Fiftieth Anniversary of the California Net Energy System Symposium: What are the energy coefficients for cows?

Abstract

The same model structure used to describe energy metabolism in the growing animal is often used to model energy metabolism in the cow. Energy requirements of the cow are modeled as the summation of energy required for maintenance and recovered energy, where recovered energy is the summation of energy for the conceptus, milk, and tissue energy. Energetic requirements of the cow fluctuate throughout the production cycle depending on whether they are pregnant, lactating, or both. The current model requires energy cost to be associated with either net energy of maintenance or the partial efficiencies of conceptus growth, milk production, and tissue energy change. Mathematically, they are not independent. Incorrectly estimating one will result in an erroneous estimate in the other. Most of the current models in production agriculture allocate energy use into maintenance, and synthesis of tissues making it difficult to assign energy utilization by tissues that provide support functions to pregnancy, lactation, and weight fluctuation. The consequence is the assignment of partial efficiencies that reflect whole animal efficiencies rather than tissue efficiencies. Historically, these models have been predictive of energy metabolism, but caution should be used when inferring the energetic efficiency at the tissue level. Alternative modeling approaches more thoroughly describe tissue energy metabolism and have been used to estimate whole animal metabolism. These models resolve the problems associated with developing coefficients that lack biological meaning but are more complex. There is a critical need for independent data sets to test new components of the model for cows.

INTRODUCTION

The same model structure used to describe energy metabolism in the growing animal is used to model energy metabolism in the cow (BCNRM, 2016). Like the growing animal, energy requirements of the cow are modeled as the summation of energy required for maintenance and recovered energy. The cow differs from the growing animal in that recovered energy is not only tissue energy but includes conceptus, milk, and tissue energy. In BCNRM (2016), the efficiency of lactation and pregnancy are assumed to be the same as maintenance, and their net energy requirements are expressed as net energy of maintenance (NE_m) requirements.

Maintenance

Energetic requirements of the cow fluctuate throughout the year depending on whether they are pregnant, lactating, or both. Maintenance is defined as when intake metabolizable energy (ME_i) equals heat production (zero energy balance). The majority of the ME required by a cow annually is used to maintain basic biological functions (maintenance). During the course of a year, approximately 15% is required for milk synthesis, 8% for conceptus growth, 13% for activity, and 64% for maintenance. The current BCNRM (2016) defines NE_m as 77 kcal/(kg shrunk body weight, SBW)^0.75. There has been some debate as whether this value needs to be adjusted for cows with higher milk potential. Buskirk et al. (1992) estimated recovered energy and calculated heat production in beef cows and estimated a fasting heat production (FHP) of 73 kcal/(kg body weight, BW)^0.75. Using in-direct calorimetry, Flatt et al. (1965; Table 1) reported a FHP in nonpregnant, nonlactating Holsteins that had previously been on ad libitum diets to be 76 kcal/kg BW^0.75. These findings suggest that 77 kcal/kg SBW^0.75 is a reasonable base estimate for NE_m for cows in a “normal” basal metabolic state regardless of lactation potential.

Table 1.

Fasting heat production of dry, nonpregnant Holstein cows¹

Feed level	kcal·kg−0.75·d−1
Ad libitum	76.2 ± 0.8
Maintenance	73.5 ± 1.5
1/2 maintenance	71.6 ± 1.2

¹ Flatt et al. (1965).

The problem is deciding what is the “true” maintenance of a mature cow. Cows typically fluctuate in weight throughout the year. They often lose tissue energy when dietary nutrients are unavailable and gain tissue energy when dietary nutrients are plentiful. When we change dietary levels, cows adapt to a “new” BW that matches nutrient availability. This makes it possible for a cow to be at zero energy balance (maintenance) at different BWs and body condition scores (BCS). Adaptation to new maintenance can take as long as 112 d (Freetly and Nienaber, 1998; Figure 1). It becomes difficult to define the maintenance requirement of a mature cow because it will differ if we are referring to her at a BCS 4 versus a BCS 5, for example. In practice, we may need to define what BW she would be at the desired BCS and calculate her maintenance energy requirement at our preferred conditions rather than the maintenance energy requirement at her current level.

Figure 1.

Energy balance of feed-restricted cows (Freetly and Nienaber, 1998).

When a cow expresses different maintenance requirements at different BW and BCS, we are making one of two assumptions in the current model. If we assume that NE_m stays fixed, then the partial efficiency of maintenance must change (Figure 2). However, if we assume the partial efficiency of maintenance is fixed, then NE_m must change (Figure 3). Ferrell et al. (1986) demonstrated that FHP in growing lambs was positively related to their immediate previous plane of nutrition. Flatt et al. (1965; Table 1) found a similar pattern in Holstein cows as did Birkelo et al. (1991). Burrin et al. (1989) reported that portal-drained viscera accounted for 19% to 22% and the liver 22% to 41% of the total animal oxygen consumption in lambs. Eisemann et al. (1996) report these tissues to use approximately 12% to 25% of the total oxygen consumption each. During feed restriction, oxygen consumption decreases in these tissues (Burrin et al., 1989; Freetly et al., 1995; Ortigues and Durand, 1995). The whole animal and organ data suggest that we may need to adjust NE_m for feed-restricted cattle. The question becomes if the adjustment to NE_m is additive or becomes a proportion of NE_m?

Figure 2.

If FHP is assumed to be fixed, then k_m differs.

Figure 3.

If k_m is assumed to be fixed, then FHP differs.

$${\text{NE}}_{\text{m}}=\text{ a }\times \text{ 77 Mcal}/{\text{kg SBW}}^{0.\text{75}}{\text{or NE}}_{\text{m}}=\text{ a }+\text{ 77 Mcal}/{\text{kg SBW}}^{0.\text{75}}$$

Williams and Jenkins (2003) proposed an alternative model where

$${\text{ME}}_{\text{i}}={\text{ ME}}_{\text{m}}+{\text{ H}}_{\text{i}}{\text{E}}_{\mathrm{v}}+{\text{ H}}_{\text{i}}{\text{E}}_{\mathrm{r}}+\text{ RE}$$

In their model, “H_iE_r represents costs that are directly involved in the synthesis of recovered energy, and H_iE_v represents costs that are associated with supporting energy-expending processes that are not directly part of the pathways from precursors absorbed to products synthesized.” In their model, H_iE_v is considered to be additive and is grouped with ME_m as a cost of “maintenance” and is a function of feed intake. Metabolizable energy for maintenance is ME_m, and recovered energy is RE.

Previous studies in cows suggest that NE_m for cattle in a “normal” physiological state can be predicted with 77 kcal/kg SBW^0.75, but altering nutrient level may require an adjustment. The model of Williams and Jenkins (2003) suggests that it is an additive adjustment and the coefficient can change in sign.

Pregnancy

It has long been recognized that there is an increase in heat production associated with pregnancy that exceeds the expected value of nonpregnant females that are of the same size (Brody et al., 1948; Moe and Tyrrell, 1972; Ferrell et al., 1976; Figure 4). This increase is often referred to as the heat increment of pregnancy. In the current model (BCNRM, 2016), the efficiency of conceptus energy retention is 13%. This value suggests the efficiency of growth in the conceptus is low relative to other tissues (i.e., muscle). An example of this estimate can be obtained in the study of Ferrell et al. (1976) when we calculate the first derivative of the change in energy accretion equation in the fetus, and divide by the heat increment of pregnancy plus energy accretion (Table 2) efficiency of energy accretion ranges between 7.5% and 16.2%. The efficiency of energy accretion in the fetus is most likely not this low. Reynolds et al. (1986) measured oxygen consumption by the fetus using net flux measures in the arterial-venous blood, and when combined with the equation for rate of energy accretion from Ferrell et al. (1976), efficiency of energy accretion by the fetus ranges between 36% and 40% (Table 3). These values are consistent with the efficiency we expect from other tissues. We can use the same approach to calculate efficiency of the gravid uterus (Table 4) and find the estimate of efficiency is lower than that of the fetus. The lower efficiency is most likely the higher metabolic rate of the placenta relative to tissue energy accretion. The placenta is a highly specialized tissue providing nutrients to the fetus and acting as a barrier against harmful compounds. These data suggest that tissues other than those associated with the gravid uterus are contributing to the heat increment of pregnancy.

Figure 4.

Heat increment of pregnancy in beef cows (Ferrell et al., 1976).

Table 2.

Energetic efficiency of fetal growth as a function of heat increment (kcal/d)¹

Days pregnant	Energy accreted	ME for conceptus (heat increment pregnancy)	Efficiency, %
134	15	113	13.3
189	59	787	7.5
237	137	845	16.2
264	207	2,047	10.1

¹ Ferrell et al. (1976).

Table 3.

Energetic efficiency of fetal growth as a function of tissue heat production (kcal/d)

Days pregnant	Energy accreted1	Heat production2	Efficiency, %
137	40	65	38
180	162	246	40
226	458	679	40
250	644	1,133	36

¹ Ferrell et al. (1976).

² Reynolds et al. (1986).

Table 4.

Energetic efficiency of gravid uterus growth as a function of tissue heat production (kcal/d)

Days pregnant	Energy accreted1	Heat production2	Efficiency, %
137	86	327	21
180	215	507	30
226	503	1,164	30
250	745	2,048	27

¹ Ferrell et al. (1976).

² Reynolds et al. (1986).

Freetly and Ferrell (1997) demonstrated that in the ewe the liver could account for 20% of the increase in heat production associated with pregnancy. Rosenfeld (1977) reported a 75% increase in cardiac output associated with pregnancy in the ewe. Overall, about 40% of the heat increment of pregnancy can be contributed to the gravid uterus, 20% to the maternal liver, and the remainder to other maternal tissues (Figure 5). Cumulatively, these studies suggest that in the current model (BCNRM, 2016), efficiency is underestimated, and maintenance and NE_m are underestimated. The level of increase in NE_m is mostly likely dependent on stage of pregnancy and would follow the growth of the gravid uterus. In production systems where lactation ends before cows reach the final 40% of gestation, adjustments are most likely additive; however, they may not be additive if lactation overlaps with late gestation.

Figure 5.

Tissue contribution to the heat increment of pregnancy.

Lactation

Like pregnancy, there is an increase in heat production over that of nonlactating animals that was originally defined as the heat increment of lactation (Brody et al., 1948). There are few studies in beef cattle that have measured the efficiency of milk production (Patle and Mudgal, 1977; Reynolds and Tyrrell, 2000; Freetly et al., 2006). The absence of studies is, impart, driven by the difficulty of measuring milk production and milk composition in a “normal environment.” The partial efficiency of lactation has traditionally been calculated as the slope of recovered energy on ME_i. The efficiencies of recovered energy in Reynolds and Tyrrell’s (2000) study (0.641) and Freetly et al.’s (2006) study (0.693) in beef cattle are similar to that calculated by Moe and Tyrrell (1972) for Holsteins (0.608), suggesting the capacity to produce milk does not affect the efficiency of production. The efficiency of energy retention in milk reported in these studies is lower than the theoretical calculations reported by Baldwin (1968). Baldwin (1968), using metabolic stoichiometric arguments, determined the efficiency of net energy of lactation (NE_l) was 0.76. Using an energy balance study, Freetly et al. (2006) determined the efficiency of NE_l synthesis was 0.73 after accounting for maintenance and tissue energy recovery. This experimental study is within good agreement with the theoretical value. Combined, these studies suggest, like pregnancy, the efficiency of milk synthesis is underestimated as a result of energetic cost associated with support tissues contributing to the heat increment of lactation.

Predicting FHP by regressing recovered energy on ME_i has had mixed results in beef cattle. The intercept from Reynolds and Tyrrell’s (2000) study suggests no increase in NE_m where that of Freetly et al. (2006) suggests an increase. Increased maintenance requirement between lactating and nonlactating Holsteins has been observed (Moe et al., 1970; Moe, 1981). In a weight stasis study, Neville and McCullough (1969) calculated the feed required to maintain BW after adjusting for milk yield increased in lactating Herefords compared with nonlactating cows. Cumulatively, these studies suggest an increase in metabolic rates of tissues that support milk synthesis.

The liver is one of the tissues that supports lactation (Freetly et al., 1993). The increase in energy consumption by support tissues is probably a function of the level of milk production rather than simply lactation. Benson et al. (2002) reported that hepatic oxygen consumption was greater in early lactation compared with middle lactation. Freetly and Ferrell (1997) demonstrated that liver oxygen consumption increases with increased milk production in the ewe.

The BCNRM (2016) sets NE_m for cows to 77 kcal/BW kg^0.75 SBW, but the user has the option to change the value. The BCNRM (2016) attempts to account for the increased energetic cost of milk production by increasing NE_m by 20% in high lactating breeds, but this global increase fails to capture the relationship between energy expenditure of support tissues and the level of milk energy production. It may be more appropriate to make the adjustment to NE_m as a function of NE_l rather than simply applying a breed adjustment.

The study of Ferrell and Jenkins (1984) in nonpregnant, nonlactating cows and the study of Montaño-Bermudez et al. (1990) concluded that feed required for weight maintenance was greater in breeds with moderate to high lactation potential, suggesting that selecting for milk production results in an increase in maintenance even in the nonlactating cow. The hypothesis is that cattle selected for milk production will also be selected for larger support tissues that increase the maintenance. However, Flatt et al. (1965) determined that the NE_m in nonpregnant, nonlactating Holsteins was 76.2 ± 0.8 kcal/kg BW^0.75, suggesting that even in high milk potential cows, the NE_m returns to a value that is similar to that adapted by BCNRM (2016) of 77 kcal/kg BW^0.75.

Unlike dairy production systems, the nature of beef production systems makes it difficult to develop quantitative model inputs for milk yield and composition. The greatest limitation in predicting energy requirements for lactating beef cows is establishing milk yield. Milk yield in the beef cow is a function of the phenotypic capacity of the cow to produce milk and the ability of the calf to consume milk. In early lactation, the capacity of the cow to produce milk is greater than the ability of the calf to consume milk and that relationship reverses in late lactation. The current model (Jenkins and Ferrell, 1984) used by the BCNRM (2016) assumes the same lactation curve for all cows that vary based on peak yield. Although this model most likely captures most of the behavior, it does not allow for differences in persistency in late lactation that is observed in some breeds (Freetly and Cundiff, 1998). The other component in calculating NE_l is estimating the energy content based on the milk constituents. There has been a general trend for milk production to increase in beef cattle (Kuehn and Thallman, 2017). Milk fat is often inversely related to milk yield. Earlier reports of milk composition may not be applicable to current beef cows.

Tissue Energy

A new component in BCNRM (2016) accounts for changes in energetic density of tissue loss or gain dependent on the BCS score on the cow. The new approach increases energy density of the tissue as the BCS increases. Freetly et al. (2006) estimated the partial efficiency of tissue energy gain in nonpregnant, nonlactating cows to be 0.50. Patle and Mudgal (1977) reported a partial efficiency of 0.65 in lactating beef cows, and Freetly et al. (2006) reported a partial efficiency of 0.68 in lactating cows. The BCNRM (2016) assumes that for every Mcal above NE_m that a Mcal of tissue energy is gained. This assumption may be too high based on the previously reported partial efficiencies. It is also assumed that 1 Mcal of tissue energy is converted to 0.8 Mcal of NE_m when cows are losing tissue energy. This conversion ratio is similar to the values reported for the conversion of tissue energy to milk energy (Moe et al., 1970; Freetly et al., 2006).

As previously discussed in the Maintenance section, there appears to be an adaptation in NE_m when cows lose or gain energy. Adaptation to feed restriction can take up to 112 d (Freetly and Nienaber, 1998). Adaptation of visceral tissue may be one of the underlying mechanisms; however, during feed restriction, these adaptations are verily acute (Burrin et al., 1989; Freetly et al., 1995; Ortigues and Durand, 1995) and do not explain the long-term adaptation. The rate of energy accretion increases rapidly regardless of the refeeding level and decreases over time. Williams and Jenkins (2003) modeled a lag that adjusted the NE_m component of their model to describe this behavior. Like pregnancy and lactation, the partial efficiency and NE_m are interrelated and assumptions made in one will have consequences of interpretation of the other.

Is It the Coefficients or the Model?

At the root of the question “What are the energy coefficients for cows?” is the assumption that we have the correct model. An attribute of the model is its simplicity, but this simplicity has resulted in the development of coefficients that describe the whole animal and do not describe tissue-level energy metabolism. In order for the model to predict energy metabolism, coefficients need to be modified in ways that are not often intuitive. The level of aggregation of the biology in the coefficients limits their usefulness beyond prediction of energy metabolism. Genetic selection on polygenic traits slows progress and, not having defined physiological mechanisms, prevents informed research on development of management strategies and pharmaceuticals to improve efficiency. Alternative models have been proposed to address the limitations in the coefficients. Two primary approaches have been used. The first approach has been to use the structure of the California Net Energy system and subdivide the coefficients into more pools (Williams and Jenkins, 2003), and the second approach has been to develop mechanistic models that predict whole animal energy expenditure based on underlying biology. This second approach was described by Baldwin (2005). The most progress in the development of mechanistic models has been with the dairy cow. Because the underlying biology is the same between beef and dairy cows, it is reasonable to assume that these models could be adapted for use in beef cows. Mechanistic models have been useful tools integrating an array of research data and have had success in predicting whole animal nutrient utilization (Hanigan et al., 2006). While they are powerful research tools, their adoption in applied agriculture has been slow due to their complexity and typical need for many inputs. Understanding how the model will be used will dictate the preferred approach. The simplest model that will predict the desired outputs is preferred.

CONCLUSION

The current model requires energy cost to be associated with either NE_m or the partial efficiency, and they are not independent. Incorrectly estimating one will result in an erroneous estimate in the other. Many of the current models allocate energy use into maintenance, and synthesis of tissues making it difficult to assign energy utilization by tissues that provide support functions to pregnancy, lactation, and weight fluctuation. The consequence is the assignment of partial efficiencies that reflect whole animal efficiencies rather than tissue efficiencies. Historically, these models have been predictive of energy metabolism, but caution should be used when inferring the energetic efficiency at the tissue level. Other modeling approaches such as mechanistic models that integrate tissue metabolism to predict whole animal energy utilization do a better job of describing the energetic utilization by tissues because they are not bound by the limitation of assigning energy utilization to either maintenance or efficiency. There is a critical need of independent data sets to test new components of the BCNRM’s (2016) model for cows.