Functions

Extracellular calcium (Ca) is essential for formation of skeletal tissues, transmission of nervous tissue impulses, excitation of skeletal and cardiac muscle contraction, blood clotting, and as a component of milk. Intracellular Ca, while 1/10,000th of the concentration of extracellular Ca, is involved in the activity of a wide array of enzymes and serves as an important second messenger conveying information from the surface of the cell to the interior of the cell. About 98 percent of the Ca in the body is located within the skeleton, where Ca, along with phosphate anion, serves to provide structural strength and hardness to bone. The other 2 percent is found primarily in the extracellular fluids. Normally, the concentration of Ca in plasma is 2.2 to 2.5 mM (9 to 10 mg/dL) in the adult cow, with slightly higher values in calves. Between 40 and 45 percent of total Ca in plasma is bound to plasma proteins, primarily albumin, and another 5 percent is bound to organic components of the blood such as citrate or inorganic elements. From 45 percent (at higher blood pH) to 50 percent (at lower pH) of total Ca in plasma exists in the ionized, soluble form. The ionized Ca concentration of plasma must be maintained between 1 and 1.25 mM to ensure normal nerve membrane and muscle end-plate electric potential and conductivity, which has forced vertebrates to evolve an elaborate system to maintain Ca homeostasis. This system attempts to maintain a constant concentration of extracellular Ca by increasing Ca entry into the extracellular fluids whenever there is a loss of Ca from the extracellular compartment. When the loss of Ca exceeds entry, hypocalcemia can occur, resulting in loss of nerve and muscle function, which can lead to recumbency and the clinical condition referred to as milk fever.

Calcium Homeostasis

Ca leaves extracellular fluids during bone formation, in digestive secretions, sweat, and, in specific situations, urine. In lactating cows, secretion of Ca in milk is by far the greatest loss, accounting for 50 to 75 percent of Ca losses. Ca lost via these routes can be replaced from dietary Ca, from resorption of Ca stored in bone, or by resorbing the Ca filtered across the renal glomerulus (i.e., reducing urinary Ca loss). Under most circumstances urinary Ca losses are no more than 1 to 2 percent of absorbed Ca intake (Knowlton and Herbein, 2002; Taylor et al., 2009). When the loss of Ca from extracellular fluids exceeds the amount of Ca entering the extracellular fluids, plasma concentrations decrease. The parathyroid glands monitor the concentration of Ca in carotid arterial blood and secrete parathyroid hormone (PTH) when they sense a decrease in blood Ca. Release of PTH immediately increases renal reabsorption mechanisms for Ca and will stimulate processes to enhance intestinal absorption of Ca and resorption of Ca from bone.

Absorption of dietary Ca can occur by passive (paracellular) transport between epithelial cells across any portion of the digestive tract whenever ionized Ca in the digesta directly over the mucosa exceeds 6 mM (Bronner, 1987). These concentrations are reached when calves are fed all-milk diets and when cows are given oral Ca drenches for prevention or treatment of hypocalcemia (Goff and Horst, 1993). Following drenching, elevated Ca concentrations are short-lived as a result of dilution with digesta and the formation of complexes and chelates that reduce ionized Ca concentrations.

In nonruminants, as much as 50 percent of dietary Ca absorption may be passive (Nellans, 1988), but the amount of passive absorption of Ca that occurs in dairy cattle is unknown. The diluting effect of the rumen would likely reduce the degree to which passive Ca absorption occurs. When chelates and less-soluble salts such as calcium carbonate (CaCO₃) and calcium sulfate (CaSO₄) move out of the rumen and interact with hydrochloric acid secreted in the abomasum, ionized Ca increases (Goff, 2018). Ca absorption is tightly regulated and is one of the primary means by which Ca homeostasis is maintained. This suggests that active transport of Ca is the major route for Ca absorption in mature ruminants, and this process is controlled by 1,25-dihydroxyvitamin D, the hormone derived from vitamin D. By carefully regulating the amount of 1,25-dihydroxyvitamin D produced, the amount of dietary Ca absorbed can be adjusted to maintain a constant concentration of extracellular Ca (DeLuca, 1979; Wasserman, 1981; Bronner, 1987). Regulation of Ca absorption in the intestine occurs as 1,25-dihydroxyvitamin D in blood binds vitamin D receptors in the intestinal enterocyte that initiate transcription and translation of several key proteins needed for active transport of Ca (Goff, 2018). These include an apical membrane channel protein that facilitates movement of ionized Ca through the cell membrane; production of calbindin, which binds Ca ions as they pass through the enterocyte membrane; and plasma membrane Ca ATPase that works with a sodium/calcium exchanger that pumps intracellular Ca through the enterocyte basolateral membrane into blood by exchanging 2 moles of Ca for 3 moles of sodium (Na).

When dietary Ca is insufficient to meet the requirements of the animal, Ca will be withdrawn from bone to maintain a normal concentration of extracellular Ca. If dietary Ca is severely deficient for a prolonged period, the animal will develop osteoporosis to the point of developing fractures, but plasma Ca will only be slightly lower than normal. A sudden large increase in loss of Ca from the extracellular pool (e.g., initiation of lactation) can result in acute hypocalcemia before the Ca homeostatic mechanisms can act. This is discussed in the section on milk fever (see Chapter 12).

Requirement for Absorbed Calcium

A factorial system is used to estimate the amounts of Ca required for maintenance, growth, pregnancy, and lactation as in the previous NRC (2001) publication.

Body Tissue Mobilization and Replenishment

Mobilization of body tissue in support of lactation includes the mobilization of bone Ca to support the demand for milk Ca secretion. Each kilogram of body tissue mobilized includes 21 g ash, and Ca accounted for 53 percent of total bone ash in samples taken at 8 days and 11 weeks postpartum (Taylor et al., 2009). This suggests that 11 g of Ca would be provided for each kilogram of body tissue mobilized. Ca balances of −11 to −15 g/d were observed in cows fed 0.52 percent dietary Ca during the first 8 weeks postpartum (Taylor et al., 2009). Ca mobilized at the beginning of lactation needs to be replenished as cows regain mobilized tissue stores over the course of the lactation. Mobilization of skeletal Ca is almost inevitable during early lactation, and cows could lose 800 to 1,300 g of bone Ca in early lactation (Ellenberger et al., 1931). This would require up to 8 g of absorbed Ca/d during the last 20 weeks of lactation to replenish. However, because of the uncertainties involved, no provision for replenishment of skeletal Ca mobilized during early lactation was included in the model.

Calcium Absorption Coefficient

The amount of Ca that must be fed to meet the requirement for absorbed Ca is dependent on the availability of Ca from the diet. The amount of Ca absorbed will generally equal the requirement for Ca if the diet contains enough available Ca. The proportion of dietary Ca absorbed will decrease as dietary Ca increases above requirements. As vitamin D–mediated Ca absorption from the intestine is tightly regulated, the determination of the efficiency of absorption of Ca from a diet requires that animals be fed at or near their Ca requirement. This will ensure that intestinal Ca absorption mechanisms are fully activated. Few studies fulfill this criterion, suggesting that published absorption data may often underestimate the availability of Ca. Furthermore, measuring Ca availability of specific ingredients is extremely difficult, and data must be extrapolated from total diets.

Ca absorption is usually measured using digestion experiments in which apparent Ca absorption equals Ca intake minus fecal Ca, both expressed in g/d. Since fecal Ca includes both undigested feed and metabolic fecal Ca, true Ca absorption is estimated by adding metabolic fecal Ca to apparently absorbed Ca. The absorption coefficient (AC) for a diet or feed is calculated as true Ca absorbed (g/d) divided by Ca intake (g/d).

Previous committees that authored Nutrient Requirements of Dairy Cattle publications (NRC, 1971, 1978, 1989) used a single AC for all diets. The 1971 and 1978 publications assumed an AC of 0.45, whereas the 1989 publication used 0.38. The AC was reduced in the 1989 Nutrient Requirements of Dairy Cattle partly in response to reports that cows in early lactation were less able to utilize dietary Ca (Ramberg, 1974; van't Klooster, 1976), making use of a lower coefficient prudent. The 0.38 AC was based largely on a summary of 11 experiments with lactating dairy cows in which the average apparent absorption of dietary Ca was 0.38 (Hibbs and Conrad, 1983). In the majority of these 11 experiments, cows were fed diets well in excess of their Ca requirements, but in 3 of the experiments, the cows were in negative Ca balance, and the AC was still less than 0.40. In those experiments, alfalfa and brome hay supplied most of the Ca. The 2016 Nutrient Requirements of Beef Cattle uses a mean true Ca availability from the diet of 0.50, and the INRA system (INRA, 2018) uses values between 0.30 and 0.55. The 2001 Dairy NRC adopted a system where Ca availability was based on AC of individual feeds rather than using a single dietary average because availability of Ca from forages and individual mineral supplements varies widely (Hansard et al., 1957; Ward et al., 1972; Martz et al., 1990).

To accommodate a system based on AC of individual feeds, AC based on work of Hansard et al. (1957) and summaries of the relative availabilities of mineral supplements compared to either calcium chloride (CaCl₂) or calcium carbonate (CaCO₃) were adopted. A major factor limiting Ca absorption is the solubility of the Ca from the mineral source, and CaCl₂ represents a source of highly soluble Ca. In NRC (2001), CaCl₂ was assigned an AC of 0.95. That value was based on studies in which ⁴⁵CaCl was used as a tracer to measure Ca absorption by young calves (Hansard et al., 1954). Hansard et al. (1957) demonstrated that CaCl₂ is between 1.2 and 1.32 times more absorbable than CaCO₃. Therefore, the efficiency of absorption of Ca from CaCO₃ was set at 75 percent. Finally, a list of common supplemental sources of Ca and an estimate of the efficiency of absorption of Ca from each source were developed using data summarized by Soares (1995a) based on their efficiency of absorption relative to CaCO₃. In summary, ACs for all mineral supplements were based the assumption that CaCl₂ had an AC of 0.95. Unfortunately, there were very little data in the literature at the time with diets fed at near estimated Ca requirements to determine whether the adopted ACs agreed with actual measured values. However, since the 2001 publication, several appropriate experiments have been conducted.

To evaluate NRC (2001) ACs, 45 treatment means from experiments with high-producing cows in early lactation were assembled (Wohlt et al., 1986; Martz et al., 1999; Knowlton and Herbein, 2002; Moreira et al., 2009; Taylor et al., 2009). Apparent Ca absorption from those studies was converted to true Ca absorption by subtracting metabolic fecal Ca from fecal Ca output (metabolic fecal Ca, g/d = 0.9 × DMI, kg/d). Diet ingredient information from those experiments was entered into NRC (2001) software to generate a predicted AC for Ca for each diet, and those were compared with the actual measured values.

The mean 2001 NRC predicted and actual true ACs (based on measured apparent absorption and estimated metabolic fecal Ca) were 0.60 and 0.45, respectively. Interestingly, the mean (± SD) measured value (0.45 ± 0.047) for true Ca absorption was identical to the average value adopted by the 1971 and 1979 committees (NRC, 1971, 1978). The range in Ca concentration in the diets summarized was from 0.17 to 1.03 percent, and it is known that excess Ca intake will reduce Ca absorption. However, the within-study change in true Ca availability was from 0.8 to 5.9 g/100 g Ca intake, which is equivalent to a reduction in the ACs of 0.008 to 0.059 units/100 g Ca intake. This suggests that the effect of Ca intake on absorption was small. The predicted intercept at zero Ca intake was 50.2 (± 1.5) equivalent to a true AC of 0.50 and varied little among studies. These observations suggest that the previously adopted ACs for feedstuffs and mineral supplements were too high.

Part of the overprediction of Ca absorption likely stemmed from the use of 0.95 true AC for CaCl₂ based on values observed in young (<30 days old) calves prior to weaning (Hansard et al., 1954). In that same study, subsequent measurements of absorption from CaCl₂ were much lower in animals ranging in age from 6 to 190 months. The measured AC for Ca from CaCl₂ was 0.60 and 0.53 in young and mature steers, respectively (Hansard et al., 1957). The relative availability of CaCO₃ was based on the 0.95 AC of CaCl₂ (which was too high), and the ACs for other mineral supplements were based on their availability relative to CaCO₃, resulting in an overestimation of the Ca availability for all mineral supplements.

To correct the observed overprediction of Ca availability, the ACs for each of the feed classes and mineral supplements were reviewed and adjusted where deemed appropriate. A comparison of adjusted AC to the other coefficients and literature values (Hansard et al., 1957; NRC, 2001; Kiarie and Nyachoti, 2010) is shown in Table 7-1. The availability of CaCl₂ was reduced from 0.95 to 0.60, which is more in line with measured values in functioning ruminants (Hansard et al., 1957). For most supplements, the ACs were reduced by about 25 percent. In the 2001 NRC, the AC for corn silage was set at 0.60, but the mean measured value across three treatments (Martz et al., 1990, 1999) was 0.425. Therefore, a more conservative value of 0.40 was assigned. Legume silages and hays because of their high Ca concentration can be significant contributors of Ca, but for reasons discussed above, the AC was held at 0.30. For all other forages, the AC was raised to 0.40, a value that is more consistent with values observed for grasses and hays reported by Hansard et al. (1957). The coefficient for cereal grains and protein supplements was maintained at 0.60. Although there is little experimental basis for assigning this value, these feedstuffs are generally low in Ca and are only minor contributors to overall Ca intake. Most nonforage feedstuffs will contain only small amounts of Ca. A notable exception is for Ca soaps of palm oil fatty acids (FAs), which can be 7 to 9 percent Ca. Although the FAs in this product are approximately 76 percent digestible and digestion can only occur following dissociation of the Ca from the palmitate in the small intestine, it is not likely that Ca absorption would exceed that of CaCl₂, so an availability of 0.60 was adopted. The accuracy of the new ACs was examined by entering the values in the 2001 NRC software for each of the feed ingredients in the diets from experiments with high-producing cows in early lactation (Wohlt et al., 1986; Martz et al., 1999; Knowlton and Herbein, 2002; Moreira et al., 2009; Taylor et al., 2009) and then comparing true Ca absorption (measured apparent plus estimated metabolic fecal Ca) with predicted absorbed Ca (see Figure 7-1). Both sets of coefficients were correlated with measured Ca absorption. As discussed, use of the 2001 NRC ACs overpredicted Ca absorption at high intakes of absorbed Ca (slope = 0.67). After adjustment of the ACs (see Table 7-1), measured and predicted Ca absorption were in good agreement where the slope (0.97) and intercept of the predicted versus actual regression equation did not differ from 1 and 0, respectively. The Ca-to-phosphorus (P) ratio was once thought to affect absorption of Ca and P, but data suggest that the ratio is not critical, unless the ratio is >7:1 or <1:1 (ARC, 1980; Miller, 1983) and the model does not adjust AC for that ratio.

FIGURE 7-1

Comparison of true Ca absorption (g/d) (measured apparent absorption adjusted for estimated metabolic fecal Ca) versus predicted 2001 NRC values (open circles) and the adjusted (solid squares) absorption coefficients for individual feeds as shown in Table (more...)

Although adjustments in the ACs improved the accuracy of prediction of Ca absorption, the committee recognizes that there is a dearth of measured availability data on individual feeds and mineral supplements. This is particularly important for comparing mineral supplements and for feeds that can provide substantial dietary Ca such as legume forages and canola meal. Major sources of variation for the AC of supplements have not been quantified. For example, particle size of limestone can affect rumen pH (Keyser et al., 1985), but it is unknown whether particle size (within typical ranges) affects the AC. The increased availability of stable isotopes of Ca such as ⁴²Ca and ⁴⁴Ca and higher sensitivity of mass spectrometer measurements should allow for improved estimates of Ca availability and metabolic fecal Ca losses.

Effects of Physiologic State

The amount of available Ca that will be absorbed varies with the physiologic state of the animal. Hansard et al. (1954) and Horst et al. (1978) reported that the efficiency of absorption of Ca decreases markedly as animals become older. As animals age, there is a decline in vitamin D receptors in the intestinal tract (Horst et al., 1990), which is thought to reduce the ability to respond to 1,25-dihydroxyvitamin D. The difference in efficiency of Ca absorption in beef steers from 1 to 6 years of age is nearly negligible (Hansard et al., 1954). Age was not included as a factor to adjust dietary Ca requirement in cattle >200 kg BW.

In early lactation, nearly all cows are in negative Ca balance (Ellenberger et al., 1931; Ender et al., 1971; Ramberg, 1974). As feed and Ca intake increase, most cows transition into positive Ca balance about 6 to 8 weeks into lactation (Ellenberger et al., 1931; Hibbs and Conrad, 1983). Cows in the first 10 days of lactation are at greatest risk of being in negative Ca balance (Ramberg, 1974), and many are subclinically hypocalcemic during this period (Reinhardt et al., 2011). Ramberg (1974) reported that the rate of entry of Ca into the extracellular fluid pool from the intestine increased about 1.55-fold from the day before parturition until 10 days in milk. Thereafter, the rate of entry of Ca into the extracellular pool from the intestine was constant. A study by van't Klooster (1976) demonstrated that Ca absorption increased from 22 percent in late gestation to 36 percent by day 8 of lactation, after which it remained relatively constant. This represented a 1.6-fold increase in efficiency of Ca absorption over this 8-day period. Regression analysis of data of Ward et al. (1972) predicted that cows need to be fed 5 g Ca/kg milk in early lactation to avoid negative Ca balance. However, there was no evidence to demonstrate that negative Ca balance in early lactation was detrimental to the cow provided the concentration of Ca in plasma remained normal (i.e., lactational osteoporosis ensures adequate entry of Ca from bone into the extracellular Ca pool).

Calcium Deficiency

A deficiency of dietary Ca in young animals leads to a failure to mineralize new bone and contributes to retarded growth. Rickets is more commonly caused by a deficiency of vitamin D or P, but a deficiency of Ca can contribute to rickets as well. In older animals, a deficiency of dietary Ca forces the animal to withdraw Ca from bone, which causes osteoporosis and osteomalacia, making bones prone to spontaneous fractures. The concentration of Ca in milk is not altered even during a severe dietary deficiency of Ca (Becker et al., 1933).

Excess Dietary Calcium

Feeding excessive dietary Ca is generally not associated with any specific toxicity. The maximum tolerable level (MTL) for Ca in ruminants was set at 1.5 percent of dietary dry matter (DM; NRC, 2005). Feeding excessive Ca could interfere with trace mineral absorption (especially zinc [Zn] and selenium [Se]) and dilutes energy and protein the animal might better utilize for increased production.

Physiologic Roles

P has more known biological functions than any other mineral. About 80 percent of the body's P is in bones and teeth principally as apatite salts and as calcium phosphate. It is in every cell of the body, and almost all energy transactions involve formation or breaking of high-energy phosphate bonds (such as those in adenosine triphosphate [ATP]). Phosphorylation is a primary regulator of numerous enzymes. P also is involved in acid-base balance of blood and other bodily fluids, as well as in cell differentiation, and is a component of cell walls and cell contents as phospholipids and nucleic acids.

Normal P concentration in blood plasma of dairy animals is 1.3 to 2.6 mmol/L (4 to 8 mg/dL), and whole blood concentrations are six to eight times greater (Goff, 2004). Plasma concentrations decrease with increasing age and are lower in early lactation than later lactation (Forar et al., 1982). For a 600-kg cow, approximately 1 to 2 g of inorganic phosphate is circulating in blood plasma, 5 to 8 g of inorganic P is in the extracellular pool, and total intracellular P is about 155 g (Goff, 1998). The intracellular concentration of P is about 10 times greater than the concentration in plasma (Goff, 1998).

Rumen microorganisms also require P (Burroughs et al., 1951; Breves and Schröder, 1991), and it is supplied to the rumen by the diet and recycling via saliva. Using various techniques, estimates of P recycling in lactating dairy cows fed adequate or excess P range from about 30 to 75 g/d (Kebreab et al., 2005; Puggaard et al., 2011). Inadequate supply of P to the rumen can reduce fiber digestibility. A diet with 0.24 percent P had lower neutral detergent fiber (NDF) digestibility than a similar diet with 0.34 percent P when fed to dairy cows (Puggaard et al., 2011). Digestibility of NDF was reduced even though the concentration of P in rumen fluid was 3 to 4 mmol/L, which is greater than the concentrations (0.6 to 2.5 mmol/L) that maximized cellulose digestibility in vitro (Hall et al., 1961; Chicco et al., 1965). However, no evidence is available showing improved ruminal digestion once the cow's P requirement is met.

Phosphorus Homeostasis

Blood plasma P concentrations are controlled via alterations in intestinal absorption, P recycling via saliva, renal excretion, and bone resorption. Absorption of P from the intestines is much less regulated than absorption of Ca. Net absorption of P occurs mainly in the small intestine (Grace et al., 1974; Reinhardt et al., 1988), with only small amounts absorbed from the rumen, omasum, and abomasum. Absorption is thought to occur mainly in the duodenum and jejunum (Care et al., 1980; Scott et al., 1984); however, little is known about absorption anterior to the small intestine (Breves and Schröder, 1991). Presumably, as in nonruminants, absorption occurs via two distinct mechanisms. A saturable vitamin D–dependent active transport system is operative when animals are fed low P diets. Synthesis of 1,25-dihydroxyvitamin D can be stimulated when blood P is low, resulting in more efficient absorption (Horst, 1986). Feeding 25-hydroxyvitamin D increases circulating 1,25-dihydroxyvitamin D and plasma P concentrations (Wilkens et al., 2012; Weiss et al., 2015b), which could indicate increased intestinal absorption of P. Passive absorption predominates when adequate or excessive amounts of potentially absorbable P are consumed, and absorption is proportional to the concentration gradient between the lumen of the small intestine and blood plasma (Wasserman and Taylor, 1976). However, data suggest that this process is saturable (Mogodiniyai Kasmaei and Holtenius, 2013), and current ruminant P models use Michaelis–Menten kinetics to describe intestinal absorption (Kebreab et al., 2004; Hill et al., 2008).

Renal clearance is usually a minor contributor to P homeostasis, but both urinary concentration and excretion of P increase as the supply of absorbable P to dairy cows increases (Knowlton and Herbein, 2002; Guyton et al., 2003; Knowlton et al., 2005; Puggaard et al., 2011; Mogodiniyai Kasmaei and Holtenius, 2013). The increase in urinary P excretion as P intake increases is usually less than 10 percent of the increase in P intake. P recycling via saliva is the major homeostatic mechanism for P (Breves and Schröder, 1991). P absorbed from the intestine in excess of requirement elevates blood P, which is then transferred to saliva and reenters the rumen. Salivary and plasma concentrations of P have a strong positive correlation (Valk et al., 2002), but the mass of P recycled via saliva is not necessarily correlated with plasma P concentrations when cows are fed marginal amounts of P (Puggaard et al., 2011). Recycled P can be used by ruminal microorganisms; a portion of it will be reabsorbed by the intestines, and a portion will pass out in feces. Fecal excretion of recycled P is one reason why apparent absorption of P does not reflect true absorption of dietary P.

Requirements for Absorbed Phosphorus

As described previously (NRC, 2001), the factorial approach was used to estimate the requirement for absorbed P by summing requirements for maintenance, growth, pregnancy, and lactation.

Dietary Requirement and Efficiency of Absorption

The dietary requirement is the total requirement for absorbed P divided by the AC for P from the diet. The use of feed (or feed class)–specific AC was introduced in NRC (2001), and that approach has been expanded. The AC for P in NRC (2001) was set at 0.64 for all forages except corn silage and 0.70 for all other feeds. The AC for P supplements ranged from 0.30 to 0.90, and the AC for total diets (weighted average from the dietary ingredients) was usually around 0.70.

To accurately determine the AC for a specific feedstuff or mineral source, P must be fed in an amount close to the animal's true requirement, and P recycling must be accurately quantified. Most studies do not satisfy these experimental specifications. Furthermore, even simple diets will contain multiple sources of P, and accurately partitioning the overall dietary AC into AC for ingredients is not possible. In the previous edition (NRC, 2001), the AC for P for all feedstuffs other than mineral supplements was based on data for alfalfa hay and corn silage. An alternative to assuming all feedstuffs have essentially the same AC is to analytically partition dietary P into fractions and estimate the AC for each fraction via modeling (Hill et al., 2008; Feng et al., 2015, 2016). This is the approach used for basal ingredients (described below).

The AC for P supplements from NRC (2001) was retained because newer data are not available. These values were tabulated from Soares (1995b), Peeler (1972), and other sources in the literature and are used in the model. Values determined using ruminants, especially cattle, were given preference whenever possible in tabulation. Dicalcium phosphate (calcium phosphate dibasic) with an AC of 0.75 in cattle (Tillman and Brethour, 1958; Challa and Braithwaite, 1988), phosphoric acid with an AC of 0.90 in cattle (Tillman and Brethour, 1958), and monosodium phosphate with an AC of 0.90 in sheep (Tillman and Brethour, 1958) were taken as reference standards. The ACs of P in other mineral sources were set based on these reference standards (Soares, 1995b).

Form of Dietary Phosphorus

For the current model, feed P is analytically partitioned into inorganic P (blue molybdovanadate method; AOAC, 2000) and organic P (total P − inorganic P). A model on P metabolism and absorption (Hill et al., 2008; Feng et al., 2015, 2016) also included a phytate P fraction, but its absorption coefficient was similar to that of the nonphytate organic P fraction (0.66 versus 0.7). Therefore, those two fractions were combined into organic P, which simplifies analytical requirements. Using the above P model, the AC is 0.84 for inorganic P and 0.68 for organic P fraction (i.e., average for phytate and nonphytate organic P).

The weighted average AC is then calculated based on the size of the two P fractions, which is the AC values in the feed library. For feeds that did not have P fraction data, AC values from similar feeds were used, or the AC was set at the default of 0.72. Additional analytical data are needed regarding P fractions of different feedstuffs. Feeds can be assayed for total P and inorganic P and those values entered in the feed library, but at the time of publication, most commercial labs did not conduct those assays. Factors other than form of P can affect AC; however, these effects have not been adequately quantified and cannot be modeled.

Phosphorus Intake

Although not as tightly regulated as Ca, true absorption of P decreases as P intake increases above requirements (Challa and Braithwaite, 1988; Challa et al., 1989; Martz et al., 1999); however, adequate data are not available to accurately quantify that effect. Because salivary P typically supplies at least 2-fold greater amounts of P to the lumen of the small intestine than does dietary P, the efficiency of absorption of salivary P is important. Salivary P is in the form of sodium and potassium phosphate salts. The AC of salivary endogenous P recycled to the small intestine was 0.68 to 0.81 in bull calves (Challa et al., 1989). Excessive dietary P relative to the requirement reduced the efficiency of absorption of inorganic or salivary P (Challa et al., 1989). The AC shown in Table 19-3 for mineral supplements and the AC values used for the various P fractions outlined above should be considered maximum absorption. If P is fed in excess of requirements, those ACs will overestimate actual absorption; however, because this occurs once the P requirement is met, it will not affect the amount of dietary P needed to meet requirements for absorbed P.

Use of Phytase

Phytate phosphorus (inositol polyphosphate) is the common storage form of P in many plants and usually comprises the largest proportion of organic P in concentrates (Nelson et al., 1968; Morse et al., 1992). Forages (or vegetative matter) usually have low concentrations of phytate. Normal ruminal metabolism breaks down most of the phytate; however, exogenous phytase can increase phytate breakdown in the rumen (Brask-Pedersen et al., 2013). Feeding supplemental phytase to dairy cows has not consistently reduced fecal excretion of P, and most studies reported no effect (Guyton et al., 2003; Kincaid et al., 2005; Knowlton et al., 2005; Knowlton et al., 2007).

Dietary Calcium

When cows are fed P at or above requirements, Ca intake ranging from deficient to excess usually has not affected efficiency of P absorption (Hibbs and Conrad, 1983; Moreira et al., 2009; Taylor et al., 2009; Herrera et al., 2010). Solubility of supplemental Ca (CaCl₂ versus limestone) did not affect P absorption by dairy cows (Herrera et al., 2010). However, Ca and P apparent digestibility are positively correlated (Hibbs and Conrad, 1983).

Animal Responses to Varying Dietary Phosphorus

Production responses by growing and lactating cattle to differing dietary P concentrations were reviewed in the previous edition; therefore, only more recent studies will be reviewed in detail in this version. In growing heifers, diets with 0.3 to 0.34 percent P generally resulted in maximum gain, adequate blood P concentrations, and adequate bone strength compared with animals fed diets with lower concentrations of P (NRC, 2001). Newer data support that conclusion. A study with Holstein and Holstein × Jersey crossbred heifers that started at 4 months of age and ended at 22 months of age found no differences in growth (weight and stature), reproductive measures, or bone strength between heifers fed 0.3 or 0.4 percent P (Esser et al., 2009; Bjelland et al., 2011).

The review conducted previously (NRC, 2001) concluded that for lactating cows, diets with 0.32 to 0.42 percent P for the entire lactation were sufficient depending on milk yield potential. Furthermore, they concluded that no benefits on lactational performance occurred when cows were fed diets with >0.42 percent P. Because of the ability to mobilize P from bone, longer-term performance studies evaluating effects of differing concentrations of dietary P on lactating cows are more meaningful than short-term studies. Newer studies lasting from 9 weeks to two lactations largely support the conclusions reached by the previous committee. Grazing dairy cows were fed diets with approximately 0.22 or 0.31 percent P starting at about 30 days in milk through about 90 days in milk (Reid et al., 2015), and no effects on milk yield, milk composition, or feed intake were observed. Dietary P concentration of 0.33 or 0.42 percent did not affect milk yield (35.1 versus 35.4 kg/d), DMI, or milk composition of mid-lactation Holstein cows fed diets for 14 weeks (Wu et al., 2003). However, Holstein cows fed diets with 0.32 percent P had reduced yields of fat-corrected milk (40.3 versus 44.3 kg/d) and DMI (25.0 versus 26.5 kg/d) compared with cows fed diets with 0.44 percent P for 10 weeks. The diets with 0.32 percent P did not meet the P requirement based on the current model. In a 23-week experiment (Lopez et al., 2004a,b, DMI, milk yield (35.1 versus 34.9 kg/d), milk composition, health disorders (except occurrence of eye inflammation, which was statistically greater in cows fed high P), and reproductive measures did not differ between Holstein cows fed diets with 0.37 or 0.57 percent P starting immediately after parturition. Similar results were obtained when Swedish Red and White cattle were fed diets with 0.32 or 0.42 percent P during the first 4 months of lactation (Ekelund et al., 2006). Based on bone markers, cattle in both groups exhibited bone P resorption, but resorption was similar between treatment groups.

Two multilactation studies evaluated effects of varying dietary P concentrations on long-term health and production of dairy cows. In one study, dairy cows (breed not reported, approximately 600 kg BW) were fed diets with 0.24, 0.28, or 0.33 percent starting in mid-lactation and continuing through a dry period and then for the entire next lactation and the subsequent dry period (Valk and Sebek, 1999). No treatment effects were observed in the first lactation period on milk yield (26.8, 25.9, and 27.5 kg/d, respectively), milk composition, or DMI. During the first dry period, cows fed the lowest P diet had reduced DMI. During the second lactation, cows fed the lowest P diet produced significantly less milk, consumed less DM, and were losing BW, and because of animal welfare concerns, that treatment was terminated after cows were on the treatment for approximately 12 months. No differences in milk yield (33.0 versus 34.1 kg/d), DMI, milk composition, or BW were observed between cows fed 0.28 or 0.33 percent P during the second lactation of the experiment. In another study, milk production (36.4 versus 35.4) and milk composition did not differ between Holstein cows fed diets with 0.35 or 0.42 percent P (Odongo et al., 2007) over two lactations. However, first-lactation, but not multiparous, cows fed the low P diet had lower DMI than first-lactation cows fed 0.42 percent P. BW and body condition were also lower for first-lactation cows fed the low P diet, indicating 0.35 percent dietary P was not adequate for first-lactation cows. Data from that experiment could not be evaluated with the current model because adequate parity data were not included in the study. But overall, data from longer-term production studies support the P requirements calculated using the current model.

Phosphorus and Reproduction

The previous edition (NRC, 2001) reviewed published research reports from 1923 through 1999 to assess the effects of dietary P on reproductive performance of cattle, and studies published after 1999 have been added to this review. In some studies, but not all, severe deficiency of dietary P caused infertility or reduced reproductive performance of cattle (Alderman, 1963; Morrow, 1969; McClure, 1994). Typically, P concentration was <0.20 percent of dietary DM, the deficient diet was fed for an extended length of time (1 to 4 years), and where measured, feed intake was depressed, causing coincidental deficiencies of energy, protein, and other nutrients. Low body condition generally is considered the main cause of reduced reproductive efficiency in P-deficient cows (Holmes, 1981). Little (1975) demonstrated that deficiencies of P and protein were additive on failure to exhibit first postpartum estrus in grazing multiparous beef cows.

In growing heifers, experimentally induced reproductive failure caused by a dietary P deficiency has been very difficult to produce. The studies reviewed by NRC (2001) reported no adverse effects on reproduction in heifers when they were fed diets with as little as 0.15 percent P for several months (some studies lasted more than 1 year). With lactating dairy cows, evidence from available research to support feeding P in excess of requirements to improve reproduction is virtually nonexistent. Results of 10 studies can be summarized very succinctly (Steevens et al., 1971; Carstairs et al., 1980; Call et al., 1987; Brodison et al., 1989; Brintrup et al., 1993; Valk and Sebek, 1999; Wu and Satter, 2000; Wu et al., 2000; Lopez et al., 2004a,c Odongo et al., 2007). All measures of reproductive performance compared within each study were not affected by the concentration of dietary P with one exception. In the study by Steevens et al. (1971), services per conception were greater in the second year for cows fed 0.40 versus 0.55 percent P, but not in the first year of study. Among these seven studies, dietary P ranged from 0.24 to 0.62 percent of dietary DM, length of feeding different dietary P concentrations ranged from the first 12 weeks of lactation to as long as three consecutive lactations, and average milk yields ranged from 15 to 37 kg/d. As long as dietary P was greater than 0.31 percent, reproductive performance was normal and not improved with increased concentrations of P. Cows in some of the studies would not be considered high-producing cows by modern standards. However, the more recent studies used cows producing more than 35 kg, and no effects of dietary P on reproduction were observed in those studies. The preponderance of data does not support feeding dietary P at concentrations in excess of those needed to meet dietary requirements to improve reproductive performance.

Phosphorus Deficiency

Detailed description of occurrence, etiology, clinical pathology, diagnosis, and treatment of P deficiency in ruminants has been described (Goff, 1998). Signs of deficiency may occur rather quickly if dietary P is insufficient. Deficiency is most common in cattle grazing forages on soils low in P or in animals consuming excessively mature forages or crop residues with low P content. Dairy cows do not seem to have the ability to self-select appropriate intakes of P or other minerals (Muller et al., 1977). Hypophosphatemia can also occur when a cow develops a displaced abomasum (Grünberg et al., 2005). Nonspecific chronic signs of deficiency include unthriftiness, inappetence, poor growth, and reduced milk yields, but signs are often complicated by coincidental deficiencies of other nutrients such as protein or energy. Animals may be chronically hypophosphatemic (<4 mg/dL in plasma), but the concentration of P in milk remains within the normal range. Hemoglobinuria (Jubb et al., 1990) and liver dysfunction (Grünberg et al., 2005) are associated with hypophosphatemia. In severe deficiency cases, bone mass is lost, and bones become weak. Severe clinical manifestations of P deficiency include acute hypophosphatemia, rickets in young growing animals, and osteomalacia in adults. Cows may also exhibit pica.

Acute hypophosphatemia (less than 2 mg P/dL of plasma) may occur when cows are fed marginally low dietary P and challenged by extra demand for P in late pregnancy with accelerated fetal growth, especially with twin fetuses and with colostrum and milk formation during early lactation. The disease usually is complicated with concurrent hypocalcemia, hypomagnesemia, and possibly hypoglycemia.

Concentrations of P in plasma often fall below the normal range in the periparturient period (Grünberg, 2008). In other mammals, physiologic correction can occur rather rapidly as P absorption is responsive to renal production of 1,25-dihydroxyvitamin D, which is stimulated by low P in the blood (Reinhardt et al., 1988; Goff, 1998). Feeding peripartum dairy cows 25-hydroxyvitamin D increased plasma 1,25-dihydroxyvitamin D and elevated plasma P (Wilkens et al., 2012; Weiss et al., 2015b). Secretion of cortisol around parturition may depress concentrations of P in plasma. Intravenous Ca to correct hypocalcemia usually results in a rise in P in plasma because parathyroid hormone secretion is lowered, reducing urinary and salivary loss of P. It also stimulates resumption of gut motility, recycling of salivary P, and absorption. Oral or intravenous administration of a soluble form of P such as sodium monophosphate can help correct hypophosphatemia. In some cows with severe cases of clinical milk fever, protracted hypophosphatemia (P in plasma <1 mg/dL) occurs with recumbency; even with successful treatment for hypocalcemia, P in blood remains low. This disorder is not well understood. However, increasing the amount or concentration of P in the diet in excess of requirement in late pregnancy or early lactation will probably not correct hypophosphatemia in the periparturient period, as this disorder seems to occur secondary to hypocalcemia.

When young calves are fed P-deficient diets, rickets occurs from a failure of mineralization in osteoid and cartilaginous (growth plate) matrices during bone remodeling. In contrast, in mature animals (no active growth plates), osteomalacia occurs over time with P deficiency with failure of mineralization of the remodeled osteoid matrix. In the adult, P in bone released during remodeling is used to maintain concentrations of P in blood rather than being reincorporated into bone. In young animals, bone cartilage remains unmineralized, resulting in bone that can be flexed without breaking.

Maximum Tolerable Level

NRC (2005) set the MTL of P for cattle at 0.7 percent of diet DM. That concentration was chosen because studies feeding higher concentrations were lacking, not because data were available showing negative effects when cattle were fed diets with >0.7 percent P. Long-term feeding of excess P can cause problems with Ca metabolism, inducing excessive bone resorption and urinary calculi, secondary to the elevated concentrations of P in blood (NRC, 2005). Most often, P toxicity is complicated with low dietary Ca, but ruminants can tolerate a wide ratio of Ca-to-P as long as P and Ca are adequate. Supplemental phosphates given in large oral doses are not considered highly toxic but can result in mild diarrhea and abdominal distress. Dairy cattle are quite adept at excreting excess absorbed P to maintain concentrations of P in blood within a normal range via salivary secretion and fecal excretion (Challa et al., 1989). Urinary excretion of P also may increase, although its quantitative importance is small relative to fecal excretion. Feeding 0.69 percent P to Holstein–Friesian cows for 14 weeks prepartum through 22 weeks of lactation caused no problems or signs of toxicity (De Boer et al., 1981). In contrast, a meta-analysis determined that even moderate overfeeding of P during the prepartum period was a risk factor for hypocalcemia (Lean et al., 2006). High P intake (>80 g/d) by cows approaching parturition increased blood P and incidences of milk fever and hypocalcemia (Reinhardt and Conrad, 1980). High (0.64 percent versus 0.22 percent) dietary P reduced apparent absorption of magnesium (Mg) in pregnant dairy heifers (Schonewille et al., 1994).

Magnesium Requirement

A factorial approach was taken to describe the Mg requirements of dairy cattle.

Absorption and Dietary Requirements

Dietary requirements, not absorbed requirements, are generally similar to NRC (2001); however, the previous version included a substantial safety factor. If a similar safety factor was included, dietary requirements would be approximately 25 percent greater than the previous version.

Mg is absorbed primarily from the small intestine of young calves. As the rumen and the reticulum develop, they become the main site for Mg absorption (Pfeffer et al., 1970; Martens and Rayssiguier, 1980), but some absorption may occur in the large intestine. In adult ruminants, the small intestine is a site of net secretion of Mg, but absorption may still occur in that site (Greene et al., 1983). Absorption of Mg from the rumen mostly occurs via two active mechanisms (Leonhard-Marek et al., 2010; Fach, 2015; Martens et al., 2018). One mechanism (potential difference-dependent uptake) is driven by an electrical gradient at the apical (luminal) membrane and is an active process inhibited by elevated potassium (K) concentrations in rumen fluid (Leonhard-Marek and Martens, 1996; Leonhard-Marek et al., 2010; Fach, 2015). This is a high-affinity, low-capacity transporter system. The second system (low affinity, high capacity) is driven by the Mg concentration gradient that can exist between the rumen contents and the epithelial cell and is independent of the electrical potential difference and not sensitive to K concentrations. This transport system is active and electrically neutral; therefore, it involves either cotransport of an anion (e.g., Cl⁻ or HCO₃⁻) or an exchange with intracellular protons (Leonhard-Marek et al., 2010; Fach, 2015). The exact mechanism is not known at this time.

Factors Affecting Absorption

Absorption of Mg does not appear to be under any type of hormonal regulation; excess absorbed Mg is filtered by the kidney and excreted. A major driver of Mg absorption is the gradient between intracellular Mg and rumen contents. An increase in Mg intake usually linearly increases the concentration of Mg in rumen fluid, which usually increases apparent and calculated true absorption of Mg (Jittakhot et al., 2004a,b,c). However, Mg absorption might be saturable. Increasing dietary Mg concentrations above 1.1 percent continued to increase the concentration of soluble Mg in rumen fluid but did not increase apparent or true absorption of Mg by dry dairy cows (Jittakhot et al., 2004b).

Martens et al. (2018) reviewed Mg absorption by ruminants and antagonists to absorptions in great detail. Dietary K is a significant antagonist to Mg absorption because ruminal K disrupts the electrical gradient needed to drive Mg absorption (Fisher et al., 1994; Ram et al., 1998; Schonewille et al., 1999, 2008; Jittakhot et al., 2004c; Weiss, 2004). Inadequate intake of Na increases the concentration of K in rumen fluid (Bailey, 1961; Martens et al., 1987) and reduces absorption of Mg (Martens et al., 1987). However, once the Na requirement is met, dietary Na does not appear to affect Mg absorption. High dietary P concentration (ca. 0.6 percent) reduced apparent Mg absorption in heifers by about 18 percent (Schnewille et al., 1994), but within typical dietary concentrations, effects of dietary P are probably small.

Abrupt elevation in concentrations of ruminal ammonia reduces Mg absorption; however, chronic elevation (i.e., several days) did not affect Mg absorption (Gäbel and Martens, 1986). High concentrations of ruminal ammonia reduce the electrical potential, but the change probably is not great enough to affect Mg absorption. The adaption response suggests the involvement of inducible transport proteins, and alteration of the Na/proton pump has been implicated (Fach, 2015). Dietary changes that cause an abrupt increase in ruminal ammonia (e.g., initial turnout onto high-protein pasture) should be avoided; however, once animals are adapted, high ruminal ammonia does not appear to affect Mg absorption.

Rumen pH is negatively correlated with Mg solubility and under in vitro and other experimental situations, a small drop in pH within the normal physiological range (6.5 to 5.5) has increased Mg solubility by more than 50 percent (Dalley et al., 1997). When rumen pH was reduced by more realistic diet manipulation (i.e., increased starch concentrations), ruminal Mg concentrations increased (Schonewille et al., 2000a), but the effect was less consistent than with in vitro systems. Furthermore, the effect of ruminal pH on absorption of Mg was less dramatic than changes in solubility (Horn and Smith, 1978). In addition to Mg solubility, pH may have direct effects on Mg absorption systems. In cell culture experiments, Mg permeability through a protein channel increased markedly as pH decreased below 7 (Li et al., 2007). Increasing the dietary concentration of readily fermentable carbohydrates can increase apparent absorption of Mg. Adding 30 percent starch to a diet increased apparent Mg absorption by 50 percent (0.37 versus 0.24) or 28 percent (0.25 versus 0.20) when goats were fed low (0.8 percent) or high (3.4 percent) K diets, respectively (Schonewille et al., 1997). However, apparent Mg absorption did not differ when dairy cows were fed diets with 10 or 20 percent starch (Schonewille et al., 2000a). More data with cattle are needed before the effects of dietary starch can be modeled.

Supplemental dietary fat can reduce apparent digestibility of Mg, but the reduction was not related to the concentration of supplemental fat in the diet (Jenkins and Palmquist, 1984; Rahnema et al., 1994). Apparent Mg absorption decreased about 20 percent when cattle were fed diets that contained 2.5 to 5 percent added fat compared with control diets with no added fat. Supplementing up to 5 percent added fat from whole cottonseed did not affect apparent Mg absorption (Smith et al., 1981). Although data are limited, assuming a 20 percent reduction in absorption of Mg when supplemental fat is fed is recommended but was not included in the software. Feeding ionophores increased apparent absorption of Mg by beef cattle and dairy cattle by 10 to 28 percent when magnesium oxide (MgO) was fed (Greene et al., 1986a; Spears et al., 1989; Tebbe et al., 2018). However, monensin reduced absorption of Mg by 23 percent when magnesium sulfate was fed (Tebbe et al., 2018). Effects of monensin on Mg absorption are not included in the model.

The availability of Mg from MgO is affected by particle size (smaller particles enhance absorption), calcination temperature, and origin (Jesse et al., 1981; van Ravenswaay et al., 1989; Xin et al., 1989; Hemingway et al., 1998). Particle size also likely affects Mg availability from magnesium carbonate (MgCO₃), magnesium hydroxide (Mg(OH)₂), and dolomitic limestone.

Quantifying Absorption

In NRC (2001), inadequate data were available for a rigorous evaluation of Mg absorption, but a substantial number of studies have since been published. However, quantitative estimates of the true absorption of Mg are still difficult to obtain because of the uncertainty regarding the daily loss of endogenous fecal Mg. Endogenous fecal Mg has been expressed relative to BW, and typical estimates were 2 to 5 mg Mg/kg BW (Greene et al., 1986b; NRC, 2001; Schonewille et al., 2008). However, saliva and digestive secretions are important contributors to endogenous fecal Mg, and these are related more to DMI than BW, especially when comparing across physiologic states (e.g., dry versus lactating cow). Therefore, data from two meta-analyses (Weiss, 2004; Schonewille et al., 2008) were used to estimate endogenous fecal Mg as a function of DMI. Dietary Mg (g/kg of diet DM) was regressed on concentration of apparently digested Mg (g/kg) with trial as a random effect, but because of the negative effect of K, only studies with dietary K ≤2 percent were used. The absolute value of the intercept, 0.3 g Mg/kg DMI, is an estimate of endogenous fecal Mg. In sheep, loss of endogenous fecal Mg was positively correlated with serum concentrations of Mg (Allsop and Rook, 1979). If this is true for dairy cattle, cows consuming less than adequate Mg could have a lower loss of endogenous fecal Mg than cows fed adequate Mg, but no adjustment was made to endogenous fecal Mg loss based on Mg status of the cow.

Adequate data were available to quantify the relationship between dietary K concentration and Mg absorption by dairy cows. Data from studies using heifers, dry cows, and lactating cows (Weiss, 2004; Holtenius et al., 2008; Schonewille et al., 2008) were combined (see Table 7-2). If the amount of supplemental Mg as a percentage of total diet Mg could not be calculated, the study was deleted. The final data set contained 97 treatment means from 23 studies. True absorption of Mg was calculated as described above, and only dietary K concentration and percentage of total Mg provided by supplemental sources (MgO was the source of supplemental Mg in all studies except for three) were statistically related to it. The effect of dietary K was not linear; transforming to the natural logarithm provided the best fit. The resulting equation (trial was included as a random effect) was

where K is expressed as g/kg total diet and Supplemental = percentage of dietary Mg provided by MgO. Standard errors associated with the coefficients are 4.8, 1.54, and 0.034 for intercept, K, and supplemental coefficients, respectively.

A potential problem with this equation is the collinearity between dietary Mg concentration and supplementation (r = 0.70); however dietary Mg concentration was not statistically related with true absorption of Mg. Setting supplemental Mg at 0 and basal dietary K as 12 g/kg of diet DM (approximate K requirement), true absorption of Mg from basal diet = 0.31, which was assigned as the default for all feeds. Setting supplemental Mg at 100 percent and dietary K at 12 g/kg yields an estimate of 0.23 as the default availability for Mg from MgO, which is 26 percent lower than true absorption of Mg from basal feeds. This agrees with individual studies (van Ravenswaay et al., 1989; Davenport et al., 1990; Holtenius et al., 2008) in which apparent absorption of Mg was measured for diets with and without supplemental MgO and with <20 g of K/kg DM. In those studies, true absorption of Mg from MgO (calculated using the difference method) was 22 to 45 percent lower than the true absorption of Mg from the basal diet. The prediction error associated with Equation 7-13 is high (95 percent prediction interval associated with estimated ACs is + 0.16); users may wish to adjust ACs based on risk tolerance. In the previous NRC, the default AC for basal ingredients was reduced by 1-SD unit from the mean.

Absorption coefficients for common Mg supplements are in Table 19-3 (see Chapter 19). The default value for MgO reflects the average of the MgO used in the experiments; however, substantial variation exists among MgO sources, which can influence Mg availability as discussed above. High-quality MgO (e.g., small particle size and proper calcination procedures) may have greater availability than the default value. The proportion of particles <0.25 mm in MgO is positively correlated, and the proportion of particles >1.0 mm is negatively correlated with apparent absorption of Mg. Solubility of Mg from MgO in various solutions (water, citric acid, weak hydrochloric acid, buffered rumen fluid) is positively correlated with Mg absorption, but current data are not adequate to use solubility to quantify or adjust the ACs.

Few data are available for other Mg supplements. Relative to MgO, calculated true absorption of Mg was 1.7 times greater (van Ravenswaay et al., 1989) for magnesium sulfate (MgSO₄), about the same for Mg(OH)₂ (Davenport et al., 1990; Hemingway et al., 1998) and reagent-grade MgCO₃ (Ammerman et al., 1972), about 0.5 times for dolomite limestone (Gerken and Fontenot, 1967), and 0.2 times for magnesite (Ammerman et al., 1972). However, Tebbe et al. (2018) reported that in diets without monensin, apparent absorption of Mg when MgSO₄ was fed was only about 10 percent greater than that from MgO. Based on available data and because efficiency of Mg absorption differs between sheep and cattle, data from dairy cows were given more weight than data from sheep, and the true absorption of Mg from MgSO₄ was assumed to be 20 percent greater than that from MgO. With monensin, apparent absorption of Mg when MgSO₄ was fed was about 30 percent lower than when MgO was fed (this effect is not included in the model). Data are not available for MgCl₂, but because of similar solubility to MgSO₄, they were assigned the same AC.

Magnesium Deficiency

A deficiency of Mg is of greater practical concern than deficiency of most other minerals because of the limited labile stores of Mg within the body and because of the commonly occurring antagonists of Mg absorption discussed above. A clinical deficiency of Mg results in muscle twitching, hyperexcitability, convulsions, and often death (Martens et al., 2018) and is commonly referred to as grass or lactation tetany because it often occurs in spring when cattle are first let out to graze, and it is more common in lactating than nonlactating cattle. The direct cause of clinical signs is low concentration of Mg in cerebrospinal fluid. Low concentrations of Mg in plasma (less than approximately 0.7 mmol/L) are not associated with any specific clinical signs but are a risk factor for clinical hypocalcemia (discussed in more detail in Chapter 12).

Maximum Tolerable Level

Cattle can excrete large amounts of Mg in urine, so Mg toxicity is not a practical problem in dairy cattle. Although an MTL of 0.6 percent has been established (NRC, 2005), negative effects in cattle have been observed only when dietary concentrations are >1 percent. The negative effects of high Mg are generally reduced feed intake, reduced diet digestibility, and osmotic diarrhea.

Fecal Sodium, Potassium, and Chloride

Ruminants evolved consuming forages that were high in K (>20 g K/kg DM), low in Na (≤1 g Na/kg DM), and moderate in Cl (3 to 6 g Cl/kg DM). Therefore, their requirements reflect the differences in relative K, Na, and Cl concentrations of feeds. Dairy cow feces contain approximately 85 percent water. Fecal water output was strongly related to the sum of Na, K, and Cl fecal excretion when expressed on an equivalent weight basis in 122 balance experiments with dairy cows with a mean fecal strong ion excretion rate of 3.47 (± 1.24) equivalents per day, where Fecal H₂O, L/d = 15.5 (± 1.78) + 5.88 (± 0.385) × Fecal Strong Ions (Eq/d); RMSE = 3.89; R² = 0.861; P< 0.001. Because of the relationship between strong ion and fecal water excretion, the committee suggests that metabolic fecal requirements for Na, K, and Cl are likely due to the need to maintain osmotic balance and consistent fecal moisture content.

Sodium

Cattle evolved on feeds that are low in Na; hence, they developed efficient absorptive processes and a tenacious ability to conserve Na via the kidney, but they have only a small reservoir of Na in a form that is readily available for metabolism.

Physiologic Roles

Na is the primary extracellular cation (Aitken, 1976). In addition, 30 to 50 percent of total body Na is in a nonexchangeable fraction in the crystalline structure of bone (Eldman et al., 1954). The exchangeable fraction of Na modulates extracellular fluid volume and acid-base equilibrium (Stewart, 1983; McKeown, 1986). Heart function and nerve impulse conduction and transmission are dependent on the proper balance of Na and K. Na also plays an indispensable role in sodium–potassium adenosine triphosphate enzyme (Na-K ATPase) responsible for creating electrical gradients for nutrient transport. The Na–K pump is essential for all eukaryotic cells, enabling transport of glucose, amino acids (AAs), and phosphate into cells and hydrogen (H), Ca, bicarbonate, K, and Cl ions out of cells (Lechene, 1988). Sodium bicarbonate (NaHCO₃) is a major component of saliva that helps buffer acids produced during rumen fermentation (Erdman, 1988).

Typical Na concentrations in blood plasma are 150 mEq/L and 160 to 180 mEq/L in saliva. Na is the predominant cation in rumen fluid with a typical content of 80 to 90 mEq/L, but the range can be from 50 to 140 mEq/L (Bennink et al., 1978; Catterton and Erdman, 2016). Ruminal concentrations of Na and K are strongly negatively correlated (Catterton and Erdman, 2016). Increased dietary K results in increased rumen K concentrations, which stimulate Na absorption across the rumen wall and reduce rumen Na concentration to maintain electrical and osmotic neutrality (Martens and Blume, 1987).

Sodium Utilization and Homeostasis

Absorption occurs throughout the digestive tract, and dietary Na generally is assumed to be almost completely available. Absorption occurs by an active transport process in the reticulorumen, abomasum, omasum, and duodenum. Passive absorption also occurs through the intestinal wall, so there is a tendency toward equal concentrations in intestinal and fecal fluids. However, substantial active absorption against a sizable concentration gradient occurs in the lower small intestine and large intestine (Renkema et al., 1962).

Na and K can interchange such that in Na-deficient animals, K excretion is increased, providing a mechanism that helps ensure that ruminants can subsist on feeds low in Na over long periods of time. Na concentrations in blood and tissues are maintained principally via reabsorption and excretion by the kidneys. Excretion of Na, K, and Cl− is closely synchronized. Na is the central effector of ion excretion, and changes in renal reabsorption are chief determinants of Na excretion. Endocrine control via tissue receptors and the renin–angiotensin system, aldosterone, and atrial natriuretic factor monitor and modulate Na concentrations in various tissues, which consequently control fluid volume, blood pressure, K concentrations, and renal processing of other ions. When cattle are depleted of Na, salivary glands decrease secretion of Na in saliva. The decrease in Na content is replaced reciprocally by nearly the same concentration of K (van Leeuwen, 1970; Morris and Gartner, 1971).

Requirement for Absorbed Sodium

Dietary Requirement and Efficiency of Absorption

The regression coefficient of 0.98 for absorbed Na versus dietary Na was not statistically different from 1, implying that true absorption is 100 percent. The previous committee (NRC, 2001) set the Na absorption rate at 90 percent, which seemed too low given the ruminant animals' ability to survive on extremely low Na diets. In addition, Na from typical feeds is solubilized and released in the liquid matrix of digesta and is readily available for absorption. Feedstuffs commonly used in diets for dairy cattle do not contain enough Na to meet requirements, and supplemental sources typically account for the majority of an animal's total Na intake.

NaCl is the most often used supplement, and its Na is considered 100 percent available. The efficiency of absorption of Na from other salts (e.g., NaHCO₃, sodium carbonate [Na₂CO₃], sodium sesquicarbonate [Na₃H (CO₃)₂]) is also considered essentially 100 percent. When some animal byproduct feedstuffs containing bone are fed, Na would be less available as it is tightly bound in the crystalline structure. However, these sources represent rare circumstances and are minor Na sources compared to typical supplemental Na salts. Therefore, the committee set the AC for Na at 1.00 for all feeds.

For a 650-kg cow consuming 28 kg/d of feed DM, the metabolic fecal requirement for Na (28 kg × 1.45 g Na/kg) would be 41 g of dietary Na/d. The milk production requirement for a cow producing 45 kg/d milk would be 18 g/d (45 kg milk × 0.4 g Na/kg in milk) for a total Na requirement of 59 g/d or 0.21 percent Na in the diet DM. This compares with the previous requirement of 36 g (50 × 0.65 / 0.90) for milk production and 27 g (700 × 0.038 / 0.90) for maintenance for a total of 63 g Na/d (0.25 percent of diet DM). While the maintenance requirement for Na has increased, the reduced requirement for Na in milk more than compensated, such that total Na requirements are slightly lower than in the 2001 NRC.

Lactational Responses to Varying Dietary Sodium Concentrations

Kemp and Geurink (1966) reported that 0.14 percent Na in grazed forage was sufficient to support more than 30 kg of milk production per day. However, feeding lactating dairy cows a diet with no supplemental NaCl (0.16 percent Na, dry basis) resulted in marked depressions in DMI and milk yield after just 1 to 2 weeks of feeding (Mallonee et al., 1982a). Empirical modeling of data from 15 experiments with lactating cows conducted in either cool or warm seasons suggested that DMI and milk yield were improved by dietary concentrations of Na well above those needed to meet requirements (Sanchez et al., 1994b,c. DMI and milk yield responses over a range of dietary Na concentrations (0.11 to 1.20 percent, dry basis) were curvilinear, with maximum performance at 0.70 to 0.80 percent Na. Concentrations of Na, K, Cl⁻, Ca, and P in diet DM ranged from below those needed to meet requirements to concentrations considerably higher. Thus, there is a potential confounding between Na and DCAD, and it is not known whether the optimal Na concentration would vary if the dietary concentrations of other macrominerals would have been closer to requirements. There were interactions of Na with K, Cl⁻, and P on DMI, indicating that responses to Na differed over the range of dietary concentrations to those minerals. In addition, interactions of dietary Na with K, Cl⁻, and P on DMI differed in experiments conducted in the cool or warm season. In hot weather, milk yield and DMI increased when Na increased from basal (0.18 percent Na, dry basis) to 0.55 percent dietary Na with either NaCl or NaHCO₃; Cl− was equalized among diets (Schneider et al., 1986).

Little evidence exists for a Na-by-K interaction when dietary Cl⁻ was held constant, and Na and K were fed at or above their estimated requirement. Na and K were equally effective when dietary anion cation difference was increased by addition of either cation. In only one study (Iwaniuk and Erdman, 2015) was Na more effective than K in maintaining milk fat, but cation had no effect on milk yield or intake. In other experiments (West et al., 1992; Sanchez et al., 1997; Hu and Kung, 2009), the K/Na ratio did not affect intake or milk production. Wildman et al. (2007) showed a quadratic effect of the K/Na ratio on milk production but no effect on intake or milk composition.

Sodium Deficiency

Babcock (1905) fed a diet very low in Na to dairy cows and described intense craving for salt, licking and chewing various objects, and general pica. Deficiency signs were manifested within 2 to 3 weeks. Na deficiency signs may not develop for weeks to months, depending on rate of milk production. However, feed intake and milk yield began to decline 1 to 2 weeks after cows were fed a diet without supplemental NaCl (0.16 percent Na), and pica and drinking of urine of other cows were observed (Mallonee et al., 1982a). Although dietary Cl− concentration was not measured in that study, potassium chloride (KCl) was supplemented (1.0 percent total dietary K), so Cl⁻ deficiency was probably not the cause of the condition. The condition was reversed quickly by inclusion of NaCl in the diet. Other deficiency signs include loss of appetite; rapid loss of BW; an unthrifty, haggard appearance; lusterless eyes; and rough hair coat (Underwood, 1981). More extreme signs of deficiency include incoordination, shivering, weakness, dehydration, and cardiac arrhythmia leading to death.

Free-Choice Feeding of Sodium Chloride and Sodium (Sodium Chloride) Toxicity

Cattle consume salt liberally when it is available. Smith et al. (1953) found that lactating cows consumed more salt when provided free-choice in granular versus block form, but consumption of block was sufficient to meet needs for lactation. Demott et al. (1968) fed lactating cows 4 percent NaCl in a grain mix at 1 kg of grain for each 2 kg of 4 percent fat-corrected milk yield for 2 weeks without ill effects on milk yield, BW, or general health. Although total DMI was not measured, the Na concentration of the total diet DM would have been about 0.8 to 1.0 percent. High intake of NaCl can increase the incidence and severity of udder edema (Randall et al., 1974). Feeding diets with 0.88 percent Na from NaCl or NaHCO₃ to mid-lactation Holstein cows did not cause toxicity or reduce feed intake and milk yield compared with 0.55 percent Na (Schneider et al., 1986).

A major factor influencing the degree of exhibition of NaCl toxicosis is the availability and quality of drinking water. Extensive discussion of the effects of high Na and Cl− concentrations in drinking water is provided in Chapter 9. NRC (2005) set the MTL of NaCl at 3 percent of diet DM for lactating cattle and 4.5 percent for growing cattle.

Physiologic Roles

Cl− is the major anion in the body involved in regulation of osmotic pressure, making up more than 60 percent of the total anion equivalents in the extracellular fluid. As a strong anion, it always is dissociated in solution. It is essential for transport of carbon dioxide and oxygen, the chief anion in gastric secretions, and accompanied by H⁺ in nearly equivalent amounts. It is needed for activation of pancreatic amylase, and chlorinated compounds are produced by some phagocytic cells to kill pathogens. Typical concentrations of Cl⁻ are from 90 and 110 mEq/L in blood plasma and 10 to 30 mEq/L in ruminal fluid. The concentration of Cl− in cattle was estimated to be about 1.2 to 1.4 g/kg over the range of 100 to 500 kg empty body weight (EBW; ARC, 1980).

Utilization and Homeostasis

About 80 percent of the Cl⁻ entering the digestive tract arises from digestive secretions in saliva, gastric fluid, bile, and pancreatic juice. Cl− is absorbed throughout the digestive tract. It, like Na, is absorbed mainly from the upper small intestine by passive diffusion following Na along an electric gradient. Cl− is transported across the ruminal wall to blood against a wide concentration gradient (Sperber and Hyden, 1952). Martens and Blume (1987) showed that Cl⁻ was co-transported actively with Na across the rumen wall, although the exact mechanism is unclear. Substantial absorption of Cl− from gastric secretions (hydrochloric acid) occurs in the distal ileum and large intestine by exchange with secreted bicarbonate. Appreciable quantities of Cl⁻ are excreted in the feces, in part to maintain osmatic balance along with the other strong ions (Na⁺ and K) to maintain fecal moisture content. In the short term, relatively large day-to-day differences in dietary intake of Cl− have little effect on the total Cl⁻ entering the digestive tract. Much smaller amounts of Cl⁻ are lost in sweat mainly as NaCl or KCl. Cl− fed in excess of needs for maintenance and milk production is primarily excreted in the urine.

Tight regulation of the concentration of Cl− in extracellular fluid and its homeostasis is coupled intimately to that of Na. The role of Cl− in maintaining ionic and fluid balance was thought to be passive to that of Na and K. However, Fettman et al. (1984b) showed that during Cl⁻ deficiency, the ion functioned independently to mediate Cl− conservation. Cl− was conserved by reducing excretion by the kidney, as well as in feces and milk. Excess Cl− intake is excreted mainly in urine of steers and sheep (Nelson et al., 1955), but in lactating cows, a significant amount of Cl− is excreted via feces (Coppock, 1986). Normally, anion concentration in extracellular fluid is regulated secondarily to cation concentrations, and when the amount exceeds reabsorption capability of the kidney, excess Cl− is excreted in urine (Hilwig, 1976). Cl− excretion is tied to excretion of strong cations, acid-base balance, and maintenance of electrochemical neutrality of the urine (Stewart, 1981). Under normal circumstances, excess cations are secreted in conjunction with Cl⁻, and bicarbonate ion excretion increases to maintain electrochemical balance, resulting in alkaline urine. If bicarbonate or electrolyte cations need to be conserved in relation to Cl⁻, Cl− excretion is accompanied by ammonium ions and urine pH decreases to maintain systemic acid-base balance.

Requirement for Absorbed Chloride

Dietary Requirement and Efficiency of Absorption

Little research has been done in ruminants to measure the true AC for Cl⁻ principally due to the widespread availability of good, inexpensive inorganic sources (e.g., NaCl). Cl− from inorganic sources and common feedstuffs is freely released into the liquid phase of the digesta and readily absorbed (Underwood, 1981). Apparent absorption of Cl− in lactating cows fed fresh forage ranged from 71 to 95 percent and averaged 88 percent (Kemp, 1966). This is comparable to other estimates of absorption efficiency of 85 to 91 percent in cattle and sheep fed mixed diets (ARC, 1980). Paquay et al. (1969b) found that apparent absorption of Cl⁻ was not influenced by intake of Cl− but was correlated negatively with intakes of DM, energy, and pentosan, as well as positively correlated with intakes of K and N. Factors such as lactation, pregnancy, and growth affecting the requirement for Cl− do not appear to alter the efficiency of Cl⁻ absorption. Overall, the absorption efficiency for Cl− in ingredients commonly fed to dairy cattle is usually ≥90 percent (Henry, 1995b). The committee estimated AC from 144 Cl− balance studies was 92 percent. Therefore, an AC for Cl− of 0.92 was assigned for all dietary ingredients.

For a 650-kg cow consuming 28 kg/d DMI and producing 45 kg/d milk, the new requirement for Cl− includes 34 g/d for maintenance (28 kg DMI × 1.11 / 0.92) plus 49 g Cl⁻ required for milk production (45 kg milk × 1 / 0.92) for a total of 83 g dietary Cl⁻. This compares with the previous dietary Cl⁻ requirement of 81 g/d (NRC, 2001). While the maintenance requirement has increased, milk production requirements have decreased such that the total Cl⁻ requirement for lactating cows has changed little compared to the previous NRC.

Lactation and Growth Responses to Varying Dietary Chloride

Coppock (1986) reviewed the estimated requirements of dietary Cl− for lactating dairy cows in studies in which milk production ranged from 24 to 32 kg/d. Holstein cows fed a diet with 0.18 percent Cl− conserved Cl⁻ by dramatically reducing excretion of Cl− in urine and feces and tended to reduce Cl⁻ output in milk; however, intakes of feed and water, as well as milk yield and composition, did not differ from cows fed 0.40 percent Cl⁻ (Coppock et al., 1979). Half of the cows in each treatment group had free access to a trace-mineral salt block, and cows fed the diet low in Cl− consumed more of the salt block. Fettman et al. (1984a) fed diets containing 0.10, 0.27, and 0.45 percent Cl− for the first 8 to 11 weeks of lactation. Cows fed 0.10 percent Cl⁻ rapidly exhibited clinical signs of Cl⁻ deficiency and poor performance compared with those fed medium and high concentrations of dietary Cl⁻. Health, feed intake, and yield and composition of milk by cows fed the medium and high concentrations of dietary Cl⁻ were similar. Empirical models with a large data set showed that increasing dietary Cl⁻ over a range of 0.15 to 1.62 percent decreased DMI and milk yield of mid-lactation cows (Sanchez et al., 1994b). The negative effects of increasing dietary Cl⁻ were more dramatic in hot summer weather than in winter. This is consistent with the results of Escobosa et al. (1984) showing profound exacerbating effects of high dietary Cl⁻ on acid-base balance (metabolic acidosis) and lactation performance during heat stress. Adding Cl⁻ to the diet when the other strong ions (Na⁺ and K⁺) are held constant reduces DCAD, and the decrease in DCAD was likely the cause of the negative effect of high Cl⁻ (discussed in the DCAD section below).

Feeding diets with 0.038 percent Cl− for 7 weeks to male Holstein calves did not produce clinical deficiency or depress feed intake, growth rate, or digestibility of feed compared with calves fed 0.50 percent Cl⁻ (Burkhalter et al., 1979). Calves fed the low Cl− diet adapted by reducing urinary excretion of Cl⁻, and their water intake and urine output were greater than that of calves fed more Cl⁻. Calves fed a low Cl⁻ (0.038 percent) diet developed mild alkalosis, but it did not affect growth, and calves adapted to the low intake of Cl⁻ (Burkhalter et al., 1980).

If NaCl is used to meet the Na requirement, generally the Cl⁻ requirement is met or exceeded. However, if NaHCO₃ or some other Na-containing salt is used to supply Na, it may be necessary to meet the Cl⁻ requirement with another supplement (e.g., KCl). Research is needed to establish more accurate requirements and appropriate dietary concentrations of Cl⁻ (and Na) for all classes of dairy cattle. If current estimates are too high, it could contribute to soil salinity when manure is applied (Coppock, 1986). Cl− in drinking water also may make a major contribution to intake of Cl⁻. In a survey of 39 California dairy herds by Castillo et al. (2013), inclusion of the Cl− in water increased estimated total Cl− intake by 6.5 percent.

Chloride Deficiency

Cl⁻ deficiency was created in young calves (100 kg BW) by feeding a diet with 0.063 percent Cl− and removing about 600 g of abomasal contents daily (Neathery et al., 1981). Clinical signs were anorexia, weight loss, lethargy, mild polydipsia, and mild polyuria. In latter stages, severe eye defects and reduced respiration rates occurred, and blood and mucus appeared in feces. Deficiency of Cl− resulted in severe alkalosis and hypochloremia, which manifested in secondary hypokalemia, hyponatremia, and uremia. Control calves also had abomasal contents removed daily but were fed a diet with 0.48 percent Cl⁻, and they grew normally and showed no signs of deficiency. During the first 8 to 11 weeks of lactation, dairy cows fed low (0.1 percent, dry basis) Cl− exhibited dramatic and progressive declines in intakes of feed and water, BW, milk yield, and electrolyte concentrations in blood serum, saliva, urine, milk, and feces (Fettman et al., 1984b).

A significant decline of Cl− in blood serum was found within 3 days after switching cows from a diet containing 0.42 percent to a diet with 0.10 percent Cl⁻ (Fettman et al., 1984b). Clinical signs of deficiency were depraved appetite, lethargy, hypophagia, emaciation, hypogalactia, constipation, and cardiovascular depression. Metabolic alterations were severe primary hypochloremia, secondary hypokalemia, and metabolic alkalosis (Fettman et al., 1984a,bc). Cl⁻ deficiency, resulting from an inadequate dietary supply or loss of gastric juices, can lead to alkalosis due to an excess of bicarbonate, because inadequate Cl− is partially compensated for by bicarbonate.

Chloride Toxicity

High systemic concentrations of Cl⁻, in the absence of a neutralizing cation (e.g., Na⁺), can cause disturbance of normal acid-base equilibrium (Stewart, 1981; Escobosa et al., 1984), but the maximum tolerable concentration of Cl− in the diet has not been determined. The maximum tolerable concentration of dietary NaCl was set at 3.0 percent (dry basis) for lactating dairy cows and 4.5 percent for growing cattle (NRC, 2005).

Physiologic Roles

K is the third most abundant mineral in the body. It must be supplied daily because there is little storage in the body and the animal's requirement for K is high. K is involved in osmotic pressure and acid-base regulation, water balance, nerve impulse transmission, muscle contraction, and oxygen and carbon dioxide transport; as an activator or cofactor in many enzymatic reactions; in cellular uptake of AAs and synthesis of protein; in carbohydrate metabolism; and in maintenance of normal cardiac and renal tissue (Stewart, 1981; Hemken, 1983). It is the major intracellular electrolyte with concentrations in the range of 150 to 155 mEq/L. In contrast to Na⁺ and Cl⁻, extracellular concentrations of K⁺ are low (about 5 mEq/L). Saliva typically contains <10 mEq/L, whereas concentrations in ruminal fluid range from 40 to 100 mEq/L (Bennink et al., 1978; Catterton and Erdman, 2016). Blood plasma contains 5 to 10 mEq/L. The vast majority of K in blood is located within red blood cells (Aitken, 1976; Hemken, 1983). About 80 percent of the K in the body is associated with lean tissue and bone. Gastrointestinal contents account for an additional 15 percent of body K and are affected by the K content of the diet (Belyea et al., 1978).

Potassium Utilization and Homeostasis

K is absorbed primarily in the duodenum by simple diffusion, and some absorption occurs in the jejunum, ileum, and large intestine. The main excretory route of excess absorbed K is via urine. This route is primarily under regulation by aldosterone, which increases Na reabsorption in the kidney with the concomitant excretion of K. Blood acid-base status also affects urinary excretion of K (McGuirk and Butler, 1980). With the onset of an alkalotic condition, intracellular H⁺ are exchanged with K⁺ in plasma as part of the regulatory mechanisms to maintain blood pH. A large gradient exists between intracellular renal tubule concentrations of K and that of luminal fluid (urine). This gradient affects the passage of K from the tubular cells into urine.

There is a distinct relationship between excess cations such as Na⁺ and K⁺ and urinary pH (Hu and Murphy, 2004). Excess K and Na are excreted in the urine and result in increased urinary bicarbonate secretion. Because K⁺ is the primary cation in dairy cattle diets, intake responses to K may be directly related to changes in urinary acid-base balance. Fecal K is primarily from endogenous losses as true digestibility of K approaches 100 percent.

Requirement for Absorbed Potassium

The factorial method was used to derive the absorbed K requirement. In the previous report (NRC, 2001), the maintenance requirement for K was set at 0.038 g K/kg BW plus 6.1 g K/kg DMI. These requirements were based on suggested endogenous urinary losses of 0.038 g K/kg BW and endogenous fecal losses of 2.6 g K/kg of dietary DM (Gueguen et al., 1989) coupled with an empirical adjustment of the endogenous fecal losses of an additional 3.5 g K/kg diet DM for a total 6.1 g K/kg. The empirical adjustment to fecal losses was added because based on production responses, the initial requirement was not adequate (Dennis et al., 1976; Dennis and Hemken, 1978; Erdman et al., 1980; Sanchez et al., 1994b,c. Generally, as dietary K increased from 0.5 to 1.2 percent of dietary DM, feed intake was consistently increased. The previous committee suggested that a higher maintenance requirement for absorbed K for lactating cows compared with nonlactating animals was justified based on K's role in dynamic processes associated with ruminal function at higher levels of feed intake and maintenance of systemic acid-base balance.

However, the adjustment was applied to the metabolic fecal K requirement. Applying the adjustment in this way implied much greater fecal losses than the actual measured losses. A metabolic fecal requirement of 6.1 g K/kg diet DM in a diet containing 1.2 percent K would have implied an apparent AC of 0.49, which is far below any measured values in the literature.

As it is difficult to formulate a diet with less than 1 percent K using traditional forages, the studies used to determine the intake and milk production responses to K often fed atypical diets that were high in cereal grains, by-product feeds such as brewers dried grains, distillers grains, and cottonseed hulls as a forage substitute to achieve a low K basal diet (Dennis et al., 1976; Dennis and Hemken, 1978; Erdman et al., 1980; Sanchez et al., 1994b,c. In most cases, the calculated DCAD of the basal diets was low (0 to 50 mEq per kilogram diet DM) using the Ender et al. (1971) equation, which includes sulfide (S²⁻). In addition, dietary K was increased by addition of KCl, which would not change the DCAD concentration. Therefore, the committee is uncertain whether these results could be applied to more typical diets where basal DCAD is 200 mEq/kg or greater. Since only dietary K and Na can be used to increase DCAD, the question is whether Na⁺ could replace K⁺ as a urinary cation to maintain an alkaline urine once the needs for metabolic fecal and milk K secretion have been met.

Metabolic fecal excretion of K was estimated from the results of 149 individual K digestibility measurements from nine experiments in which cows were fed diets ranging from 0.96 percent to 1.86 percent K. The metabolic fecal requirement was determined by regression of apparently absorbed K on K intake, both expressed as grams per kilogram of feed DM. The regression equation was as follows: Absorbed K = −2.48 (± 0.74) + 1.02 (± 0.056) K Intake; RMSE = 0.52; R² = 0.93; P < 0.001. This equation suggests a metabolic fecal requirement of 2.48 g K/kg DMI and a true absorption of 1.02, which was not different from 1. Therefore, the metabolic fecal requirement was set at 2.5 g K/kg DMI, similar to previous estimates (Paquay et al., 1969a; Gueguen et al., 1989), and an AC of 1.0 was assigned to dietary K.

Urinary excretion of K is dependent on the relative excretion rates of Na and Cl⁻ (Stewart, 1981). Dairy cattle typically secrete an alkaline urine with a pH of ~8.0 (Hu and Murphy, 2004) because of the need to excrete excess cations (K⁺ and Na⁺) in relation to anions (Cl⁻), which in turn increases urinary bicarbonate secretion to maintain the electrochemical balance in the urine. As previously indicated, a maintenance requirement based on the measured metabolic fecal K alone would result in a diet that is too low in K compared to the observed experimental responses to dietary K with respect to feed intake and milk production (Dennis et al., 1976; Dennis and Hemken, 1978; Erdman et al., 1980; Sanchez et al., 1994b,c. Therefore, the committee arbitrarily set an AI to meet the endogenous urinary K needs at 0.2 g/kg BW. This will usually result in a minimum dietary K concentration of 1.00 percent of diet DM for lactating dairy cows. The AI to meet endogenous urinary K need for growing heifers and dry cows was set at 0.07 g K/kg BW to maintain a minimum total dietary K of 0.60 percent.

Dietary Requirement and Efficiency of Absorption

Hemken (1983) indicated that K is almost completely absorbed with a true digestibility of 95 percent or greater for most feedstuffs. Paquay et al. (1969a) found that the apparent absorption of K by dairy cows fed alfalfa silage, clover silage, and cabbage silage ranged from 87 to 94 percent. Apparent absorption was slightly lower in four tropical forages fed to sheep, but efficiency of absorption was not affected by maturity of the forage (Perdomo et al., 1977). Average apparent absorption of K in eight forages fed to cattle and sheep was 85 percent (Miller, 1995).

Because K is excreted mainly in urine, urinary excretion and apparent absorption are reliable criteria for estimation of efficiency of absorption. Supplemental K from inorganic sources such as potassium carbonate, KCl, and potassium sulfate is highly soluble and readily available for absorption (Peeler, 1972; Miller, 1995). In the model, an AC value of 1.00 for K was used for all feedstuffs and mineral sources.

For growing heifers weighing 300 kg, gaining 1 kg BW/d and consuming 7 kg DM/d, the previous requirements (NRC, 2001) was 35 g (0.49 percent K in diet DM). The new requirement is 41 g (0.59 percent K in diet DM). The dietary K requirement from the previous report (NRC, 2001) for a 650-kg lactating dairy cow producing 45 kg milk and consuming 28 kg diet DM/d was 265 g (0.94 percent K in diet DM). The new dietary requirement is 268 g (0.96 percent K in diet DM). Essentially, the total requirements for K have not changed substantially, but the route of excretion has shifted from metabolic fecal to endogenous urinary excretion.

Production Responses to Varying Concentrations of Dietary Potassium

Potassium Deficiency

Signs of severe K deficiency were manifested in lactating dairy cattle fed diets with 0.06 to 0.15 percent K (Pradhan and Hemken, 1968; Mallonee et al., 1982b). Marked decline in feed and water intake, reduced BW and milk yield, pica, loss of hair glossiness, decreased pliability of the hide, lower concentrations of K⁺ in plasma and milk, and higher blood hematocrit readings occurred within a few days to a few weeks after cows were offered the K-deficient diets. Rate of occurrence and severity of deficiency signs appear to be related to rate of milk production, with higher-yielding cows affected more quickly and severely than lower-yielding cows. With severe K deficiency, cows will be profoundly weak or recumbent with overall muscular weakness and poor intestinal tone (Sielman et al., 1997). In this case, hypokalemia syndrome was associated with treatment of ketosis.

When diets contained 0.5 to 0.7 percent K, the only apparent sign of inadequacy in lactating cows was reduced feed intake with corresponding lower milk yield compared with cows fed adequate K. Because many forages contain high concentrations of K, severe K deficiency would be extremely rare. However, marginal deficiency can occur if corn silage is the sole forage and no supplemental K is fed.

Potassium Toxicity

The dietary concentration of K that leads to toxicity is not well defined (Ward, 1966b). K toxicosis is unlikely to occur under natural conditions but could occur as a result of excess supplementation. Acute toxicity (Ward, 1966a) and death (apparently cardiac arrest) occurred when 501 g of K as KCl was given by stomach tube to a cow (475 kg BW). This amount was approximately the daily amount consumed by similar cows fed 15 kg of alfalfa that was consumed without ill effects. Dennis and Harbaugh (1948) administered 182 and 240 g of K as KCl without detectable clinical signs of toxicity, but 393 g by stomach tube to cattle weighing about 300 kg resulted in one death, two that required treatment, and two exhibiting no signs of toxicity. When 4.6 percent dietary K (via supplemental potassium carbonate) was fed to cows during early lactation, feed intake and milk yield were reduced, and water intake and urinary excretion were increased (Fisher et al., 1994). NRC (2005) set the maximum tolerable concentration at 2.0 percent of diet DM based on indexes of animal health. However, cattle are known to tolerate high concentrations (>3 percent of DM) of K such as are seen in early spring pastures for extended periods of time. Dietary K depresses Mg absorption and is a risk factor for grass tetany. Feeding K in excess of that needed to meet requirements can present metabolic and physiologic challenges to cattle and will increase excretion of K into the environment.

Dietary Cation–Anion Difference

DCAD was discussed in the 2001 NRC, but no requirements were suggested. The DCAD is calculated as the difference between the sum of the major cations (Na⁺ and K⁺) and the sum of the major anions (Cl⁻ and sometimes S²⁻) and is expressed in milliequivalents (mEq) per kg or per 100 g of diet DM. The simplest calculation of the DCAD equation (Mongin, 1981) includes dietary Na, K, and Cl− and was developed for use in poultry diets. Ender et al. (1971), in early work related to the use of DCAD for prevention of milk fever, proposed a DCAD equation that included sulfur (S) assumed to be S²⁻, as the second anion. When discussing specific dietary values, DCAD calculated using Ender et al. (1971) will be referred to as DCAD-S, whereas the term DCAD refers to the Mongin (1981) equation. For a diet containing 1.11 percent K, 0.25 percent Na, 0.33 percent Cl⁻, and 0.20 percent S, the DCAD and DCAD-S concentrations would be 300 and 185 mEq/kg DM, respectively.

Inclusion of other dietary cations (Ca²⁺ and Mg²⁺) and anions (P³⁻) have been suggested to be included in the DCAD equations. However, the contribution of those ions to urine net acid excretion is dependent on their relative absorption rates and the degree of urinary excretion. Constable et al. (2009) found that K⁺, Na⁺, Ca²⁺, Mg²⁺, Cl⁻, and sulfate (SO₄²⁻) ion accounted for most of the strong ion effects on urine pH in cattle. However, urinary Ca and Mg losses in lactating cows are minimal such that the relative amounts of K, Na, Cl⁻, and SO₄²⁻ excretion have the greatest impact on urine pH.

Cattle routinely consume diets that are high in DCAD, resulting in urinary excretion of excess strong cations (primarily K⁺) relative to strong anions (Cl⁻). In high DCAD diets, urinary electrochemical neutrality is maintained by increased bicarbonate secretion. Addition of a dietary cation source such as NaHCO₃ (Erdman et al., 1982) to lactating dairy cows resulted in decreased urinary net acid excretion, increased urinary bicarbonate secretion, and increased urine pH. Feed intake declined as urine pH dropped from 8 to 7 (Hu and Murphy, 2004) with decreasing DCAD. When negative DCAD diets are fed, urinary excretion of excess Cl− relative to K⁺ and Na⁺ occurs, resulting in decreased urinary bicarbonate secretion (Tucker et al., 1988) and reduced urinary pH. Reduction in urinary pH below 7 is associated with increased Ca excretion in the urine (Constable et al., 2009). Feeding low DCAD diets to prepartum dairy cows reduces prevalence of milk fever (see Chapter 12).

Much of the research on the effects of DCAD in lactating dairy cows has used cation sources such as sodium and potassium bicarbonate, carbonate, and sesquicarbonate salts to increase the DCAD concentration in the diet. The anion components of these salts can also act as buffering agents in the rumen and contribute to the increase in rumen pH and acetate/propionate ratio in part because of their buffering effect. A 100 mEq/kg DM in DCAD resulted in a linear (0.03) increase in rumen pH (Iwaniuk and Erdman, 2015). Thus, the overall effect of DCAD effects on animal performance is likely due to a combination of their effects on rumen fermentation and acid-base status of the cow.

Lactation Responses to Dietary Cation–Anion Difference

Maximum DMI, milk yield, and 4 percent fat-corrected milk yield occurred at a DCAD of 380 mEq/kg diet DM (Sanchez et al., 1994b). Similar results were observed when data were subdivided by season with maximal DCAD for maximal DMI and 4 percent fat-corrected milk yield. However, maximum response appeared to occur at a slightly higher DCAD in winter-fed versus summer-fed cows.

Since the last report (NRC, 2001), two meta-analyses of published literature on DCAD effects on lactating dairy cattle (Hu and Murphy, 2004; Iwaniuk and Erdman 2015) have been conducted. Hu and Murphy (2004) reported that DCAD had significant effects on production and acid-base responses in lactating cows. The data set included 17 experiments and 69 treatment means using DCAD calculated with the Mongin (1981) equation. Comparison of response curves for DMI and urine pH revealed a 1.5-kg/d decline in feed DMI and a 1 pH-unit drop in urine pH as DCAD decreased from 340 to 120 mEq/kg diet DM. The close relationship between reduced DMI and urine pH suggests that urinary acid-base status is related to DMI. The maximal intake and yields of milk, fat-corrected milk, and fat occurred at DCAD of 396, 336, 489, and 550 mEq/kg diet DM, respectively, and the DCAD at which 80 percent of the maximum response occurred was 200, 185, 190, and 305 mEq/kg DM, respectively.

Iwaniuk and Erdman (2015) conducted a much larger meta-analysis of data collected from 43 published studies that included 196 treatment means. This analysis included data previously summarized by Hu and Murphy (2004) and incorporated earlier and later published work where supplements such as sodium and potassium bicarbonate, carbonate, and sesquicarbonate salts were fed. Data for DMI, milk production, and 3.5 percent fat-corrected milk were fitted to an asymptotic model where Y = a + b (1 − e^(−k × ^DCAD-S)), where a = intercept, b = maximal response to DCAD-S, and k is the rate constant for the effect of DCAD-S (mEq/kg diet DM) on the response.

For DMI, the response equation was DMI, kg/d = 18.44 (± 0.389) + 1.11 (± 0.468) (1 − e⁽−0.0038 (±0.002) × ^DCAD-S)), R² = 0.41, RMSE = 0.53. Eighty percent and 66 percent of the maximal intake response to DCAD-S occurred at 425 and 290 mEq/kg diet DM, respectively. Milk production responses to DCAD-S were relatively small with a maximal response of 1.1 kg/d.

Milk fat percentage and milk fat yield increased linearly with increased DCAD with 0.1 percentage unit and a 39-g/d increase in fat percentage and fat yield per 100-mEq/kg diet DM increase in DCAD-S. Rumen pH (83 treatment means) increased linearly with increasing DCAD-S, with the changes in pH being consistent with the effects on biohydrogenation of FA intermediates that are known to inhibit milk fat synthesis (see Chapter 4).

The response function for 3.5 percent fat-corrected milk was 3.5 percent FCM, kg/d = 25.49 (± 0.751) + 4.82 (± 1.57) × (1 − e⁽−0.0024 (±0.001) × ^DCAD-S), R² = 0.48, RMSE = 0.73. For 3.5 percent FCM, 80 percent and 66 percent of the maximal response to DCAD occurred at a DCAD-S of 675 and 450 mEq/kg diet DM, respectively. However, 675 mEq/kg was outside of the range of inference in the data set. The changes in 3.5 percent fat-corrected milk production reflected the curvilinear response in milk yield and the linear response in fat yield to increasing DCAD-S.

With fewer observations (52 and 42, respectively), a 100-mEq/kg diet DM increase DCAD-S resulted in a 0.73 and 1.54 percentage unit increase in DM and NDF digestibility, respectively. The change in NDF digestibility accounted for approximately two-thirds of the increase in DM digestibility.

Dietary Cation–Anion Difference Requirements for Lactating Cows

There is no fixed requirement for DCAD in lactating cows. Rather, the feeding level chosen should be determined based on the incremental production responses (milk and fat yield) in relation to the incremental added costs (DMI and mineral salt supplementation) according to the response equations outlined by Hu and Murphy (2004) and Iwaniuk and Erdman (2015).

An absolute practical minimum or AI for DCAD would be based on the minimum requirements for the minerals, K, Na, Cl⁻, and S. For example, a 700-kg dairy cow producing 50 kg milk will require a diet containing 1.11 percent, 0.25 percent, 0.33 percent, and 0.20 percent K, Na, Cl⁻, and S, respectively. The calculated DCAD-S using those requirements would be 174 mEq/kg diet DM or 301 mEq/kg using the Mongin (1981) equation. If measured Cl− and S concentration of the diet exceeds the minimum requirement, then additional K or Na may be needed for an acceptable DCAD and DCAD-S.

Growth Responses to Dietary Cation–Anion Difference

Few studies have examined the influence of DCAD on growth of calves. Calves (1 to 12 weeks of age) grew faster when fed a 200-mEq DCAD-S diet than calves fed a −100-mEq diet (Xin et al., 1991). In another study, Holstein and Jersey calves averaging 56 to 70 days of age were fed diets containing −180, 45, 225, and 383 mEq/kg DCAD-S (Jackson et al., 1992). Feed intake and ADG responded quadratically, being greatest at 225 mEq and lowest with −180 mEq. In a follow-up study, intake, growth rate, and Ca metabolism were compared for Holstein calves (56 to 70 days of age) fed diets with −180 or 130 mEq DCAD-S/kg of dietary DM in a factorial arrangement of treatments with 0.42 and 0.52 percent dietary Ca (Jackson and Hemken, 1994). Feed intake did not differ due to DCAD, but growth rate was increased with the 130 mEq/kg DCAD-S; dietary Ca had no effect. Urinary Ca excretion was greater for calves fed diets with −180 mEq compared with diets with 130 mEq. Breaking strength of the ninth rib was greater for calves fed the 130-mEq treatment compared with the −180-mEq treatment; breaking strength of the seventh rib was greater when calves were fed either higher DCAD or higher Ca. Based on these studies, the AI of DCAD-S for growing calves is 150 to 200 mEq/kg of diet DM, which is the approximate value obtained when calves are fed their minimum requirements for K, Na, and Cl⁻.

No studies were identified in which DCAD was varied in growing dairy heifers. Growing (Ross et al., 1994b) and finishing (Ross et al., 1994a) beef steers were fed diets containing 0, 150, 300, and 450 mEq/kg DCAD (Mongin, 1981 equation), and a DCAD of 150 mEq/kg maximized feed intake and growth rate. In the absence of experiments with growing dairy heifers, an adequate DCAD (Mongin, 1981 equation) would be 150 mEq/kg diet DM.

Function

About 0.15 percent of the body is S, predominantly in the form of S AAs (S-AAs) and the amino sulfonic acid, taurine, but S is also a component of thiamin and biotin, structural compounds (e.g., chondroitin sulfate), and other biologically important molecules. Met, thiamin, and biotin cannot be synthesized by cattle; they must either be supplied in the diet or synthesized by ruminal microbes. The sulfate ion (SO₄²⁻) is found in cellular and extracellular spaces, and concentrations are likely under homeostatic control via renal clearance and perhaps other mechanisms (Markovich, 2001). S is not a major factor in acid-base balance, but the primary function of SO₄²⁻ is likely acid-base balance.

Requirement

The dietary requirement for S by the cow is primarily to provide adequate substrate to ensure maximal microbial protein synthesis, which in turn will increase the supply of the S-containing compounds required by cows. Based on in vitro and in vivo studies, maximum fiber digestibility usually occurs when diets contain 0.15 to 0.25 percent total S (Guardiola et al., 1983; Qi et al., 1994). Bouchard and Conrad (1973a,b) determined that 0.20 percent dietary S (supplemental S provided by Na, Ca, K, or MgSO₄) was adequate to sustain maximal S retention in mid-lactation dairy cows producing 30 to 37 kg milk/d. Based on the lack of any newer data, the S requirement for all classes of dairy cattle (excluding preruminant calves) remained at 0.2 percent of diet DM or

where DMI is kg/d, and S is total dietary, not absorbed.

Historically, a dietary N to S ratio of 10:1 to 12:1 has been considered optimal (Bouchard and Conrad, 1973a). However, no evidence is available indicating that the ratio is important when the dietary S requirement is met. Low-protein diets may benefit from S supplementation, but that is because dietary S probably was also low.

Sources

The S content of feedstuffs is positively correlated to the protein concentration; however, the use of S-based fertilizers can increase S concentrations of forages without a concomitant increase in protein concentrations (Spears et al., 1985; Arthington et al., 2002). S is not routinely assayed by many feed-testing labs, and table averages will likely underestimate concentrations for forages that have been fertilized with S. Most of the S in plants is in S-containing AAs, and those AAs reach the intestine to be absorbed, are used by rumen microbes to synthesize bacterial AAs, or are degraded. Within the rumen, S is released from degradation of S-AAs, but the rate of release is much greater for cysteine than for Met (Bird, 1972b). Furthermore, in vitro ruminal degradation of Met is slower than breakdown of other AAs (Mbanzamihigo et al., 1997). Theoretically, a diet with very low protein degradability could be limiting in rumen-available S, even though total dietary S is adequate; however, under practical situations, this is unlikely to occur.

Cows can consume inorganic S (usually as sulfate, SO₄²⁻) via forages that have been fertilized with S-containing fertilizers, distillers grains (Nietner et al., 2015), water (see Chapter 9), and from S supplements (e.g., MgSO₄, CaSO₄, and K₂SO₄). Based on in vitro rumen measures, in vivo S balance, and ruminant growth studies (mostly with sheep), the different sources of inorganic Sr have similar biological value (Henry and Ammerman, 1995).

Within cells and extracellular spaces of the cow, SO₄²⁻ is involved with acid-base balance and perhaps other functions. Oxidation of S-AAs within cells is a major source of SO₄²⁻, but SO₄²⁻ transporters also exist in the intestine (Markovich, 2001), and SO₄²⁻ can be absorbed by ruminants (Bird and Moir, 1971). However, much of the ingested SO₄²⁻ is probably reduced to hydrogen sulfide within the rumen.

Excess Sulfur

Excess ingested S (includes S from the diet and drinking water) causes indirect and direct negative effects on cow health and productivity. Excess intake of S can lead to deficiencies or reduced status of many trace minerals. Providing approximately 0.2 percent added S from SO₄²⁻ (total diet S at approximately 0.4 percent) reduces the absorption of copper (Cu) and Se (van Ryssen et al., 1998; Ivancic and Weiss, 2001; Richter et al., 2012), and newer data suggest it may also negatively affect manganese (Mn) and Zn retention in cattle (Pogge et al., 2014). Negative effects of dietary S probably occur at concentrations less than 0.4 percent of the diet. Increasing dietary S from 0.13 percent up to 0.35 percent by increasing dietary inclusion of distillers grains linearly decreased liver Cu concentrations in feedlot lambs (Felix et al., 2012).

Excess SO₄²⁻ added to rations can reduce feed intake and performance without eliciting any signs of clinical toxicity (Kandylis, 1984) perhaps mediated via a reduction in DCAD (discussed above). Diets with 0.2 percent added SO₄² − S reduced DMI by lactating dairy cows (Ivancic and Weiss, 2001; Tebbe et al., 2018). High concentrations of SO₄²⁻ in water can reduce water intake (see Chapter 9).

Clinical S toxicity causes neurologic changes, including blindness, coma, muscle twitches, and recumbency (Kandylis, 1984). Many of those clinical signs are consistent with polioencephalomalacia (Gould, 1998). Postmortem examination reveals severe enteritis, peritoneal effusion, and petechial hemorrhages in many organs, especially kidneys (Bird, 1972a). Often the breath will smell of hydrogen sulfide (H₂S)—which is likely the toxic principal in S toxicosis. Much of the ingested SO₄²⁻ and S from ruminally degraded S-AAs is reduced to H₂S. When dietary S concentrations are close to requirement (i.e., 0.2 percent), the H₂S is used by ruminal bacteria to synthesize S-AAs and other organic S-containing compounds (i.e., assimilatory pathway), resulting in low ruminal concentrations of H₂S. However, when higher concentrations of dietary S are fed (usually from SO₄²⁻ sources), dissimilatory reduction of SO₄²⁻ by SO₄²⁻ reducing bacteria occurs, and ruminal H₂S concentrations become elevated (see review by Drewnoski et al., 2014). The generally accepted etiology of H₂S toxicity is that at low rumen pH (pKa of H₂S is about 7), much of the H₂S produced remains as H₂S, which is volatile and can be eructated. After eructation, it can be inhaled, enter the circulation, and reach the brain, causing brain damage and polioencephalomalacia (Bird, 1972a). In beef cattle fed high-grain finishing diets without any forage, the risk of polioencephalomalacia increases greatly when total dietary S is greater than 0.42 percent (water assumed to provide trivial amounts of S), but when the diet contained 8 percent forage NDF, dietary S had to be closer to 0.6 percent to increase the risk (Nichols et al., 2012). Higher dietary fiber should increase ruminal pH, causing more of the H₂S to be dissociated (HS⁻) and not volatile. Because diets fed to dairy cows typically have substantially more forage than feedlot diets, polioencephalomalacia is not likely to be observed in dairy cattle even at very high concentrations of dietary S. NRC (2005) set the MTL of dietary S (water assumed to be a trivial source of S) at 0.3 percent for diets with 85 percent concentrate and at 0.5 percent for diets with at least 40 percent forage (more representative of dairy cow diets).

Sulfate anions have been added to rations of dry cows before calving to decrease the DCAD to help prevent milk fever (see Chapter 12), often to levels above 0.5 percent S. That concentration of S in high-forage diets should not cause clinical toxicity; however, Se and Cu absorption will likely be reduced, but because these diets are typically only fed for a few weeks, the overall effect on Se and Cu status is small. Longer-term feeding of high S diets (e.g., 0.5 percent) is not recommended.

Biochemistry and Absorption

In mammalian tissues, the primary active form of Cr is as a component of a small peptide called chromodulin. Chromodulin contains only four AA residues but can bind 4 Cr (Cr³⁺) ions and has been isolated from bovine liver (Davis and Vincent, 1997) and colostrum (Yamamoto et al., 1988). This compound is thought to bind to insulin-activated insulin receptors and stimulate tyrosine kinase, resulting in enhanced responses to insulin (Vincent, 2000). It may have other functions related to insulin activity. Cr also is likely involved in gene regulation.

Little is known regarding the absorption mechanism for Cr, but in nonruminants, absorption is usually <1 percent of Cr intake (Lukaski, 1999). Absorption of Cr from organic sources in nonruminants is greater (~3 percent of intake) than that for inorganic sources (Cefalu and Hu, 2004). Potential antagonists to Cr absorption include high concentrations of dietary Fe and perhaps Zn and phytate (Pavlata, 2007). Organic sources of Cr may also be converted more quickly within the body to biologically active forms, and organic forms usually have greater biological effects than inorganic sources (Vinson and Hsiao, 1985; Balk et al., 2007). Quantitative data on absorption of any form of Cr by ruminants are not available.

Cattle Responses to Supplemental Chromium

An extensive review of animal responses to Cr was conducted in 1997 (NRC, 1997), and another review limited to cattle was published in 1999 (Kegley and Spears, 1999). This section will concentrate on more recent publications. Studies have evaluated the effects of supplemental Cr on production measures, glucose and lipid metabolism, and immune function. Source of Cr and supplementation rate varied among studies, but most studies used chromium picolinate, chromium Met, or chromium propionate, and supplementation rates were usually 4 to 10 mg/d (approximately 0.5 mg Cr/kg diet DM when fed to lactating dairy cows). The effect of supplemental Cr on insulin sensitivity depends on the physiological state of the animal. Within 1 week or so prior to parturition, supplemental Cr often reduces insulin sensitivity in multiparous cows, as measured by an increased plasma insulin to plasma glucose ratio (Hayirli et al., 2001; Pechova et al., 2002). With primiparous animals, supplemental Cr starting 6 weeks prepartum increased insulin sensitivity when measured 2 weeks prepartum but reduced it when measured 2 weeks postpartum (Subiyatno et al., 1996). Increased plasma insulin prepartum can stimulate lipogenesis and suppress lipolysis, which can explain why supplemental Cr prepartum often reduces plasma nonesterified fatty acids (NEFAs) (Hayirli et al., 2001; Bryan et al., 2004). Elevated plasma NEFA prepartum is a risk factor for several health disorders in dairy cows (Roberts et al., 2012; McArt et al., 2013; Qu et al., 2014). Cr supplementation of growing beef and dairy animals enhances insulin sensitivity (Sumner et al., 2007; Spears et al., 2012). This should result in increased glucose uptake and increased protein synthesis by skeletal muscle. In early lactation dairy cows, supplemental Cr increased insulin responsiveness in one study (Hayirli et al., 2001), but in a limited study, supplemental Cr reduced insulin sensitivity (Subiyatno et al., 1996). Supplementation rates were similar between studies, but source of Cr differed. Supplemental Cr (approximate intake of 6 to 9 mg/d) usually increases milk yields in early lactation, higher-producing (>30 kg/d) cows (NRC, 1997; Hayirli et al., 2001a; AlSaiady et al., 2004; Smith et al., 2005; Sadri et al., 2009; Vargas-Rodriguez et al., 2014), but not in lower (ca. 30 kg/d)–producing cows (Bryan et al., 2004).

Supplemental Cr has improved certain measures of immune function in beef and dairy cattle (reviewed by Weiss and Spears, 2005). The most consistent effect has been increased blastogenesis of cytotoxic T-lymphocytes, which may be modulated via reduced cortisol concentrations. Neutrophil function has not been affected by Cr (Weiss and Spears, 2005); however, concentrations of proinflammatory cytokines in activated neutrophils were greater when cows were supplemented with Cr (Yuan et al., 2014). This may enhance overall immune response, but in a clinical trial, Cr did not affect incidence of mastitis (Chang et al., 1996).

Although production and immune function responses to supplemental Cr are often positive and Cr has been shown to have metabolic effects, establishing a requirement for Cr is not possible because total intakes of Cr (basal plus supplemental) have not been measured. Glucose and insulin often respond to Cr supplementation, but results are not linear with dose. The lowest rate of supplementation (0.01 mg Cr/kg BW for growing cattle or 0.006 mg Cr/kg BW for lactating cow) usually was adequate for maximal response (Hayirli et al., 2001; Spears et al., 2012). Whether the changes observed in insulin and glucose represent improvements in metabolism are not known. Milk yield has responded linearly and quadratically to increasing supplemental Cr (Hayirli et al., 2001; Smith et al., 2008). The maximal response occurred at a supplementation rate of approximately 0.01 mg Cr/kg BW (~6 mg/d) but in the study that reported a linear response that was the highest rate tested. Although an AI for Cr cannot be established based on only two titration studies with lactating cows, supplementation of approximately 0.01 mg Cr/kg BW often increases milk yield in early lactation.

Maximum Tolerable Level

The maximum tolerable dietary concentration for Cr³+ from soluble forms was set at 100 mg Cr/kg of diet DM (NRC, 2005), but data are very limited regarding adverse responses to high dietary concentrations of Cr³⁺. Hayirli et al. (2001) reported that milk yield responded quadratically to increasing Cr and feeding cows 0.025 mg Cr/kg BW (15 mg/d) reduced milk yields to values similar to the control. Concentrations of Cr in milk, muscle, and body fat were not greater in cows fed 2 mg Cr/kg of diet DM compared with cows fed no supplemental Cr; however, Cr concentrations in liver and kidney were two to three times greater (Lloyd et al., 2010). In vitro, Cr picolinate can increase production of the hydroxyl radical, which can negatively affect immune function, damage DNA, and oxidize membrane FAs (Vincent, 2000). The concentration of Cr at which this occurs in vivo is unknown, and it is not clear whether this effect is unique to Cr picolinate.

Function

The primary function of cobalt (Co) is to serve as a precursor for vitamin B₁₂ (cobalamin) synthesis in the rumen. Rumen microbes can usually produce adequate vitamin B₁₂ if adequate Co is available in the diet (see Chapter 8 for details). In addition to synthesizing vitamin B₁₂, bacteria can synthesize vitamin B₁₂ analogues, which are not biologically active. The presence of these vitamin B₁₂ analogues in liver and blood reduces the utility of vitamin B₁₂ determination to assess the status of dietary Co (Halpin et al., 1984). However, hepatic vitamin B₁₂ concentrations below 0.1 μg/g wet weight are indicative of Co deficiency (Smith, 1987). A portion of dietary Co can be absorbed in the cation form (Smith, 1987); however, it has no known function and, once absorbed, does not appear capable of reentering the rumen so microbes could use it. Most is excreted in the urine, and a smaller amount exits with the bile (Underwood, 1981).

Cobalt carbonate, chloride, nitrate, sulfate, and glucoheptonate all appear to be suitable sources of Co for ruminants. Cobalt oxide, which is much less soluble, resulted in much lower vitamin B₁₂ concentrations during in vitro rumen fermentation (Kawashima et al., 1997), and was somewhat less available (Henry, 1995a). Cobalt oxide pellets and controlled-release glass pellets containing Co that remain in the rumen–reticulum have been used successfully to supply Co over extended periods of time to ruminants on pasture, although regurgitation can cause loss of some types of pellets (Poole and Connolly, 1967).

Deficiency

Ruminants appear to be more sensitive to vitamin B₁₂ deficiency than nonruminants (see Chapter 8 for details). This is likely because vitamin B₁₂ is a key component in the pathways for gluconeogenesis and de novo methyl group synthesis. Ruminants are dependent on gluconeogenesis for meeting needs of tissues for glucose with ruminally derived propionate serving as a primary glucose precursor. A breakdown in propionate metabolism where methylmalonyl-CoA is converted to succinyl-CoA is a primary defect arising from a deficiency of vitamin B₁₂. Inadequate dietary Co elevates plasma homocysteine concentrations in growing beef cattle, suggesting a deficiency in methyl group availability for resynthesis of Met from homocysteine (Stangl et al., 2000). Stores of vitamin B₁₂ in the liver of adult ruminants are usually sufficient to last several months when they are placed on a Co-deficient diet. Hepatic concentrations of vitamin B₁₂, urinary and plasma concentrations of methylmalonic acid, and serum concentrations of homocysteine have been used to evaluate Co (and vitamin B₁₂) status of cattle (Stangl et al., 2000).

Young animals are more sensitive to dietary insufficiency of Co because they have lower reserves of vitamin B₁₂ in the liver. Early signs of a deficiency of Co include failure to grow, unthriftiness, and loss of weight (Smith, 1997). More severe signs include fatty degeneration of the liver, anemia with pale mucous membranes (Underwood, 1981), and reduced resistance to infection as a result of impaired neutrophil function (MacPherson et al., 1987; Paterson and MacPherson, 1990). Although cows may have adequate stores of vitamin B₁₂ to last several months, ruminal microbes do not. Within a few days of feeding a diet deficient in Co, ruminal concentrations of succinate rise. This may be the result of a blockade of microbial conversion of succinate to propionate, or a shift in ruminal bacterial populations toward succinate production rather than propionate production (Kennedy et al., 1996).

Requirement

The dietary requirement for Co was previously estimated to be 0.11 mg/kg of dietary DM. This was based on the amount of Co that must be supplied to keep plasma concentrations of vitamin B₁₂ above 0.3 μg/L (Marston, 1970). However, depending on the biochemical response criteria, adequate dietary Co (basal plus supplemental) ranged from 0.13 to 0.25 mg Co/kg diet DM for growing beef cattle (Stangl et al., 2000). Maximal growth in beef cattle occurred when the diet contained between 0.15 and 0.18 mg Co/kg DM (Schwarz et al., 2000; Tiffany et al., 2003). Liver vitamin B₁₂ and folate concentrations were maximized at 0.24 and 0.19 mg Co/kg, respectively, whereas plasma vitamin B₁₂ was maximized at a dietary Co concentration of 0.26 mg/kg (Stangl et al., 2000). The dietary Co concentrations required to minimize plasma homocysteine and methyl malonic acid concentrations were 0.16 and 0.12 mg/kg of diet DM, respectively. Milk yield responses to increasing dietary Co have not been finely titrated. Few positive production responses have been reported when Co was supplemented to dairy cows, but in those studies, the lowest concentration evaluated was approximately 0.20 mg Co/kg of diet DM (Kincaid et al., 2003; Kincaid and Socha, 2007; Akins et al., 2013). Feeds are not commonly assayed for Co (sensitivity is an issue), but in the studies above, basal concentrations averaged about 0.1 mg Co/kg DM. Because of the lack of adequate data on basal feed and the variable concentrations required for maximal growth and biochemical indicators responses, the AI of total Co (basal + supplemental) is set at 0.2 mg Co/kg of diet DM or

where DMI is kg/d.

Beef steers fed barley in high-grain diets (85 percent of diet DM) had lower concentrations of vitamin B₁₂ in rumen fluid and in plasma than steers fed a corn-based diet at similar concentrations of supplemental Co (Tiffany and Spears, 2005). Whether type of grain affects Co conversion to vitamin B₁₂ when fed in typical dairy cow diets (usually<40 percent starch grains) is unknown.

Special Properties of Cobalt

Dietary Co may also have some effects independent of its necessity for production of vitamin B₁₂. Co fed at 0.25 to 0.35 mg/kg of dietary DM, well above those required for sufficient vitamin B₁₂ synthesis, seems to enhance ruminal digestion of feedstuffs, especially lower-quality forages (Saxena and Ranjhan, 1978; Lopez-Guisa and Satter, 1992). This effect may be due to selection of certain microbial populations with a higher Co requirement or may be a result of the divalent Co cation forming crosslinks between negatively charged bacteria and negatively charged forage particles, which allows bacteria to attach to forage particles more efficiently (Lopez-Guisa and Satter, 1992). Cu, Ca²⁺, and Mg²⁺ are divalent cations that may have some of the same ability (Storry, 1961; Somers, 1983). Addition of Co has increased total anaerobic bacteria in the rumen by 50 percent and lactic acid production in the rumen by 86 percent (Young, 1979).These results suggest that ruminal microbes may require greater concentrations of dietary Co than the cow. However, the general lack of effects on DMI, milk composition, and milk yield when diets contained 0.15 to 0.20 mg Co/kg DM suggests that the AI of 0.20 mg Co/kg is also adequate for ruminal bacteria.

Toxicity

Co toxicity causes reduced feed intake, loss of BW, and eventually anemia—signs similar to those seen in Co deficiency (Ely et al., 1948; Keener et al., 1949; NRC, 2005). The maximal tolerable dietary Co concentration has been set at 25 mg/kg of dietary DM (NRC, 2005).

Function

Cu is a component of several proteins, including cytochrome c oxidase (required for aerobic respiration), lysyl oxidase (required for formation of collagen and elastin), and tyrosinase (necessary for production of melanin pigment). Cu is required for hemoglobin synthesis and is involved with iron (Fe) metabolism (e.g., as a component of ceruloplasmin). Cu, along with Zn, is a component of cytosolic superoxide dismutase, which protects cells from the toxic effects of reactive oxygen species (ROS). This is particularly important in phagocytic cells and may be a primary mode of action for reduced infectious disease when adequate Cu is fed.

Absorption of Dietary Copper

Intestinal absorption of Cu by humans and rodents appears to be under homeostatic control and is upregulated when low Cu diets are fed and downregulated with high intakes of Cu (Lönnerdal, 2008). Whether homeostatic regulation of absorption occurs in ruminants has not been determined. A weak negative relationship was observed between initial concentration of Cu in liver and rate of accumulation of liver Cu, but over a wide range of initial concentrations (60 to 230 mg Cu/kg of liver dry weight), rate of liver accumulation was essentially constant in nonlactating dairy cows (Balemi et al., 2010). That range encompasses what is considered adequate Cu status. Age of the animal, chemical form of dietary Cu, and the presence of antagonists affect intestinal absorption of Cu. In calves without functioning rumens, the Cu AC can be as high as 0.70. Bremner and Dalgarno (1973a,b) found that 50 to 60 percent of dietary Cu (supplied as Cu sulfate) was retained in liver for calves between 3 and 14 weeks of age. As the rumen starts functioning, absorption of Cu decreases substantially, and the AC for Cu is usually ≤0.05 in adult cattle.

Effect of Other Minerals on Copper Absorption

S, molybdenum (Mo) in combination with S, and Fe antagonize Cu absorption in ruminants. Zn concentrations need to be 10 to 20 times requirement before antagonism is observed in ruminants (Miller et al., 1989), so Zn antagonism is not of practical importance. Antagonism of Cu by Fe could occur via competition for intestinal binding sites (e.g., divalent metal transporter or DMT1), and Fe may exacerbate the reaction between Cu and S within the rumen (Gould and Kendall, 2011). Fe antagonism of Cu absorption usually requires supplemental Fe at concentrations exceeding 250 mg Fe/kg of diet DM (Chase et al., 2000; Mullis et al., 2003). Dietary Fe concentrations can be that high because of the Fe in forages. High Fe concentrations in forages are likely caused by soil contamination and may not greatly affect Cu absorption because the Fe is mostly in the form of ferric oxide, which is essentially inert. However, high-Fe soil has been implicated in reducing Cu status of grazing sheep (Suttle et al., 1984). Because of acid conditions in silages, over time, the Fe in silages may become more reactive (Hansen and Spears, 2009) and might cause increased antagonism. Data are not available showing lower Cu status in cattle fed silages with high concentrations of Fe.

Increased consumption of S (dietary and via drinking water) in the absence of high Mo concentrations reduces Cu status of cattle (Arthington et al., 2002; Pogge et al., 2014). The antagonism may occur because of formation of Cu sulfides within the rumen. The dietary concentration of S required to reduce Cu absorption or Cu status has not been titrated, but liver concentrations of Cu in beef cattle decreased when diets contained 0.5 to 0.6 percent total S (Arthington et al., 2002; Pogge et al., 2014). Water that contained approximately 500 mg S (as sulfate)/L also reduced liver Cu concentrations in growing beef heifers (Wright et al., 2000). Mo interacts with S and exacerbates the antagonism. Within the rumen, S and Mo can form thiomolybdates and bind soluble Cu, but thiomolybdates can be absorbed into the circulation and bind Cu compounds within the animal (Gould and Kendall, 2011). Evidence that dietary Mo in the absence of high dietary S interferes with Cu absorption is lacking (Gardner et al., 2003). Many of the negative responses observed when dietary Mo is elevated may actually be signs of molybdenosis. An equation to estimate absorption of Cu based on dietary S and Mo has been developed using data from sheep experiments (McLauchlan and Suttle, 1976):

where Cu absorption = g absorbed/g of total Cu; S = dietary S (g/kg of diet DM) and Mo = dietary Mo in mg/kg of diet DM.

The accuracy of this equation has not been evaluated with cattle and was not included in the software. However, based on that equation, Cu absorption from diets with low Mo (≤1 mg Mo/kg) and dietary S at concentrations of 0.3, 0.4, and 0.5 percent would be reduced approximately 15, 30, and 43 percent compared with a diet with 0.22 percent S. With 4 mg Mo/kg, estimated absorption would be reduced another approximately 25 percent at each of those S concentrations.

Other Factors Affecting Absorption of Copper

Soil consumption at rates that may occur during grazing reduced absorption of Cu by 50 percent in sheep (Suttle, 1975) and cattle (Dewes, 1996). The effect of grazing on Cu absorption is likely a function of stocking rate and height of the sward (lower sward and greater stocking rate will increase soil consumption) and type of soil (Grace et al., 1996). Soils with clays may have a greater negative effect on Cu absorption. The model does not include an adjustment of Cu availability for grazing cattle because of the numerous uncertainties; however, users may wish to increase intake of Cu by cattle grazing short swards on clay soils.

Breed differences exist among cattle in susceptibility to Cu toxicity. Jersey cattle fed the same diet as Holstein cattle accumulated more Cu in their livers (Du et al., 1996; Morales et al., 2000). Whether this reflects differences in feed intake, efficiency of Cu absorption, hepatic storage, or biliary excretion of Cu is not known. However, differences in abundance of a Cu transporter in intestinal cells are likely a major cause of differences in Cu metabolism among beef cattle breeds (Fry et al., 2013). Because of a lack of data on Cu absorption by different breeds, the model does not include breed effects when calculating Cu requirements or Cu supply. Requirements and supply were calculated largely using data from Holsteins; therefore, absorbed supply of Cu may be underestimated, and requirements may be overestimated for Jersey cattle.

Effect of Source of Copper

Cu is found in most common feedstuffs in the range of 4 to 15 mg Cu/kg DM, and true absorption of Cu in those feeds was set at 0.05 (Buckley, 1991), assuming total diet had <0.22 percent S and <1 mg Mo/kg. In the seventh revised edition (NRC, 2001), the AC for feedstuffs was set at 0.04. Inorganic Cu is usually supplemented in the sulfate, chloride, carbonate, or oxide forms. Several commercial products in which the Cu is chelated or associated with organic compounds (e.g., AAs or carbohydrates) are also available. Very little data on actual true absorption of Cu from these sources are available, but several studies have been conducted to evaluate relative bioavailability by comparing changes in hepatic Cu concentrations when cattle are fed different Cu sources. Other biomarkers have been used to calculate relative bioavailability, but their accuracy and sensitivity are uncertain. The AC for Cu from Cu sulfate (CuSO₄) was set at 0.05 in the previous edition (NRC, 2001), and that value was retained. The AC for other Cu supplements was based on relative bioavailability studies using liver Cu concentrations when available.

The bioavailability of Cu from dietary copper oxide (CuO) is almost zero (Langlands et al., 1989; Kegley and Spears, 1994); however, ruminal boluses containing CuO can be effective sources of Cu (Parkins et al., 1994). The long-term exposure (weeks or months) of CuO in a bolus to the ruminal environment increases its availability, whereas finely ground CuO does not stay in the rumen long enough to be solubilized. The bioavailability in Cu from copper chloride (CuCl₂) is similar to that of CuSO₄ (Ivan et al., 1990). In many studies, indicators of Cu status were not different in cattle fed CuSO₄ or proprietary Cu supplements, including Cu proteinates, Cu-AA complexes and tribasic CuCl₂ (Wittenberg et al., 1990; Ward and Spears, 1993; Du et al., 1996; Arthington et al., 2003; Spears et al., 2004; Correa et al., 2014). However, in other studies, proprietary products were significantly more available than CuSO₄ (Kincaid et al., 1986; Rabiansky et al., 1999; Hansen et al., 2008). Reasons for the inconsistent results are not clear, but relative availability of proprietary forms of supplemental Cu probably depends on the presence of antagonists (e.g., S and Mo), Cu status, and specific product being fed. For example, tribasic CuCl₂ was almost twice as effective at increasing liver Cu concentration as CuSO₄ in the presence of S and Mo, but when those Cu sources were fed to cattle in low Cu status without antagonists, they had equal bioavailability (Spears et al., 2004). Inadequate data are available to quantify factors affecting relative bioavailability of proprietary Cu supplements; therefore, the generic commercial Cu supplement in the feed library was assigned the same AC as CuSO₄. Users can modify the value based on available data for the specific product and situation.

Copper Requirements

Endogenous losses of Cu were assumed to be predominantly via bile and are expressed relative to BW. Using Cu isotopes (Buckley, 1991), biliary and urinary loss of Cu was approximately 0.0145 mg Cu/kg BW. This is more than twice the value used in the seventh revised edition (NRC, 2001). Cu content of growing tissues, when the liver is included as part of the carcass, is 2 to 2.5 mg Cu/kg based primarily on studies of sheep and beef cattle (Grace, 1983; Miranda et al., 2006; García-Vaquero et al., 2011). Because of concerns with the consumption of excessive Cu, the growth requirement was set at 2.0 mg Cu/kg live weight change. Cu content of milk is about 0.04 mg/kg when diets contain typical concentrations of Cu (Castillo et al., 2013; Faulkner et al., 2017), but it can be as high as 0.2 mg Cu/kg when animals are fed a high Cu diet (Schwarz and Kirchgessner, 1978). The requirement for absorbed Cu for lactation was set at 0.04 mg Cu/kg milk produced. This is about a 70 percent decrease compared to the 0.15 mg Cu/kg milk produced used by the seventh revised edition (NRC, 2001). The conceptus at 190 days of gestation of Holstein cows (average BW = 715 kg) contained approximately 20 mg Cu (House and Bell, 1993). Because data are lacking, the committee assumed no Cu accumulated during the first trimester, and accumulation was linear between 90 and 190 days of gestation, so that gestation requirement for absorbed Cu was (0.2 mg Cu accumulated/d from 90 to 190 days of gestation or 0.3 μg Cu/kg maternal BW/d). From 190 days of gestation until calving, Cu accumulation in conceptus was 1.6 mg/d or 2.3 μg Cu/kg of maternal BW/d (House and Bell, 1993).

For an average lactating Holstein cow (35 kg of milk, 650 kg BW, 150 days pregnant), the total requirement for absorbed Cu is 11.0 mg/d compared with 11.4 mg obtained using the seventh revised edition (NRC, 2001). Assuming an intake of 23 kg and a dietary AC of 0.045, dietary Cu concentration of about 11 mg/kg would meet her requirement. Requirements for a 700-kg nonlactating cow at 260 days of gestation is 11.7 mg/d (approximately 20 mg dietary Cu/kg assuming an intake of 13.5 kg) compared with 7 mg/d calculating using the seventh revised edition (NRC, 2001).

Signs of Copper Deficiency

Clinical signs of Cu deficiency are generally nonspecific (i.e., reduced growth rate, ill-thrift, increased prevalence of disease, reduced reproductive efficiency), but Cu deficiency can result in a loss of hair pigment or loss of hair. Diarrhea can also occur with Cu deficiency, but that may be related to excess Mo. Anemia (hypochromic macrocytic), fragile bones and osteoporosis, and cardiac failure also are observed in Cu deficiency (Underwood, 1981). Inadequate supply of Cu reduces the killing ability of phagocytic cells of cattle, but responses to supplemental Cu on cellular and humoral immunity have been inconsistent (Weiss and Spears, 2005). Impaired immune function is likely the reason cows fed inadequate Cu have more severe mastitis than those fed adequate Cu (Scaletti et al., 2003). The control diets in essentially all studies that showed improved immune function and reduced infectious disease would not have met current Cu requirements.

Assessing Copper Adequacy

Dietary concentrations of Cu are of limited value in assessing adequacy of Cu supply because of variable dietary and water concentrations of antagonists (discussed above). Plasma concentrations of Cu <0.5 mg/L are generally considered indicative of clinical Cu deficiency. However, other than confirming overt Cu deficiency, plasma concentrations are not useful in assessing status, including situations with excessive stores of Cu in the liver (López-Alonso et al., 2006). Activities of Cu-containing proteins (e.g., ceruloplasmin or superoxide dismutase) are generally considered unreliable biomarkers of Cu status in cattle (Mulryan and Mason, 1992; López-Alonso et al., 2006; Hepburn et al., 2009). Concentrations of the Cu chaperon protein may have potential as a marker of Cu status, but additional research is needed (Hepburn et al., 2009). Concentrations of Cu in liver are the standard for assessing Cu status in cattle, although recommended reference values vary. Concentrations of Cu in liver <10 mg/kg on a DM basis are generally considered indicative of impending clinical deficiency, and values less than about 35 mg Cu/kg DM are generally considered suboptimal (Smart et al., 1992; Underwood and Suttle, 1999). However, in the presence of high dietary Mo and S, which promote formation and absorption of thiomolybdates into the blood, Cu in liver may not accurately reflect Cu status (Suttle, 1991). The concentration of Cu in liver indicative of excess has not been definitively identified, but field reports indicate Cu toxicity occurs with liver concentration of 300 to 350 mg Cu/kg dry weight (Auza et al., 1999; Grace and Knowles, 2015).

Copper Toxicity

Cu toxicosis can occur in cattle that consume excessive amounts of supplemental Cu or feeds that have been contaminated with Cu compounds used for other agricultural or industrial purposes (Underwood and Suttle, 1999). When cattle consume excessive Cu, they accumulate large amounts of Cu in the liver before toxicosis becomes evident. Stress or other factors may result in the sudden liberation of large amounts of Cu from the liver to the blood, causing a hemolytic crisis. Such crises are characterized by considerable hemolysis, jaundice, methemoglobinemia, hemoglobinuria, generalized icterus, widespread necrosis, and often death (Steffen et al., 1997; Underwood and Suttle, 1999; Johnston et al., 2014). Because of antagonists (e.g., S and Mo) and because Cu continues to accumulate in the liver when excess Cu is fed (Balemi et al., 2010), defining a dietary concentration that will result in toxicity is not possible. NRC (2005) set the MTL for dietary Cu for cattle at 40 mg Cu/kg DM. However, growth rate and feed conversion were reduced in beef cattle with liver Cu concentrations of 290 mg/kg dry weight and fed diets with 20 mg supplemental Cu/kg DM for a total dietary concentration of 30 mg Cu/kg (Engle and Spears, 2000), and fiber digestibility by dairy cows was reduced when CuSO₄ was supplemented to increase total dietary Cu to 20 mg/kg (Faulkner and Weiss, 2017).

Function

The sole role of iodine (I) as a required nutrient is for the synthesis of the thyroid hormones thyroxine (T4) and triiodothyronine (T3) that regulate energy metabolism. The amount of I incorporated into thyroid hormones was about 0.25 mg/d in calves weighing 45 kg and increased to 1.4 mg I/d in nonpregnant heifers weighing 400 kg (Mixner et al., 1966). Late-gestation cows incorporate about 1.5 mg I/d into thyroid hormone (Miller et al., 1988). Thyroid hormone production increases during lactation, especially in high-producing cows, and I incorporation into thyroid hormones may reach 4 to 4.5 mg I/d (Sorensen, 1962). Thyroid hormone production also is increased during cold weather to stimulate an increase in basal metabolic rate (Goodman and Middlesworth, 1980). About 80 to 90 percent of dietary I is absorbed, and most of the I not taken up by the thyroid gland is excreted in urine and milk (Miller et al., 1988). The I content of milk is a reasonable indicator of I status because it increases as dietary I intake increases (Berg et al., 1988). The availability of assays for thyroxine and thyroxine-stimulating hormone (TSH) might provide an accurate assessment of thyroid function and the causes of thyroid dysfunction. Alternatively, blood TSH concentrations might be used as a biomarker for thyroid function to reevaluate the minimum I requirements in dairy cattle.

When the I content of the diet is adequate or excessive, less than 20 percent of the dietary I will be incorporated into the thyroid gland (Sorensen, 1962). Under conditions where intake of dietary I is marginal, the thyroid gland will incorporate about 30 percent of the dietary I into thyroid hormones (Miller et al., 1975). When severely I deficient, the hyperplastic thyroid can bind up to 65 percent of the I consumed by the cow (Lengemann and Swanson, 1957).

Adequate Intakes

I requirements in previous reports were based on a limited number of thyroxine production rates measured in cattle during the 1960s and 1970s. Due to the time that has elapsed since those studies were conducted and the limited number of measurements, the committee concluded that there were insufficient data to determine an estimated average requirement for I. Therefore, AIs, rather than requirements, were determined.

Total daily thyroxine secretion rate (TSR) increased at a decreasing rate in growing heifers ages 77 to 686 days (Mixner et al., 1962) and varied from 0.008 to 0.0030 mg/kg BW. In lactating cows, the mean thyroxine secretion rate was about 0.30 mg/100 kg BW (Mixner et al., 1962; Miller et al., 1975). The previous report listed separate requirements for maintenance, pregnancy, and lactation. It was suggested that the thyroxine secretion rate was 2.5-fold greater in lactating cows than in dry cows (Sorenson et al., 1962), yet limited evidence (Swanson et al., 1972) showed only a 25 percent increase from late pregnancy to lactation. The primary determining factor for TSR was BW. Therefore, the AI for maintenance for all groups of animals is based on TSR related to BW. In a summary of the studies of Swanson et al. (1972) and Mixner et al. (1962), TSR, mg/d = 0.0653 × BW^0.528 (R² = 0.96). Thyroxine (T4) contains 66 percent I and Miller et al. (1975) suggested that under conditions where I was not limiting, 20 percent of dietary I was used by the thyroid gland to synthesize thyroxine. Therefore, the maintenance AI for I can be predicted by the following equation: I, mg/d = 0.216 × BW^0.528.

In addition to thyroxine synthesis, I is also secreted in milk (discussed below), and for lactating cows, this loss must be replaced by dietary intake. At low I intake, milk contains about 0.05 mg I/L, and transfer of dietary I to milk is about 0.5 (Swanson et al., 1972). Therefore, the AI for lactation was set at 0.1 mg/L of milk. The equation for total AI for all classes of cattle except calves is

where BW is kg and milk is kg/d.

The AI for I for nonruminating calves was based on the AI established for human infants and was set at 0.8 mg I/kg DMI. Once calves are ruminating, Equation 7-34 should be used.

This amount of I will likely not be adequate when diets contain goitrogenic feeds such as canola meal (Pappas et al., 1979). Diets with canola meal decreased transfer of I into milk by 50 percent (discussed below). Assuming thyroxine synthesis is affected similarly, the AI for animals fed diets with goitrogenic feeds would be twice that calculated above. However, because of limited data, this adjustment is not included in the model, but users should consider adjusting supplementation when goitrogens are fed. Based on typical DMI, a dry cow (700 kg BW) and an average lactating cow (650 kg BW and 35 kg/d of milk) would need to be fed diets with 0.51 and 0.48 mg I/kg DM when diets did not contain goitrogens and 1.02 and 0.96 mg I/kg DM with goitrogenic diets. Because of human health concerns related to excess I intake and the fact that milk I concentrations increase with increased I concentration in the cow's diet, supplemental I should not exceed amounts deemed as AI for the cow (see toxicity section).

Factors Affecting Iodine Needs

Goitrogens are compounds that interfere with the synthesis or secretion of thyroid hormones and cause hypothyroidism. Goitrogens fall into two main categories: (1) cyanogenic goitrogens impair iodide (I⁻¹) uptake by the thyroid gland. Cyanogenic glucosides can be found in many feeds, including raw soybeans, beet pulp, corn, sweet potato, white clover, and millet, and once ingested are metabolized to thiocyanate and isothiocyanate. These compounds alter I⁻¹ transport across the thyroid cell membrane, reducing I retention; (2) Progoitrins and goitrins found in cruciferous plants (rape, kale, cabbage, turnips, and mustard) and aliphatic disulfides found in onions that inhibit thyroperoxidase prevent formation of mono- and diiodotyrosine (Ermans and Bourdoux, 1989). With goitrins, especially those of the thiouracil type, hormone synthesis may not be restored by dietary I supplementation, and the offending feedstuff needs to be reduced or removed from the diet.

Canola meal derived from low erucic acid varieties of rapeseed contains glucosinolates that can be converted into thiocyanate during seed processing. Recent studies (Franke et al., 2009a,b; Weiss et al., 2015a) demonstrated that when canola meal substitutes for soybean meal in diets, milk I concentrations are reduced. While milk I continued to increase with increasing dietary I concentration (up to 5 mg/kg diet DM) when diets contained canola meal, the rate of increase in milk I was reduced by 50 percent or more depending on the amount of canola meal fed (Franke et al., 2009a; Weiss et al., 2015a). However, blood serum I concentrations were similar to controls (Weiss et al., 2015a), and urinary I excretion was increased, suggesting I absorption is not impaired (Franke et al., 2009b). The negative effects of dietary goitrogens can be overcome by increasing the concentration of dietary I or removal of the feeds containing goitrogens. Diets that contain goitrogenic ingredients may need more I than the recommended AI.

Supplemental I in the form of ethylenediamine dihydriodide (EDDI) has been used to decrease foot rot in beef cattle (Maas et al., 1984) and more recently has been suggested as a possible treatment for digital dermatitis in dairy cattle (Gomez et al., 2014). Dietary EDDI may also have value in treating ringworm in young cattle (Cam et al., 2007). However, concerns about excessive milk I concentrations with supplemental I would prevent the use of EDDI for these purposes in lactating dairy cows.

Sources of Iodine

Most sources of I are readily absorbable, and the iodides of Na, K, and Ca are commonly used. Potassium iodide is easily oxidized and volatilizes before the animal can ingest it. Pentacalcium orthoperiodate and EDDI are more stable and less soluble and are commonly used in mineral blocks and salt licks exposed to the weather. Concentrations of I in forage are variable and dependent on the I content of the soil. Soils near the oceans tend to provide adequate I in plants. However, in the Great Lakes regions and Northwest United States, I concentrations in forages are generally low enough to result in a deficiency of I unless I is supplemented.

Milk Iodine

Typical range in milk I concentration is 100 to 300 μg/L (Flachowsky et al., 2014), and I in dairy products is readily absorbed by humans. Milk and dairy products are major sources of I intake in the United States and Europe, where dairy products make up a significant portion of the average diet. Dairy products may account for 30 to 74 percent of I intake in the United States (Murray et al., 2008). The recommended dietary allowance for I intake ranges from 70 μg/d in infants to 290 μg/d in pregnant and nursing women (Swanson et al., 2012). However, the upper tolerable limit for I for humans is only 2.2- to 3.5-fold greater than recommended intake. A suggested limit for milk is 500 μg I/kg. While there have been concerns about excess I intake from milk, more recent evidence based on urinary I excretion suggests that a significant portion of individuals with high I requirements (pregnant and nursing women) may not be consuming adequate I. In part, this is due to the reduced consumption of iodized table salt in the United States. Milk and dairy products will continue to make an important contribution to I intake in the general population.

Flachowsky et al. (2014) reviewed factors that affected milk I concentration. By far, dietary concentration of I is the primary factor that affects milk I concentration. Berg et al. (1988) demonstrated a linear increase in milk I excretion with increasing supplementation of I in the form of EDDI with 26 to 39 percent of the supplemental I excreted in the milk. Other factors such as I source, I antagonists (discussed above), farm management practices including teat dipping with I-containing substances, and the use of I sanitizers in milk processing can affect milk I concentrations (Flachowsky et al., 2014). I present in iodophor-containing teat dips and udder disinfectants were shown to be absorbed through the teat skin and markedly increased milk I concentrations. However, use of iodophors was discontinued in dairy teat dips and disinfectants in the late 1980s such that milk I concentrations have gradually declined (Flachowsky et al., 2014).

Deficiency Symptoms

I deficiency reduces production of thyroid hormones, slowing the rate of oxidation of all cells. Often the first indication of I deficiency is enlargement of the thyroid (goiter) of newborn calves (Miller et al., 1968). Calves also may be born hairless, weak, or dead. Fetal death can occur at any stage of gestation, but often the cows will appear normal (Hemken, 1970). In adult cattle, I deficiency can cause enlarged thyroid glands, reduced fertility (males and females), and increased morbidity. Under conditions of marginal or deficient dietary I, the maternal thyroid gland becomes extremely efficient in removing I from the plasma and recovering I during the degradation of spent thyroid hormone and thyroglobulin. Unfortunately, this leaves little I for the fetal thyroid gland, and the fetus becomes hypothyroid. The goiter condition is the hyperplastic response of the thyroid gland to increased stimulation of thyroid growth by thyroid-stimulating hormone produced in the pituitary gland. Under mild I deficiency, the hyperplastic thyroid gland can compensate for the reduced absorption of I (Hetzel and Welby, 1997).

Toxicity

The maximum tolerable limit was set at 50 mg I/kg of diet DM (NRC, 2005). Clinical signs of I toxicity in ruminants include excessive nasal and ocular discharge, hyperthermia, salivation, decreased milk production, coughing, and dry, scaly coats (Paulikova et al., 2002). As discussed earlier, high concentrations of dietary I in the diet increase I concentrations in milk, and because humans are much more sensitive to I toxicity than cows, the danger of excess dietary I fed to cattle is a public health issue (Hetzel and Welby, 1997). Current U.S. Food and Drug Administration (FDA) regulations set the maximum limit of I supplementation in cattle from EDDI at 50 mg/d, which in a lactating cow consuming 25 kg of DM/d would correspond to a maximum concentration of 2.0 mg I/kg DM, roughly three times the amount considered to be needed.

Functions and Measures of Adequacy

Fe has a multitude of functions within the body, including oxygen transport (component of heme found in hemoglobin and myoglobin), electron transport (e.g., ferredoxins and cytochrome P-450 enzymes), immunity (e.g., myeloperoxidase and catalase), energy metabolism (e.g., aconitase), and gene regulations (Beard, 2001; Templeton and Liu, 2003). Fe deficiency results in hypochromic microcytic anemia due to failure to produce hemoglobin. Light-colored veal is due to low muscle myoglobin as a result of restricted dietary Fe. In young dairy cattle (<24 months old), serum ferritin had a stronger correlation with concentrations of Fe in liver and spleen than did hemoglobin (Miyata and Furugouri, 1987). However, over a lactation cycle, serum ferritin did not appear to reflect changes in Fe stores in dairy cows (Furugouri et al., 1982). Other measures of Fe status (e.g., Fe binding capacity of serum, plasma Fe) can change in response to factors that are independent of Fe status such as inflammation (Baydar and Dabak, 2014) and parturition (Miltenburg et al., 1991).

Because Fe deficiency is very rare in adult cattle, data evaluating the accuracy and sensitivity of Fe status indicators are not available. Common measures of Fe status did not change over the dry and early lactation (up to 60 days in milk) periods and were not affected when dry and lactating cows were fed 30 mg/kg of supplemental Fe from and Fe-AA complex (Weiss et al., 2010). Fe supplementation did not affect any production measure other than a small, but statistically significant, decrease in milk SCC. In Fe-deficient calves, morbidity and mortality associated with depressed immune responses occurred prior to any observed change in packed cell volume (Mollerberg and Moreno-Lopez, 1975). In lactating goats, hemoglobin concentrations were negatively related to SCC (Atroshi et al., 1986). Beard (2001) suggested that changes in functional measurements such as immune measures and performance during exercise in humans occurred prior to changes in hemoglobin and other measures of Fe status.

Requirement

Because of the uncertainty regarding the true absorption of Fe from feeds and the amount of Fe needed for maintenance, the factorial approach was used to generate an AI. During tissue and protein turnover, the majority of Fe is effectively recovered and recycled so that maintenance requirements for Fe are negligible. Milk contains about 1 mg Fe/kg (Aleixo and Nóbrega, 2003; Schnell et al., 2015). The absorbed Fe requirement of the conceptus of the pregnant cow between day 190 of gestation and the day of calving has been estimated to be 0.025 mg Fe/kg maternal BW or 18 mg/d for a 715-kg late-gestation Holstein cow (House and Bell, 1993). Increased vascularization of the uterus and other reproductive organs and fetal hematopoiesis would increase the gestation requirement for Fe earlier than 190 days of gestation, but data are unavailable to quantify this. Estimates of Fe content in the body range from 18 to 34 mg Fe/kg BW of calves (Bremner and Dalgarno, 1973c). The absorbed Fe requirement for growth of cattle has been set at 34 mg Fe/kg ADG.

Absorption

Fe in the ferric form (Fe³⁺) is poorly absorbed from the intestinal tract; however, reductases exist on the surface of enterocytes that convert ferric Fe to ferrous (Fe²⁺), allowing for absorption. Fe-deficient calves absorbed Fe provided by ferric chloride but not from ferric oxide (Ammerman et al., 1967). Some Fe³⁺ can be reduced to Fe²⁺ on reaction with the acid of the abomasum (Wollenberg and Rummel, 1987). The concentration of soluble Fe in corn silage also increased markedly following in vitro simulated abomasal exposure (Hansen and Spears, 2009).

Form of dietary Fe affects absorption, but Fe absorption by animals is a tightly regulated process, and Fe status interacts with form of Fe to ultimately determine the amount of Fe absorbed. Absorption of radioactive Fe from ferric chloride was three to five times greater by Fe-depleted calves than by Fe-sufficient calves (Ammerman et al., 1967). Calves fed diets with 750 mg Fe (from ferrous sulfate)/kg DM downregulated expression of a duodenal Fe importer (DMT1) found on the luminal side of the enterocyte and ferroportin, an exporter of Fe found on the basal-lateral membrane of enterocytes (Hansen et al., 2010a). The DMT1 may also be involved in absorption of Cu and Mn, which could be a reason high Fe can antagonize absorption of those two metals. Based on changes in expression of many proteins involved with Fe metabolism and a decrease in plasma Fe concentrations (Hansen et al., 2010b), a diet that is deficient in Cu but extremely excess in Mn may induce Fe deficiency in ruminants. In nonruminants, dietary Zn (Lind et al., 2003), phytate (Gillooly et al., 1983), phosphate, and Ca (Monsen and Cook, 1976) can reduce Fe absorption, and dietary βcarotene, vitamin A, and vitamin C can increase Fe absorption (García-Casal et al., 1998). Elevated dietary P reduces hepatic Fe concentrations in steers (Standish et al., 1971), but whether the other factors listed above affect Fe absorption in ruminants is unknown.

Assigning ACs for Fe to basal ingredients and supplements is plagued by a lack of data and because absorption is dependent on Fe supply and Fe status of the animal. Maximum absorption efficiency occurs when animals are deficient in Fe, which is very rare in adult ruminants. Maximum absorption can occur in calves because diets are inherently low in Fe, and deficiency is possible if animals are not supplemented. Absorption of Fe by deficient calves fed low Fe liquid-based diets ranges from 0.55 to 0.72 (Matrone et al., 1957; Bremner and Dalgarno, 1973a,b; Miltenburg et al., 1993). An analysis of balance studies done in growing calves suggests that the AC of soluble Fe from liquid diets declined from 0.40 to 0.15 as dietary Fe increased from 40 to 100 mg/kg (ARC, 1980).

Once the animal is ruminating, the efficiency of Fe absorption is considerably lower, in part because diets are generally excess in Fe and because much of the dietary Fe is provided by forages. Mechanically harvested forages can be high in Fe because soil contamination occurs during harvest. Much of that Fe is likely ferric oxide, which, based on data from calves and sheep, is essentially unavailable (Ammerman et al., 1967; van Ravenswaay et al., 2001). Based on in vitro tests and solubility, Fe from soil contamination has low bioavailability, but after soil-contaminated forage is stored as silage (pH ~4), solubility of Fe was about 20 times greater than in preensiled samples (Hansen and Spears, 2009). Whether that results in increased absorption of Fe is unknown.

Studies using radioactive Fe determined that Fe absorption efficiency was less than 0.02 in adult cattle fed a diet that supplied much more Fe than was needed (van Bruwaene et al., 1984). When pregnant ewes were fed diets that contained 20 mg Fe/kg diet (much lower than most practical diets fed dairy cows), absorption of dietary Fe was 0.21 (Hoskins and Hansard, 1964). Because data are not available to contradict the value used previously (NRC, 2001) for Fe absorption from basal ingredients (0.10), that value was retained. However, because of the typically high concentrations of Fe in most dairy diets and because much of the Fe is likely from soil contamination, an AC less than 0.10 is probably more accurate. Based on limited data from calves and sheep, ferrous sulfate has the greatest relative bioavailability, followed by ferrous carbonate (Ammerman et al., 1967; van Ravenswaay et al., 2001). In NRC (2001), Fe supplements were given ACs of 0.4 to 0.6, which may be appropriate for preruminant cattle but are likely much too high for adult ruminants fed practical diets. No data are available on actual absorption of Fe by ruminants; ferrous sulfate was assumed to have an absorption of 0.20 based on data from Hoskins and Hansard (1964), and based on relative availability, ferrous carbonate was assigned an absorption value of 0.10. Many mineral supplements are contaminated with low concentrations of Fe, and because of a total lack of data, the Fe was assumed to be mostly Fe oxide and given an AC of 0.01.

A 650-kg cow producing 25 kg of milk/d at 205 days of gestation and consuming 20 kg/d DM needs to absorb only 41 mg Fe/d or be fed a diet (AC = 0.1) with just 20 mg Fe/kg DM. Most feedstuffs will contain adequate Fe to meet the Fe requirements of adult cattle. Milk-fed calves are the only group of cattle that ordinarily require Fe supplementation. Feeding veal calves <15 weeks of age as little as 39 mg Fe/kg DM will allow calves to grow at a normal rate, but the muscles remain pale and the animals remain slightly anemic (Bernier et al., 1984). A study by (Lindt and Blum, 1994) suggests that a 50-mg Fe/kg diet is adequate to maintain physiological function in growing veal calves.

Toxicity

NRC (2005) set the MTL for cattle at 500 mg Fe/kg of diet DM; however, that value is dependent on source. The MTL assumes the Fe is from a readily available source; therefore, if the Fe is predominantly from forage, diets in excess of that value likely will cause little problem. However, when readily available sources of Fe are fed, the MTL established by NRC (2005) may be too high. Milk yield and DMI decreased linearly when cows were fed diets with 0, 250, or 400 mg supplemental Fe from Fe sulfate (McCaughey et al., 2005). Diets supplemented with 500 mg Fe (from Fe sulfate) decreases measures of Cu status (Chase et al., 2000). Excess Fe can lead to oxidative stress by increasing the generation of ROS and reactive N species (Meneghini, 1997). High Fe (from Fe sulfate) reduced in vitro ruminal digestibility of DM (Harrison et al., 1992). Fe in water can affect water intake and increase measures of oxidative stress in dairy cows (see Chapter 9).

Function and Measures of Adequacy

Mn is a cofactor in a host of enzymes and other proteins that are needed for normal metabolism of AAs, carbohydrates, and lipids and is required by every organ system in the body. Mn-dependent transferases are vital for cartilage production and bone development, and a common sign of Mn deficiency is skeletal abnormalities. Clinical signs of these abnormalities include enlarged joints, shortened and weak bones, superior brachygnathism, ataxia, and deafness and equilibrium problems caused by improper development of bones in the middle ear. The skeletal system of the gestating fetus is especially sensitive to Mn deficiencies, and the newborn calf can show clinical signs of deficiency while its dam appears normal (Hansen et al., 2006a). The majority of calves born from beef cows (Rojas et al., 1965) and heifers (Hansen et al., 2006a) fed diets with approximately 16 mg Mn/kg diet DM for the entire gestation had bone deformity signs indicative of a Mn deficiency, whereas the dams appeared clinically normal.

Reproductive efficiency can be reduced when cows are fed diets with inadequate Mn; however, the concentration of dietary Mn at which this occurs is poorly defined (Hidiroglou, 1979). In an old study (Bentley and Phillips, 1951), heifers fed diets with approximately 10 mg Mn/kg had poor fertility, and the calves that were born had bone disorders. When the diet contained 30 to 40 mg Mn/kg DM, fertility improved and no bone disorders were noted. More recently, beef heifers fed a basal diet with 16 mg Mn/kg had similar reproductive measures as did heifers fed diets with 26 mg Mn/kg; however, heifers fed diets with 56 to 66 mg Mn/kg had a substantial but statistically insignificant increase in pregnancy rate compared with heifers fed 16 or 26 mg Mn/kg (Hansen et al., 2006b).

At the current time, no sensitive indicators of Mn status have been identified for cattle. Manganese superoxide dismutase (Mn-SOD) is in the mitochondria and works in concert with other antioxidants to minimize accumulation of ROS. In some species, Mn-SOD activity responds to changes in Mn intake (de Rosa et al., 1980). Data are lacking relating Mn intake to Mn-SOD activity in cattle, but steers injected with a mix of trace minerals including Mn had higher Mn-SOD activity in red blood cells than steers injected with saline (Genther and Hansen, 2014). In humans, Mn-SOD activity of lymphocytes responded to changes in Mn intake (Davis and Greger, 1992), and in rats, Mn intake was related to several measures of immune function (Son et al., 2007). Arginase is a Mn-dependent enzyme in the urea cycle, and rats fed Mn-deficient diets had reduced arginase activity, which resulted in higher concentrations of plasma ammonia and lower concentrations of plasma urea (Brock et al., 1994). Plasma urea concentrations in cattle have not been shown to respond to changes in dietary Mn. Arginase activity also affects concentrations of nitric oxide in certain tissues, which could explain some of the effects Mn has on immune function (Chang et al., 1998). Concentrations of Mn in plasma or whole blood did not differ between cattle fed diets with wide ranges in Mn concentrations (Weiss and Socha, 2005; Hansen et al., 2006a,b), indicating that it is not a good indicator of status in clinically normal cattle. Newborn calves that were displaying clinical signs of Mn deficiency had lower whole-blood Mn than calves born from heifers fed adequate Mn during gestation (Hansen et al., 2006a). In beef cattle, the concentration of Mn in liver responded linearly to increasing dietary concentrations of Mn (16 to 66 mg/kg), but liver Mn was only about 10 percent greater (9.4 versus 8.2 mg Mn/kg of liver DM), whereas intake of Mn varied more than 4-fold (Hansen et al., 2006b). The lack of sensitive status indicators and established reference ranges for liver concentrations make quantifying the Mn requirement difficult.

Manganese Absorption and Homeostasis

Intestinal absorption of Mn is thought to occur mainly via DMT1, a protein that is also involved with Cu, Zn, and especially Fe absorption (Garrick et al., 2003). Expression of DMT1 in the duodenum of humans experiencing an Fe deficiency is increased (Zoller et al., 2001), and it is downregulated in calves fed high Fe (Hansen et al., 2010a). Whether Mn status affects expression is unknown. Although data with cattle are lacking, high Co, Fe, and Cu can inhibit uptake of Mn in other species (Garrick et al., 2003). In humans, high Ca reduced absorption of Mn, but P and Mg did not (Davidsson et al., 1991). In chickens, high Ca had no effect on Mn absorption, whereas phosphate significantly reduced Mn absorption (Wedekind et al., 1991). Whether any of these antagonisms of Mn absorption occur in cattle are unknown. Elevated dietary S reduced apparent absorption of Mn by steers (Pogge et al., 2014).

Measuring the true absorption of Mn is especially difficult because the majority of dietary Mn that is absorbed is removed from the portal circulation by the liver and is excreted back into the intestine via bile. Using radioactive Mn (MnCl₂), Mn absorption by humans ranged from 1.0 to 2.5 percent (Davidsson et al., 1991; Finley et al., 1994). In rats, Mn (MnCl₂) absorption averaged 1.8 percent (Chua and Morgan, 1997). In a very limited study (two cows), absorption of radioactive Mn (from MnCl₂) averaged about 0.5 percent (Vagg, 1976), and in another limited study (two cows) based on portal vein data, Sansom et al. (1978) calculated an absorption rate of 0.5 percent for MnCl₂. Across studies, the AC in cattle was about one-fourth the value measured in nonruminants. That may be a true species difference or it could be caused by differences in intake of Mn. Absorption and turnover of Mn are positively correlated (Britton and Cotzias, 1966). When Mn intake increases, more Mn is absorbed, but biliary excretion also increases so that very little net change in body Mn content occurs.

Data are not available regarding absorption of Mn from basal feeds or from inorganic sources other than MnCl₂. In the previous edition (NRC, 2001), the AC for Mn from the Cl− and SO₄⁻² salts was set at 0.01 and 0.0075 for all basal feedstuffs. The same relative differences between supplemental Mn sources and basal feedstuffs were retained in this edition; however, AC of Mn from MnCl₂ was set at 0.005 based on the limited cow data that are available. Manganese sulfate was assigned the same AC value (0.005), and basal ingredients were assigned an AC of 0.004 (i.e., 0.005 × 0.75). The ACs of other sources of supplemental Mn were set at Mn carbonate (0.0015) and Mn monoxide (0.003) based on relative bioavailability studies (Henry, 1995c).

Manganese Requirements

An AI is used for Mn because inadequate data are available to establish a requirement. The amount of Mn needed by a dairy cow for maintenance has not been experimentally measured. Approximate intakes of 250 to 300 mg total Mn by gestating cattle were not adequate to prevent deficiency signs in their offspring (Hansen et al., 2006a; de Carvalho et al., 2010). Assuming an AC for dietary Mn of 0.004 and that approximately 50 mg Mn was needed for gestation (see below), the maintenance requirement for absorbed Mn must be greater than 0.0016 to 0.002 mg Mn/kg BW. Weiss and Socha (2005) determined an intake of approximately 580 mg Mn/d was necessary to maintain lactating dairy cows in zero Mn balance. In that study, average secretion of Mn in milk was 0.6 mg/d, which equals 150 mg dietary Mn (assumed AC = 0.004). Subtracting that value from 580 mg yields an estimated maintenance requirement of 430 mg for a 650-kg cow or 0.0026 mg absorbed Mn/kg BW. Because of a lack of other data, that value was set as the maintenance AI for absorbed Mn. This is 30 percent higher than the requirement estimated in the previous version (i.e., 0.002 mg Mn/kg BW).

House and Bell (1993) determined that 0.3 mg Mn/d was deposited into the fetus from 190 days of gestation until parturition for Holstein cows weighing 715 kg. Therefore, the daily gestation requirement (starting at 190 days of gestation) for absorbed Mn was set at 0.00042 mg/kg BW of the dam. The lactation requirement is equal to the amount of Mn secreted in milk daily. The concentration in milk ranges from about 0.016 to 0.050 mg Mn/kg (Gunshin et al., 1985; Erdogan et al., 2004; Pechova et al., 2008; Castillo et al., 2013). The weighted average was 0.027, which is essentially equal to 0.03 mg Mn/kg, the value used in the previous edition (NRC, 2001). Therefore, the lactation requirement for absorbed Mn was set at 0.03 mg Mn/kg of milk. The concentration of Mn in carcasses of calves averages about 2.5 mg/kg of total carcass on a DM basis (Suttle, 1979). Assuming the carcasses used in these experiments were 27 percent DM, the Mn requirement for growth can be estimated to be 0.7 mg Mn/kg BW gain. No new data were found on tissue accretion of Mn or whole-body Mn concentrations in cattle; therefore, the growth requirement was not changed.

The AI for absorbed Mn for a late-gestation Holstein cow weighing 700 kg would be 1.8 mg/d for maintenance plus 0.3 mg/d for gestation for a total of 2.1 mg absorbed Mn. This is about 490 mg total dietary Mn (assumed AC = 0.0042). Assuming a DMI of 12.5 kg/d, a concentration of 40 mg total Mn/kg of diet DM would meet the requirement. This is substantially greater than in the previous version. For a 650-kg cow producing 45 kg of milk/d, the maintenance requirement would be 1.7 mg and 1.35 mg for lactation for a total absorbed Mn requirement of 3.1 mg/d. Assuming a DMI of 26 kg/d, dietary concentration of total Mn to meet the requirement would be approximately 27 mg Mn/kg (assumed AC of 0.0042).

Maximum Tolerable Level

Because absorption of dietary Mn is extremely low, Mn toxicity in ruminants is unlikely to occur, and the few documented incidences with adverse effects are limited to reduced feed intake and growth (Jenkins and Hidiroglou, 1991). The maximum tolerable amount of Mn as given by NRC (2005) is 2,000 mg Mn/kg of diet DM. However, diets with 500 mg Mn/kg exacerbated the negative effects of feeding a Cu-deficient diet (Hansen et al., 2009). Furthermore, cattle fed a Cu-deficient diet with 500 mg Mn/kg DM displayed some indications of an Fe deficiency (Hansen et al., 2010b).

Functions and Animal Response

The only known nutritional function of Se is as a component of specific selenoproteins. Essentially, any protein can contain Se because of nonspecific incorporation of selenomethionine (replacing Met) into the polypeptide chain. However, selenoproteins require selenocysteine in a specific location within the peptide chain to be functionally active. Selenocysteine is identical to cysteine, except a Se molecule replaces the sulfur molecule. To be inserted into the proper location in a peptide, selenocysteine must be synthesized from serine that is joined to a specific transfer RNA (tRNA; UGA codon). Selenocysteine that is absorbed by the intestine cannot be directly inserted into the active site within a polypeptide chain. In humans, at least 25 genes for selenoproteins have been identified (Lu and Holmgren, 2009), but the functions of many of the resulting selenoproteins are unknown. Glutathione peroxidases (GPx) are a family of selenoproteins that reduce hydrogen peroxide to water or phospholipid hydroperoxides to phospholipids, and increasing Se intake of cattle often increases the activity of these enzymes. Iodothyronine deiodinases are a family of selenoenzymes that activates thyroxin by deiodinating T₄ into T₃, and Se supplementation has increased serum T₃ concentrations in cattle (Awadeh et al., 1998; Contreras et al., 2002). The enzyme can also inactivate T₃ by further deiodination. Thioredoxin reductases (TRx) are selenoenzymes that reduce thioredoxin, which is involved in regulation of the redox potential of cells. Two other selenoproteins (selenoprotein P and selenoprotein W) have been found in bovine tissue, but their functions are unclear.

White muscle disease or nutritional muscular dystrophy is the classical sign of a clinical Se deficiency. Signs of this disease include leg weakness and stiffness, flexion of the hock joints, and muscle tremors (NRC, 1983). Cardiac and skeletal muscles have chalky striations and necrosis, and animals often die from cardiac failure. In cattle, improved Se status has increased growth rates (Wichtel et al., 1994; Reis et al., 2008) and reduced prevalence of retained fetal membranes (reviewed by Hemingway [1999] and Jovanovic et al. [2013]). Most studies evaluating effects of Se on clinical and subclinical mastitis have reported positive results (Smith et al., 1985; Erskine et al., 1987, 1989, 1990; Weiss et al., 1990; Wichtel et al., 1994; Jukola et al., 1996; Malbe et al., 2003; Kommisrud et al., 2005). Other health problems that have responded to Se supplementation include metritis, cystic ovaries (Harrison et al., 1984; Enjalbert et al., 2006), and udder edema (Miller et al., 1993). The likely mode of action of Se for these health disorders is via regulation of cellular concentrations of ROS via GPx and TRx. Se supplementation of cattle has improved the function of immune cells, including neutrophil (Hogan et al., 1990; Cebra et al., 2003), macrophage (Ndiweni and Finch, 1995), and lymphocyte (Stabel et al., 1990; Cao et al., 1992; Pollock et al., 1994). Concentrations of specific ROS within cells affect inflammatory responses, arachidonic acid metabolism, and production of prostaglandins and numerous cytokines (Salman et al., 2009; Sordillo, 2013).

Sources

Concentrations of Se in plants are positively correlated with those in the soil (Whelan et al., 1989; Hall et al., 2011). In general, feeds grown in the central plains of the United States and Canada contain more than 0.1 mg Se/kg DM, and feeds grown east of the Mississippi River and west of the Rocky Mountains typically contain <0.1 mg Se/kg DM (NRC, 1983). Except for Se accumulator plants (e.g., Astragalus bisulcatus), the predominant form of Se in plants is selenomethionine plus minor amounts of selenocysteine and selenite (Whanger, 2002). Se concentration of feeds is positively correlated with protein concentrations, and plant parts that are higher in protein also are higher in Se. Leaves of forage plants contain 1.5 to 2 times more Se than do stems (Gupta and Winter, 1989).

Based on current regulations of the U.S. FDA (1997, 2003), the only forms of Se that can be added legally to diets in the United States are sodium selenite, sodium selenate, and Se-enriched yeast at levels not to exceed 0.3 mg supplemental Se/kg DM. Other sources of supplemental Se that have been evaluated in cattle include barium selenate (Ceballos et al., 2010) and Se dioxide (Grace et al., 1995).

Efficiency of Absorption

Apparent absorption of Se from diets without supplemental Se is between 0.30 and 0.60 for sheep, goats, and nonlactating dairy cows (Harrison and Conrad, 1984a,b; Aspila, 1988; Koenig et al., 1997; Gresakova et al., 2013). Apparent absorption of Se in diets that contain selenite and selenate in diets ranged from 0.36 to 0.51 (Harrison and Conrad, 1984b; Ivancic and Weiss, 2001; Gresakova et al., 2013), and for diets with Se yeast, apparent absorption ranged from about 0.57 to 0.62 (Walker et al., 2010; Gresakova et al., 2013). In a direct comparison, apparent absorption of Se from Se yeast was about 24 percent greater (0.62 versus 0.50) than that of selenite when fed to sheep (Gresakova et al., 2013). Because of endogenous fecal losses, true absorption of Se is greater than apparent absorption. True absorption estimated using the regression method averaged about 0.5 in dairy cows fed inorganic Se (Harrison and Conrad, 1984a; Ivancic and Weiss, 2001). No data are available on estimated true absorption of Se from yeast. Because Se from Se yeast is retained in cellular proteins to a greater extent than that from inorganic Se, endogenous fecal losses of Se when Se yeast is fed may be greater, but the true absorption would also be greater.

Factors Affecting Absorption

Intestinal absorption of selenate is greater than selenite in sheep and rat intestinal in vitro models (Ardüser et al., 1986) and in the human Caco-2 cell model (Gammelgaard et al., 2012); however, because most but not all of the selenate is reduced to selenite within the rumen, absorption of selenate is probably only slightly greater than selenite in cattle (Podoll et al., 1992). The predominant form of Se in Se yeast is selenomethionine, which is absorbed via the same mechanism as Met. Intestinal absorption of selenomethionine is greater than absorption of inorganic Se sources (Gammelgaard et al., 2012). Assuming differences in apparent absorption accurately reflect differences in true absorption, absorption of Se from high-quality (i.e., high proportion of Se as selenomethionine) Se yeast is at least 1.2 times that of inorganic Se. Se uptake by ruminal microorganisms is much greater when Se yeast is fed compared with selenite (Mainville et al., 2009), but form of Se did not affect measures of ruminal fermentation (Panev et al., 2013).

Dietary sulfate added to increase concentrations of dietary S by 0.2 and 0.4 percentage units (Ivancic and Weiss, 2001) and low (<0.6 percent of DM) and high (>1.0 percent of DM) concentrations of dietary Ca (Harrison and Conrad, 1984a) reduce absorption of inorganic Se by cattle. Supplementation of S from anionic salts (approximately 0.6 percent total diet S) for the last 3 weeks of gestation did not influence Se status of nonlactating cows (Gant et al., 1998). Long-term feeding of diets that contained approximately 0.3 percentage units of added sulfate–sulfur to beef cattle did not affect concentrations of Se in blood or the activity of GPx (Khan et al., 1987). In an in vitro sheep intestinal model, thiosulfate and molybdate reduced absorption of inorganic Se (Ardüser et al., 1986). In nonruminants, Se absorption was not affected by dietary Cu, Fe, Mo, and Mn over a wide range of concentrations (Abdel Rahim et al., 1986). In dairy cows, increased dietary Cu did not affect measures of Se status (Koenig et al., 1991). Rats that were Mg deficient had significantly lower Se absorption than rats adequate in Mg (Jiménez et al., 1997), and elevated dietary Zn reduced Se absorption in rats (House and Welch, 1989). Whether Mg or Zn affects Se absorption in cattle is unknown. Antagonists to absorption of inorganic Se may affect absorption of Se yeast differently.

Indicators of Selenium Status

Se status can be assessed by concentrations of Se in tissue and blood, activity of glutathione peroxidase, and various immune cell assays. Few differences in blood and tissue concentrations of Se occur between different inorganic sources of Se when fed to dairy cattle (Podoll et al., 1992; Gibson et al., 1993; Ortman and Pehrson, 1999; Ortman et al., 1999). Se from Se yeast or from basal ingredients with higher concentrations of Se usually increases concentrations of Se in blood and activity of GPx more than diets with inorganic Se (Conrad and Moxon, 1979; Ortman and Pehrson, 1997; Knowles et al., 1999; Ortman et al., 1999; Weiss and Hogan, 2005; Juniper et al., 2008; Phipps et al., 2008; Koenig and Beauchemin, 2009). Blood concentrations and GPx activity average 20 to 25 percent greater when Se yeast is fed (Weiss, 2003; Juniper et al., 2006, 2008; Phipps et al., 2008; Koenig and Beauchemin, 2009) similar to the difference in apparent absorption. On average, milk Se is about 1.9 times greater when Se yeast is fed compared with inorganic Se, but a meta-analysis determined that depending on supplementation rate, the difference could be more than three times (Ceballos et al., 2009). Feeding rumen-protected Met reduces the concentration of Se in milk when Se yeast is fed (Weiss and St-Pierre, 2009).

Requirements and Factors Affecting Requirements

The seventh revised edition (NRC, 2001) defined the Se requirement as 0.3 mg/kg of dietary DM for all classes of dairy cattle. No new data are available to dispute this requirement. However, most of the data supporting this requirement were generated from experiments in which selenite or selenate/kg of dietary DM (DM basis) was fed, and total dietary Se ranged from 0.35 to 0.40 mg/kg.

Establishing requirements for Se using the factorial approach is difficult because the deposition of Se in body tissues, conceptus, and milk is dependent on Se intake and Se source, and essentially no data are available on endogenous fecal and urinary losses. Assuming a cow is fed a diet with approximately 0.3 mg Se from inorganic sources/kg of dietary DM, the conceptus will accumulate 0.055 mg Se/d during the last trimester of gestation (House and Bell, 1994). Comparable data when Se yeast is fed are not available, but accumulation in swine fetuses when Se yeast was fed was approximately 1.3 times greater than when selenite was fed (Ma et al., 2014).

Se concentrations vary across tissues in cattle, with kidney usually having the highest concentrations and muscle having lower concentrations (Lawler et al., 2004; Juniper et al., 2008), but muscle contained about 0.3 mg Se/kg DM when growing cattle were fed a diet with 0.3 mg Se/kg from selenite (Juniper et al., 2008) and 0.4 to 0.8 mg Se/kg dry weight when Se yeast was fed (Juniper et al., 2008; Richards et al., 2011). Concentrations of Se in various tissues of growing beef animals were 1.25 times greater when Se yeast was fed compared with selenite (Juniper et al., 2008). Se concentration of milk averages about 0.02 mg/kg and 0.04 mg/kg when selenite and Se yeast is fed at approximately 0.3 mg Se/kg of diet, respectively (Ceballos et al., 2009). Using the regression method, endogenous fecal and urinary losses varied by more than a factor of two depending on the source of data (Harrison and Conrad, 1984b; Ivancic and Weiss, 2001). Endogenous cells sloughed by cows that are deficient in Se would likely have lower concentrations of Se than cells sloughed by cows adequate in Se. No data are available on endogenous losses when Se yeast is fed but would likely be greater because of greater Se concentrations in cells.

Current FDA regulations limit Se supplementation to 0.3 mg/kg of diet (assumed air dried or approximately 90 percent DM basis) in the United States (FDA, 1997), and in most situations, that amount of supplemental Se will maintain dairy cattle in good Se status. Based on the effect of Se on mastitis, concentrations of Se in whole blood should be greater than about 0.18 μg/mL or approximately 0.075 μg/mL for plasma when inorganic Se is fed (Jukola et al., 1996). Intake of approximately 6 mg/d of inorganic Se should maintain those blood concentrations (Maus et al., 1980). Less Se may be needed when Se yeast is fed; however, more Se is also secreted in milk and retained in the body (nonspecific proteins) when Se yeast is fed. Based on available data, the AI of Se for all classes of cattle was set at

where DMI is in kg/d, and basal diet is generally assumed to provide no Se.

Most of the studies with Se for dairy cattle were conducted in areas with low soil Se so that basal Se was usually <0.1 mg Se/kg DM. This means that intake of supplemental Se was approximately equal to intake of total Se. In areas where soil has higher Se concentrations (e.g., North and South Dakota), users are advised to analyze locally grown forages for Se and include basal Se in the calculation of Se supply. Since Se concentrations in feedstuffs are low, specialized equipment is needed for the assay, which many labs do not have. The Se concentrations in the feed library included in the model are means and can differ greatly from feeds grown on specific farms.

Toxicity

Alkali disease and blind staggers result from Se toxicity. Clinical signs include sloughing of hooves, lameness, loss of hair, and emaciation. Most cases of Se toxicity have been related to consumption of Se-accumulating plants (e.g., Astragalus sp.), and much of the Se in those plants is found in methyl-selenium compounds. Similar clinical signs were also caused by feeding high doses of selenomethionine (10 mg Se/kg diet DM) or selenite (25 mg Se/kg diet DM) to yearling cattle for 120 days (O'Toole and Raisbeck, 1995). Acute toxicity can occur when cows are fed 10 to 20 mg Se/kg BW from selenite. An injection of about 0.5 mg Se/kg BW to young cattle (ca. 200 kg) resulted in a 67 percent mortality rate (NRC, 1983). The current MTL for dietary Se is 5 mg/kg of diet DM (NRC, 2005) or about 16 times greater than the recommended dietary concentration.

Function

Zn is a component of more than 200 enzymes, including oxidoreductases (e.g., Cu–Zn superoxide dismutase), transferases (e.g., RNA polymerase), hydrolases (e.g., alkaline phosphatase and carboxypeptidase), lyases (carbonic anhydrase), and ligases (e.g., tRNA synthetase) (Kidd et al., 1996). Zn is involved with macronutrient metabolism, numerous aspects of the immune system, gene regulation, hormonal regulation, neurotransmission, and apoptosis. The effects of Zn on immune function have received substantial attention (Fraker and King, 2004). Historically, these effects were thought to be manifested via Zn-containing enzymes such superoxide dismutase (which usually does not show a change in activity when dietary Zn intake is altered). However, accumulating data indicate that changes in Zn concentrations within immune cells are a major regulatory mechanism that may affect the entire immune system (Haase and Rink, 2009). The ubiquitous functions of Zn are likely the primary reason identifying markers of Zn status has been so difficult.

Absorption

Molecular and cellular mechanisms of absorption of dietary Zn have not been studied to any degree in the bovine, but substantial information is available from rodents and poultry models. Whether information from those species reflects mechanisms in cattle is unknown. In rats and poultry, Zn absorption occurs throughout the small intestine and perhaps in the large intestine by two different mechanisms: a saturable, transport-mediated absorption system and nonsaturable diffusion. In poultry, transport-mediated absorption occurred primarily in the duodenum and jejunum, and nonsaturable diffusion occurred primarily in the ileum (Yu et al., 2008). In rats, saturable absorption of Zn was found in all segments of the small intestine. The saturable transporters probably belong to the Zip family, and expression is downregulated when dietary Zn supply is high and upregulated when supply is low (Liuzzi and Cousins, 2004; Lichten and Cousins, 2009). Based on poultry and rodent data, at low dietary concentrations of Zn, high-affinity transporters become important (jejunum and ileum) but at high concentrations of Zn, diffusion in ileum and colon likely would predominate because of transporter saturation and downregulation. Although some regulation of Zn absorption probably occurs, based on the putative absorption processes, when excess-absorbable Zn is fed, cows will absorb Zn in excess of requirement. Export of Zn out of enterocytes appears to be regulated and is used to maintain Zn homeostasis. Expression of metallothionein genes or synthesis of the protein in intestinal cells is upregulated when excess Zn is provided (Tran et al., 1998; Shen et al., 2008), which likely is one mechanism of increased fecal excretion of endogenous Zn.

Factors Affecting Zinc Absorption

Measuring the true absorption of Zn or relative availability of Zn is exceedingly difficult because fecal excretion of Zn is used to maintain Zn homeostasis, and good markers of Zn status are lacking. Lactating cows that were adapted to a Zn-deficient diet (6 mg Zn/kg diet DM) absorbed nearly 50 percent of dietary Zn (Kirchgessner and Schwarz, 1976). However, maximizing Zn absorption by feeding deficient diets is clearly not desirable. Based on isotope studies, Zn absorption by ruminating cattle fed practical diets that were likely adequate in Zn ranged from 12 to 33 percent (Miller and Cragle, 1965; Hansard et al., 1968; Miller et al., 1968). However, these studies are decades old and were done with cattle with low DMI compared with modern lactating cows. Whether DMI affects Zn absorption is unknown, but a positive correlation between DM digestibility and Zn absorption has been observed (Miller and Cragle, 1965). Because new data are lacking, the AC for zinc chloride (ZnCl₂; the Zn source used in the above studies) used in the previous version (NRC, 2001) was retained (i.e., 0.20). Absorption of Zn by rats fed radioactive Zn from ZnCl₂ or from soy flour produced by plants fertilized with radioactive ZnCl₂ were the same (Stuart et al., 1986). Similar results were found when preruminant calves were fed ZnCl₂ or corn plants grown with radioactive Zn (Neathery et al., 1972). These studies, at least for nonruminants, indicate that Zn contained in basal ingredients has similar absorbability as ZnCl₂ (ca. 0.20). In agreement, true absorption of Zn was 0.182 by adult goats fed a diet in which about 50 percent of the Zn was from basal ingredients and 50 percent from ZnCl₂ (Hattori et al., 2010). Therefore, Zn in basal ingredients was assigned an AC of 0.20. Absorption coefficients for Zn from other supplements were estimated from studies measuring relative bioavailability. ZnCl₂ and zinc sulfate (ZnSO₄) had similar bioavailability in calves with a functioning rumen (Kincaid, 1979). Zinc oxide had a bioavailability of approximately 80 percent that of ZnSO₄ (Sandoval et al., 1997) so that its AC was set at 0.16. Zinc carbonate had the same bioavailability as ZnSO₄ in sheep (Sandoval et al., 1997). Several proprietary supplemental Zn sources are available, but published data on measures of relative bioavailability are limited (Cao et al., 2000). The limited data indicate slightly higher (10 to 20 percent) greater bioavailability for some proprietary compounds compared with ZnSO₄. A problem with most relative bioavailability studies is that very high concentrations of Zn are fed, which can affect results.

Several dietary components can interfere with Zn absorption or increase body losses of Zn, but most of the studies have been done with nonruminants (Lönnerdal, 2000). Phytate clearly reduces Zn absorption in nonruminants. Zn absorption by calves fed milk averaged about 50 percent (Miller and Cragle, 1965) but was reduced by more than half when soybean protein was included in the diet, likely because of the phytic acid in the soybean product (Miller et al., 1968). Phytase has little effect on P absorption (see P section), suggesting that in functioning ruminants, phytic acid probably does not greatly affect Zn absorption. High-fiber diets can reduce Zn absorption in nonruminants, but that affect is often confounded with effects of phytate. Diets that differed greatly in concentration of undigested NDF did not affect apparent Zn absorption in dairy cows (Faulkner et al., 2017). Ca can reduce Zn absorption in nonruminants, but that may be an indirect effect caused by the effects of Ca on phytate (Lönnerdal, 2000). In cattle, supplementation of Ca was associated with a reduction of Zn in serum of yearling steers and calves (Mills et al., 1967; Perry et al., 1968), but no deleterious effects of increased dietary Ca on metabolism of Zn or growth in sheep were observed (Pond and Wallace, 1986).

Cadmium markedly reduces Zn absorption in rats (Evans et al., 1974). High dietary Fe (1,000 mg Fe/kg of diet provided by ferrous sulfate) reduced liver Zn concentration in steers by about 18 percent (Standish et al., 1971). Zn and Cu are antagonistic to one another. Very high Cu can interfere with Zn metabolism; however, very low Zn-to-Cu ratios (0.15:1) are likely necessary to produce antagonism (Oestreicher and Cousins, 1985). Diets with 40 mg Cu/kg and 50 mg Zn/kg (0.8:1 Cu-to-Zn ratio) did not reduce plasma Zn in growing steers (Gooneratne et al., 1994). Elevated dietary S (approximately 0.5 percent) has increased urinary loss of Zn and reduced Zn absorption in beef cattle (Gooneratne et al., 2011; Pogge et al., 2014). Conversely, feeding monensin may increase absorption of Zn (Spears, 1990), and with nonruminants, certain proteins such as whey or beef increase the absorption of Zn, but other proteins such as casein and isolated soy protein (phytase treated) reduce Zn absorption (Lönnerdal, 2000). Based on currently available information, most practical diets should not contain adequate concentrations of antagonistic substances to reduce Zn absorption with the possible exception of excess S.

Dietary Zinc Requirement

A factorial approach was taken to determine the dietary requirement for Zn. In the previous edition (NRC, 2001) endogenous fecal and the obligate urinary loss of Zn was set at 0.033 mg Zn/kg BW and 0.012 mg Zn/kg BW, respectively, for a total maintenance requirement of 0.045 mg Zn/kg BW. The data used to generate those equations came from a study with growing heifers (ca. 300 kg BW) fed radioactive Zn (Hansard et al., 1968). Newer data do not support the value used for the obligate urinary loss and bring into question the value used for endogenous fecal loss. Growing cattle (ca. 300 kg) fed low but not deficient Zn diets excreted 0.003 mg Zn/kg BW in the urine (Gooneratne et al., 2011; Pogge et al., 2014), and lactating cows fed low Zn diets (30 mg Zn/kg DM) excreted 0.0016 mg Zn/kg BW in urine (Faulkner et al., 2017). Because urinary excretion of Zn is so low, the obligate urinary loss was set at zero. A stable isotope study with goats fed at maintenance estimated endogenous fecal loss of Zn at 0.17 mg/kg BW (Hattori et al., 2010). An earlier study with 300-kg beef heifers estimated endogenous fecal loss at 0.027 mg Zn/kg BW (Hansard et al., 1968). On a DMI basis (intake was <2 percent of BW in both studies), endogenous fecal Zn ranged from 2.0 (heifers) to 8.6 (goats) mg Zn/kg DMI. The extremely limited data and the great differences between studies make it difficult to estimate this requirement. In addition, the diets were not typical of what is fed to dairy cows. One reason for the disparate results is that endogenous fecal loss of Zn reflects Zn status. As more Zn is fed, Zn bound to metallothionein is excreted via feces to maintain Zn homeostasis. Based on Zn intake, the goats (Hattori et al., 2010) were likely in greater Zn status than the heifers (Hansard et al., 1968). Data on endogenous fecal excretion of Zn (and most trace minerals) by dairy cows are clearly needed to improve estimates of maintenance requirements. Therefore, the committee decided to use the mean value (rounded to the nearest whole number) from the two experiments and set the endogenous fecal Zn requirement as

No new data are available on Zn accretion by the conceptus; therefore, the gestation requirement was not changed and set at 0.017 mg Zn/kg of maternal BW, which equals 12 mg Zn/d between day 190 of gestation and the end of gestation for a 715-kg Holstein cow (House and Bell, 1993). Newer data support retaining the lactation requirement for Zn at 4 mg/kg of milk, but concentrations can range from about 3 to 6 mg Zn/kg (Schwartz and Kirchgessner, 1975; Kinal et al., 2007; Castillo et al., 2013; Faulkner et al., 2017). The amount of Zn retained during growth of body tissues averages 24 mg Zn/kg ADG (range, 16 to 31 mg) (Miller, 1970; Kirchgessner and Neesse, 1976). Zn accretion in growing sheep averaged about 28 mg Zn/kg of empty BW but was 24 mg Zn/kg in young sheep (15 kg BW) and increased to 30 mg Zn/kg in sheep weighing about 50 kg (Bellof and Pallauf, 2007). Whether the growth requirement (per kg of BW) increases as growing cattle get larger is not known; therefore, a single growth requirement (24 mg Zn/kg daily gain) was used.

Deficiency

Cattle that are deficient in Zn quickly exhibit reduced DMI and growth rates. With a more prolonged deficiency, the animals exhibit reduced growth of testes, weak hoof horn, and parakeratosis of the skin on the legs, head (especially nostrils), and neck. On necropsy, thymic atrophy and lymphoid depletion of the spleen and lymph nodes are evident (Miller and Miller, 1962; Mills et al., 1967; Mayland et al., 1980). A genetic defect that greatly reduces absorption of Zn has been identified in Dutch–Friesian cattle, and they become severely deficient in Zn unless fed extremely large amounts of dietary Zn (Flagstad, 1976). Marginal deficiency of Zn may increase the risk of mastitis and other infectious disease. Dairy cows fed diets with approximately 41 mg Zn/kg diet DM had higher milk SCC than cows fed diets with 63 mg/kg Zn (Cope et al., 2009). The diet with 41 mg Zn/kg would not meet the current Zn requirements.

Concentrations of Zn in serum are normally between 0.7 and 1.3 μg/mL, and concentrations below 0.5 μg/mL are often considered deficient. However, stress or disease can cause a rapid redistribution of Zn out of extracellular fluids, causing concentrations of Zn in serum to fall into the “deficient” range even when dietary Zn is adequate (Goff and Stabel, 1990). Liver Zn concentrations are not reflective of Zn intake but will decline with prolonged periods of dietary deficiency (Herdt and Hoff, 2011). Increased liver Zn concentrations have been observed in sheep fed supplemental Zn compared to no added dietary Zn (Cao et al., 2000), and the effects are greater when fed zinc lysine compared to zinc sulfate, zinc oxide, and zinc Met (Rojas et al., 1995). However, no differences were observed in liver Zn concentrations of adult cattle fed 360 mg/d of supplemental Zn compared to no supplemental Zn (Rojas et al., 1996), and source of supplemental Zn had no effect on liver Zn concentrations (Rojas et al., 1996; Siciliano-Jones et al., 2008). Liver Zn concentrations in cattle are also affected by age and perhaps other factors (Puschner et al., 2004). Carbonic anhydrase and alkaline phosphatase activities in blood have been used to assess Zn status, but these are difficult to interpret because concurrent disease can affect these enzymes as much as a deficiency of Zn (Mills, 1987). No widely agreed-on status indicator is available for Zn.

Toxicity

Cattle can generally tolerate high concentrations of dietary Zn. Clinical toxicity was observed in cattle fed a 900-mg Zn/kg diet (Ott et al., 1966a,b. Feed intake, milk production, and Cu status were reduced when cows were fed diets with 2,000 mg Zn/kg (from ZnSO₄) but not when fed diets with 1,000 mg Zn/kg (Miller et al., 1989). NRC (2005) established an MTL for cattle at 500 mg Zn/kg diet DM. However, dairy cows can likely tolerate greater concentrations.