Abstract:
146 TRUE METABOLISABLE ENERGY (TME) AND THE ALTERNATIVE D.J. FARRELL* SUMMARY Reappraisal is made of two recently developed methods for determining metabolizable energy of poultry feeds and ingredients. The true metabolizable energy (TME) method appears to be useful for the measurement of the energy content of oils and fats. The major criticism of the TME method is that endogenous excreta energy of, starved cockerels used in computing' TME may not always correspond to that of fed birds. - Experimental data are presented to illustrate that-feedstuffs do not always give the same intercept value for endogenous excreta energy; this may be different from that of starved birds. The rapid method for determining metaboliz'able energy (ME) compares for a range of feeds and ingredients, with other methods of favourably, determination. Collection of excreta is not extended from 24 'to 32 hours. In a recent publication, independent comparisons made between the TME method and the rapid ME method favoured the latter. INTRODUCTION At a previous Nutrition School (Farrell 1977) I outlined the basis . of Sibbald's true metabolizable energy (TME) system (Sibbdld 1976) and. . introduced a new, rapid method of determining apparent metabolizable . energy (ME) of poultry feeds and feedstuffs (Farrell 1978). Modifications have been made to both methods in the light'of further research. In this papera critical assessment is made of TME and its usefulness as a method of assay. There is discussion of the rapid meth,od which includes comparison made with the conventional method of determining ME, The current situation is that energy requirements of poultrv are still expressed in terms of apparent metabolizable energy (ARC 1975; MRC 1977) as are the energy concentrations of feed ingredients. . TRUE METBOLIZABLE ENERGY In recent years Sibbald (1977a,b) has argued for the introduction of a TME system (Sibbald 1976) in which an attempt is made to separate the excreta originating from the feed from that which is endogenous in origin. In fowl, the latter stems from endogenous urine and metabolic faeces. At this stage the two important reasons for making a correction for endogenous excreta energy appear to be (i) that if food intake of the bird is low due to poor .acceptability, then the endogenous excreta become a disproportion- . . ate component of the total excreta and the apparent metabolizable energy (ME) value is depressed (Guillaume and Summers 1970), and (ii)' endogenous excreta are probably not constant but are to some extent characteristic of the bird and may be of the diet. Using the endogenous excreta output (if it can'be determined accurately) of the same bird when starved to correct its total excreta output although plausible, may not give the true correction value. *Department of Biochemistry and Nutrition, Universitv of New Enqland, Armidale, NS.W. 2351, r 147 One of the key questionsin the TME system lies in the validity of the estimation of the endogenous excreta voided. For the adult, starved cockerel there is considerable variation in excreta energy. Values may range from 7.9 to 19.6 kcal/day (Farrell 1978) and differences in body weight and weight loss during starvation can only explain a small amount 'of this variation (Sibbald and Price 1978). Indeed we have found that even for the same cockerel, day to day variationin endogenous excreta .'output is considerable (Farrell 1977, unpublished data). Because endogenous excreta are probably related to metabolic rate which increases with decreasing ambient temperature it would be expected thatendogenous excreta increases with decreasing temperature.. The observations of' Farrell and Swain (1977) are given for starved broilers in Table 1. TABLE 1 Effect of ambient temperature on endoqenous excreta of starved individual broiler chickens (1 kg) Another important question is the validity of using endoqenous excreta of starved birds as representative of the endogenous corn-ponent of birds when fed, and its verification by experimentation. Sibbald (1976) regressed energy voided as excreta against feed consumption for a range of feedstuffs. The zero intercepts which give an estimate of endogenous excreta energy, 'were similar for the 12 feedstuffs tested at 9.8 kcal/day. The reason for the common intercept forthe 12 equations was that 48 values .from birds starved for 48 hours were included in the regression analysis. For each of the 12 feedstuffs tested observations of fed birds ranged from 17 to 7, thusxthe addition of the 48 values for birds receiving no feed 'forced' theintercepts of the equations through a common origin. This approach did not allow meaningful statistical testing of individual equations to determine if the Y intercepts gave different values for individual equations. Such an approach would provide answers to the proposal that specific ingredients may influence endogenous excreta, and the use of starved birds to provide this value would therefore be inappropriate. Edmunds.on (personal communication 1977) observed that at least for one feedstuff, soybean meal, the intercept was different from other diets tested, and gave a value that tias higher. Furthermore, the level of intake of Sibbald's (1976) birds was low and did not usually exceed 30s ,per cockerel because of the force-feeding procedure. used. As a consequence, the regression ecruations derived bv Sibbald (1976) for the various feedstuffs did nothave particularly-imnressive correlation coefficients (r) although the number of observations exceeded 50 for each feedstuff. For example, for corn r was .0.85. Without the inclusion of the 48 observations of starved birds variation would have been even greater. As indicated the TME method as originallv described by (1976) involved the force-feeding of a pelleted feedstuff or to adult cockerels that had been starved for 21 hours. After 'excreta voided are collected quantitatively. This in itself Sibbald diet (30-40~~) 24 hours the is difficult when one considers that for some feeds the dry excreta originating from the feed ma? be only '5 or 6 g from an intake.of 30 g, together with #3' to 4 g of ,endogenous excreta. A weight-paired, unfed cockerel, was used in the original method to provide a measure of endogenous excreta energy voided during the 24 hour collection period. The calculation is as follows: TME. (kcal/g) = (G'E of feed x intake) - (GE of excreta of fed bird' - GE of excreta of starved bird) Weight of feed fed It is self-evident that the force-feeding procedure allows a precise but limited feed input of about 30-40 g. Because total excreta output is comparatively small, say 25 kcal from a cerealyrain, it is absolutely .essential therefore to make a correction for endoTenous excreta, which. may amount to 40% of excreta energy, if meaningful values are to result. The only meaningful energy system under these rather special circumstances is TME. . To illustrate the point,'shown in Fig. 1 is the relationship between the apparent metabolizable energy obtained at various levels of intake for the same ingredient. The.smaller the amount of ingredient offered the ' lower will be the ME value: Correction for endogenous excreta at any intake will allow calculation 'of the TME of the ingredient. Below about 50 g of feed there is a rapid decline in the ME value for the same ingredient, but above this amount there appears to be no good reason for such a correction since TME is greater than ME by essentially a constant amount. It is clear from the example given that TME is extremely sensitive to the value used for ,endogenous excreta. If the value does not apply specifically to the fed,bird under the experimental conditions, then a orecise TME of that feed ingredient can not be obtained. FIGURE 1 Relationship between apparent metabolizable energy value (y) and food intake (x) Although in principle the conventional method of determining . metabolizable energy of feeds with growing chickens fed the test diet .has many disadvantages, because of cost, large sample size, high labour input etc. (Sibbald 1975a) the only real advantaqe'of.TME, as developed by Sibbald (197'6) is that fats and oils can be f&e.-fed in relatively large amounts (Sibbald 1978; Halloran and Sibbald 1979; Sibbald and Kramer 1978, 1980b). In contrast, only relatively small amounts (1045%) can be added to a basal diet using conventional methods of determination: consequently ME is likely to be variable (Guirguis 1976). It would seem that TME has got considerable advantage over other methods when fats and oils are being investigated. Not only is it possible to introduce exact amounts of lipid into the crop but it is possible to examine effects of combinations of different fats and oils on TME (Sibbald and Kramer 1978). Even then there is doubt about the validity of correction for metabolic lipid excretion since this 'may be influenced by the nature and amount of diet consumed' (Sibbald and Kramer 1978). A rather disturbing aspect ofthe TME method is that it has been modified such that endogenous excreta are not now collected from a cockerel weight-paired, at the same time as the fed bird. Recently six or seven control birds have been used to establish the endogenous excreta loss (Sibbald and Price 1977; Sibbald and Kramer 1980) of the fed . cockerels. The real advantage of this TME system appears to be now .largely lost in that a constant value is being applied to endogenous -. excreta which is not only known to be variable but may tend to be characteristic of individual birds and perhaps of the diet. As stated previously there is when force-fed to adult cockerels, to allow the digestive tract to be Muztar and Slinger 1979). Clearly ingredients. now evidence that for some feedstuffs, a period of 24 h is not sufficient completely emptied (Sibbald 1979; it is necessary to identify these . There is some suggestion that TME values are less variable than ME values for a range of feedstuffs (Sibbald 1976). The basis of this ccanparison was bound to give misleading results, since the same cockerels given only 30-45 g of feed were used to make both measurements of ME and TME simultaneously. Sibbald (1975b) in an experiment in which birds were given different amounts of wheat, observed that as the intake declined below 45 g/day there was a decline in the ME of the wheat due to the contribution of endogenous excreta (see Fig. 1). Thus when a comparison is made between TME and ME with cockerels.force-fed only 30 g/day-; then TME is bound to give values with less variation than ME. Furthermore many of the ME values obtained by Sibbald (1976) do not correspond with published data. For example, the ME value for oats was 1.97 kcal/g, In recent ' fish meal was 2.07 kcal/g and dextrose 2.59 kcal/g. publications for many ingredients the standard error of the TME mean appears to be increasing. Reasons for this increase'are. that 'regurgitation of feed by force-fed cockerels can occur at high levels of input (Sibbald 1975,. 1976; Sibbald and Kramer 1980b) and excreta of food origin ,are not always voided during the next 24 hours (Sibbald 1979, 1980a; Muztar and Slinger 1979). As mentioned previously a key question that has notbeen satisfactorily answered is to what extent, if any, does diet influence. endogenous excreta output? An experiment was therefore designed in an attempt to answer this question. 150 ENDOGENOUS EXCRETA AND DIET Starved adult White Leghorn x Black.Australorp cockerels trained to consume their daily allowance inone hour were given each of nine pelleted' ingredients in amounts that ranged from 20 to'110 g. Excreta,were collected for the next 32 hours. The regression of excreta eneray (kcal,y) in 32 hours on feed intake '(g,x) was calculated for each feedstuff. Comparisons were made of the equations using analysis of variance and' covariance analysis to determine if the slope&and intercepts of the equations were different. To determine endogenous excreta of starved cockerels, twelve birds were starved for 32 hours and excreta were collected for the next 32 hours. Regression equations, calculated for each of the 9 ingredients are shown in Table 2. TABLE 2 Regression of excreta energy (Y) on feed intake (X) The TME value was calculated for each ingredient by subtracting from the gross energy of the ingredient th.e appropriate regression -. When the lines were tested statistically they were coefficient. significantly'different (PcO.01) in both slope and intercept. Thus the ingredient influenced significantly endogenous excreta. The mean value 'for the 12 starved cockerels was 18.2 kcal. Although these results are preliminary and more m.easurements are currently being made, it does appear that the use of 'endogenous excreta of starved birds to provide the corrector factor to obtain TME is under some circumstances inappropriate. RAPID DETERMINATION OF ME There is some discussion as to whether true metabolizable en&gy of a feedstuff is the most meaningful description of its usefulness in poultry nutritibn. As stated, the energy requirements of poultry are expressed in terms of ME (ARC 1975; NRC 1977). Secondly, it is doubtfiil whether a really valid argument can be raised to justify correction for endogenous excreta under normal circumstances. The requirement for energy is for metabolic processes, production and the replacement of endogenous losses. It may be useful to know the latter, -but it is an integral part of the bird's energy requirement. Althduqh it is of s interest to know the true metabolizable energy of a feedstuff the present method of determining TME may not provide the answer. 151 The need for such a correction forendoqenous excreta has been eliminated in the rapid method developed by Farrell (1978) in which adult cockerels are trained to consume their daily feed allowance within an hour. Not only does this remove the trauma of force-feeding, .which may influence digestive processes, but ,intake is sufficientlv high (80-110 g) .e to remove any necessity to correct for excreta of endogenous origin (see Fig. 1). Although both this and Sibbald's method are rapid and low cost, we have found that we canuse adult cockersleach day on different diets because'there is apparently no need for a period of adjustment. The low .cost and minimum time required to obtain resul'ts are major arguments in favour of Sibbald's TME system. These arguments also apply to the recent method developed by Farrell (1978). We have now extended our collection of exc,reta from the trained cockerels to 32 hours for all feedstuffs,,since some feeds do not clear the digestive tract in 24 hours. In order to standardize the procedure within normal working hours we are using this elapsed time for all feedstuffs and feed ingredients. The use of adult cockerels in the determination of metabolizable energy has a number of advantages. The birds can be used repeatedly, and we have been using the.same birds for almost three years. Because they are maintaining constant weight they are in nitrogen equilibrium se and this dubious correction to values measured with growing chickens for protein retention (Farrell 1979) is avoided. Values obtained are in terms of ME and compare favourably with those determined with growing chickens and those predicted from chemical composition. In Table 3 are given ME data on both calculated, predicted and determined on formulated diets and collated by Dr. T.R. Walker (personal communication). TABLE 3 Apparent metabolizable energy (Meal/kg) calculated, predicted and determined using the rapid method (UNE) and conventional biological method (Sydney University) for four diets It is quite clear from Table 3 that for formulated diets the rapid method of determination gives va1ue.s that are in close agreement with those predicted from chemical composition, and those determined using grotips of chickens 'and collecting excreta over several days in the conventional manner. More recently we have completed a comparison between the rapid method using a 32 hour collection period, the conventional method (Dr. D. Balnave, University of Sydney) and ME calculated from chemical analvsis There were seven diets, of (Dr. Dai Suter, N.B. Love Industries'Ltd.). these two were formulat'ed diets used in production,.one diet was a basal diet consisting of 87% corn'and 8% fish meal, two diets of the basal diet (50%) + meat meal (50%) or sunflower meal (50%) and the other two diets 152 were all-sorghum,or all-barley with mineral arid vitamin additions. All diets were pelleted. The preliminarv results are given in Table 4. ., TABLE 4 Comparison of the metabolizable energy (Mcal/kq) o.f feedstuffs and diets using two biological and one predicted-method It is interesting that the major discrepancy between the two . biological methods existed for diets 4 and 5 which contained unusually .. large amounts (50% of meat meal.and sunflower meal. Intake on all diets was at least 80 g which is well above the minimum for adult cockerels necessary to obtain valid measurement of ME (Fig. 1). Sibbald (19'77a) pointed out that 'if the t.est diet is un-palatable then feed intake will be less than that of the basal diet? This . difference in feed intake-can have a profound effect on the observed A.M.E. [apparent metabolisable energy]'. 'It is interesting that for these two diets (4 .and 5) ME values obtained by the rapid method agree closely with those predicted from chemical composition. . CONCLUDING REMARKS There is some doubt'about the correction used for endogenous excreta . to obtain TME.. It has been shown that some feedstuffs give amounts of .. endogenous excreta that are different from ,others; these are also different from the mean'value obtained for starved birds. Because fats and oils can,be included in diets in only relatively small amounts, there appears to be a real advantage in the use of TME to evaluate these feedstuffs which can be force fed both singly and in various combinations. One of the difficulties with the TME procedure as used bli Sibbald For is that there is no easy method of checking the values obtained. the rapid method this is easily and frequently done. Moreover because cockerels are hungry'at the commencement of their one-hour feed acceptability has never been a problem. If an intake of over 70 g is not achieved measurement is,discarded. Perhaps the most objective way of assessing both methods discussed here is to cite the recent work of Chami et al. `(1980) who compared TME and ME. Their conclusion was that 'Sibbald's method (TME) is fast, but not accurate for all feed ingredients. It is questionable to regard the TME method as scientifically more accurate than conventional ME methods. The method of Sibbald (1976) is no faster than the ME method of Farrell (1978). It may be more desirable to dilute .a basal diet with the test ingredient and determine ME according to the procedure of Farrell (1978) because excreta collection over 24,. 48 and 72 hour post-feeding gave the same ME values*'. ACKNOWLEDGEMENTS ' . I wish to thank Drs. T.R. Walker, D. Sut'er and D. Balnave for their cooperation in some of the experiments reported here. The technical assistance of Sherilee Ford and E. Thomson is gratefully acknowledged. REFERENCES 'The Nutrient Requirements of Farm Livestock. No. 1, ARC (1.975). (Agricultural Research Council: London). ' Poultry'. CHAMI, D.B.I VOHRA, P. and KRATZER, F.H. (1980). Poult. Sci., 59: 569. FARRELL, D.J. (1977). in 'Recent Advances in Animal Nutrition 1977', p. 1541 ed, D.J. Farrell (University of New England Publishing Unit). FARRELL, D.J. (1978). Br. Poult. Sci., 19: 303. FARRELL, D.J. (1979). Aust. J. Inst. Agric. Sci., 45: 21. FARRELL, D.J. and SWAIN, S. (1977). B.r . 'Poult. scr, 18: 725. GUILLAUME, J. and SUMMERS, J.D. (1970). Can. J. Anim. Sci., ,50: 363. GUIRGTJIS, N. (1976). Aust. J. Expt, Agric. Anim. Hus., 16: 6E. HALLORAN, H.R. and SIBBALD, 1-R. (1979). Poult. Sci., 5T 1299. MUZTAR, A.J. and'SLINGER, S.J. (1979). Nutrition Repor= International, 19: 689. 'Nutrient Requirements of Domestic Animals. No. 1: NRC (1-7). Nutrient Requirements of Poultry'. 7th ed. (National Academy of Sciences: Washington D-C.). SIBBALD, 1-R. (1975a). Feedstuffs, 47:(7), 22. SIBBALD, I-R.' (1975b): Poult. Sci-,54: 1990. SIBBALD, 1-R. (1976). Poult. Sci., 55: 303. 'SIBBALD, 1-R. (1977a). ' Feedstuffs, 49: (42), 21. 49: SIBBALD, 1-R. (197733). Feedstuffs, Z (43), 23. Poult. Sci., 57: 473. SIBBALD, 1-R. (1978). I SIBBALD, 1-R. (1979).. Poult, Sci., 58: 896. Poult. Sci-,59: 374. SIBBALD, 1-R. .(1980a). SIBBALD, R. and KRAMER, J.K.G. (1977)? Poult. Sci., 56: 2079. .SIBBALD, 1-R. and KRAMER, J.K.G. (1978). Poult. Sci.y57:'685. SIBBALD, 1-R. and KRAMER, J.K.G. (1980). Poult, Sci., 59: '316. SIBBALD, 1-R. and PRICE, K. (1977). Poult..Sci., 56: 2070. ?% SIBBALD, 1-R. and PRICE, K. (1978). Poult. Sci.,E 556.