Clin Blochem, Vol. 23, pp. 3-10. 1990 P r i n t e d in Canada. All r i g h t s reserved,
0009-9120/90 $3.00 - .00 C o p y r i g h t ¢ 1990 T h e C a n a d i a n Society of C l i n i c a l C h e m i s t s .
Calcium Homeostasis B. E. C. NORDIN Division of Clinical Chemistry, Institute of Medical and Veterinary Science, Frome Road, Adelaide, South Australia The regulation of plasma calcium concentration and bone homeostasis is briefly described. The components of plasma calcium and the method for deriving these components are presented. The factors which regulate plasma calcium concentration are discussed with particular reference to the role of the kidney. The mechanisms responsible for the maintenance of the skeleton are described and the failure of these mechanisms in the pathogenesis of osteoporosis is outlined. In postmenopausal women there is increased excretion of calcium and recent works suggest that this is due to an increase in the complexed fraction of calcium as a result of an increase in plasma bicarbonate and anion gap. The significance of the increase in urine calcium excretion in postmenopausal women and its impact on dietary calcium requirement are discussed.
KEY WORDS: calcium; osteoporosis; menopause; absorption.
Introduction his review will focus on two aspects of calcium homeostasis -- plasma calcium homeostasis and bone homeostasis. Other aspects of calcium homeostasis, including the regulation of intracellular calcium, are outside the scope of this review. The methodology has been described in detail elsewhere (1-3!. References to clinical cases and normal volunteers will be made at various points. The definition of such cases will become a p parent in the text; a distinction will be made t h r o u g h o u t between normal volunteers and subjects who, though attending a metabolic clinic, have been found to be normal from the bone point of view and have therefore been tr eate d as controls in certain situations.
T
Plasma calcium homeostasis It is well known t h a t the plasma calcium concentration, and therefore the extracellular calcium, is m a i n t a i n e d within very narrow limits. The plasma calcium is composed of three fractions -- proteinbound, complexed and ionized -- which are in a state of internal equilibrium. The ionized calcium is assumed to be the "active" fraction which regulates
Correspondence: Prof. B. E. C. Nordin, Division of Clinical Chemistry, Institute of Medical and Veterinary Science, Frome Road, Adelaide, South Australia. Manuscript received February 3, 1989; revised July 31, 1989; accepted August 15, 1989. CLINICAL BIOCHEMISTRY, VOLUME 23, FEBRUARY 1990
and is regulated readily by parathyroid gland activity. The magnitude of the other fractions follows the usual mass action equilibria whereby the amounts bound to the various ligands are determined by the ionized calcium concentration on the one hand, and the concentration of the rel evant anions on the other. If the ionized calcium is held constant, the total calcium concentration is determined by the concentrations of albumin, globulin, bicarbonate and various organic acids, whereas if the anion concentrations are held constant then the total calcium is a simple function of the ionized calcium. In the normal situation, the ionized calcium is virtually 50% of the total, the complexed calcium 5-10% of the total, and the protein-bound calcium 40-45% of the total. Of these the complexed fraction can only be calculated as the difference between the ionized and ultrafiltrable components, has never attracted much attention, and is not well characterised. The operation of mass action equilibria makes it erroneous to correct the plasma calcium for ligand concentrations by adding or subtracting a fixed a m o u n t for each unit of ligand concentration. Yet this is widely done. Since albumin is the main single binder of calcium in the plasma, it is a common practice to add or subtract to the plasma calcium 0.02 mmol/L × (44 - plasma albumin concentration) (4). This may be acceptable in a data set in which all the ionized calciums are equal -- or fall within a vary narrow range -- and the main determ i n a n t of the variance in the set is the ligand concentrations, in particular the albumin. But it is not correct when one is dealing -- as is so often the case -- with a set of observations in subjects with varyi ng albumin and ionized calcium concentrations. The albumin (and the other rel ev a n t anions) do not bind a fixed amount of calcium per unit but a relatively constant proportion of the total calcium available. When the ionized calcium is high, the albumin being far below its sat urat i on level binds more calcium per gram t h a n when the ionized calcium is low. It follows t h a t if the plasma calcium is to be corrected for plasma albumin (at best a very partial correction) it should be done by assuming t h a t a certain proportion of the total r a t h e r t h a n a certain amount is bound to each gram of albumin. These considerations are exemplified in Table 1. In a set of 556 normal postmenopausal women, total plasma calcium was regressed on albumin, globulin,
NORDIN TABLE 1 Multiple Regression Analysis of Plasma Calcium on Plasma Ligands in 556 Normal Postmenopausal Women t
Plasma calcium =
0.015 x albumin (g/L) +0.0059 x globulin (g/L) +0.0091 x bicarbonate (retool/L) +0.010 x anion gap (retool/L) + 1.21 mmol/L
anion gap and bicarbonate..The regression yielded a highly significant correlation and highly significant regression coefficients for each of the four potential ligands. In this homogeneous set, in which ionized calcium might be presumed to be relatively constant, the coefficients represent the actual amount of calcium bound per unit of ligand. The regression coefficient for albumin is very close to the widely used value of 0.02 mmol/L (4). But it is clear t h a t there is much more binding of calcium t h a n can be accounted for by albumin alone and therefore correction for albumin, even if done correctly, would be quite inadequate. The residual constant is 1.21 mmol/L which probably represents the mean ionized calcium concentration in this set. These data suggest t h a t the calcium fractions might be calculated from a knowledge of total calcium, albumin, globulin, anion gap and bicarbonate concentration. For this purpose we have developed a formula based on the assumption t h a t each unit of ligand will bind a certain proportion of the total calcium. The proportional coefficients were arrived at by dividing the coefficients in Table 1 by the mean total plasma calcium of 2.42 mmol/L. Thus, the binding of calcium by albumin amounts to 0.00613 of the total calcium per gram of albumin per litre. The same transformations applied to the other coefficients in Table 1 yield the formula for calculation of ionized calcium shown in Table 2 (3). This formula has been applied to the plasma analyses in 105 consecutive inpatients in whom the ionized calcium was also measured. The calculated values were then regressed on the measured values to yield the comparison shown in Figure 1. The regression coefficient was 1.00 and the intercept - 0.005. The coefficient of correlation was 0.84 (p < 0.001). Since these observations appeared to validate the formula, it was then used to calculate the
12.1 7.6 5.8 6.6
p
0009-9120/90 $3.00 - .00 C o p y r i g h t ¢ 1990 T h e C a n a d i a n Society of C l i n i c a l C h e m i s t s .
Calcium Homeostasis B. E. C. NORDIN Division of Clinical Chemistry, Institute of Medical and Veterinary Science, Frome Road, Adelaide, South Australia The regulation of plasma calcium concentration and bone homeostasis is briefly described. The components of plasma calcium and the method for deriving these components are presented. The factors which regulate plasma calcium concentration are discussed with particular reference to the role of the kidney. The mechanisms responsible for the maintenance of the skeleton are described and the failure of these mechanisms in the pathogenesis of osteoporosis is outlined. In postmenopausal women there is increased excretion of calcium and recent works suggest that this is due to an increase in the complexed fraction of calcium as a result of an increase in plasma bicarbonate and anion gap. The significance of the increase in urine calcium excretion in postmenopausal women and its impact on dietary calcium requirement are discussed.
KEY WORDS: calcium; osteoporosis; menopause; absorption.
Introduction his review will focus on two aspects of calcium homeostasis -- plasma calcium homeostasis and bone homeostasis. Other aspects of calcium homeostasis, including the regulation of intracellular calcium, are outside the scope of this review. The methodology has been described in detail elsewhere (1-3!. References to clinical cases and normal volunteers will be made at various points. The definition of such cases will become a p parent in the text; a distinction will be made t h r o u g h o u t between normal volunteers and subjects who, though attending a metabolic clinic, have been found to be normal from the bone point of view and have therefore been tr eate d as controls in certain situations.
T
Plasma calcium homeostasis It is well known t h a t the plasma calcium concentration, and therefore the extracellular calcium, is m a i n t a i n e d within very narrow limits. The plasma calcium is composed of three fractions -- proteinbound, complexed and ionized -- which are in a state of internal equilibrium. The ionized calcium is assumed to be the "active" fraction which regulates
Correspondence: Prof. B. E. C. Nordin, Division of Clinical Chemistry, Institute of Medical and Veterinary Science, Frome Road, Adelaide, South Australia. Manuscript received February 3, 1989; revised July 31, 1989; accepted August 15, 1989. CLINICAL BIOCHEMISTRY, VOLUME 23, FEBRUARY 1990
and is regulated readily by parathyroid gland activity. The magnitude of the other fractions follows the usual mass action equilibria whereby the amounts bound to the various ligands are determined by the ionized calcium concentration on the one hand, and the concentration of the rel evant anions on the other. If the ionized calcium is held constant, the total calcium concentration is determined by the concentrations of albumin, globulin, bicarbonate and various organic acids, whereas if the anion concentrations are held constant then the total calcium is a simple function of the ionized calcium. In the normal situation, the ionized calcium is virtually 50% of the total, the complexed calcium 5-10% of the total, and the protein-bound calcium 40-45% of the total. Of these the complexed fraction can only be calculated as the difference between the ionized and ultrafiltrable components, has never attracted much attention, and is not well characterised. The operation of mass action equilibria makes it erroneous to correct the plasma calcium for ligand concentrations by adding or subtracting a fixed a m o u n t for each unit of ligand concentration. Yet this is widely done. Since albumin is the main single binder of calcium in the plasma, it is a common practice to add or subtract to the plasma calcium 0.02 mmol/L × (44 - plasma albumin concentration) (4). This may be acceptable in a data set in which all the ionized calciums are equal -- or fall within a vary narrow range -- and the main determ i n a n t of the variance in the set is the ligand concentrations, in particular the albumin. But it is not correct when one is dealing -- as is so often the case -- with a set of observations in subjects with varyi ng albumin and ionized calcium concentrations. The albumin (and the other rel ev a n t anions) do not bind a fixed amount of calcium per unit but a relatively constant proportion of the total calcium available. When the ionized calcium is high, the albumin being far below its sat urat i on level binds more calcium per gram t h a n when the ionized calcium is low. It follows t h a t if the plasma calcium is to be corrected for plasma albumin (at best a very partial correction) it should be done by assuming t h a t a certain proportion of the total r a t h e r t h a n a certain amount is bound to each gram of albumin. These considerations are exemplified in Table 1. In a set of 556 normal postmenopausal women, total plasma calcium was regressed on albumin, globulin,
NORDIN TABLE 1 Multiple Regression Analysis of Plasma Calcium on Plasma Ligands in 556 Normal Postmenopausal Women t
Plasma calcium =
0.015 x albumin (g/L) +0.0059 x globulin (g/L) +0.0091 x bicarbonate (retool/L) +0.010 x anion gap (retool/L) + 1.21 mmol/L
anion gap and bicarbonate..The regression yielded a highly significant correlation and highly significant regression coefficients for each of the four potential ligands. In this homogeneous set, in which ionized calcium might be presumed to be relatively constant, the coefficients represent the actual amount of calcium bound per unit of ligand. The regression coefficient for albumin is very close to the widely used value of 0.02 mmol/L (4). But it is clear t h a t there is much more binding of calcium t h a n can be accounted for by albumin alone and therefore correction for albumin, even if done correctly, would be quite inadequate. The residual constant is 1.21 mmol/L which probably represents the mean ionized calcium concentration in this set. These data suggest t h a t the calcium fractions might be calculated from a knowledge of total calcium, albumin, globulin, anion gap and bicarbonate concentration. For this purpose we have developed a formula based on the assumption t h a t each unit of ligand will bind a certain proportion of the total calcium. The proportional coefficients were arrived at by dividing the coefficients in Table 1 by the mean total plasma calcium of 2.42 mmol/L. Thus, the binding of calcium by albumin amounts to 0.00613 of the total calcium per gram of albumin per litre. The same transformations applied to the other coefficients in Table 1 yield the formula for calculation of ionized calcium shown in Table 2 (3). This formula has been applied to the plasma analyses in 105 consecutive inpatients in whom the ionized calcium was also measured. The calculated values were then regressed on the measured values to yield the comparison shown in Figure 1. The regression coefficient was 1.00 and the intercept - 0.005. The coefficient of correlation was 0.84 (p < 0.001). Since these observations appeared to validate the formula, it was then used to calculate the
12.1 7.6 5.8 6.6
p
