Contents - Previous - Next

Adequacy of dietary mineral supply for human bone growth and mineralisation

1. Body content, biological role and childhood accretion rates
2. Likely manifestations of mineral deficiencies in children
3. Dietary intakes of children in developing countries
4. Supplementation studies

A. Prentice and C.J. Bates

Correspondence to: A. Prentice, MRC Dunn Nutrition Unit, Downhams Lane, Milton Road, Cambridge CB4 1XJ, UK.

MRC Dunn Nutrition Unit, Cambridge, UK, and Keneba, The Gambia

The evidence on the relationship between dietary mineral supply and bone development in children has been extensively reviewed. Data from children and primates suggest that overt deficiencies of Ca, P and Zn are likely to produce rickets and growth retardation, while the effects of Mg deficiency on human bone are unknown. The manifestations of marginal deficiencies are little understood. The biological needs for Ca, P. Mg and Zn in childhood have been calculated based on mineral deposition rates, using published values for the mineral content of the human body, and on obligatory endogenous losses. As a rough guide, the estimated biological requirements for the Ca, P. Mg and Zn can be taken as 200, 100, 4 and 1 mg/d respectively. A comparison of measured daily intakes of children in developing countries with biological requirements was made. This revealed that P and Mg intakes were many times higher than estimated needs. Ca intakes at all ages were found to be close to the biological requirement for children in many Third World societies, before any allowance for possible poor absorption. Zn intakes approach estimated needs in breast-fed infants, particularly during weaning, but are 4-5 times higher in older children. Poor absorption from phytate-rich diets could affect Zn supply. Supplementation studies indicate that raising Zn intakes can increase height gains in certain vulnerable groups, such as infant and adolescent boys. In conclusion, the evidence suggests that inadequate dietary intakes of Ca and Zn may contribute to linear growth retardation in children of developing countries but more research is needed.

We were asked to review the evidence on whether "deficiency or malabsorption of calcium, phosphorus, magnesium or zinc (Ca, P. Mg, Zn) affect linear growth and deposition of bone mineral in the human, given that the diets of Third World children tend to be low in these elements, even when breast-fed, and that this may be an important contributory factor in their poor growth performance."

This statement raises many fundamental questions that need to be examined:

1. What are the biological requirements of Ca, P. Mg and Zn for normal growth in the human?

2. What are the likely manifestations of an inadequate supply of Ca, P. Mg and Zn in the growing child? Do children in developing countries show signs that could be attributed to mineral deficiencies?

3. Are the mineral intakes of children in developing countries low in relation to the biological requirement or in comparison with well-nourished children in developed countries? Are diet composition and illness likely to affect mineral bioavailability? What is the contribution of breast-milk to mineral intakes in early childhood?

4. What evidence do we have that increasing the intakes of these minerals would improve growth or bone development in Third World children?

In the following paper, each of these questions will be discussed in detail and the evidence that is currently available will be reviewed.

1. Body content, biological role and childhood accretion rates

Table 1 provides details of the body content of Ca, P. Mg and Zn in a new-born baby born at term, a typical man and a typical woman. As can be seen, an adult contains approximately 1 kg and 0.5 kg of Ca and P respectively, while Mg and Zn are present in smaller quantities.

The body compartments of the four minerals are summarised in Table 1. Approximately 99% of calcium and 80% of phosphorus in the body are contained in the inorganic phase of bone and teeth, imparting structure and strength. The crystal structure of bone salt resembles hydroxyapatite [Ca10(PO4)6(OH)2], which contains Ca and P in the proportion 2.15:1 g/g (Russell et al., 1986). However, crystallisation of bone salt occurs in several stages, proceeding from amorphous calcium phosphate through intermediate crystalline structures, such as octocalcium phosphate (Russell et al., 1986). These compounds have lower Ca:P ratios than hydroxyapatite and consequently the proportion of Ca to P in young bone is generally between 1.7:1 - 2.14:1 g/g (Specker & Tsang, 1987). The measured ratio in adult human bone ash is around 2.3:1 g/g (Mitchell et al., 1945).

Bone salt is not pure hydroxyapatite, since it contains anions, such as carbonate and citrate, and cations, such as Mg and Zn (Widdowson & Dickerson, 1964). These ions either substitute within the crystal lattice or are absorbed onto the crystal surface (Russell et al. 1986). Approximately 60% of body Mg and 30% of body Zn are present in the skeleton where their concentrations are higher than in other tissues in the body (Department of Health, 1991; Schwartz, 1990). Zn is associated with alkaline phosphatase at calcification sites and is also deposited within the inorganic matrix (Hambidge, Casey & Krebs, 1986). The function of Mg and Zn in bone is largely unknown. Mg may play a role in the control of crystal formation and in crystal stability (Schwartz, 1990) while Zn is thought to be involved in chondrogenesis, collagen synthesis, osteoblastic function and calcification (Hambidge, Casey & Krebs, 1986).

All four minerals have important functions outside the bone compartment and are widely distributed throughout the soft tissues and fluids (Department of Health, 1991). The remaining 1% of total body Ca is involved in processes such as nerve and muscle function, blood-clotting and intracellular signalling. Non-osseous P is a component of many essential compounds, such as phospholipids and those with high-energy phosphate bonds like ATP. Non-skeletal Mg is involved in DNA replication, RNA synthesis, and is a co-factor for enzymes requiring ATP. Zn is essential for cell division, nucleic acid and protein synthesis and is a component of many enzymes. Unlike the other three minerals, the major portion of total body Zn occurs not in bone but in the soft tissues, primarily in muscle (Table 1), although the concentration of Zn in bone is high.

Table 1. Whole body mineral content and compartments of calcium, phosphorus, magnesium and zinc in the human a


Baby b (gm)


Body compartment

Male c (gm)

Female d (gm)

Bones (%)

Soft tissues (%)

























To convert mg to mmol divide by 40, 31, 24.3, 65.4 for Ca, P, Mg, Zn respectively.
a Chemical data from Widdowson & Dickerson (1964).
b Based on 3.5 kg full-term infant.
c Based on 60 kg fat-free mass (e.g. man 70 kg body weight, 15% fat).
d Based on 45 kg fat-free mass (e.g. woman 60 kg body weight, 25% fat).

Considerable quantities of all four minerals are deposited in the body between birth and maturity. The accretion of mineral is greater than the increase in body weight over the same period. For example, the amount of Ca expressed relative to body weight increases from approximately 8 g Ca/kg at birth to 19 g Ca/kg in a 70 kg man (Table 1; Widdowson & Dickerson, 1964). Table 2 gives estimated values, based largely on the compositional data in Table 1, for mineral accretion rates during childhood. The continuous rates have been obtained by assuming that maturity is reached by 18 years of age in both sexes and that the accretion rate is constant throughout childhood (British Nutrition Foundation, 1989). Such calculations indicate, for example, that a boy who at maturity has a fat-free mass of 60 kg (equivalent to a 70 kg man with 15% body fat) has to retain 200 mg Ca, 107 mg P, 4 mg Mg and 0.25 mg Zn every day for eighteen years in order to achieve the required mineral deposition (Table 2). In reality, of course, growth is not uniform and is greatest soon after birth and during adolescence. At these times accretion rates will be considerably higher than average, while rates will be somewhat less in the intervening years. In addition, mineral deposition during periods of catch-up growth, when children are recovering from illness or malnutrition, will be substantially above average. Estimates of likely accretion rates in infancy and in adolescence are given in Table 2, based on the arguments in British Nutrition Foundation (1989), Kanis & Passmore (1989), Fomon (1974) and Leitch & Aitken (1959). The values in infancy for Ca, P, Mg are somewhat lower than the continuous accretion rates due to the relatively low mineral content of young bone (Fomon, 1974).

In addition to the requirements for growth, losses of the four minerals occur in urine, sweat, gastrointestinal fluids, skin, hair and nails. Quantitative data on mineral losses in infants and children are very limited. There is evidence, however, that mineral losses are greatly reduced in individuals habituated to very low intakes (Widdowson & Dickerson, 1964; Begum & Pereira (1969); Nicholls & Nimalasuriya, 1939; Taylor et al. 1991) and it is unclear what figures should be used as estimates of obligatory losses. The British COMA Committee in their recent evaluation of dietary reference values assumed that there are no obligatory Ca losses in children and gave no figures for P or Mg (Department of Health, 1991). Endogenous Zn losses in infants have been estimated at 0.07 mg/kg/d (Ziegler et al. 1989) in faeces and 0.02 mg/kg/d in urine and sweat (Krebs & Hambidge, 1989), producing a total estimated requirement for accretion + losses of 0.9-1.2 mg/d (Department of Health, 1991; King & Turnlund, 1988).

The number of studies on the composition of the human body are very limited, involving the chemical analyses of only a small number of possibly atypical individuals (Widdowson & Dickerson, 1964; British Nutrition Foundation, 1989). In addition, differences in chemical composition may exist between the races; for example adult Blacks in the United States are known to have higher total body Ca and P content than Whites of the same height (Cohn et al. 1977). The figures in Tables 1 and 2, therefore, can be used only to provide an approximate assessment of mineral deposition during childhood. Similarly, the data on endogenous losses in children are insecure and there are likely to be considerable variations between individuals. However, these data provide a useful basis on which to discuss the likely adequacy of dietary supply for children in Third World countries, and for this purpose the following figures, based on the continuous accretion rate for boys (Table 1) + losses for Zn, are a useful rough guide: Ca 200 mg/d, P 100 mg/d, Mg 4 mg/d, Zn 1 mg/d. These figures will be referred to in the rest of the text as the 'biological requirement'.

Table 2. Estimated mineral accretion rates in childhood


Continuous a

Infancy b

Peak c



0-4 months

4-12 months

Calcium mg/d






Phosphorus mg/d






Magnesium mg/d






Zinc mg/d






a Average accretion rate in childhood based on assumption of continuous growth and maturity at eighteen years in both sexes (British Nutrition Foundation, 1989).

b Accretion in infancy as calculated in Fomon (1974) and Krebs & Hambidge (1986).

c Peak rate in adolescence based on calcium calculation of Kanis & Passmore (1989) and assuming proportions of Ca P, Mg, Zn are the same as during continuous growth.

2. Likely manifestations of mineral deficiencies in children

It is well known that severe protein-energy malnutrition in children results in linear growth retardation and reduced bone mineral contents which resemble juvenile osteoporosis (Garn et al. 1964; Adams & Berridge, 1969). Skeletal maturation, as assessed by the appearance of ossification centres, is sometimes delayed (Adams & Berridge, 1969). In contrast, very little is known about the manifestations of specific deficiencies of bone-forming minerals in the human. As all the Ca and P needed for building bones must originate from the diet, and as there are no significant extra-skeletal reservoirs of these minerals, there must be intakes which cannot support normal growth and bone development.

There is evidence to suggest that very low Ca intakes in children may induce rickets and osteomalacia (Pettifor, 1991). This is based on a number of case reports from the United States and South Africa of infants and children presenting with radiological rickets, growth retardation and biochemical signs of hyperparathyroidism, but with normal vitamin D status (Pettifor, 1991; Maltz, Fish & Holliday, 1970; Kooh et al., 1977; Pettifor et al., 1978; 1981b). The previous diets of these children were very low in Ca, and all the children responded to Ca-rich hospital diets. Iliac crest biopsies from some of the South African children revealed osteopenia (undermineralisation), as characterised by a reduced amount of calcified trabecular bone, severe osteomalacia, as evidenced by increased osteoid volume, surface and thickness, a reduced calcification front and a prolonged mineralisation lag time, and lesions suggestive of hyperparathyroidism, such as increased numbers of osteoblasts and enhanced osteoclastic bone resorption (Pettifor, 1991; Marie et al., 1982). The mechanism by which Ca deficiency may cause rickets is unknown but may involve increased catabolism of 25-hydroxy vitamin D in the liver (Fraser, 1988a, b; Clements, Johnson & Fraser, 1987).

Detailed experiments with young vitamin D-replete baboons fed diets low in Ca but adequate in P for 8-16 months produced mild radiological rickets and histological features of osteomalacia, such as increases in growth plate thickness, osteoid seam thickness, volume and surface, and delays in calcification rate and mineralisation lag time (Pettifor et al., 1984). Biochemical abnormalities suggestive of hyperparathyroidism, such as a transient fall in serum Ca and raised alkaline phosphatase levels, also developed. However, similar experiments with juvenile cinnamon ringtail monkeys produced no radiological, biochemical or histological effects of a low Ca but adequate P diet over a 7 year period (Anderson et al., 1977). The observations in children and baboons are in contrast to the situation with growing laboratory rats where calcium deficiency leads to osteoporosis with no impairment of linear growth until the bone mineral content is less than 50% of normal (Fraser, 1988a; Moore et al., 1963). Young mice appear to be less sensitive to a low Ca diet than rats (Ornoy, Wolinsky & Guggenheim, 1974).

There is evidence that phosphorus deficiency also precipitates rickets in children: the rickets-like metabolic bone disease of premature babies is currently believed to be primarily a problem of phosphorus supply (Pettifor, 1991; Bishop, 1989). It is characterised by hypophosphataemia, raised serum alkaline phosphatase levels, low urinary phosphorus excretion and a reduced ability to retain Ca as shown by hypercalciuria. The disorder can be ameliorated largely by phosphate supplementation (Pettifor, 1991; Bishop, 1989).

It is possible that the effects of Ca and P deficiencies may arise from an imbalance of the two bone-forming minerals in the diet. The children with rickets in the United States and South Africa had diets low in Ca but adequate or high in P (Maltz, Fish & Holliday, 1970; Kooh et al., 1977; Pettifor et al., 1978, 1981b). Diets containing low ratios of Ca:P are known to produce secondary hyperparathyroidism in young horses, dogs and other animals (Marie et al., 1982; Fraser, 1988a). Young baboons fed low Ca and low P diets for 8-16 months did not exhibit the histological or biochemical abnormalities that were observed in animals fed low Ca but adequate P diets (Pettifor et al., 1984). However, studies in young ringtail monkeys showed that altering the Ca:P ratio in the range 1:4 to 1:0.4 mg/mg had no consequences (Anderson et al. 1977).

The effects of marginally low Ca intakes in children are difficult to assess. Studies in South African village children showed that a significant number of individuals had biochemical signs of hyperparathyroidism (low serum Ca, raised alkaline phosphatase) (Pettifor et al., 1979; Eyberg, Pettifor & Moodley, 1986) which normalised after calcium supplementation (Pettifor et al. 1981a; see section 4). Raised alkaline phosphatase levels have also been observed in underprivileged Brazilian children (Linhares, Round & Jones, 1986). South African children aged 8-16 years with low Ca intakes and low serum Ca levels have reduced forearm bone densities and metacarpal cortical thickness compared with other children in the community of the same age, weight and height (Eyberg, Pettifor & Moodely, 1986). Low forearm bone mineral contents relative to body size have also been noted in Gambian children (Prentice et al., 1990; Lo et al., 1990). However, chemical analysis of skeletons from children and adults in Ceylon did not reveal marked differences in the amount of Ca per unit dry weight of bone (Nicholls & Nimalasuriya, 1939) and no indications that low Ca intakes were associated with bone rarefaction or reduced compact bone width were noted in X-rays of children in Surinam (Luyken & Luyken-Konig, 1969). Rickets, however, is a common problem in many tropical countries and in children consuming macrobiotic diets in Western countries, and it has been speculated that low calcium intakes may be a predisposing factor (Fraser, 1991; Dagnelie et al. 1990).

Marginal P intakes have been implicated in the incidence of bladder stones in young children of Northeast Thailand (Valyasevi, Dhanamitta & Van Reen, 1969). Infants and children living in a stone endemic area excrete low levels of P and high amounts of Ca in their urine and crystalluria is common (Valyasevi & Dhanamitta, 1967; Valyasevi et al., 1967), features which were ameliorated with phosphate supplementation (Valyasevi, Dhanamitta & Van Reen, 1969). Breast-milk contains relatively low amounts of P, with a Ca:P ratio of approximately 2:1 mg/mg. Phosphate supplementation of breast-fed children during the first two weeks of life has been shown to enhance Ca, P and Mg retentions (Widdowson et al., 1963). However, prior to the development of modern formula milks, excessive phosphorus intakes by neonates given feeds based on unmodified cow's milk were associated with hyperphosphataemia, hypocalcaemia and tetany (Oppe & Redstone, 1968).

Severe magnesium deficiency in man is characterised by muscle weakness, neuromuscular dysfunction and cardiac disturbances (Department of Health, 1991), normally in association with debilitating diseases such as diabetes and alcoholism (Department of Health, 1991; Wacker & Vallee, 1964). Failure to thrive, growth retardation, bone abnormalities and disturbances of Ca metabolism have been described in Mg-depleted animals (Department of Health, 1991; Schwartz, 1990). The Mg content in bone of growing Mg-deficient animals can be 80% below normal, although calcium contents are often increased (Schwartz, 1990).

In contrast to Ca, P and Mg, the consequences of human Zn deficiency have been well documented due to problems arising from acrodermatitis enteropathica (an inherited disorder affecting Zn absorption), sickle-cell anaemia, chronic renal disease and other conditions (Aggett, 1988). Moderate-severe Zn deficiency in children depresses growth, appetite, skeletal maturation and gonad development which can be reversed with Zn treatment (Aggett, 1988; Prasad, 1991). Zn deficiency is associated with metabolic disturbances of a wide range of hormones, cytokines and enzymes involved in growth and bone development [e.g. IGF-1 (somatomedin C), growth hormone, thyroid hormone, insulin, prolactin, alkaline phosphatase and prostaglandins] (Hambidge, Casey & Krebs, 1986). In addition, Zn deficiency affects the immune system, the structure of the skin and intestinal mucosa, taste perception, wound healing, and dark adaptation (Department of Health, 1991; Hambidge, Casey & Krebs, 1986; Aggett, 1988; Prasad, 1991; O'Dell & Reeves, 1988). Children with severe malnutrition show clinical signs and immunological deficits which are correctable by Zn (Golden et al., 1978; Golden & Golden, 1979; see section 4). Whether the effects of Zn deficiency on growth and bone development are a direct consequence of inadequate Zn supply for bone formation or are secondary to the effects of Zn on appetite, the action of growth-promoting factors or cell division is not known (Leek et al., 1988).

Marginal Zn deficiency in rhesus monkeys during gestation and postnatal life has been shown to cause growth retardation, delayed skeletal maturation and defective mineralisation (Leek et al., 1984; 1988). The bone abnormalities, which were most severe at 6 months of age and gradually improved thereafter, were similar to those seen in human rickets (Leek et al. 1984; 1988).

There have been a number of observations that suggest that Zn deprivation may be implicated in human growth retardation. Adolescent nutritional dwarfism in Middle-Eastern countries, characterised by poor growth and delayed sexual maturity, has been related to Zn deficiency in association with deficiencies of other nutrients (Prasad, 1991). In addition, poor Zn status, as suggested by low Zn levels in blood or hair, has been described in growth-retarded Chinese, Mexican, Thai and Papua New Guinean children among others (Chen et al., 1985; Udomkesmalee et al., 1990; Gibson et al. 1991).

Contents - Previous - Next