5

Phosphorus

BACKGROUND INFORMATION

Overview

Phosphorus is most commonly found in nature in its pentavalent form in combination with oxygen, as phosphate (PO43−). Phosphorus (as phosphate) is an essential constituent of all known protoplasm and its content is quite uniform across most plant and animal tissues. Except for specialized cells with high ribonucleic acid content, and for nervous tissue with high myelin content, tissue phosphorus occurs at concentrations ranging approximately from 0.25 to 0.65 mmol (7.8 to 20.1 mg)/g protein. A practical consequence is that, as organisms consume other organisms lower in the food chain (whether animal or plant), they automatically obtain their phosphorus.

Phosphorus makes up about 0.5 percent of the newborn infant body (Fomon and Nelson, 1993), and from 0.65 to 1.1 percent of the adult body (Aloia et al., 1984; Diem, 1970). Eighty-five percent of adult body phosphorus is in bone. The remaining 15 percent is distributed through the soft tissues (Diem, 1970). Total phosphorus concentration in whole blood is 13 mmol/liter (40 mg/dl), most of which is in the phospholipids of red blood cells and plasma lipoproteins. Approximately 1 mmol/liter (3.1 mg/dl) is present as inorganic phosphate (Pi). This inorganic phosphate component, while a tiny fraction of body phosphorus (< 0.1 percent), is of critical importance. In adults this component makes up about 15 mmol (465 mg) and is located mainly in the blood and extracellular fluid.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride 5 Phosphorus BACKGROUND INFORMATION Overview Phosphorus is most commonly found in nature in its pentavalent form in combination with oxygen, as phosphate (PO43−). Phosphorus (as phosphate) is an essential constituent of all known protoplasm and its content is quite uniform across most plant and animal tissues. Except for specialized cells with high ribonucleic acid content, and for nervous tissue with high myelin content, tissue phosphorus occurs at concentrations ranging approximately from 0.25 to 0.65 mmol (7.8 to 20.1 mg)/g protein. A practical consequence is that, as organisms consume other organisms lower in the food chain (whether animal or plant), they automatically obtain their phosphorus. Phosphorus makes up about 0.5 percent of the newborn infant body (Fomon and Nelson, 1993), and from 0.65 to 1.1 percent of the adult body (Aloia et al., 1984; Diem, 1970). Eighty-five percent of adult body phosphorus is in bone. The remaining 15 percent is distributed through the soft tissues (Diem, 1970). Total phosphorus concentration in whole blood is 13 mmol/liter (40 mg/dl), most of which is in the phospholipids of red blood cells and plasma lipoproteins. Approximately 1 mmol/liter (3.1 mg/dl) is present as inorganic phosphate (Pi). This inorganic phosphate component, while a tiny fraction of body phosphorus (< 0.1 percent), is of critical importance. In adults this component makes up about 15 mmol (465 mg) and is located mainly in the blood and extracellular fluid.

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride It is into this inorganic phosphate compartment that phosphate is inserted upon absorption from the diet and resorption from bone and from this compartment that most urinary phosphorus and hydroxyapatite mineral phosphorus are derived. This compartment is also the primary source from which the cells of all tissues derive both structural and high-energy phosphate. Structurally, phosphorus occurs as phospholipids, which are a major component of most biological membranes, and as nucleotides and nucleic acids. The functional roles include: (1) the buffering of acid or alkali excesses, hence helping to maintain normal pH; (2) the temporary storage and transfer of the energy derived from metabolic fuels; and (3) by phosphorylation, the activation of many catalytic proteins. Since phosphate is not irreversibly consumed in these processes and can be recycled indefinitely, the actual function of dietary phosphorus is first to support tissue growth (either during individual development or through pregnancy and lactation) and, second, to replace excretory and dermal losses. In both processes it is necessary to maintain a normal level of Pi in the extracellular fluid (ECF), which would otherwise be depleted of its phosphorus by growth and excretion. Physiology of Absorption, Metabolism, and Excretion Food phosphorus is a mixture of inorganic and organic forms. Intestinal phosphatases hydrolyze the organic forms contained in ingested protoplasm, and thus most phosphorus absorption occurs as inorganic phosphate. On a mixed diet, net absorption of total phosphorus in various reports ranges from 55 to 70 percent in adults (Lemann, 1996; Nordin, 1989; Stanbury, 1971) and from 65 to 90 percent in infants and children (Wilkinson, 1976; Ziegler and Fomon, 1983). There is no evidence that this absorption efficiency varies with dietary intake. In the data from both Stanbury (1971) and Lemann (1996), the intercept of the regression of adult fecal phosphorus on dietary phosphorus is not significantly different from zero, and the relationship is linear out to intakes of at least 3.1 g (100 mmol)/day. This means that there is no apparent adaptive mechanism that improves phosphorus absorption at low intakes. This is in sharp contrast to calcium, for which absorption efficiency increases as dietary intake decreases (Heaney et al., 1990b) and for which adaptive mechanisms exist that improve absorption still further at habitual low intakes (Heaney et al., 1989). A portion of phosphorus absorption is by way of a saturable, active transport facilitated by 1,25-dihydroxyvitamin D (1,25(OH)2D) (Chen et al., 1974; Cramer, 1961). However, the fact that fractional

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride phosphorus absorption is virtually constant across a broad range of intakes suggests that the bulk of phosphorus absorption occurs by passive, concentration-dependent processes. Also, even in the face of dangerous hyperphosphatemia, phosphorus continues to be absorbed from the diet at an efficiency only slightly lower than normal (Brickman et al., 1974). Phosphorus absorption is reduced by ingestion of aluminum-containing antacids and by pharmacologic doses of calcium carbonate. There is, however, no significant interference with phosphorus absorption by calcium at intakes within the typical adult range. Excretion of endogenous phosphorus is mainly through the kidneys. Inorganic serum phosphate is filtered at the glomerulus and reabsorbed in the proximal tubule. The transport capacity of the proximal tubule for phosphorus is limited; it cannot exceed a certain number of mmol per unit time. This limit is called the tubular maximum for phosphate (TmP). TmP varies inversely with parathyroid hormone (PTH) concentration; PTH thereby adjusts renal clearance of Pi. At filtered loads less than the TmP (for example, at low plasma Pi values), most or all of the filtered load is reabsorbed, and thus plasma phosphate levels can be at least partially maintained. By contrast, at filtered loads above the TmP, urinary phosphorus is a linear function of plasma phosphate (Bijvoet, 1969; Lemann, 1996; Nordin, 1989). In the healthy adult, urine phosphorus is essentially equal to absorbed diet phosphorus, less small amounts of phosphorus lost in shed cells of skin and intestinal mucosa. This regulation of phosphorus excretion is apparent from early infancy. In infants, as in adults, the major site of regulation of phosphorus retention is at the kidney. In studies of infants receiving different calcium intakes (DeVizia et al., 1985; Moya et al., 1992; Williams et al., 1970; Ziegler and Fomon, 1983), phosphorus retention did not differ even with high amounts of dietary calcium (calcium:phosphorus [Ca:P] molar ratios of 0:6, 1:1, or 1.4:1). Any reduction in absorption of phosphorus due to high amounts of dietary calcium were compensated for by parallel reductions in renal phosphorus excretion (DeVizia et al., 1985; Fomon and Nelson, 1993; Moya et al., 1992). The least renal excretory work to maintain normal phosphorus homeostasis would be achieved with human milk as the major source of minerals during the first year of life. Regulation of the Serum Inorganic Phosphate Concentration Pi levels are only loosely regulated. Normal Pi levels decline with age from infancy to maturity (Table 5-1). The most likely reason for the

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride TABLE 5-1 Normative Values for Serum Inorganic Phosphorus (mmol/liter) for Age Age (y) Mean 2.5 Percentile 97.5 Percentile 0–0.5 2.15 1.88 2.42 2 1.81 1.43 2.20 4 1.77 1.38 2.15 6 1.72 1.33 2.11 8 1.67 1.29 2.06 10 1.63 1.24 2.01 12 1.58 1.19 1.97 14 1.53 1.15 1.92 16 1.49 1.10 1.88 20 1.39 1.01 1.78 Adult 1.15 0.87 1.41 higher Pi in newborn infants than in older children and adults is the lower glomerular filtration rate (GFR) of infants. GFR is about 32 ml/min/1.73 m2 at about 1 week of age, and rises to 87 at 4 to 6 months (Brodehl et al., 1982; Svenningsen and Lindquist, 1974). In the first months of life, plasma Pi concentration appears to be a reflection both of renal glomerular maturity and of the amount of dietary intake. Mean serum Pi appears to decline by about 0.3 mmol/liter (0.9 mg/dl) across the second half of the first year of life (Specker et al., 1986). Human milk-fed, compared with formula-fed, infants have a slightly lower plasma Pi (2.07 versus 2.25 mmol/liter or 6.4 versus 7.0 mg/dl) which is simply a function of differences in intake (Greer et al., 1982c; Specker et al., 1986); Gaucasian compared with African American infants have a slightly higher plasma Pi irrespective of type of milk feeding (Specker et al., 1986). The general relationship between absorbed phosphorus intake and plasma Pi in adults is set forth in Figure 5-1, derived by Nordin (1989) from the infusion studies of Bijvoet (1969). (In Bijovet's studies, a neutral phosphate solution was infused intravenously at a steadily increasing rate and produced a controlled hyperphosphatemia.) The achieved plasma Pi could thus be directly related to the quantity entering the circulation. Plasma Pi rises rapidly at low intakes, since the filtered load will be below the TmP, and little of the absorbed phosphorus will be lost in the urine (Figure 5-1). The steep, ascending portion of the curve thus represents a filling up of the extracellular fluid space with absorbed phosphate. At higher intakes, urinary excretion rises to match absorbed input and plasma levels change much more slowly.

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride FIGURE 5-1 Relation of serum Pi to absorbed intake in adults with normal renal function. (See Nordin [1989] for further details.) The solid curve can be empirically approximated by the following equation: Pi = 0.00765 × AbsP + 0.8194 × (1 − e(−0.2635 × AbsP), in which Pi = serum Pi (in mmol/liter), and AbsP = absorbed phosphorus intake (also in mmol). Solving this equation for the lower and upper limits of the normal range for Pi, as well as for its midpoint, yields the values presented in Table 5-1. The dashed horizontal lines represent approximate upper and lower limits of the normal range, while the dashed curves reflect the relationship between serum Pi and ingested intake for absorption efficiencies about 15 percent higher and lower than average. (© Robert P. Heaney, 1996. Reproduced with permission.) The relationship shown in Figure 5-1 holds only in adult individuals with adequate renal function; that is, the slow rise of plasma Pi with rising phosphorus intake over most of the intake range applies only so long as excess absorbed phosphate can be spilled into the urine. However, in individuals with reduced renal function, phosphorus clearance remains essentially normal so long as GFR is at least 20 percent of mean adult normal values, largely because tubular reabsorption is reduced to match the reduction in filtered load. Below that level, excretion of absorbed phosphate requires higher and higher levels of plasma Pi to maintain a filtered load at least

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride equal to the absorbed load. This is the reason for the hyperphosphatemia typically found in patients with end-stage renal disease. Another process depleting the blood of Pi is mineralization of nucleated bone and cartilage matrix. The amorphous calcium phosphate formed in the first stages of mineralization exhibits a Ca:P molar ratio of about 1.33:1, or very close to the molar ratio of Ca:P in adult ECF. While outside of a mineralizing environment, ECF calcium and phosphorus concentrations are indefinitely stable at physiological pH and pCO2, ECF is supersaturated in the presence of the hydroxyapatite crystal lattice. Hence, ECF supports calcium phosphate deposition only in the presence of a suitable crystal nucleus. As a consequence, in nonosseous tissues, ECF [Ca2+] and Pi concentrations will be essentially what can be measured in peripheral venous blood. However, at active bone-forming sites, ECF is depleted of both its calcium and its phosphate. Osteoblast function seems not to be appreciably affected by ECF [Ca2+], but like other tissues, the osteoblast needs a critical level of Pi in its bathing fluid for fully normal cellular functioning. Local Pi depletion both impairs osteoblast function and limits mineral deposition in previously deposited matrix. Finally, it should be noted that ECF Pi levels are indirectly supported by two mechanisms that amount to a weak, negative feedback type of control. One occurs via release of phosphate from bone, and the other via regulation of the renal 1-α hydroxylase. Still Pi concentration affects the responsiveness of the osteoclast to PTH: for any given PTH level, resorption is higher when Pi levels are low, and vice versa (Raisz and Niemann, 1969). High Pi values, by reducing bony responsiveness to PTH, lead to increased PTH secretion in order to maintain calcium homeostasis, and thereby to a lowering of ECF Pi. Similarly, high plasma Pi levels suppress renal synthesis of 1,25(OH)2D (Portale et al., 1989), thereby slightly reducing net phosphorus absorption from the diet. Both mechanisms reduce phosphorus input into the ECF when Pi is high and augment it when Pi is low. However, in neither circumstance is the effect on plasma Pi more than modest. (See below for other special circumstances.) Factors Affecting the Phosphorus Requirement Bioavailability Most food sources exhibit good phosphorus bioavailability. There is, however, one major exception. All plant seeds (beans, peas, cereals, nuts) contain a nonprotoplasmic, storage form of phosphate,

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride phytic acid. The digestive systems of most mammals cannot hydrolyze phytic acid, and hence its phosphorus is not directly available. However, some foods contain phytase, as do some colonic bacteria. Thus, a variable amount of phytate phosphorus becomes available. However, in gnotobiotic animals, no phytate phosphorus is absorbed (Wise and Gilburt, 1982). In at least one study (Parfitt et al., 1964), up to 50 percent of phytate phosphorus was absorbed, representing the combined effect of food phytases and bacterial enzymes. Also, yeasts can hydrolyze phytate, and thus whole grain cereals incorporated into leavened bread products have higher phosphate bioavailability than cereal grains used, for example, in unleavened bread or breakfast cereals. Finally, unabsorbed calcium in the digestate complexes with phytic acid and interferes with bacterial hydrolysis of phytate (Sandberg et al., 1993; Wise and Gilburt, 1982). This may be a part of the explanation for calcium's interference with phosphorus absorption. In infants, both the quantity of ingested phosphorus and the dietary bioavailability vary by type of milk fed. The efficiency of absorption is highest from human milk (85 to 90 percent) (Williams et al., 1970), intermediate from cow milk (72 percent) (Williams et al., 1970; Ziegler and Fomon, 1983), and lowest from soy formulas, which contain phytic acid (~59 percent) (Ziegler and Fomon, 1983). Because infant formulas contain substantially greater amounts of phosphorus than human milk, the absorbed phosphorus from cow milk and soy formulas is twice that attained by human milk-fed infants (Moya et al., 1992). The higher amounts of phosphorus (and also calcium and other mineral elements) in formulas that are based on cow milk or soy isolate protein effectively offset the lower mineral absorption of these formulas relative to human milk. Relatively low intakes of phosphorus, as occur with human milk, may actually confer an advantage to the infant by virtue of the low residual phosphorus in the lower bowel. Low intestinal phosphorus concentrations lower the fecal pH (Manz, 1992), which in turn may reduce proliferation of potentially pathogenic microorganisms and provide an immune protective effect. Nutrient-Nutrient Interactions Calcium. In the past, considerable emphasis was placed on the Ca:P ratio of the diet (for example, Chinn, 1981), particularly in infant nutrition (for example, Fomon and Nelson, 1993). The concept has some utility under conditions of rapid growth (in which a

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride large share of the ingested nutrients is converted into tissue mass), but it has no demonstrable relevance in adults. An optimal ratio ensures that, if intake of one nutrient is adequate for growth, the intake of the associated nutrient will also be adequate without a wasteful surplus of one or the other. However, the ratio by itself is of severely limited value, in that there is little merit to having the ratio “correct” if the absolute quantities of both nutrients are insufficient to support optimal growth. Furthermore, the intake ratio, by itself, fails to take into account both differing bioavailabilities and physiological adaptive responses. For example, in term-born infants during the first year of life, a higher calcium content of soy-based formulas was found to reduce phosphorus absorption, but phosphorus retention was similar because of offsetting changes in renal phosphorus output (DeVizia et al., 1985). Substitution of lactose with sucrose and/or hydrolyzed corn syrup solids had no effect on the efficiency of phosphorus absorption from formulas based on cow milk (Moya et al., 1992) or soy protein (Ziegler and Fomon, 1983). Estimates of optimal Ca:P intake ratios have frequently been based on the calcium and phosphorus needs of bone building. The molar ratio of Ca:P in synthetic hydroxyapatite is 1.67:1; in actual bone mineral, usually closer to 1.5:1; and in amorphous calcium-phosphate (the first mineral deposited at the mineralizing site), 1.3:1 (Nordin, 1976). However, during growth, soft tissue will be accreting phosphorus as well. On average, lean soft tissue growth accounts for about 1 mmol (31 mg) phosphorus for every 5 mmol (155 mg) added to bone (Diem, 1970). Since soft tissue accretion of calcium is negligible compared with skeletal calcium accretion, an absorbed Ca:P molar ratio sufficient to support the sum of bony and soft tissue growth would be ~1.3:1 (assuming equivalent degrees of renal conservation of both nutrients). The corresponding ingested intake ratio must consider the differing absorption efficiencies for dietary calcium and phosphorus. In infants, with net absorptions for calcium and phosphorus of approximately 60 and 80 percent respectively, the intake ratio matching tissue accretion would be ~2:1 (see also the section “Birth through 12 Months, ” below). This value is somewhat higher than the Ca:P molar ratio of human milk, which in most populations is in the range of 1.5:1 (Fomon and Nelson, 1993). Human milk must be presumed to be optimal for the infant's nutritional needs. The disparity between its ratio of ~1.5:1 and the ingested ratio calculated above reflects both the uncertainties in

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride the estimates on which the calculation is based and the limitations inherent in using such calculations. If growth were the only consideration, the intake ratio would have to be substantially higher than 2:1 after infancy, because calcium absorption drops more sharply with age than does phosphorus absorption (Abrams et al., 1997b; Fomon and Nelson, 1993). However, as larger fractions of ingested food are used for energy (and a correspondingly smaller proportion for growth), the notion of a dietary Ca:P molar ratio has little meaning or value, particularly since, on a mixed diet, there is likely to be a relative surplus of phosphorus. Under such circumstances it would be inappropriate to conclude, simply on the basis of a departure from some theoretical Ca:P ratio, either that calcium intake should be elevated or, phosphorus intake reduced. In balance studies in human adults, Ca:P molar ratios ranging from 0.08:1 to 2.40:1 (a 30-fold range) had no effect on either calcium balance or calcium absorption (Heaney and Recker, 1982; Spencer et al., 1965, 1978a). Thus, for the reasons cited, there is little or no evidence for relating the two nutrients, one to the other, during most of human life. Intake of Phosphorus The USDA Continuing Survey of Food Intake of Individuals (CSFII) in 1994, adjusted by the method of Nusser et al. (1996), indicated that the mean daily phosphorus intake from food in males aged 9 and over was 1,495 mg (48.2 mmol) (fifth percentile = 874 mg [28.2 mmol]; fiftieth percentile = 1,445 mg [46.6 mmol]; ninety-fifth percentile = 2,282 mg [73.6 mmol]) (see Appendix D for data tables). The mean daily intake in females aged 9 and over was 1,024 mg (33 mmol) (fifth percentile = 620 mg [20 mmol]; fiftieth percentile = 1,001 mg [32.3 mmol]; ninety-fifth percentile = 1,510 mg [48.7 mmol]). In both sexes, intakes decreased at ages 51 and over. The NHANES III data show similar median intake values (Alaimo et al., 1994). National survey data for Canada are not available. Both extent of usage of phosphate salts as additives and the amount per serving have increased substantially over the past 20 years, and the nutrient databases may not reflect these changes (Calvo and Park, 1996). For that reason, phosphorus intake may be underestimated for certain individuals who rely heavily on processed foods. However, one comparison of calculated intakes with analyzed intake data from the U.S. Food and Drug Administration's Total Diet Study found slight overestimates of phosphorus intake (by an

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride average of 61 mg [2 mmol]/day or 5 percent of the total intake) when intakes were calculated using USDA's nutrient database (Pennington and Wilson, 1990). Because of the uncertainty about phosphorus values for processed foods in nutrient databases, trends in phosphorus intake may be difficult to ascertain. Daily intakes of women aged 19 to 50 years from USDA 's national surveys averaged 965 mg (31.1 mmol) in 1977, 1,039 mg (33.5 mmol) in 1985, and 1,022 mg (33.0 mmol) in 1994 (Cleveland et al., 1996; USDA, 1985). Thus, it appears that intakes from foods increased about 8 percent between 1977 and 1985, but then decreased slightly between 1985 and 1994. Food supply data show a larger increase in phosphorus consumption: 12 percent from 1980 through 1994 (from 1,480 to 1,680 mg [47.7 to 54.2 mmol]/day per capita) (USDA, 1997). However, disappearance data may be unreliable for detecting trends because phosphate additives (such as those in cola beverages) are not included. Disappearance data on phosphorus-containing additives show that the use of these additives has increased by 17 percent over the last decade (Calvo, 1993). These figures also do not reflect actual consumption, because not all phosphates included in disappearance data are actually consumed, (for example, blends of sodium tripolyphosphate and sodium hexametaphosphate are used in brines for curing meat, but the brine is rinsed off and not consumed). Nevertheless, taken together, these data suggest a substantial increase in phosphorus consumption, in the range of 10 to 15 percent, over the past 20 years. Food Sources of Phosphorus Phosphates are found in foods as naturally occurring components of biological molecules and as food additives in the form of various phosphate salts. These salts are used in processed foods for nonnutrient functions, such as moisture retention, smoothness, and binding. In infants, dietary intake of phosphorus spans a wide range, depending on whether the food is human milk, cow milk, adapted cow milk formula, or soy formula (see Table 5-2). Moreover, the phosphorus concentration of human milk declines with progressing lactation, especially between 4 and 25 weeks of lactation (Atkinson et al., 1995). By contrast, more of the variation in dietary intake of phosphorus in adults is due to differences in total food intake and less to differences in food composition. Phosphorus contents of adult diets average about 62 mg (2 mmol)/100 kcal in both sexes (Carroll et al., 1983), and phosphorus:energy ratios exhibit a coefficient of variation of only

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride TABLE 5-2 Average Phosphorus Content and Calcium:Phosphorus Molar Ratio of Various Infant Feedings Feeding Type P (mmol/liter) Ca:P Molar Ratio Human milka   1 week 5.1 ± 0.9 1.3:1 4 weeks 4.8 ± 0.8 1.4:1 16 weeks 3.9 ± 0.5 1.5:1 Cows' milk formula 12 1.0:1 Soy formulab 15 1.2:1 Whole cows' milk 30 1.0:1 about one-third that of total phosphorus intake. Nevertheless, individuals with high dairy product intakes will have diets with higher phosphorus density values, since the phosphorus density of cow milk is higher than that of most other foods in a typical diet. The same is true for diets high in colas and a few other soft drinks that use phosphoric acid as the acidulant. A 12-ounce serving of such beverages contains about 50 mg (< 2 mmol), which is only 5 percent of the typical intake of an adult woman. However, when consumed in a quantity of five or more servings per day, such beverages may contribute substantially to total phosphorus intake. a Milk phosphorus at three different weeks of lactation (Atkinson et al., 1995). b Phosphorus content of soy formula includes about 3 mmol/L, present as phytate phosphorus which is likely not to be bioavailable (DeVizia and Mansi, 1992). Intake from Supplements Phosphorus supplements are not widely used in the United States. Based on a national survey in 1986, about 10 percent of U.S. adults and 6 percent of children aged 2 to 6 years took supplements containing phosphorus (Moss et al., 1989). Usage by men and women was similar, as was the dose taken by users: a median of about 120 mg (3.9 mmol)/day and a ninety-fifth percentile of 448 mg (14.5 mmol)/day. Young children who took supplements had a median supplemental intake of only 48 mg (1.5 mmol)/day and a ninety-fifth percentile of intake of 200 mg (6.5 mmol)/day. Effects of Inadequate Phosphorus Intake Hypophosphatemia and Phosphorus Depletion Inadequate phosphorus intake is expressed as hypophosphatemia.

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride FIGURE 5-2 Serum phosphorus concentrations in lactating and non-lactating women. The dotted lines represent the normal range for non-lactating women. The circles represent the serum phosphorus levels of lactating women (Bryne et al., 1987; Chan et al., 1982b; Cross et al., 1995a; Dobnig et al., 1995; Kalkwarf et al., 1996; Kent et al., 1990; Lopez et al., 1996; Specker et al., 1991a). increased during lactation. Apparently, increased bone resorption and decreased urinary excretion of phosphorus (Kent et al., 1990), which occur independent of dietary intake of phosphorus or calcium, provide the necessary phosphorus for milk production. Therefore, the EAR and RDA are estimated to be similar to that obtained for nonlactating women of their respective age groups. EAR for Lactation   14 through 18 years 1,055 mg (34.0 mmol)/day 19 through 30 years 580 mg (18.7 mmol)/day 31 through 50 years 580 mg (18.7 mmol)/day RDA for Lactation   14 through 18 years 1,250 mg (40.3 mmol)/day 19 through 30 years 700 mg (22.6 mmol)/day 31 through 50 years 700 mg (22.6 mmol)/day

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride Utilizing the 1994 CSFII intake data, adjusted for day-to-day variations (Nusser et al., 1996), the median intake of phosphorus for the 16 lactating women in the survey is 1,483 mg (47.8 mmol)/day and the lowest intake reported was at the first percentile of 1,113 mg (35.9 mmol)/day (see Appendix D). All women consumed above the phosphorus EAR of 580 to 1,055 mg (18.7 to 34 mmol)/day during lactation. Special Considerations Adolescent Mothers and Mothers Nursing Multiple Infants. As with calcium, phosphorus requirements may be increased in lactating adolescents and in mothers nursing multiple infants. Requirements during these periods may be higher because of the adolescent mother 's own growth requirements and the requirement for increased milk production while nursing multiple infants. It is not known whether decreased urinary phosphorus and increased maternal bone resorption provides sufficient amounts of phosphorus to meet these increased needs. TOLERABLE UPPER INTAKE LEVELS Hazard Identification As shown in Figure 5-1, serum Pi rises as total phosphorus intake increases. Excess phosphorus intake from any source is expressed as hyperphosphatemia, and essentially all the adverse effects of phosphorus excess are due to the elevated Pi in the ECF. The principal effects that have been attributed to hyperphosphatemia are: (1) adjustments in the hormonal control system regulating the calcium economy, (2) ectopic (metastatic) calcification, particularly of the kidney, (3) in some animal models, increased porosity of the skeleton, and (4) a suggestion that high phosphorus intakes could reduce calcium absorption by complexing calcium in the chyme. Concern about high phosphorus intake has been raised in recent years because of a probable population-level increase in phosphorus intake through such sources as cola beverages and food phosphate additives. It has been reported that high intakes of polyphosphates, such as are found in food additives, can interfere with absorption of iron, copper, and zinc (Bour et al., 1984); however, described effects are small, and have not been consistent across studies (for example,

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride Snedeker et al., 1982). For this reason, as well as because trace mineral status may be low for many reasons, it was not considered feasible to use trace mineral status as an indicator of excess phosphorus intake. Nevertheless, given the trend toward increased use of phosphate additives in a variety of food products (Calvo and Heath, 1988; Calvo and Park, 1996), it would be well to be alert to the possibility of some interference in individuals with marginal trace mineral status. Most of the studies that describe harmful effects of phosphorus intake used animal models. In extrapolating these data to humans, it is important to note that the phosphorus density of human diets represents the extreme low end of the continuum of standard diets for pets and laboratory animals. The median human dietary phosphorus density in the 1994 CSFII was very close to 62 mg (2.0 mmol)/100 kcal for all adults and for both sexes (Cleveland et al., 1996). By contrast, the diets of standard laboratory rats and mice have phosphorus densities ranging from 124 to 186 mg (4 to 6 mmol)/100 kcal. Cats and dogs have diets with densities close to 279 mg (9 mmol)/100 kcal, and laboratory primate diets average about 155 mg (5 mmol)/100 kcal. Adjustments in Calcium-Regulating Hormones As noted earlier, Pi is not, strictly speaking, regulated. Thus, a high phosphorus diet produces a higher level of plasma Pi, especially during the absorptive phase after eating. High Pi levels reduce urine calcium loss, reduce renal synthesis of 1,25(OH) 2D, reduce serum ionized calcium, and lead to increases in PTH release (Portale et al., 1989; Wood et al., 1988). These effects reflect adjustments in the control system that regulates the calcium economy and are not in themselves necessarily adverse. As already noted, the reduction in 1,25(OH)2D synthesis may slightly mitigate the hyperphosphatemia by a small reduction in phosphorus absorption from the intestine. It is known that added dietary phosphate loads of 1.5 to 2.5 g of inorganic phosphorus (as phosphate) in humans lead acutely to very slight drops in ECF [Ca2+] and to correspondingly elevated PTH concentrations (Calvo and Heath, 1988; Silverberg et al., 1986). It has been proposed that these adjustments in circulating calcium regulating hormones (associated with any elevation of serum Pi, even though within the usual normal range) could have adverse effects on the skeleton (Calvo and Heath, 1988; Calvo et al., 1988, 1990; Krook et al., 1975). The increase in PTH is often pre-

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride sumed to be harmful to bone. The mild hypocalcemic effect seen with acute increases in phosphorus intake is often attributed to the formation of calcium phosphate (CaHPO4) complexes in plasma, which is the presumed mechanism driving down [Ca2+] (Krook et al., 1975). However, it is doubtful that this is the correct explanation. As already noted, adult plasma is undersaturated with respect to CaHPO4 H2O, and the modest elevations in plasma Pi produced by a high phosphorus diet are probably not sufficient to alter plasma [Ca2+] directly. Moreover, causes of hypercalcemia, such as primary bone wasting and simple intravenous infusion of calcium, cause serum P i to rise, not fall, as might be expected if concentrations of the two ions were reciprocally related to one another at normal concentrations (Howard et al., 1953). Rather, it is more likely that the initial fall in [Ca2+] following elevation of plasma Pi is produced by direct inhibition of PTH-mediated osteoclastic release of calcium from bone (Raisz and Niemann, 1969). Thus, rather than harmful, this inhibition of bone resorption and consequent increase in PTH may actually be beneficial to bone, since it amounts to a relative resistance to the bone-resorbing effects of PTH. This point had been made by an earlier expert panel (Chinn, 1981) and is supported by more recent evidence showing that, despite the elevated PTH, urine hydroxyproline (a measure of bone resorption) falls on high phosphorus intakes, as does urine calcium (Silverberg et al., 1986). Diets high in phosphorus and low in calcium produce a sustained rise in PTH (Calvo et al., 1988, 1990); but diets low in calcium without extra phosphorus produce the same change (Barger-Lux et al., 1995), and for that reason it is unlikely that the high phosphorus feature of the altered intake is the culprit in the first instance. In addition, with respect to high phosphorus intakes, chronic administration of 2 g (65 mmol)/day phosphorus in men for at least 8 weeks produced no effect on calcium balance or calcium absorption relative to a diet containing only 806 mg (26 mmol) phosphorus (Spencer et al., 1965, 1978a). Calcium intake (low, normal, or high) had no influence on this lack of effect, underscoring the lack of physiological relevance of the dietary Ca:P ratio in adults. Further, calcium kinetic studies performed in adult women in whom phosphorus intake was doubled from 1.1 to 2.3 g (35.5 to 74.2 mmol), showed no effect whatsoever on bone turnover processes after 4 months of treatment (Heaney and Recker, 1987). A similar conclusion was reached more recently by Bizik et al. (1996), who doubled phosphorus intake (from 800 to 1,600 mg [26 to 52 mmol]/day) for 10 days in seven healthy young men (aged 22 to 31

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride years) and found no increase in urine deoxypyridinoline excretion (a marker of bone resorption). For all these reasons, it is doubtful whether phosphorus intakes, within the range currently thought to be experienced by the U.S. population and/or associated with serum Pi values in the normal range, adversely affect bone health. Normal, healthy, term-born infants, like adults, can adjust to a relatively wide range of dietary Ca:P ratios as provided in contemporary infant formulas. However, as a result of developmental immaturity in renal handling of phosphorus by infants, ECF [Ca2+] and Pi concentrations are closer to saturation in infancy, and infants are, therefore, more at risk of developing hypocalcemia as a consequence of hyperphosphatemia (DeVizia and Mansi, 1992). However, in the first month of life, some infants exhibit unusual sensitivity to phosphorus intakes above those associated with human milk. In the past, the clinical syndrome of late neonatal hypocalcemic tetany was observed when infants were fed whole evaporated cow milk with a very high phosphorus content (DeVizia and Mansi, 1992). Surprisingly, even with the introduction of modified cow milk formulas with phosphorus content reduced to one-half or less than that of evaporated whole milk, the syndrome of hypocalcemia has still been observed in young infants (Specker et al., 1991b; Venkataraman et al., 1985). However, the phosphorus content of such formulas is still substantially higher than that of human milk (Specker et al., 1991b). If such hyperphosphatemia is allowed to persist during early infancy, parathyroid hyperplasia, ectopic calcifications, and low serum calcitriol may occur (Portale et al., 1986). In such cases, the compensatory mechanisms for handling phosphate loads, mainly renal excretion, must be overwhelmed, leading to excessive phosphorus retention and the other metabolic consequences. It has also been suggested that a high content of phosphorus and calcium in an infant's diet may be a predisposing factor for the development of retention acidosis (Manz, 1992). It is not possible to identify, in advance, infants at risk for these syndromes unless they have renal dysfunction. Based on the data of Specker et al. (1991b), the risk of hypocalcemia (serum calcium less than 1.1 mmol/liter [4.4 mg/dl]) is 30 out of 10,000 neonates fed such formulas. Human milk with its low phosphorus content is both safer and better suited to the growth needs of the infant than cow milk (Manz, 1992). Finally, one report associates high intakes of phosphoric acid-containing cola beverages with slight reductions of serum calcium in Mexican children (Mazariegos-Ramos et al., 1995), but it is not clear to what extent the effect is due to the acid load of

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride the colas, the associated low intake of calcium-rich beverages, or the phosphorus itself. Except as noted in very young infants, none of these adjustments in calcium-regulating hormones is clearly adverse in its own right, particularly if calcium intake is adequate. Hence these effects do not provide a useful basis for estimating the Tolerable Upper Intake Level (UL). Metastatic Calcification The most serious, clearly harmful effect of hyperphosphatemia is calcification of nonskeletal tissues. This occurs when the calcium and phosphorus concentrations of ECF exceed the limits of solubility for secondary calcium phosphate (CaHPO4). This critical concentration is strongly dependent on amounts of other ions in the ECF, especially HCO3− citrate, H+, and K+, and so cannot be unambiguously defined. However, tissue calcification virtually never occurs at ECF calcium × phosphorus ion products less than ~4 (mmol/liter)2 [~1 (mg/dl)2]. Although ECF in adults is normally less than half-saturated with respect to CaHPO4, elevation of plasma Pi, if extreme, can bring the ECF to the point of saturation. Although both calcium and phosphate are involved in such ectopic mineralization, ECF calcium levels are tightly regulated and are usually affected little by even large variations in calcium intake. By contrast, the sensitivity of ECF Pi to joint effects of diet and renal clearance means that an elevation in ECF Pi will usually be the cause of supersaturation. When calcification involves the kidney, renal function can deteriorate rapidly, renal phosphorus clearance drops, and ECF Pi rises yet further, leading to a rapid downhill spiral. Under saturated conditions, susceptible tissue matrices will begin to accumulate CaHPO4 crystals, particularly if local pH rises above 7.4. Saturation of ECF with respect to calcium and phosphorus almost never occurs in individuals with normal renal function, mainly because urine phosphate excretion rises in direct proportion to dietary intake. As Figure 5-1 shows, the upper limit of the normal adult range for serum Pi typically occurs at absorbed intakes above 2.2 g (71 mmol)/day. At 62.5 percent absorption, that means ingested intakes above 3.4 g (110 mmol)/day. The 1994 CSFII data indicate that the reported intake at the ninety-fifth percentile was 2.5 g (81.7 mmol)/day in boys aged 14 through 18 years (see Appendix D). Hyperphosphatemia from dietary causes becomes a problem mainly in patients with end-stage renal disease or in such conditions as vitamin D intoxication. When functioning kidney tissue mass is reduced to less than ~20 percent of normal, the GFR becomes too low to clear typical absorbed loads of dietary phospho-

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride rus, and then even sharply reduced phosphorus diets may still be excessive as they lead to hyperphosphatemia. Although metastatic calcification can occur in patients with endstage renal disease in whom ECF Pi levels are not adequately controlled, it is not known to occur from dietary sources alone in persons with adequate renal function. For that reason, calcification in previously normal kidneys produced by high phosphorus intakes has been studied mainly in rats and mice (Craig, 1959; Hamuro et al., 1970; McFarlane, 1941; NRC, 1995). Production of calcification has required very high phosphate loads over and above the animals' already high basal phosphate intakes and in several reports has required partial reduction of renal tissue mass, as well. Skeletal Porosity Skeletal lesions associated with high phosphorus intakes have been described in rabbits (Jowsey and Balasubramaniam, 1972) and bulls (Krook et al., 1975). As with kidney toxicity, the bony lesions required extremely large phosphate intakes (in rabbits, about 40-fold typical human intakes on a body weight basis, and in bulls, feeding of a ration designed to support milk production in cows). None of these situations has any evident direct relevance to human nutrition or to human dietary intake of phosphate. Krook et al. (1975) also noted that bone loss develops in household pets and zoo animals fed human table scraps and meat. Despite acknowledging that such foods are poor in calcium, they attribute the bone loss to the high phosphorus content of such diets. Lacking evidence that phosphorus would produce this effect with diets adequate in calcium, this conclusion seems unwarranted. Finally, given the evidence cited above that high phosphorus intakes in humans do not lead to negative calcium balance or to increased bone resorption, it seems likely that the bone disease in other animals is more a consequence of low effective calcium intake than of high phosphorus intake per se. Interference with Calcium Absorption As noted, some concerns have been expressed that a high phosphorus intake could interfere with calcium nutrition by complexing calcium in the chyme and reducing its absorption (Calvo and Heath, 1988; Calvo and Park, 1996). Given the relative absorption efficiencies of calcium and phosphorus, there would not be a stoichiometric excess of phosphorus relative to calcium in the chyme until the Ca:P intake ratio fell below 0.375:1. However, even this is a purely theoretical concern. In the studies of Spencer et al. (1978a), in

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride which inorganic neutral phosphate was added to the diet, and of Heaney and Recker (1982), who studied women on their habitual intakes of food phosphorus, even Ca:P ratios as low as 0.08:1 did not lower calcium absorption. Nevertheless, it must be noted that it is more difficult for the body to compensate for impaired calcium absorption at low dietary calcium intakes compared with higher intakes (Heaney, 1997). As prior expert panels have noted (Chinn, 1981), even the theoretical potential for interference with the calcium economy by high phosphorus intakes is effectively negated if calcium intake is adequate. Dose-Response Assessment Adults: Ages 19 through 70 Years Data Selection. A UL can be defined as an intake associated with the upper boundary of adult normal values of serum Pi. No reports exist of untoward effects following high dietary phosphorus intakes in humans. Essentially all instances of dysfunction (and, hence, all instances of hyperphosphatemia) in humans occur for nondietary reasons (for example, end-stage renal disease, vitamin D intoxication). Therefore, data on the normal adult range for serum Pi are used as the basis for deriving a UL for adults. Identification of a No-Observed-Adverse-Effect Level (NOAEL) (or Lowest-Observed-Adverse-Effect Level [LOAEL]) and Critical Endpoint. If the normal adult range for serum Pi is used as a first approach to estimating the UL, the upper boundary of adult normal values of serum Pi is reached at a daily phosphorus intake of 3.5 g (113 mmol) (Figure 5-1). There is no evidence that individuals consuming this intake may experience any untoward effects. As shown in Table 5-1, infants, children, and adolescents have higher upper limits for serum Pi than do adults, which indicates that their tissues tolerate the higher Pi levels well. Values of Pi above the nominal adult human normal range are also normally found in typical adult laboratory animals (for example, rats) and occur regularly in adult humans treated with the bisphosphonate, etidronate, used for the treatment of Paget's disease of the bone and osteoporosis (Recker et al., 1973). No suggestion of harm comes from any of these situations, indicating that the UL is substantially higher than that associated with the upper normal bound of serum Pi in adults. The higher values for serum Pi in infancy are manifestly tissue-safe levels, and if they are taken as an approximation of the upper normal

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride human value (on the ground that there is no basis for assuming major differences in tissue susceptibility to metastatic mineralization at different ages), the corresponding ingested intake in an adult (assuming the relationship of Figure 5-1) would be over 10.2 g (330 mmol)/day. Uncertainty Assessment. No benefit is evident from serum Pi values above the usual normal range in adults. Moreover, information is lacking concerning adverse effects in the zone between normal Pi and levels associated with ectopic mineralization. Therefore, in keeping with the pharmacokinetic practice where the relationship between intake and blood level is known (Petley et al., 1995), an uncertainty factor (UF) of 2.5 is chosen. Derivation of the UL. A UL of ~4.0 g (~130 mmol)/day for adults is calculated by dividing a NOAEL of 10.2 g (330 mmol)/day by a UF of 2.5. UL for Adults 19 through 70 years 4.0 g (130 mmol)/day Infants: Ages 0 through 12 Months As with adults, there are essentially no reports of adverse effects clearly attributable to high phosphorus intake of dietary origin in infants, children, or adolescents. Except for the sensitivity of very young infants noted above, there are no data relating to adverse effects of phosphorus intake for most of the first year of life. For that reason, it was determined that it was impossible to establish a specific UL for infants. UL for Infants 0 through 12 months Not possible to establish; source of intake should be from formula and food only Toddlers and Children: Ages 1 through 8 Years For toddlers and children, a UL of 3.0 g (96.8 mmol)/day is calculated by dividing the NOAEL for adults (10.2 g [330 mmol]/day by a UF of ~3.3 to account for potentially increased susceptibility due to smaller body size. UL for Children 1 through 8 years 3.0 g (96.8 mmol)/day

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride Adolescents: Ages 9 through 18 Years There is no evidence to suggest increased susceptibility to adverse effects during adolescence. Therefore, the same UL specified above for adults is selected for adolescents, 4.0 g (130 mmol)/day. UL for Adolescents 9 through 18 years 4.0 g (130.0 mmol)/day Older Adults: Ages > 70 Because of an increasing prevalence of impaired renal function after age 70, a larger UF of 3.3 seems prudent, and the UL for adults of this age is set at 3.0 g (96.8 mmol)/day. UL for Older Adults > 70 years 3.0 g (96.8 mmol)/day Pregnancy and Lactation During pregnancy, absorption efficiency for phosphorus rises by about 15 percent, and thus, the UL associated with the upper end of the normal range will be about 15 percent lower, that is, about 3.5 g (112.9 mmol)/day. During lactation, the phosphorus economy of a woman does not differ detectably from the nonlactating state. Hence the UL for this physiologic state is not different from the nonlactating state, 4.0 g (130 mmol)/day. UL for Pregnancy 14 through 50 years 3.5 g (112.9 mmol)/day UL for Lactation 14 through 50 years 4.0 g (130.0 mmol)/day Special Considerations It is recognized that population groups such as professional athletes, military trainees, or those whose level of energy expenditure exceeds 6,000 kcal/day, may have dietary phosphorus intakes whose distributions overlap these limits. In such individuals with phosphorus intakes above the UL, no harm is known to result. Exposure Assessment Utilizing the 1994 CSFII data, adjusted for one day food records (Nusser et al., 1996), the highest mean intake of phosphorus for any age and life stage group was reported for males aged 19 through 30 years: 1.7 g (53.5 mmol)/day. The highest reported intake at the

OCR for page 146
DRI DIETARY REFERENCE INTAKES FOR Calcium, Phosphorus, Magnesium, Vitamin D, and Fluoride ninety-fifth percentile was 2.5 g (81.7 mmol)/day in boys aged 14 through 18 years (see Appendix D), which is well below the UL of 4.0 g (130 mmol)/day. In 1986, approximately 10 percent of adults in the United States took a supplement containing phosphorus (Moss et al., 1989), and of those, the ninetieth percentile of supplemental phosphorus intake was 264 mg (8.5 mmol)/day. The ninety-fifth percentile intake for phosphorus supplements for adults was 448 mg (14.5 mmol)/day. Risk Characterization Phosphorus exposure data indicate that only a small percentage of the U.S. population is likely to exceed the UL. Because phosphorus supplements are not widely consumed, nor is the dosage high, total intake from diet plus supplements would infrequently exceed the UL. RESEARCH RECOMMENDATIONS The model that relates absorbed phosphorus intake to serum phosphorus must be evaluated in clinical studies using oral phosphorus intakes, and investigated in children and adolescents as well as adults. Bone mineral mass as a function of dietary phosphorus intake should be investigated at all stages of the life cycle. The practical effect of phosphate-containing food additives on trace mineral status (iron, copper, and zinc) should be evaluated.