Baixe Poultry...arciais - f 12 energy utilization - measurement and prediction e outras Notas de estudo em PDF para zootecnia, somente na Docsity! CHAPTER 12 Energy utilization: measurement and prediction M.G. MacLeod Roslin Institute (Edinburgh), Roslin, Midlothian EH25 9PS, UK © CAB International 2002. Poultry Feedstuffs: Supply, Composition and Nutritive Value (eds J.M. McNab and K.N. Boorman) 191 Evaluation of energy yield from the diet is of crucial importance in practical and research nutrition. One of the major reasons for its importance is that energy content has a key role in the control of food intake. This means that the intake of individual substances is strongly influenced by the nutrient:energy ratio. Reliable estimates of the bird’s energy requirement and the availability of energy from the food are therefore essential foundations for the accurate for- mulation of a diet which provides biologically, economically or environmentally optimal intakes of specific nutrients. Energy is not a chemical entity; it is a sum- mation of the biologically available energy of the chemical constituents of the food. There is a good argument for describing a diet in terms of nutrient chemi- cal contents (e.g. amino acids, carbohydrates, fats), but this should be done in the knowledge that some of the bird’s physiological control mechanisms may perceive these substrates as contributors to energy supply rather than as spe- cific chemicals. It is therefore necessary to describe a feedstuff or ingredient in terms of both its biological energy value and its chemical composition. A survey of feed evaluation in Europe (de Boer and Bickel, 1988) listed the wide range of energy evaluation systems in use. The section on poultry was con- spicuous for its comparatively high degree of consensus, with almost all coun- tries using a metabolizable energy system (metabolizable energy, ME = gross energy − [faecal + urinary energy]). ME is a reliable index of what is available to the bird for maintenance and production, but not a predictor of how efficiently the bird then uses what is available. Its low variability (Hill and Anderson, 1958) is a consequence of ignoring many of the bird’s metabolic responses to its food. ME is assumed to be linearly additive, which is practically convenient but not exhaustively tested (MacLeod et al., 1996). There have been many authoritative reviews on ME (e.g. Miller, 1974; Sibbald, 1982; Fisher, 1989; Fisher and McNab, 1989). This review will, therefore, attempt to introduce or reiterate some of the key ideas on ME but will not attempt to repeat comprehensive coverage. It will also try to introduce some alternative approaches to energy evaluation. As long as requirements for energy and provision of energy by diet ingredients are expressed in the same form, accurate relativity between energy values may be more critical than absolute accuracy in allowing substitution of ingredients INTRODUCTION Poultry Chapter 12 29/5/02 11:22 AM Page 191 during diet formulation; internal consistency is therefore essential in any feed evaluation system, although a systematic bias in both response estimates (or ‘requirements’) and ‘bioavailability’ may be tolerable. The yardstick of any feed evaluation system is how well it satisfies these requirements. Any discussion on energy evaluation must be informed by a knowledge of how energy is partitioned among different functions. A classical representation is shown in Fig. 12.1, which is followed by a list of terms used in energy evalua- tion (Table 12.1). Three general types of energy balance experiment for measuring ME were identified by Fisher and McNab (1989) and are quoted below: 1. Traditional assays which involve preliminary feeding periods to establish ‘equilibrium’ conditions. Differences in carryover in the digestive tract between the beginning and end of the assay period (‘end-effects’) are controlled by try- ing to ensure that they are the same. Complete diets must be used in most cases and substitution methods used for ingredients. 2. Rapid assays, using starvation before and after giving a known aliquot of test feed to control ‘end-effects’, but which permit the birds free access to the feed. Again, complete diets and substitution methods must be used in most cases. 3. Rapid assays, as above, but using tube-feeding (or force-feeding or preci- sion feeding) to place the feed directly in the bird’s crop. These methods usu- ally avoid the need to substitute feed ingredients into a basal diet. Examples of all three types will be briefly described below and can easily be ascribed to a category. The word ‘apparent’ refers to the fact that the droppings collected for AME measurement contain matter endogenous to the bird (e.g. gut secretions, gut epithelial cells) as well as undigested food material and urine. It should be noted that only the urine is correctly termed ‘excreta’ in a physiological sense. The combined faeces and excreta of the bird should strictly (if unexcitingly) be called ‘droppings’, although ‘excreta’ has entered common useage, presumably because it sounds more scientific. A typical protocol for AME measurement (The European Reference Method, Bourdillon et al., 1990a) is shown in Fig. 12.2. This is one of many variations on the same basic technique. If an ingredi- ent cannot be fed alone, for practical reasons or because of unpalatibility, AME is usually determined by substitution of the test ingredient for part of a basal 192 M.G. MacLeod ENERGY PARTITION ME MEASUREMENT AND PREDICTION Apparent Metabolizable Energy Poultry Chapter 12 29/5/02 11:22 AM Page 192 Marker methods Metabolizability of energy, like the digestibility of other nutrients, can be mea- sured by marker ratio methods. Markers used include chromic oxide, titanium dioxide (Peddie et al., 1982) and endogenous or added acid-insoluble ash (Scott and Hall, 1998). Nitrogen correction Correction to zero nitrogen retention makes the simplification that the feed- stuff evaluated is used entirely as an energy source. The justification for this correction is that ME is purely an energy evaluation system, so materials should be assessed only for their energy value. The correction involves cal- culating (by the following equation) the additional energy which would appear in the excreta if any nitrogen retained were instead to be catabolized and excreted. ([faecal+urinary] energy)N = ([faecal+urinary] energy) + 34.4 (N intake − [faecal+urinary] N) where 34.4 kJ g1 is the mean gross energy of the nitrogenous excretory prod- ucts in the bird. The correction to excreta nitrogen is negative in many assays because of low nitrogen intake and N-corrected ME is therefore higher than the uncor- rected value. However, in cases where there is sufficient protein intake to give a positive nitrogen retention, N-corrected ME is lower than the uncorrected value. This has sometimes prompted the criticism that high-protein raw mate- rials are ‘penalised’ by correction to zero nitrogen balance. However, it is to be expected that high-protein materials are being included in the diet as sources of amino acids and that their inclusion will be determined on this basis, rather than their energy value. Nitrogen correction is also applied to endogenous energy losses (EEL) in measuring TME; since nitrogenous matter makes up a large proportion of endogenous losses, EELN is much lower than EEL (McNab and Blair, 1988). True metabolizable energy is ME corrected for endogenous energy losses (Table 12.2). It should not be taken as inextricably linked with tube-feed- ing techniques, as it often is. A correction for endogenous losses can be applied to measurements made by any ME technique. However, TME came into vogue at the same time as Sibbald’s (1976) method, because the limited food intake consequent on this technique meant that endoge- nous losses became a more significant factor in the accuracy of the mea- surement. The relationship between AME, TME and endogenous energy loss is shown graphically in Fig. 12.3, which also contains the equation for the relationship. This relationship was explored thoroughly by Jonsson and McNab (1983). Energy utilization: measurement and prediction 195 True Metabolizable Energy Poultry Chapter 12 12/6/02 9:52 AM Page 195 Tube-feeding has the advantages of precision in the measurement and timing of intake and of allowing the feeding of substances which (although nutritious) may be unpalatable at high concentrations (Table 12.3). There have been assertions that tube-feeding may not be as suitable as longer-term ‘self-feeding’ for assessing the efficacy of feed enzymes. If there is a scientific basis for these doubts, it would prob- ably be that the gut microflora need time to adapt to the changes in substrate com- position produced by the action of the exogenous (feed) enzymes. I have been unable to find published evidence for this effect. In any case, the effect of enzymes 196 M.G. MacLeod Table 12.2. Roslin true metabolizable energy (TME) protocol (McNab and Blair, 1988). Time (h) Tube-fed birds Fasted birds (for EEL) 0 Food withdrawn Food withdrawn 24 50 ml glucose solution 50 ml glucose solution 48 50 g test material 50 g glucose (granulated) 72 50 ml water 50 ml water Palpate crop 96 Droppings collection Droppings collection Period may be extended by 24 h M E ( kJ g 1 ) 10 5 0 0 25 50 75 100 Food intake (g) TME AME AME (kJ g1) = TME (kJ g1) = EEL (kJ g1) / FI (g day1) Fig. 12.3. The relationships between apparent metabolizable energy (AME), true metabolizable energy (TME) and food intake. EEL is endogenous energy losses. The curve was drawn using a realistic estimate of EEL. The figure demonstrates how a correction for EEL becomes larger and more necessary as food input decreases. Feeding Methods Poultry Chapter 12 29/5/02 11:22 AM Page 196 on metabolizability and digestibility is often not large enough to allow the detection of a superimposed treatment (such as mode of feeding) with statistical confidence. The advantages of tube-feeding would also be applicable to the measure- ment of the net energy of raw materials or whole diets. On the other hand, tube- feeding clearly circumvents feeding activity; the absence of upper alimentary tract and central stimuli of feeding may also affect centrally mediated general locomotory activity or other energy-using processes. MacLeod (1991) used rapid-response indirect respiration calorimetry (Lundy et al., 1978; MacLeod et al., 1985), with simultaneous Doppler-radar activity measurement (MacLeod et al., 1982) to measure the effect of tube-feeding on energy metabolism and phys- ical activity (Table 12.4). The thermogenic effect of feeding was about 30% less with tube-feeding, over-estimating net energy by 4%. This difference was entirely accounted for by the reduction in feeding-related activity. The theoretical relationship between true and apparent ME can be confirmed when birds are given the same amount of food by tube or by self-feeding and there is no differ- ence due to feeding method (MacLeod, 1991). Both AME and TME were the same in tube-fed and self-fed birds. Variance of all measurements was lower in tube-fed birds, probably because of the removal of variation in food intake. Energy utilization: measurement and prediction 197 Table 12.3. Comparison of tube-feeding and ‘self-feeding’ for evalua- tion of raw materials and diets; the greater the number of asterisks, the greater the advantages of the technique. Tube-feeding Self-feeding ‘End effects’ (carry-over etc.) *** ** Accuracy in measuring food intake *** ** Accuracy in timing of feeding *** * Accuracy in timing of collection *** *** Feeding of unpalatable raw materials *** * Errors due to substitution *** * Tests for action of exogenous enzymes ** *** Table 12.4. A comparison of results obtained by tube-feeding and self- feeding the same quantities of a complete diet (MacLeod, 1991). Self-fed Tube-fed Mean SEM Mean SEM Food intake (g) 46.4 3.95 50.0 0.01 Gross energy intake (kJ day−1) 789 68.1 850 1.8 AME intake 614 54.0 664 2.5 TME intake 674 54.4 722 6.1 A metabolizability 0.78 0.005 0.78 0.003 T metabolizability 0.86 0.012 0.85 0.007 AME (kJ g−1) 13.2 0.09 13.3 0.05 TME (kJ g−1) 14.6 0.19 14.4 0.12 SEM, Standard error of mean. Poultry Chapter 12 29/5/02 11:22 AM Page 197 energy costs of ingestion, absorption and metabolism, along with any more direct effects on the bird’s metabolic rate. The heat increment, and hence the net energy, of a feedstuff varies with its chemical composition; in general terms, the net energy values of fat and protein per unit of ME are respec- tively 20% higher and 20% lower than that of carbohydrate. The net effi- ciency of utilization of ME (or net availability of ME) can be defined, for calculation purposes, as the increase in energy retention which occurs per unit increase in the quantity of food energy ingested; it can, therefore, be calculated as the slope of the relationship between retention and intake over the appropriate range of intakes (possibly just two intakes within the range of interest). This slope (net efficiency) is conventionally given the symbol k (Table 12.1). k = dRE/dIME k is often given a subscript to indicate the fate of the retained energy: km, maintenance; ko, egg production; kf, fat deposition; kp, protein deposi- tion. Because of the differing energetic efficiencies of the chemical transfor- mations involved in these processes, k for a given feedstuff can vary, depending on intake and on the physiological state of the animal. For instance, because birds control their body temperature, there are strong interactions between dietary energy and the thermal environment. The inter- action is clearest in the case of energy requirement for maintenance and, therefore, gross energetic efficiency. However, there are reasons to think that ambient temperature would also affect partial (net) efficiency of energy uti- lization. One hypothesis related to this is that the heat increment of feeding would substitute for (facultative) thermoregulatory heat production, leading to higher partial efficiency of energy retention as ambient temperature decreases, until a temperature is reached where the entire heat increment of feeding substitutes for thermoregulatory heat production, giving a measured partial efficiency of 1.00 (100%). This type of pattern has been shown in pigs, but has not been demonstrated so clearly in poultry. When fasted or underfed at any temperature, fowl and turkeys allow their metabolic rate to decrease, and make greater use of heat-conservation mechanisms, even let- ting their body temperature decrease rather than increasing thermoregula- tory heat production (MacLeod et al., 1993). The effect of this is that the heat increment of feeding, and hence net energy, may change little over a wide range of temperatures (MacLeod, 1990, 1992). (Conversely, at high ambient temperatures, poultry seem to be slow in reducing thermoregula- tory heat production to allow the heat increment of feeding to be expressed, so that a reduction in food intake is the only rapid mechanism for reduction of metabolic rate (Francis et al., 1990; MacLeod and Hocking, 1993; MacLeod et al., 1993)). The relative sensitivity of net energy to the state of the animal may be seen as an advantage or a disadvantage, depending on the viewpoint taken. It is obviously a more complete indicator of response than ME but is therefore more prone to variability, for real biological reasons. 200 M.G. MacLeod Poultry Chapter 12 29/5/02 11:22 AM Page 200 MODELLING THE UTILIZATION OF DIETARY ENERGY BY POULTRY1 The perception that it is relatively inexpensive to do experiments with poultry has had both positive and negative effects on the development of a modelling approach. The positive effects are that there is a large body of data to underpin modelling and also that validation should be less costly than with other species. The negative effect is that there may be a temptation to do a new experiment for every new set of circumstances. Digestibility and metabolizability measurements should probably be exempted from this charge, since bio-assay is sufficiently rapid and precise to remain the method of choice over predictive methods (Sibbald, 1976; Farrell, 1978; McNab and Blair, 1988; Bourdillon et al., 1990a,b). However, adherence to an empirical approach in poultry nutrition has meant that relatively few scientists have made an active contribution to the development of predictive models. This section of the chapter briefly describes some empirical and mechanistic models, as well as models which have some characteristics of both (see France and Thornley, 1984, for definitions). Advances in energy evaluation for poultry are likely to come from taking into account the differing biochemical efficiencies with which the chemical con- stituents of the diet are utilized (Millward et al., 1976). This is most likely to be achieved by a modelling-based approach, of which there have already been attempts at varying levels of mechanism and aggregation (Nehring and Haenlein, 1973; de Groote, 1974a; Emmans, 1994; MacLeod, 1994). Although demonstrably capable of distinguishing between the energetic effects of feeding different classes of chemical substrate (Tasaki and Kushima, 1979), or between grossly different compound feeds (MacLeod, 1990, 1992), calori- metric experiments need copious replication to detect the effects of more subtle changes in diet composition. These relatively small effects, which may still be of biological and commercial significance, are more likely to be detected by a valid predictive model than by any but an extremely large experiment. Net energy (NE), as defined above, is that part of the dietary energy which is available to the animal for maintenance and production (i.e. NE = ME – heat increment, where heat increment is the increase in heat production which occurs when food is ingested). A ‘productive energy’ system (in most respects analogous to NE) was used commercially from about 1946 to about 1960, fol- lowing the work of Fraps and Carlyle (1939), and Fraps (1946). This system was based on ‘comparative slaughter’ measurements of energy retention and estimates of maintenance energy requirement calculated from body weight. Energy utilization: measurement and prediction 201 1Much of the material in the remaining part of this chapter has been published previously in a different form (MacLeod, 2000). Empirical Predictive Modelling of Net Energy Poultry Chapter 12 29/5/02 11:22 AM Page 201 Productive energy was discarded in favour of ME because of its lack of preci- sion (ranges of up to 20% for a single feedstuff). Much of this variability may have resulted from the measurement technique, in which inter-individual varia- tion was combined with errors inherent in the comparative slaughter method and in the method of calculating maintenance requirement from bird weight. The Rostock Net Energy (NEF) model Work at Rostock (Nehring, 1967; Schiemann, 1967; Schiemann et al., 1971; Nehring and Haenlein, 1973) demonstrated that excessive variability is not an inevitable characteristic of net energy systems. Extensive measurements with several species showed that net energy (for fat deposition) can be predicted from digestible fat, protein and carbohydrate contents of feedstuffs with a coef- ficient of variation of about 5%. The Rostock NEF system (net energy for fat deposition) was based on a large number of calorimetric trials on the main agri- cultural species. The multiple-regression models derived from these trials pre- dict NEF from measurements of digestible crude protein (P), digestible crude fat (F) and digestible crude fibre + digestible N-free extract (largely carbohy- drate, C). The model for the domestic fowl (derived from adult cockerels) is: NEF (kJ g−1) = 10.8P + 33.4F + 13.4C ( cv 5.2%) (1) The corresponding equation for ME is: ME (kJ g−1) = 17.8P + 39.8F + 17.7C (2) The coefficients relating NEF to ME (and therefore equivalent to kf, net effi- ciency of energy utilization for fattening) are 0.60, 0.84 and 0.78 for protein, fat and carbohydrate, respectively. The NEF model was based specifically on the utilization of energy for fat deposition, a decision cogently argued by Nehring (1967). The NEF system appeared to satisfy many of the criteria for a practical NE system. Its non-use in most countries may have stemmed from its reliance on digestibility coefficients, which would have been identified as a par- ticular problem in the case of poultry. A modification of the Rostock NEF Model for poultry De Groote’s (1974a) proposed NE method for poultry by-passed the uncer- tainty associated with digestibility coefficients by deriving NE values from exist- ing ME data. This was done for each feedstuff by multiplying its ME value by the relative proportions (by weight) of crude protein, crude fat and starch+sugar, each of which was in turn multiplied by an experimentally esti- mated utilization coefficient. The coefficients were 0.60, 0.90 and 0.75 for pro- tein, fat and carbohydrate, respectively. The calculation was therefore: NE = ME(0.60P + 0.90F + 0.75C)/(P+F+C) (3) where P, F and C are protein, fat and carbohydrate contents in g kg−1. In striving for simplicity, de Groote’s system neglected differences in the digestibility of the protein, fat and carbohydrate fractions both within and between feedstuffs. Perhaps more importantly, by using proportions of feedstuff by weight, it appeared to neglect the fact that fat has about twice the ME value 202 M.G. MacLeod Poultry Chapter 12 29/5/02 11:22 AM Page 202 Energy utilization: measurement and prediction 205 Table 12.7. Relative substitution values of raw materials on the basis of ME and on the basis of net energy calculated by the 2 NE equations (3 and 4) given in the text. NE is based on proportions of protein, fat and carbohy- drate by weight, NE1 on proportions by ME value. Wheat is given a relative value of 100 units as a standard for comparison. The same raw materials, with the same names, are used as in the original paper (de Groote, 1974a). The practical economic importance of using different models of feeding value may lie in the effects on the substitution values of diet ingredients. Ingredient ME NE NE1 NE1/ME Wheat 100.0 100.0 100.0 1.00 Maize 111.4 113.0 113.4 1.02 Rye 92.9 93.0 92.8 0.99 Barley 88.6 88.9 88.8 1.00 Oats 85.1 85.3 86.3 1.01 Sorghum 107.5 108.3 108.4 1.01 Wheat shorts 65.6 64.2 65.3 1.00 Wheat flour middlings 85.7 84.1 85.0 0.99 Wheat bran 42.2 41.2 42.0 1.00 Wheat white middlings 96.4 95.6 95.7 0.99 Tapioca 96.4 98.8 98.2 1.02 Rice germ meal 89.0 91.1 94.3 1.06 Dried whey 62.0 61.5 61.2 0.99 Molasses 63.6 63.9 63.3 0.99 Sugar 120.1 122.6 122.2 1.02 Groundnut meal 85.7 75.0 74.8 0.87 Sesame meal 64.6 56.9 57.5 0.89 Fish meal 100.6 89.5 94.4 0.94 Soybean meal (440 g kg−1 CP) 72.7 64.0 63.9 0.88 Soybean meal (500 g kg−1 CP) 80.2 69.5 70.0 0.87 Sunflower meal 55.5 48.7 48.9 0.88 Herring meal 103.5 90.3 94.2 0.91 Blood meal 92.9 76.8 76.6 0.82 Meat-and-bone scraps 66.1 58.9 62.1 0.94 D,L-methionine 113.6 93.4 92.5 0.81 L-lysine.HCl 85.2 70.1 69.4 0.81 Dried skim milk 81.2 75.7 75.3 0.93 Brewer’s yeast 67.2 59.2 59.5 0.89 Lucerne (160 g kg−1 CP) 36.9 34.2 34.9 0.95 Lucerne (180 g kg−1 CP) 43.8 40.5 41.6 0.95 Lucerne (200 g kg−1 CP) 48.7 38.8 46.2 0.95 Maize gluten meal (420 g kg−1 CP) 89.3 79.6 80.3 0.90 Maize gluten meal (620 g kg−1 CP) 111.4 98.6 100.2 0.90 Hydrocarbon yeast (BP) 82.5 68.7 69.1 0.84 Cottonseed meal 59.1 50.6 50.7 0.86 Lard 263.6 324.7 321.9 1.22 Tallow 240.3 296.3 293.4 1.22 Soybean oil 292.2 360.3 356.8 1.22 Poultry Chapter 12 29/5/02 11:22 AM Page 205 much of this work was not targeted at avian species, similarities between taxo- nomic groups are more important than the differences at this level of biological organization. The model of Livesey (1984) calculates energy yield (as ATP) from carbohydrates, fats and proteins which have been absorbed across the gut wall and are available for cellular catabolism. It treats all substances purely as energy sources and therefore corresponds with an ME form of evaluation. In fact, by assuming oxidation of amino acids it simulates the correction of ME to zero nitro- gen retention as discussed earlier in the case of practical ME measurements. ATP yields calculated stoichiometrically for individual amino acids, fatty acids, glycerol and glucose were related to gross energy contents of proteins and fats calculated from bond energies to give values for the gross chemical energy corresponding with a molecule of cytoplasmic ATP. Errors in this relationship are potentially large, depending on the stoichiometry selected for mitochondrial oxidative phos- phorylation, on the degree of uncoupling of oxidation and phosphorylation, and on the proportion of amino acids oxidized via gluconeogenesis. Taking account of these sources of error reduces the error range in ATP yield to about 10%. Most of the variation in ATP yield from protein was explicable in terms of real dif- ferences in heat of combustion depending on amino acid composition. Relatively little of the variation was due to differences in efficiency of ATP generation (com- pare Schulz, 1975). The residual uncertainty in food energy equivalents of cyto- plasmic ATP could probably be reduced by better information about the energy costs of absorption from the gut and translocation across other membranes. A further paper by Livesey (1985) examined the effects of oxidation-phosphoryla- tion uncoupling on the relationships among the energy equivalents of carbohy- drate, fat and protein. For a food evaluation system, in which the prime objective is to obtain reliable relative replacement values for ingredients, these scientifically important finer details may not be of critical importance. Previous models suggested for a net energy system have been highly aggregated in chemical terms (Nehring and Haenlein, 1973; de Groote, 1974a). Such systems may have the advantage of simplicity but, with better description of diet composi- tion, we can use more information to predict response. MacLeod (1994), there- fore, set out to develop a mechanistically based model of nutrient metabolism as a basis for a net energy type of food evaluation. The design specifications were that: 1. Variations in efficiency of utilization (heat increment) should be taken into account. 2. The system should, wherever possible, be based on the chemical composi- tion of ingredients. 3. It should allow for different rates and compositions of product synthesis. 4. It should avoid incorporating more assumptions than necessary, but should allow for refinement as more information accumulates. 5. It should provide, in the first instance, a reliable standard for substitution of ingredients for one another, i.e. a standard which is correct in relative terms. This relative standard should, however, lay the foundation for absolute correctness. 206 M.G. MacLeod A Simulation Model to Predict Dietary Net Energy Yield for Poultry Poultry Chapter 12 29/5/02 11:22 AM Page 206 Energy utilization: measurement and prediction 207 A simplified flow diagram of the model is shown in Fig. 12.4. The simula- tion is structured as a number of independent program units, which can be refined individually without interacting unpredictably with other parts of the program. The model incorporates several empirical relationships predicting whole-animal responses (e.g. food intake, maintenance requirement). Since this is primarily a food evaluation model, the main reason for predicting intake is to ensure that energy intake is within reasonable limits. All the model would be able to do with a large excess of energy above requirements and above its capacity for fat deposition would be to simulate ‘burning it off ’ as heat (regula- tory diet-induced thermogenesis). This would be biologically unrealistic because the domestic fowl does not usually exhibit a capacity for regulatory diet-induced thermogenesis. Whenever possible, the simulation uses experimentally determined values of the digestibilities of chemical entities within the ingredient (Heartland Lysine, 1990; Longstaff and McNab, 1991; Shafey and McDonald, 1991; Zuprizal et al., 1993). The stoichiometric foundation of the simulation was derived largely from Schulz (1978). Because of the differences between mammalian (urea-excreting, ureotelic) and avian (uric acid-excreting, uricotelic) amino acid metabolism, however, different stoichiometric coefficients were used for amino acid break- down. The energy cost of uric acid synthesis is accounted for. Amino acid com- positions of proteins in body, feathers and egg were compiled from various sources (Lunven et al., 1973; Hakansson et al., 1978; Blair et al., 1981; Nitsan et al., 1981; Hurwitz et al., 1983). The original information on the existing ingredients was derived largely from analyses at the Roslin Institute (e.g. Blair et al., 1981; McNab and Scougall, 1982), but new ingredients can be added and existing analyses edited by the user. If an amino acid analysis for a new ingredient is not available, the simulation can make estimates from a measurement of crude protein concen- tration and a statement of the most closely related feedstuff for which a full analysis exists (e.g. another legume species, another grain species). The model predicts how consumed nutrients are partitioned between different biological processes (e.g. body growth, egg production, body maintenance) and how effi- ciently the nutrients are used, in energy terms. The model has been designed particularly with food evaluation in view, so the prediction of utilization effi- ciency is given priority in constructing the model. For instance, hierarchies of biological processes (e.g. maintenance, followed by egg synthesis, followed by body growth) are usually imposed, rather than allowing simultaneous competi- tion for resources on the basis of different affinity constants. The exception is to set up a competitive interaction between protein accretion and oxidation in the metabolic fate of amino acids; the comparative affinities of these two processes are summarized by a ‘growth potential’ term determined by the age and geno- type of the bird being simulated. The hierarchies used in the simulation are hypotheses based on the interpretation of published experiments; it is implicit in their use that, functionally if not intellectually, the animal also has priorities in the use of nutrients. Such priorities have developed through natural (or artifi- cial) genetic selection. The physiological control (endocrine, neural, etc.) of these priorities is sometimes known, but is outwith the scope of this model. Poultry Chapter 12 29/5/02 11:22 AM Page 207 210 M.G. MacLeod Emmans (1984, 1994) has published a well-argued method for estimating the ‘Effective Energy’ of a diet or ingredient, in which ME is adjusted for the heat increment of feeding by applying linear coefficients to five measurable compo- nents of the interaction between the animal and its diet. The coefficients were empirically derived but are broadly interpretable in relation to biological mech- anisms. ‘Effective energy’ is equal to ME less predicted heat increment. The concept, therefore, falls within the category of ‘net energy’ systems but the term ‘effective energy’ avoids some of the semantic and logical problems associated with using ‘net energy’ to describe diets, when it more correctly describes the animal’s response. ‘Effective energy’ can be applied across species, but I will try to draw attention to aspects that are particularly relevant to poultry. Metabolizable energy (ME) is chosen as the starting point for the effective energy calculation. MEc (subscript c for conventional) is defined as the gross energy (GE) of the diet less energy losses as faecal energy (FE), urinary energy (UE) and combustible gases (MTHE): MEc (kJ day −1) = GE − (FE + UE + MTHE). (5) The production of combustible gases (largely methane) by poultry is negligible, which allows simplification of the equation to the form used in poultry feed evaluation. Correction of ME to zero nitrogen retention (NR), to give MEn, is also permissible, corresponding with standard procedure in the case of poultry. MEn(kJ day −1) = MEc – a(6.25 NR) = (hp – a).PR + hl.LR + H (6) where PR and LR are the rates of retention of protein and lipid (g day−1), hp and hl are the heats of combustion of protein (23.8 kJ g −1) and lipid (39.6 kJ g−1) and H is heat production (kJ day−1). Although poultry excrete uric acid rather than urea, the trans-species nitrogen correction (a) of 5.63 kJ g−1 protein retained (35.2 kJ g−1 N) agrees closely with the 36.5 kJ g−1 N conventionally used for poultry. The ME obtained from the diet must either be retained in the animal’s body or lost as heat. If we know or can predict the performance of the bird in terms of protein and fat retention (carbohydrate deposition being negligible in the long term), prediction of ME requirement needs only a prediction of heat production. Most of Emmans (1994) is concerned with this. Heat production is described as having two components, fasting heat production (FHP) and heat increment of feeding (HIF). FHP is given by: FHP = – (hp – a).PR – hl.LR (7) where PR and LR are protein and fat retentions (which are negative in the fasted bird). Maintenance heat production (MH) is calculated with the simplify- ing assumption that the fasted bird is catabolizing only lipid: MH = FHP – wu.FUN (8) where wu (kJ g −1 N) is the heat production associated with the synthesis and excretion of urinary N and FUN is urinary nitrogen loss during fasting (g day−1). EFFECTIVE ENERGY Poultry Chapter 12 29/5/02 11:22 AM Page 210 Energy utilization: measurement and prediction 211 Heat increment for maintenance (HIM, kJ day−1), ignoring methane production in the case of poultry, is given by the equation: HIM = wd.FOM + wu.UN (9) where wd (kJ g −1) is heat production associated with the production of faecal organic matter, FOM. Maintenance ME requirement (MEM, kJ day−1) is given by MEM = MH + HIM (10) Under conditions of positive protein and fat retention (and ignoring methane production), HIF = wd.FOM + wu.UN + wp.PR + wl.LR (11) (where wp and wl are the heat productions associated with protein and lipid deposition, respectively) and ME requirement (kJ day−1) is given by: ME = ER + MH + HIF (12) (where ER is energy retention). For poultry, the coefficients for the different components of heat increment were estimated from the feeding and comparative slaughter experiments of Hakansson et al. (1978). Effective energy requirement (EERQ, kJ day−1), taking into account the energy contents of protein and lipid, the energy costs of depositing them and the energy cost of nitrogenous excretion was shown to be given by: EERQ = MH + 50PR + 56LR (13) The effective energy of an ingredient can be expressed as: EE = MEN – wd.FOM – 0.16 wu.DCP + 12.z.DCL (14) where DCL is digestible crude lipid (g g−1) and z is the proportion of retained lipid which comes directly from feed lipid. Although the concept of ‘effective energy’ is applicable across species and genetic lines, the actual values for raw materials do differ between genotypes and the de novo measurement of the effective energy of individual ingredients can be time-consuming (Farrell et al., 1997). Poultry industry economics are highly sensitive to relatively small changes in costs and production efficiency. It is, therefore, important to be able to predict as accurately as possible the biological and financial consequences of changes in feed formulation. The intention of this review was to provide a basis for discussion on what type of energy evaluation system can help towards this objective. Are scientific knowledge and techniques at a point where we can now make a soundly based attempt at using a net energy system and is such CONCLUSIONS Poultry Chapter 12 29/5/02 11:22 AM Page 211 212 M.G. MacLeod a system a worthwhile advance? The cost-effective and biologically support- able formulation of diets depends on the use of a reliable and accurate feed- evaluation system. In its simplest form, such a system should be reducible to two lists for each nutrient or resource, one of requirements and one of bioavail- ability from a list of ingredients. The adoption of such a simple scheme carries the penalty of ignoring the existence of between-ingredient interactions when mixtures are fed. Hesitancy in the commercial uptake of net energy (NE) mod- els is worth mentioning. Because of the potential for biochemical interactions, the degree of linear additivity of NE values of individual ingredients is often mentioned as a barrier to their use in feed formulation. However, if ME values are linearly additive, while NE values are not, it can only be because ME does not detect the interactions which are seen as complicating a NE system. Emmans (1994) appears to have chosen his words carefully in stating that ‘As effective energy values … are additive to the extent that ME values are additive, they can be used to formulate diets using linear programming’ (my italics). The same might well be said for other NE systems: even if the additivity is not perfect, prediction of animal response might still be more accurate than with a system (such as ME) that takes little account of post-absorptive metabo- lism. A possible criticism is that NE is a property not of the food but of the bird’s response to the food. This is clearly true, but must be true of any mean- ingful feed evaluation system. The accessibility of computer facilities encour- ages the suggestion that mathematical simplicity should not take priority over predictive power and biological consistency. A system with a degree of com- plexity can still be easy to use if it is well-programmed and well-documented. Also, the greater the potential control over diet composition, the more likely it is that complexity in diet description and analysis of bird response will be worth- while: poultry, and poultry diets, meet this criterion well. Balnave, D. (1974) Biological factors affecting energy expenditure. In: Morris, T.R. and Freeman, B.M. (eds) Energy Requirements of Poultry. British Poultry Science Ltd, Edinburgh, pp. 25–47. Blair, J.C., Harber, C.D., McNab, J.M., Mitchell, G.G. and Scougall, R.K. (1981) Analytical Data of Poultry Feedstuffs. 1. General and Amino Acid Analyses, 1977–1980. Occasional Publication No 1, ARC Poultry Research Centre, Roslin, Midlothian, UK, 84 pp. Boer, F. de and Bickel, H. (1988) Livestock feed resources and feed evaluation in Europe. Present situation and future prospects. Livestock Production Science 19, 1–410. Bourdillon, A., Carré, B., Conan, L., Duperray, J., Huyghebaert, G., Leclercq, B., Lessire, M., McNab, J. and Wiseman, J. (1990a) European reference method for the in vivo determination of metabolisable energy with adult cockerels: repro- ducibility, effect of food intake and comparison with individual laboratory methods. British Poultry Science 31, 557–565. Bourdillon, A., Carré, B., Conan, L., Francesch, M., Fuentes, M., Huyghebaert, G., Janssen, W.M.M.A., Leclercq, B., Lessire, M., McNab, J., Rigoni, M. and Wiseman, REFERENCES Poultry Chapter 12 29/5/02 11:22 AM Page 212 Energy utilization: measurement and prediction 215 Jonnson, G. and McNab, J.M. (1983) A comparison of methods for estimating the metabolisable energy of a sample of grass meal. British Poultry Science 24, 349–359. Kampen, M. van (1974) Physical factors affecting energy expenditure. In: Morris, T.R. and Freeman, B.M. (eds) Energy Requirements of Poultry. British Poultry Science Ltd, Edinburgh, pp. 47–60. Leeson, S. and Valdes, E.V. (1996) The prediction of metabolisable energy by near- infra-red analysis and in vitro techniques. Proceedings of the Australian Poultry Science Symposium 8, 39–52. Livesey, G. (1984) The energy equivalents of ATP and the energy values of food pro- teins and fats. British Journal of Nutrition 51, 15–28. Livesey, G. (1985) Mitochondrial uncoupling and the isodynamic equivalents of protein, fat and carbohydrate at the level of biochemical energy provision. British Journal of Nutrition 53, 381–389. Longstaff, M. and McNab, J.M. (1991) The inhibitory effects of hull polysaccharides and tannins of field beans Vicia faba L. on the digestion of amino acids, starch and lipid and on digestive enzyme activities in young chicks. British Journal of Nutrition 65, 199–216. Lundy, H., MacLeod, M.G. and Jewitt, T.R. (1978) An automated multi-calorimeter sys- tem: preliminary experiments on laying hens. British Poultry Science 19, 173–186. Lunven, P., Le Clement de St Marcq, C., Carnovale, E. and Fratoni, A. (1973) Amino acid composition of the hen’s egg. British Journal of Nutrition 30, 189–194. MacLeod, M.G. (1990) Energy and nitrogen intake, expenditure and retention at 20° in growing fowl given diets with a wide range of energy and protein contents. British Journal of Nutrition 64, 625–637. MacLeod, M.G. (1991) Effects of feeding by crop intubation on energy metabolism and physical activity in domestic cockerels. British Poultry Science 32, 1089–1095. MacLeod, M.G. (1992) Energy and nitrogen intake, expenditure and retention at 32° in growing fowl given diets with a wide range of energy and protein contents. British Journal of Nutrition 67, 195–206. MacLeod, M.G. (1994) Dietary energy utilisation: prediction by computer simulation of energy metabolism. In: Aguilera, J. (ed.) Proceedings of the 13th EAAP Symposium on Energy Metabolism of Farm Animals, Mojacar, Spain, pp. 237–240. MacLeod, M.G. (1997) Dietary energy utilisation: sensitivity of a net energy model to varying estimates of food intake, maintenance requirement and protein deposition. In: McCracken, K.J. (ed.) Proceedings of the 14th EAAP Symposium on the Energy Metabolism of Farm Animals, Newcastle, Northern Ireland, pp. 323–326. MacLeod, M.G. (2000) Modelling the utilization of dietary energy and amino acids by poultry. In: Theodorou, M.K. and France, J. (eds) Feeding Systems and Feed Evaluation Models. CAB International, Wallingford, UK, pp. 393–412. MacLeod, M.G. and Hocking, P.M. (1993) Thermoregulation at high ambient tempera- ture in genetically fat and lean broiler hens fed ad libitum or on a controlled-feeding regime. British Poultry Science 34, 589–596. MacLeod, M.G., Jewitt, T.R., White, J., Verbrugge, M. and Mitchell, M.A. (1982) The contribution of locomotor activity to energy expenditure in the domestic fowl. In: Ekern, A. and Sundstol, F. (eds) Proceedings of the 9th EAAP Symposium on the Energy Metabolism of Farm Animals. Department of Animal Nutrition, Agricultural University of Norway, pp. 297–300. MacLeod, M.G., Lundy, H. and Jewitt, T.R. (1985) Heat production by the mature male turkey Meleagris gallopavo: preliminary measurements in an automated indirect open-circuit multi-calorimeter system. British Poultry Science 26, 325–333. Poultry Chapter 12 29/5/02 11:22 AM Page 215 216 M.G. MacLeod MacLeod, M.G., Savory, C.J., McCorquodale, C.C. and Boyd, A.A. (1993) Effects of long-term food restriction on energy expenditure and thermoregulation in broiler breeder fowls (Gallus domesticus). Comparative Biochemistry and Physiology 106A, 221–225. MacLeod, M.G., McNab, J.M., Bernard, K. and Wilson, S. (1996) A possible synergistic interaction in the utilisation of blood meal and hydrolysed feather meal. British Poultry Science 37, S64. McNab, J.M. and Blair, J.C. (1988) Modified assay for true and apparent metabolisable energy based on tube feeding. British Poultry Science 29, 697–707. McNab, J.M. and Scougall, R.K. (1982) The tryptophan content of some feedstuffs for poultry. Journal of the Science of Food and Agriculture 33, 715–721. Miller, W.S. (1974) The determination of metabolisable energy. In: Morris, T.R. and Freeman, B.M. (eds) Energy Requirements of Poultry. British Poultry Science Ltd, Edinburgh, pp. 91–112. Millward, D.J., Garlick, P.J. and Reeds, P.J. (1976) The energy cost of growth. Proceedings of the Nutrition Society 35, 339–349. Morris, T.R. and Freeman, B.M. (1974) Energy Requirements of Poultry. British Poultry Science Ltd, Edinburgh, 199 pp. Moughan, P.J. (1994) Modelling amino acid absorption and metabolism in the growing pig. In: D’Mello, J.P.F. (ed.) Amino Acids in Farm Animal Nutrition. CAB International, Wallingford, UK, pp. 133–154. Nehring, K. (1967) Investigation on the scientific basis for the use of Net Energy for fat- tening as a measure of feed value. In: Blaxter, K.L., Kielanowski, J. and Thorbek, G. (eds) Proceedings of the 4th EAAP Symposium on the Energy Metabolism of Farm Animals, pp. 5–20. Nehring, K. and Haenlein, G.F.W. (1973) Feed evaluation and ration calculation based on net energy. Journal of Animal Science 36, 949–964. Nitsan, Z., Dvorin, A. and Nir, I. (1981) Composition and amino acid content of carcass, skin and feathers of the growing gosling. British Poultry Science 22, 79–84. Peddie, J., Dewar, W.A., Gilbert, A.B. and Waddington, D. (1982) The use of titanium dioxide for determining apparent digestibility in mature domestic fowls (Gallus domesticus). Journal of Agricultural Science 99, 233–236. Preez, J.J. du, Hayes, J.P. and Duckitt, J.S. (1986) A rapid method to evaluate metabo- lizable energy and availability of amino acids without fasting and force-feeding experimental animals. South African Journal of Animal Science 16, 47–53. Scheele, C.W., Janssen, W.M.M.A. and Verstegen, M.W.A. (1976) Feed energy that can be stored in high energy phosphate bonds, as a determinant for predicting the energy value of poultry feeds. In: Vermorel, M. (ed.) Proceedings of the 7th EAAP Symposium on the Energy Metabolism of Farm Animals, Vichy, France. Guy de Bussac, Clermont-Ferrand, pp. 277–280. Scheele, C.W., van Es, A.J.H. and ten Have, H.G.M. (1985) The efficiency of energy utilisation by broiler chickens in relation to dietary composition. In: Moe, P.W., Tyrrell, H.F., Reynolds, P.J., van der Honing, Y. and Close, W.H. (eds) Proceedings of the 10th Symposium on Energy Metabolism of Farm Animals. Airlie, Virginia, USA, pp. 202–205. Schiemann, R. (1967) The scientific demands made of a system for evaluating feeds on energy sources and progress made towards their realisation. In: Blaxter, K.L., Kielanowski, J. and Thorbek, G. (eds) Proceedings of the 4th Symposium on the Energy Metabolism of Farm Animals, pp. 31–40. Schiemann, R., Nehring, K., Hoffmann, L., Jentsch, W. and Chudy, A. (1971) Energetische Futterbewertung und Energienormen. Deutscher Landwirtschaft Verlag, Berlin. Poultry Chapter 12 29/5/02 11:22 AM Page 216 Energy utilization: measurement and prediction 217 Schulz, A.R. (1975) Computer-based method for calculation of the available energy of proteins. Journal of Nutrition 105, 200–207. Schulz, A.R. (1978) Simulation of energy metabolism in the simple-stomached animal. British Journal of Nutrition 39, 235–254. Scott, T.A. and Hall, J.W. (1998) Using acid insoluble ash marker ratios (diet:digesta) to predict digestibility of wheat and barley metabolizable energy and nitrogen reten- tion in broiler chicks. Poultry Science 77, 674–679. Shafey, T.M. and McDonald, M.W. (1991) The effects of dietary concentrations of min- erals, source of protein, amino acids and antibiotics on the growth of and digestibil- ity of amino acids by broiler chickens. British Poultry Science 32, 535–544. Sibbald, I.R. (1976) A bioassay for true metabolizable energy in feedstuffs. Poultry Science 55, 303–308. Sibbald, I.R. (1982) Measurement of bioavailable energy in poultry feedingstuffs: a review. Canadian Journal of Animal Science 62, 179–187. Tasaki, I. and Kushima, M. (1979) Heat production when single nutrients are given to fasted cockerels. In: Mount, L.E. (ed.) Proceedings of the 8th EAAP Symposium on the Energy Metabolism of Farm Animals, Cambridge, UK. Butterworths, London, pp. 253–256. Valdes, E.V. and Leeson, S. (1992a) Measurement of metabolizable energy in poultry feeds by an in vitro system. Poultry Science 71, 1493–1503. Valdes, E.V. and Leeson, S. (1992b) Near infra-red analysis as a method to measure metabolizable energy in complete poultry feeds. Poultry Science 71, 1179–1187. Valdes, E.V. and Leeson, S. (1994) Determination of metabolizable energy of feed grade fats based on the application of near infra-red reflectance to the multilevel test diets. Poultry Science 73, 172–178. Ward, S. and MacLeod, M.G. (1991) Doubly labelled water measurements of the energy metabolism of an avian species under different ambient temperatures and reproductive states. In: Proceedings of the 12th EAAP Symposium on the Energy Metabolism of Farm Animals. Kartause Ittingen, Switzerland, pp. 230–233. Zuprizal, Larbier, M., Chagneau, A.M. and Geraert, P.A. (1993) Influence of ambient temperature on true digestibility of protein and amino acids of rapeseed and soy- bean meals in broilers. Poultry Science 72, 289–295. Poultry Chapter 12 29/5/02 11:22 AM Page 217