietary, processing and animal factors that affect energy digestibility of pig feeds. More complete reviews on this topic have been proposed previously (Noblet et al., 2004; Noblet and van Milgen, 2004; Noblet, 2006; Noblet and van Milgen, 2013). The unit for energy in the text is Joule (or kj or MJ) with 1 Cal = 4.18 Joules. Diet composition and energy digestibility For most pig diets, the digestibility coefficient of energy (DCe) varies between 70 and 90% but the variation is larger for feed ingredients (10 to 100%). Most of the variation of DCe is related to the presence of dietary fiber (DF) which is less digestible than other nutrients (<50% vs 80-100% for starch, sugars, fat or protein) and which also reduces the apparent fecal digestibility of other dietary nutrients such as crude protein and fat. Consequently, DCe is linearly and negatively related to the DF content of the feed (Le Goff and Noblet, 2001). The coefficients relating DCe to DF are such that DF essentially dilutes the diet, at least in growing pigs (table 1). In other terms, even though DF is partly digested by the young growing pig, it provides very little DE to the animal. In addition, the digestibility of DF depends on its botanical origin with some sources almost indigestible in pigs (wheat straw) and others that are highly digested (sugarbeet pulp, soybean hulls); these differences are due, on one hand, to the degree of lignification (i.e. negative effect) and, on the other hand, to the content of pectins and the degree of solubility of DF (i.e. positive effects). Some illustrations of these variations are given in table 2. In most feeding systems, the supply of minerals (carbonate, etc.) is considered to be equivalent to a diluent of the feed. However, the addition of phosphate or carbonates reduces the digestibility of energy, organic matter or crude protein and, more generally, DCe is negatively affected by the dietary ash content (Noblet and van Milgen, 2013; equation 2 in table 1). Under practical conditions, this observation should be considered when the ash content is markedly reduced (in connection with the use of phytase, for instance). Complementary aspects of dietary factors on DCe will be considered in the section on fat digestibility. Animal factors and energy digestibility Up to 20 years ago, it was accepted that a given feed was attributed only one energy value whatever the feed was fed to young, growing or adult pigs. It is even still the case in some recent feeding 1 Paper prepared for the Southeast Asian Feed Technology and Nutrition Workshop, 3-7 June 2013, Bangkok (Thaïland); available at: http://www.asaimsea.com/ 1

tables for swine (NRC, 2012). However, DCe increases with increasing BW (Noblet 2006; Noblet et al., 2013) with, on average, a 0.5% higher DCe per 10 kg BW increase (table 1). The largest effect of BW is observed when adult sows either pregnant or lactating and (close to) ad libitum growing pigs are compared (Le Goff and Noblet, 2001; Le Gall et al., 2009). And the difference due to BW increase is the most pronounced for high fiber diets or ingredients: the negative effect of DF on DCe is then lower in adult pigs than in growing pigs and the contribution of DF to energy supply becomes then significant (table 1). As illustrated in table 3, the DE value is 1.8, 4.2, 6.0, 10.3, 16.6 and 36% higher in sows for wheat, corn, soybean meal, wheat bran, corn gluten feed and soybean hulls, respectively. This improvement of energy digestibility with BW is mainly explained by an improved digestive utilization of DF (table 2), this improvement being dependent on botanical origin. As illustrated in table 2, the improvement is higher for corn DF than for wheat DF. These changes in DF and energy digestibility with BW increase are explained by an increased hindgut capacity and a subsequent reduced rate of passage of digesta in the gut (Le Goff et al., 2002). In other words, these changes in the gut physiology allow an additional digestion of the organic matter that is undigested in young pigs. From a compilation of a large data base, we were able to calculate that the adult sow was able to get 4.2 kj of additional DE per g of undigested organic matter measured in the growing pig (Noblet et al., 2004); but this value changed with the botanical origin of the ingredient (table 3). The pigs may be fed different feeding levels, especially the reproductive pigs when gestation and lactation are compared. Even there is a tendency for an increased digestibility with reduced feeding level (Noblet and van Milgen, 2013), this effect remains rather negligible (table 4) and can be ignored, even in lactating sows compared to pregnant sows. In addition, if we consider that piglets are usually fed low-fiber diets for which the effect of BW is minimized, piglets can, from a practical point of view, be considered as growing pigs concerning the digestive utilization of energy. Overall, these results suggest that that at least two different DE values should be used in pig production, one for piglets and growing pigs and one for adult sows. Under all practical conditions, the effect of pig genotype on DCe is ignored, even some results suggest higher digestibilities in unselected native breeds. However, the interpretation of these results is complicated by the fact that their rate of growth is slower and their sexual development is quicker; the feed intake may also be different. Other results obtained recently in modern breeds indicate that DCe in growing pigs is genetically dependent with, for instance, a significant effect of boar origin on DCe in a population of Large-White pigs (Noblet et al., 2013). This finding is a promising approach for getting more efficient pigs and reducing the quantities of excreta and manure. Processing and energy digestibility Digestibility of energy can be modified by technological treatments which can be more or less complex or sophisticated and with variable objectives. Pelleting, for instance, increases the energy digestibility of feeds by about 1% (Noblet, 2006; Le Gall et al., 2009). However, the improvement is more important for some ingredients such as full fat rapeseed or (high oil) corn for which pelleting improves the digestibility of fat with subsequent marked differences in their DCe and DE values between mash and pellet forms (table 5). Some ingredients may even require a "tougher" processing such as extrusion for increasing the digestibility of nutrients and energy: this is the case of full fat linseed that provides omega 3 fatty acids for which the fat fraction is poorly digested without previous extrusion (table 5). Finally, the nutrients availability in pig feeds may be improved by supplementing enzymes or acids, for instance. The literature on these topics is rather abundant with unclear conclusions. Most of the confusion comes probably from the fact that measurement of fecal digestibility is unable to account for the partition of digestion between the small intestine and the hindgut with an important compensation of the hindgut in pigs. Fat digestibility The gross energy content of fat is much higher than the other nutrients (39.4 kj/g) and its average digestibility is also high (85%). As indicated in the NE system, the efficiency of fat ME (or fat DE) for NE is higher (90%) than for the other nutrients. Consequently, fat is a unique source for increasing the energy density of pig diets that may, for instance, compensate the depressive effect of DF on energy content. Potential variations of fat digestibility are therefore important to consider. 2

The variability in fat digestibility is first related to the fatty acids (FA) composition of the triglycerides. Measurement of ileal FA digestibility in animal fat at a relatively high inclusion rate (Jorgensen et al., 1992) indicates that medium chain FA (C12 and C14) are highly digestible (94%) while the saturated long chain FA are much less digestible (85% for C16:0 and 73% for C18:0). The corresponding mono unsaturated FA (C16:1 and C18:1) are highly digestible (95%). These differences between FA according to the chain length and the degree of unsaturation have been confirmed by other authors or from total fat digestibility measurements on fat sources that differ for the degree of unsaturation or the chain length (Table 6). One practical consequence is that animal fat sources are usually less digestible than vegetable fat sources. The amount of free FA also affects fat and FA digestibilities since free FA are less digestible than FA that are bound to glycerol. The so-called acid fat sources have then a lower digestibility and a lower energy value than nonacid fat sources (table 6). Fat digestibility is not affected by BW in growingfinishing pigs and results are quite comparable in adult sows and growing pigs (unpublished data). However, at weaning, especially at an early age, fat digestibility is lower (Cera et al., 1988; 1989). In the first study, for instance, the fat digestibility of three diets averaged 71, 75, 84 and 85% in weeks 1, 2, 3 and 4 after weaning at 21d. The digestibility of fat is then dependent on two important factors, the degree of saturation of FA and the amount of free FA. From the compilation of literature measurements, Powles et al. (1995) have proposed a general equation for predicting the DE content of fat sources (MJ/kg) according to the ratio unsaturated FA:saturated FA (U/S) and the amount of free FA (FFA, g/kg fat): DE, MJ/kg = 36.9 0.0046 FFA 7.3 e -0.906 U/S. In practice and if sources that contain non negligible levels of FFA are excluded, the average fat or energy digestibility of fat sources is close to 85%. The DE and ME values are then equal to about 33.5 (39.4 x 0.85) and 33.3 (33.5 x 0.995) MJ/kg, respectively. These recommended values may slightly underestimate the energy values of vegetable fat sources (-2 to 3%) and overestimate the energy values of animal fat sources. In any case, they can be used for acid fat sources for which the energy digestibility is quite variable and does not exceed 80% (Jorgensen and Fernandez, 2000; Powles et al., 1993; Wiseman and Cole, 1987). A correction factor of 0.50 and 0.45 MJ of DE value and NE value per 100 g increase of FFA can be proposed according to Powles et al. (1995). Metabolizable energy The ME content of a feed is the difference between DE and energy losses in urine and gases (methane). In growing pigs, average energy loss in methane is equivalent to 0.4% of DE intake and is 2-3 times this amount in adult sows (Noblet and van Milgen, 2013). Energy loss in urine represents a variable percentage of DE since urinary energy depends greatly on the urinary nitrogen excretion. At a given stage of production, urinary nitrogen excretion is mainly related to the (digestible) protein content of the diet, relative to the protein and amino acids requirement of the pig (table 1). On average, it represents about 4% of the DE value and the ME/DE ratio averages 96%. However, these mean values cannot be applied to single feed ingredients. The most appropriate solution is then to estimate to consider that a constant proportion of the N provided by each ingredient included in the diet is excreted. In growing pigs, the urinary energy (kj/kg feed DM) can be predicted from urinary nitrogen (g/kg feed DM) according to the following equation: Urinary energy = 192 + 31 x urinary nitrogen For most situations, it can be assumed that urinary nitrogen represents 50% of digestible nitrogen or 40% of total nitrogen (Noblet et al. 2004). The application of this calculation method indicates that the ME/DE ratio is above 99% for N-free ingredients (fat sources, etc.) and close to 90% for high protein ingredients, such as soybean meal (Sauvant et al., 2004). The above formula is slightly different in adult sows (Noblet et al., 2004; Noblet and van Milgen, 2013). With regard to digestive gas energy losses, a method is proposed by Noblet et al. (2004) with the highest gas energy losses in high DF ingredients in adult pigs (up to 6% of DE for soybean hulls, for instance). In most cases, they are negligible in feeds for growing pigs. 3

Feeding tables and EvaPig Apart from direct measurement on pigs that are costly, time-consuming and rather complex, the DE and ME values of raw materials can be obtained from feeding tables (Sauvant et al. 2004; EvaPig; NRC 2012). In the case of tables of Sauvant et al. (2004), two sets of DE, ME and NE values are proposed, one for piglets and growing-finishing pigs and one for reproductive sows. The interest of using a specific set of values for adult sows is the most accentuated for high fiber feeds that may be highly used in the diets of pregnant sows that are fed restrictively and then contribute to improve the welfare of such animals. But, for any table, the utilization of tabulated values should be restricted to ingredients having chemical characteristics similar or close to those in the tables. However, practically, the chemical composition and therefore the energy value of most raw materials differ from those listed in the feeding tables. It is then advised to correct the energy value for differences in chemical composition between ingredients proposed in the tables and those actually available for least-cost formulation; the EvaPig tool has been proposed for such a purpose (www.evapig.com) with recalculation of energy values for both growing pigs and reproductive sows. Such corrections/adjustments can be done for about 120 reference ingredients included in the software. Finally, a rather frequent difficulty in feed evaluation consists in estimating the energy value of compound feeds when their ingredients composition is unknown or for totally new ingredients not included in the feeding table. The best solution is then to use prediction equations based on chemical criteria (Le Goff and Noblet 2001; EvaPig) or estimates from near infrared or in vitro methods (Noblet and Jaguelin-Peyraud 2007). The EvaPig software proposes such evaluations. In conclusion, the EvaPig software is first a feeding table but its main target is the estimation of nutritional values of pig feeds according to the actual (chemical) characteristics of the feed and on concepts that have been validated. It is then usable in the feed industry to generate the nutritional value of successive batches of ingredients. It can also be used for teaching the nutritional concepts used in feed evaluation. It is so far available in 14 different languages and freely available at www.evapig.com. Conclusions As indicated in several articles, the "true" energy value of a feed is better estimated according to its NE content than its DE or ME content. But the NE content is directly dependent on DE or ME values or digestible nutrients contents. The primary factor of variation of the "true" energy value of a feed is then the digestibility of energy. This review indicates that energy digestibility of a diet or an ingredient varies firstly with its chemical characteristics (dietary fiber), secondly with animal BW and technology and thirdly with other factors such as feeding level or animal genotype. In a context of precise feeding of pigs, it is then necessary to evaluate precisely the energy content of pig feeds and consider the effects of these factors. Consequently, a feed can be given several energy values. However, simplifications are required: it is then suggested to use at least two energy values in pig production, one for piglets and growing pigs and one for reproductive sows. The impact of technologies (pelleting, extrusion, enzymes cocktails, etc.) should also be taken into account. Unfortunately, the information is not available for all ingredients and all types of technologies: knowledge is required for the future. Finally, ingredients are changing over time or new ingredients become available. In order to make decision on their purchasing and utilization, rapid methods of evaluation are required. Predictions based on chemical characteristics, NIR information or in vitro data represent a first set of possibilities but they need to be also more documented. References (selection) EvaPig: a calculator of energy, amino acid and phosphorus values of ingredients and diets for growing and adult pigs. Available at: www.evapig.com Le Gall, M., M. Warpechowski, Y. Jaguelin-Peyraud, and J. Noblet, 2009. Influence of dietary fibre level and pelleting on digestibility of energy and nutrients in growing pigs and adult sows. Animal 3: 352-359. Le Goff, G., and J. Noblet, 2001. Comparative digestibility of dietary energy and nutrients in growing pigs and adult sows. J. Anim. Sci. 79: 2418-2427. 4

Le Goff, G., J. van Milgen, and J. Noblet, 2002. Influence of dietary fibre on digestive utilization and rate of passage in growing pigs, finishing pigs, and adult sows. Anim. Sci. 74: 503-515. Noblet, J., B. Sève, and C. Jondreville, 2004. Nutritional values for pigs. In: Tables of composition and nutritional value of feed materials: pigs, poultry, cattle, sheep, goats, rabbits, horses, fish, Eds. D. Sauvant, J.M. Perez and G. Tran. Wageningen Academic Publishers, Wageningen and INRA Editions, Versailles. pp. 25-35. Noblet, J., and J. van Milgen, 2004. Energy value of pig feeds: Effect of pig body weight and energy evaluation system. J. Anim. Sci. 82: E. Suppl., E229-E238. Noblet, J., 2006. Recent advances in energy evaluation of feeds for pigs. In: Recent advances in Animal Nutrition 2005, Eds. P.C. Garnsworthy and J. Wiseman. Nottingham University Press, Nottingham. pp. 1-26. Noblet, J., and Y. Jaguelin-Peyraud, 2007. Prediction of digestibility of organic matter and energy in the growing pig from an in vitro method. Anim. Feed Sci. Technol. 134: 211-222 Noblet, J., and J. van Milgen, 2013. Chapter 2: Energy and Energy Metabolism in Swine. In Sustainable Swine Nutrition, Ed. L.I. Chiba, John Wiley & Sons, Ames, Iowa, USA, pp. 23-57. Noblet, J., H. Gilbert, Y. Jaguelin-Peyraud, and T. Lebrun, 2013. Evidence of genetic variability for digestive efficiency in the growing pig. Animal (in press; doi:10.1017/s1751731113000463) NRC, 2012. Nutrient requirements of swine. The National Academy Press, Washington, D.C. Powles, J., J. Wiseman, D.J.A. Cole, and B. Hardy, 1993. Effect of chemical structure of fats upon their apparent digestible energy value when given to growing/finishing pigs. Anim. Prod. 57: 137-146 Powles, J., J. Wiseman, D.J.A. Cole, and S. Jagger, 1995. Prediction of the apparent digestible energy value of fats given to pigs. Anim. Sci. 61: 149-154 Sauvant, D., J.M. Perez, and G. Tran, 2004. Tables of composition and nutritional value of feed materials: pigs, poultry, cattle, sheep, goats, rabbits, horses, fish. Wageningen Academic Publishers, Wageningen and INRA Editions, Versailles. 5

Table 1. Effect of diet composition (g/kg dry matter) or body weight (BW, kg) on energy digestibility (DCe, %) and ME:DE coefficient (%) in pigs a Equation RSD Source 1 DCe = 98.3-0.090 x NDF 2.0 1 2 DCe = 102.6 0.079 x NDF 0.106 x Ash 1.8 1 3 DCe = 96.7-0.064 x NDF 2.2 1 3 DCe = 76.6 + 0.061 x BW 1.2 2 5 ME/DE = 100.3-0.021 x CP 0.5 1 a CP: crude protein, NDF: Neutral Detergent Fibre,; RSD: Residual standard deviation; 1: Le Goff and Noblet 2001; n=77 diets; equations 1, 2 and 5 in 60 kg growing pigs and equation 2 in adult sows, respectively); 2: Noblet et al. 2013 (1 diet;20 pigs measured over 10 consecutive weeks between 30 and 95 kg BW). Table 2. Digestibility of fibre fractions and energy in high fibre ingredients in growing pigs (G) and adult sows (S) a Wheat bran Corn bran Sugar beet pulp G S G S G S Digestibility coefficient (%) of Non-starch polysaccharides 46 54 38 82 89 92 Non cellulose polysaccharides 54 61 38 82 89 92 Cellulose 25 32 38 82 87 91 Dietary fibre a 38 46 32 74 82 86 Energy 55 62 53 77 70 76 a Adapted from Noblet and Bach-Knudsen (1997); dietary fibre = Non-starch polysaccharides + lignin Table 3. Digestible energy value of some ingredients for growing pigs and adult sows a DE, MJ/kg b Ingredient Growing pig Adult pig a c Wheat 13.85 14.10 3.0 Barley 12.85 13.18 2.5 Corn 14.18 14.77 7.0 Pea 13.89 14.39 6.0 Soybean meal 14.73 15.61 8.0 Rapeseed meal 11.55 12.43 3.5 Sunflower meal 8.95 10.25 3.5 Wheat bran 9.33 10.29 3.0 Corn gluten feed 10.80 12.59 7.0 Soybean hulls 8.37 11.46 8.0 a Adapted from Sauvant et al. (2004a, b) As fed. c kj difference in DE between adult sows and growing pigs per g of indigestible organic matter in the growing pig (Noblet et al., 2004). 6

Table 4. Effect of feeding level on digestive utilization of energy in growing pigs and adult dry sows Stage Growing pig a Adult sow b Feeding level 1 2 1 2 3 Body weight, kg 40.1 43.3 260 260 260 Feed intake, g DM/d 1106 1478 2090 2536 2966 Energy digestibility, % 83.2 82.6 85.2 85.6 85.9 a Mean of 2 compound feeds containing 13 and 21% NDF; the effect of feeding level was more pronounced in the high NDF feed (P < 0,05) (J. Noblet, unpublished data) b Mean of 4 compound feeds based on maize, wheat, barley, peas, soybean meal and variable proportions of wheat bran, soybean hulls, sugar beet pulp, wheat straw and rapeseed oil (J. Noblet, unpublished data) Table 5. Effect of pelleting on energy digestibility of feeds in growing pigs Technology Mash Pellet Reference a Wheat-SBM diets (n=2) 88.6 * 89.2 3 Wheat-corn-barley-soybean meal diets (n=4) 75.8 ** 77.3 1 Corn-SBM diets (n=3) 88.4 ** 90.3 2 Corn (n=5) 87 ** 90 3 Full-fat rapeseed 35 ** 83 4 Linseed (extrusion) 51 ** 84 5 a 1: Le Gall et al., 2009; 2: Noblet and Champion, 2003; 3: Noblet and Jaguelin-Peyraud, 2008; 4: Skiba et al., 2002; 5: Noblet et al., 2008. Table 6. Fat and energy digestibility of some fat sources: literature survey. U/S1 FFA, % a Digestibility, % Reference b Fat Energy Rapeseed oil 15.5 1 90-1 Tallow 1 0.96 3 77-1 Tallow 2 0.94 4 87-1 Palm oil 0.85 5 82-1 Soybean oil 4.3 1 89-1 Soybean acid oil 4.9 63 85-1 Tallow acid oil 0.65 82 73-1 Animal fat 1.32 0 90 89 2 Soybean oil 5.9 0 91 91 2 Palm oil 1.04 0 85 83 2 Vegetable oil by-product 1.62 55 62 62 2 Rapeseed oil 15.2 NA 92 95 3 Soybean oil NA NA 86 87 4 Rapeseed oil NA NA 88 90 4 Sunflower oil NA NA 83 85 4 Soybean oil 5.6 NA - 85 5 Lard 1.42 NA - 84 5 Tallow 0.92 NA - 79 5 Poultry fat 2.1 NA - 87 5 a U/S: ratio unsaturated fatty acids: saturated fatty acids; FFA: free fatty acids b References: 1: Powles et al. (1993); reference 2: Jorgensen and Fernandez (2000); reference 3: Jorgensen et al. (1996); reference 4: J. Noblet, unpublished data; reference 5: Bourdon and Hauzy, 1993. 7