Introduction liveweight change and energy value per unit of Energy requirement for maintenance

198 R .E. Agnew, T. Yan Livestock Production Science 66 2000 197 –215

1. Introduction liveweight change and energy value per unit of

liveweight change. The UK metabolisable energy ME feeding sys- tem, developed by Blaxter 1962, was first proposed for use in the UK in 1965 by the Agricultural

2. Energy requirement for maintenance

Research Council ARC, 1965. This system was designed to overcome the deficiencies of the starch 2.1. Methods of estimating energy requirement for equivalent SE system a net energy NE system maintenance which was then used in the UK. The SE system assumed a simple ratio of NE values of feeds for The NE requirement for maintenance NE in maintenance, fattening and lactation and also took no m energy feeding systems presently used in Europe and account of the effect of feeding level on NE con- North America was derived from calorimetric data. centration of a feed. Using the proposals put forward In the UK ME system the NE was based on fasting by ARC 1965, a simplified ME system was m metabolism data fasting heat production FHP plus recommended to be adopted in UK by the Ministry fasting urinary energy output from beef steers and of Agriculture, Fisheries and Food MAFF, 1975. dry non-pregnant dairy cows after a prolonged period The original ME system ARC, 1965 was later of restricted feeding usually at maintenance level. substantially revised by ARC 1980 and further Using this approach ARC 1980 reported a curvi- modified by Agricultural and Food Research Council linear relationship between fasting metabolism FM AFRC, 1990 and a new working version was 0.67 and liveweight LW FM50.53?LW 1.08 published in 1993 AFRC, 1993. At the same time a from a review of eight sets of data. This relationship, number of NE systems have been developed in plus an activity allowance 0.0091?LW, is taken as Europe Van Es, 1978; Institut National de Re- NE for use at present in UK AFRC, 1990. This cherche Agronomique INRA, 1978 and Northern m approach would suggest a fasting metabolism of America National Research Council NRC, 1978. around 0.30 or NE of 0.35 if an activity allowance There is no difference in principle between the m 0.75 is included MJ kg for an adult dairy cow. The ME and NE systems, with both systems recognising ME requirement for maintenance ME is calcu- that the energy requirement of cattle is the sum of m lated as NE divided by the efficiency of utilisation their energy requirements for maintenance, product- m of ME for maintenance k ,50.35?ME GE10.503 ion milk and liveweight gain and foetal growth. m AFRC, 1990. Alternatively, the NE can be esti- The only difference between them is where the m mated using regression techniques relating ME in- energetic efficiencies are embodied within the calcu- take to milk energy output, adjusted to zero energy lation. In the ME system the energetic efficiencies balance, with dairy cows offered diets at production are used for ration formulation and the prediction of levels. Using this approach, Moe et al. 1972 and animal performance, while in the NE system the Van Es 1975 reported NE values of 0.305 and efficiencies are included as part of the energy m 0.75 evaluation of feeds. 0.293 MJ kg , respectively from large sets of Over the last 2 decades a considerable volume of calorimetric data. The former value is used to form research in the energy metabolism of dairy cows has the American NE system, with an activity allowance been undertaken. These studies have highlighted a of proportionately 0.10 being added NRC, 1988. number of concerns over current energy feeding The latter value is adopted in the European NE systems. The purpose of the present review is to systems used in the Netherlands, France, Germany reflect on the impact of these studies on the UK ME and Switzerland. No activity allowance is adopted in system, and other NE systems, and validate these the Netherlands Van Es, 1978, while an activity systems using published calorimetric data. The areas allowance of proportionately 0.10 is added for loose which will be addressed in the present review housed cows in France INRA, 1989. include the energy requirement for maintenance, The NE in the UK ME system is a curvilinear m 0.75 efficiency of ME utilisation for milk production and function of liveweight and is reduced per kg , with R .E. Agnew, T. Yan Livestock Production Science 66 2000 197 –215 199 increasing liveweight age of cattle. This is because al., 1999. This type of infusion can also result in a the metabolic rate can be higher for growing than lower heat production than FHP Chowdhury, 1992. adult cattle ARC, 1965 and light adult animals However, the maintenance metabolic rate obtained 0.75 generally have a greater proportion of internal organs by fasting metabolism 0.30 MJ kg ARC, over total liveweight than heavy adult ones NRC, 1980 is similar to that derived from regression 0.75 1988. The internal organs produce much higher heat techniques 0.305 or 0.293 MJ kg Moe et al., than muscle per unit weight. In contrast, the NE for 1972; Van Es, 1975. It thus seems unlikely that the m 0.75 cattle is constant per kg in the NE systems used detriment of fasting to animal health influences in Europe and North America. In these systems the greatly the heat production. NE values were derived from data on mature dairy m cows. It is therefore likely that these latter systems 2.2. Recent research on energy requirement for may theoretically underestimate the energy require- maintenance ments of young cattle. The use of fasting metabolism data to determine The NE values currently used in Europe and m NE may have limitations. It has been suggested that North America were developed from data published m fasting after a long period of restricted nutrition can 30 years ago. However, a recent study reported by result in deamination of amino acids from tissue Birnie 1999 revealed a FHP value of 0.39 MJ 0.75 protein for the supply of essential glucose Chow- kg for dry, non-pregnant dairy cows fed at dhury and Ørskov, 1994. This can induce a range of maintenance level prior to measurement of FHP. metabolic disorders in the animal, such as hypo- Assuming a fasting urinary energy output of 0.05 of glycaemia, hyperlipidaemia, hyperketonaemia and FHP Van Es, 1972, the derived fasting metabolism hypoinsulinaemia. The deamination caused by fast- is proportionately 0.36 higher than that adopted in ing can however be reduced, as evidence of a lower current energy systems Van Es, 1978; NRC, 1988; N output in urine, after infusing a small amount of INRA, 1989; AFRC, 1990. Similar higher FHP volatile fatty acids VFAs or glucose with or values were also reported by Yan et al. 1997b and without casein Ku Vera et al., 1987, 1989; Ørskov et Birnie 1999 when dairy cows were offered, respec- Table 1 Fasting heat production of steers and dairy cows offered dried forage-based diets published since 1985 Reference Feeding Live weight Forage Fasting HP 0.75 level kg proportion MJ kg Beef and dairy steers Birkelo et al. 1991 1.2?Maintenance 344 0.40 0.357 2.7?Maintenance 344 0.40 0.383 Hotovy et al. 1991 Near maintenance 373 0.55 0.337 Carstens et al. 1989 Maintenance 236 0.40 0.350 Maintenance 372 0.55 0.331 Gill et al. 1989 Near maintenance 210 1.00 0.353 Smith and Mollison 1985 Near ad libitum 200 0.15 0.398 Near ad libitum 200 0.30 0.389 Dairy cows Birnie 1999 Maintenance 571 1.00 0.408 Maintenance 557 0.14 0.382 2?Maintenance 614 1.00 0.414 2?Maintenance 613 0.14 0.410 Yan et al. 1997b Near ad libitum 501 1.00 0.454 Near ad libitum 550 0.80 0.452 200 R .E. Agnew, T. Yan Livestock Production Science 66 2000 197 –215 Table 2 Summary of the ME requirement for maintenance ME and the efficiency of ME utilisation for lactation k by lactating dairy cattle, m l calculated by a range of authors using regression techniques and pooled calorimetric data Reference Cow Forage Method ME k m l 0.75 no. MJ kg Moe et al. 1970 350 Lucerne grass hay Multiple 0.51 0.64 Van Es et al. 1970 198 Hay silage Linear 0.49 0.62 Van Es 1975 1148 A range of forages Linear 0.49 0.60 Unsworth et al. 1994 108 Grass silage Linear 0.64 0.67 Hayasaka et al., 1995 53 Hay silage Linear 0.59 0.64 Yan et al. 1997a 221 Grass silage Linear multiple 0.67 0.65 Present study . 1500 A range of forages Linear multiple 0.62 0.66 Mean 0.57 0.64 S.D. 0.075 0.024 tively diets at near ad libitum and at twice mainte- MEI 5 0.664 MW 1 1.452 E 0.0471 0.0755 l nance levels prior to fasting Table 1. FHP data 2 1 1.079 E R 5 0.92 2 0.1200 g published using studies with beef and dairy steers since 1985, as presented in Table 1, are in accord where E , E , MEI and MW are, respectively milk with the above dairy cow results. The mean derived l g energy output, energy balance, ME intake and meta- fasting metabolism, when assuming a fasting urinary 0.75 bolic liveweight kg ; E 5 E 1 E for positive energy output of 0.05 of FHP Van Es, 1972, is l 0 l g E or 5E 2 0.84 ? E for negative E . The units for proportionately 0.19 higher than that proposed by g l g g 0.75 Eqs. 1 and 2 are, respectively MJ kg and ARC 1980. MJ day. The values in brackets are standard errors. The ME values derived from regression tech- m The mean ME derived from these two equations is niques have also been reported to be higher in recent m 0.75 studies. The ME values obtained in 6 studies are 0.62 MJ kg , a value which is proportionately m presented in Table 2. The mean ME value derived 0.27 higher than that derived from Van Es 1975, or m in recent studies Unsworth et al., 1994; Hayasaka et calculated from AFRC 1990. al., 1995; Yan et al., 1997a is proportionately 0.28 higher than that reported over 20 years ago Moe et al., 1970; Van Es et al., 1970; Van Es, 1975. As Van Es 1975 used a total of 1148 data from across the world, the NE derived from his study is thus m adopted in a number of European NE systems. The present authors have reviewed calorimetric studies with lactating dairy cows published since 1976. A total of 42 studies more than 1500 individual animal data were selected, in which the liveweights of the animals and the energy intake and outputs were available for references see appendix. The ex- perimental mean data from these 42 studies were used to examine the relationship between ME intake and energy outputs. The linear Fig. 1 and multiple regression equations obtained are: 2 Fig. 1. Relationship between ME intake and milk energy output E 5 0.637 MEI 2 0.371 R l 0 0.0358 0.0557 adjusted to zero energy balance using calorimetric data of dairy cows published since 1976 experimental means, n 542. 5 0.89 1 R .E. Agnew, T. Yan Livestock Production Science 66 2000 197 –215 201 The higher ME may reflect differences in both feed intake, total heat production MJ day was m the diet and the cow now used, particularly the found to be similar between fat and lean rats at a considerable improvement in cow genetic merit similar body lean mass although total liveweight of during the last two decades Coffey, 1992. The fat rats was heavier Ramsey et al., 1998. The latter has led to an increase in milk yield of estimated ME reported by Pullar and Webster m approximately 62 kg lactation per year Agnew et 1977 in Zucker rats and by Toutain et al. 1977 in al., 1998. Indeed, high producing dairy cows were sheep also showed no significantly difference in 0.75 found 30 years ago by Flatt et al. 1969b to require ME MJ kg when related to protein mass, but m proportionately 0.20 more ME for maintenance than was higher with lean than fat animals when related to cows producing moderate yield, as reported at the total liveweight. same time by Moe et al. 1970 and Van Es et al. A series of fasting studies, in dry and non-preg- 1970. The higher ME obtained in the recent nant Holstein–Friesian cows, carried out recently at m studies may be attributable to a higher proportion of this Institute would support the above results. Cattle liveweight as body protein mass. This is evidenced were fattened from condition scores CS Mul- in that high genetic merit cows had a lower backfat vanny, 1977 below 2.0 to over 4.5, or restricted thickness at a similar liveweight to medium and low feeding to reverse CS change. While FHP MJ 0.75 genetic merit animals Ferris et al., 1999a, and a kg was significantly higher for cattle with low higher estimated lipid-free empty body weight as a than high CS, the former animals required a similar proportion of empty liveweight Veerkamp et al., amount of estimated ME for maintenance MJ day 1994. The ME has been reported to be a function although they had a much heavier weight Birnie, m 0.75 of body protein mass discussed later. On the other 1999. A regression of FHP MJ kg against CS 2 hand, high genetic merit cows obviously require from 1.0 to 5.0 R 50.83, n 528 indicated that 0.75 greater nutrient intakes and this could stimulate the FHP was 0.483 MJ kg at CS of 1.0 and an activity of internal organs with greater digestive increase of CS by 1.0 would reduce FHP by 0.029 0.75 load, cardiac output and blood flow required to MJ kg . digest, absorb and deliver nutrients to the mammary The above findings support the view of Oldham gland and a greater oxygen consumption Reynolds, and Emmans 1990 that the major part of the energy 1996. These activities in return can enlarge the cost associated with tissue ‘maintenance’ results internal organ size. Liver and other internal organs from the continual process of synthesis, degradation can produce much more heat MJ kg than that of and replacement of those parts of body tissues which muscle Baldwin et al., 1985; Johnson et al., 1990. ‘turnover’. This is particularly the case with body protein for which the process of ‘turnover’ is sub- 2.3. Factors affecting the energy requirement for stantial, although variable Reeds, 1989. It has been maintenance suggested that fat tissue does not ‘turnover’ at all in animals fed regularly, although there does appear to 2.3.1. Body condition fat vs. lean be an extent of fatty acid turnover which is obligat- In all currently used energy systems the energy ory and which might be presumed to represent a requirement for maintenance is related to the degree of turnover of body fat Oldham and Em- liveweight of animals. However, there has been mans, 1990. The energy cost of maintaining body increasing evidence to suggest that maintenance protein would, however, be expected on stoichio- metabolic rate depends on body lean mass, rather metric grounds to exceed that of fat even if their than whole liveweight. For example, Noblet et al. rates of turnover were similar. Against this back- 1998 reported a significantly lower FHP per unit of ground it is biologically unreasonable to expect liveweight in fat than lean pigs of different breeds. maintenance to be directly related to scaled While this finding could partially be due to breed liveweight when the composition of the body may differences, a similar result has been obtained be- vary. tween fat and lean rats within a breed Pullar and It may not be realistic to exactly measure protein Webster, 1974. On the other hand, with the same mass in a live animal. However, some modern 202 R .E. Agnew, T. Yan Livestock Production Science 66 2000 197 –215 techniques can be used to indirectly predict body fat expensive process four phosphate bonds required and protein masses. For example, the ultrasonic per molecule urea synthesis Martin and Blaxter, scanning technique can be used to measure the back 1965. fat thickness of an animal Ferris et al., 1999a. The Dietary fibre concentration can also influence k , m prediction of body protein mass can also be derived because k is predicted from energy metabolisability m from slaughter techniques. Fat-free mass of dairy in the UK ME system. Increasing fibre concen- cows was reported to be similar during the dry trations in diets can reduce energy metabolisability period, early and late lactation stages Andrew et al., by reducing energy digestibility Flatt et al., 1969b; 1994. Protein mass of cattle is curvilinearly related Beever et al., 1988 and increasing methane energy to their liveweight and liveweight is the best single output as a proportion of total DE intake Yan et al., predictor of body protein mass Wright and Russel, 1999. The effects of dietary fibre fraction on 1984. Gibb and Ivings 1993 reported a linear maintenance metabolic rate and k can thus result in m relationship between body protein mass P and a higher ME for a high compared to a low fibre m liveweight LW for Holstein–Friesian cows P 5 diet. This has been demonstrated in a number of 0.0997 LW122.37, but cow body condition is not studies with lactating dairy cows Flatt et al., 1969b; considered in this equation. Tyrrell and Moe, 1972; Yan et al., 1997a and with beef cattle Beever et al., 1988; Reynolds et al., 2.3.2. Dietary concentration of fibre fraction 1991a. All currently used energy systems for cattle as- sume that dietary fibre concentration has no effect on 2.3.3. Grazing activity NE , although it may influence k in the UK ME Grazing of dairy cows on pasture is a common m m system since k is set to be positively related to practice in many parts of the world. Grazing cattle m energy metabolisability AFRC, 1990. Recent expend more energy in consuming the same amount studies have however suggested that increasing of feed when compared with prehension from a dietary fibre concentration could increase mainte- trough for housed cattle. The energy expenditures in nance metabolic rate. Cattle offered a high fibre diet eating 1 kg DM of pelleted feed, hay dried grass or have been shown to consume more feed to have a in simulated grazing were reported to be 0.23, 1.03 similar amount of ME intake to those given a low or 3.42 kJ kg liveweight in cattle Adam et al., fibre diet Reynolds et al., 1991a; McLeod and 1984. Heat production was nearly double in grow- Baldwin, 1998. The higher DM intake with high ing cattle for eating a total of 1 kg DM of forage in fibre diets to achieve similar levels of ME intake standing pasture rather than as cut pasture fresh and obviously results in greater gut fill, greater work of dried Holmes et al., 1978. The higher energy rumination and digestion and a greater production of expenditure during grazing can result from greater acetic acid in the rumen. All of these factors can physical effort for eating feed from the sward and contribute to an increase in gut mass Reynolds, more time in selecting and getting into its mouth the 1996 and consequently a higher maintenance meta- feed that it eats. For example, on a pasture providing bolic rate. Feeding of high fibre diets has been also abundant good quality herbage a sheep often spends shown to increase the metabolic activity in organs of 6 h day grazing, or about 5 h more than a similar animals. Reynolds et al. 1991a reported higher housed sheep would spend eating the same amount blood flow-rates at portal-drained viscera PDV, of DM when feed of similar quality was presented liver and kidneys, and greater oxygen use at PDV Standing Committee on Agriculture SCA, 1990. and whole body in heifers offered a high compared Expenditure of time and energy on rumination would to a low fibre diet. In addition, cattle offered a high probably be similar, but the grazing sheep could fibre diet may suffer from a deficiency of ferment- walk 3 km day in feeding and for other reasons e.g., able organic matter energy in the rumen, resulting drinking water. The NE costs of these extra ac- in a higher urea concentration in blood and urine tivities at pasture are estimated to be 20.3 kJ kg Huntington, 1989; Reynolds et al., 1991b. Syn- liveweight which for a weaned sheep represents thesis of urea from ammonia can be an energetically about proportionately 0.20 increase in NE SCA, m R .E. Agnew, T. Yan Livestock Production Science 66 2000 197 –215 203 1990. The increase in NE cost for grazing would For individual calorimetric studies, k has often m l be greater 0.30–0.40 under extensive grazing con- been determined by assuming a ME value which is m ditions, since the animals would spend more time deducted from ME intake to provide the ME avail- eating and walking Langlands et al., 1963. In able for production ME , and then relating this to p addition, lactating dairy cows grazing on pasture are adjusted milk energy output E , k 5 E ME . l 0 l l 0 p usually required to walk a distance twice a day for The k estimated from this approach is thus in- l milking. The energy costs for walking are suggested fluenced by the accuracy of ME . Using this meth- m to be 2.0 or 28 kJ kg liveweight per km for od, and ME of AFRC 1990, k values have often m l horizontal and vertical movement, respectively for been estimated to be low 0.50–0.58 with data of cattle ARC, 1980. Brody 1945 suggested a dairy cows given diets based on either grass silage proportionately 0.08 increase in maintenance energy Unsworth et al., 1994; Gordon et al., 1995; Yan et requirement for each km of walking. al., 1996; Ferris et al., 1999b, or maize silage or NRC 1988 recommend a proportionately 0.10 whole crop wheat Beever et al., 1998; Sutton et al., increase in maintenance allowance for cows grazing 1998a,b, 1999. However, the above k values could l on good pasture and up to 0.20 on sparse pasture. be increased 0.59–0.65, mean of 0.62 by using a 0.75 SCA 1990 suggest that increases in maintenance ME of 0.62 MJ kg , as obtained from Eqs. 1 m requirement for grazing cows are in a range of and 2. The latter mean k 0.62 is similar to that l 0.10–0.20 in intensive grazing conditions, to approx- 0.64 predicted from AFRC 1990 k 50.35?ME l imately 0.50 for animals grazing extensive, hilly GE10.42 using ME GE obtained in these studies. pastures where they walk considerable distances to A further method has been adopted by some preferred grazing areas and to water. However, these workers mainly in Germany, in which k is derived l allowances are not included in the current UK ME as a proportion of NE of total ME intake, where NE system AFRC, 1993 and there is a requirement to is the sum of milk energy, retained energy multiplied address this deficiency in the future. by a constant and assumed NE . However, the m efficiency calculated using this method is a mixture of efficiencies of ME utilisation for lactation k and l

3. Efficiency of ME utilisation maintenance k .