Unlike most other mammalian milk, human milk is exceptionally low in phosphorus content. In infants, there are three reasons to suggest utilization of formulas that provide low but adequate phosphorus intake and a balanced ratio of calcium to phosphorus: (1) a low intestinal phosphorus concentration is an essential condition for generating an acid pH of the feces, which inhibits the growth of potentially pathogenic organisms; (2) the immature infantile kidney cannot readily excrete a surplus of phosphorus, and higher phosphorus intake during the neonatal period is a significant risk factor for hypocalcemic tetany in both preterm¹ and full-term infants²; and (3) high phosphorus and calcium intake in the newborn can result in metabolic acidosis secondary to a decreased renal capacity for acid excretion. It is thought that all three pathophysiologic mechanisms were effective in the biochemical evolution of humans, selecting women with low phosphorus milk and infants with a constantly high intestinal absorption rate of phosphorus.⁸

In comparison with human milk, the standard infant formulas prepared from cow milk have relatively high phosphorus and calcium concentrations and a relatively low calcium-to-phosphorus ratio. Studies have shown lower bone mineral content (BMC) of the radius throughout the first 6 months of life of infants fed human milk compared with cow milk formula.^9,10 No studies have been reported on the effect of varying mineral intakes on total body bone mass accretion during the first year of life, which should be a better indicator of mineral nutriture. In addition, no studies have been designed in such a way as to determine whether early feeding during the first 6 months of life has a persistent effect on differences in bone mass accretion.

In the current study, we hypothesized that during the first year of life a higher-mineral content formula would result in higher total body bone mass accretion. We used a factorial design to determine whether bone mass accretion during the first and second 6 months of life was influenced by the mineral intake of commercially available formula during those periods and whether bone mass accretion during the second 6 months differed according to the mineral intake during the first 6 months.

METHODS

In phase II, a second randomization schedule was used to determine the type of feeding from 6 to 12 months of age. All formula- and human milk-fed infants were randomized, after stratifying by initial feeding assignment to receive a moderate-mineral cow milk-based formula (Similac), a high-mineral cow milk-based formula (Carnation Follow-Up formula), or whole cow milk. A whole-cow milk feeding group was originally included to investigate the effect of high mineral and high protein concentrations on bone mass accretion during the second 6 months of life. Mineral and protein contents of the different feedings are given in Table 1. Because of concerns regarding low iron intake for infants receiving whole cow milk,¹¹ all mothers of cow milk-fed infants were supplied with iron supplements (Fer-In-Sol; Mead Johnson Nutritionals, Evansville, IN). If there were physician concerns about the infant being randomized to cow milk at 6 months of age, the infant was rerandomized to either the moderate- or high-mineral formula. Human milk-fed infants also were randomized to these feeding groups, with five infants still receiving some human milk at 12 months of age.

Table 1. Mineral and Protein Content of Different Feeding Regimens* [View Table]

Mothers were recruited from private obstetric practices and maternity wards in area hospitals and by referrals from other study participants. Infants who met the following criteria were eligible for the study: (1) full-term infants (38 to 42 weeks of gestation based on last menstrual period); (2) birth weight, length, and head circumference above the 10th percentiles, as established by National Center for Health Statistics data¹²; (3) apparent good health without evidence of cardiac, respiratory, gastrointestinal, or other systemic disease; (4) no maternal medical history of diabetes, use of antihypertensives, anticonvulsants, anticoagulants, steroids, or ritodrine, and/or a condition with proven adverse effects on the fetus; and (5) a 1-minute Apgar score of 7 or greater and a 5-minute Apgar score of 8 or greater. These exclusions are factors that may influence normal growth or calcium homeostasis in the infant.^13,14

Anthropometric measurements, nutrient intake, and total body BMC were evaluated at 1, 3, 6, 9, and 12 months of age. The weight of the infant was measured with a digital scale, and crown-to-heel length was measured with a length board. Three-day diet records on the infants were obtained during the week preceding each visit, and nutrient content was computed using the Minnesota Nutrition Data System (Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN).

Total body BMC was measured using dual-energy x-ray, absorptiometry (QDR1000W; Hologic, Inc, Waltham, MA). The Hologic system was operated in the single-beam mode and uses fast-kilovolt (peak) switching of a radiographic tube to generate both high- and low-energy photon beams (140 and 70 kV[p], respectively). The beam is transmitted in a rectilinear fashion throughout the body, and a detector measures the transmitted photon intensity on a pixel-by-pixel basis, with each pixel measuring 2 × 5 mm. Data from a rotating reference wheel containing known amounts of bone- and soft tissue-equivalent materials is used for continuous internal calibration during the scan. The relative attenuation of the two photon beams can be related to the mass of bone mineral and soft tissue. All infants were scanned wearing light clothing and after being placed on an "egg crate" foam pad restraint system. The Hologic pediatric software (version 5.56) was used for the analysis of the scan results. Our coefficient of variation, determined from duplicate scans on 17 infants 3 to 17 months of age, is 4.5% for total body BMC. Radiation exposure from a total body scan is less than 0.5 mrem per scan, well within acceptable ranges for use in pediatric populations.

A total of 101 infants were enrolled; 14 infants were withdrawn from the study, and 87 completed the study through 12 months of age. Nine infants were withdrawn before 6 months of age (3 in the human milk, 2 in the low-mineral, and 4 in the moderate-mineral feeding groups), and 5 withdrew between 6 and 12 months of age (1 in the moderate-mineral formula, 2 in the high-mineral formula, and 2 in the cow milk feeding groups). Eight of the 14 infants were withdrawn because of recommended changes in feeding. Thirty-one human milk-fed infants completed the study through 6 months of age. However, supplemental feedings with low-mineral formula were begun before 6 months of age in 27 of the 31 infants. Thirty and 31 infants in the low- and moderate-mineral formula groups, respectively, completed the study through 6 months of age.

The final numbers of infants completing the study are given in Table 2 by their final feeding group assignments. Seventeen of the infants initially in the cow milk group were rerandomized to formula groups because of parental or physician concerns (8 to moderate- and 9 to high-mineral formula). Therefore, a total of 38 and 39 infants received the moderate- and high-mineral formula, respectively, and only 10 infants remained in the cow milk group through 12 months. There were no differences in any of the anthropometric measurements at 6 or 9 months of age between infants who were rerandomized and the 10 remaining infants in the cow milk feeding group. Analyses were performed using the final group feeding assignment.

Table 2. Numbers of Infants Completing the Study [View Table]

Data analyses were conducted to determine the effects of mineral intake during the first 6 and second 6 months on bone mass accretion during those intervals. Additional analyses were conducted to determine the effect of early diet (first 6 months) on anthropometric measurements and bone mass accretion during the second 6 months. Analysis of variance (ANOVA) was used to determine group differences in baseline characteristics or changes in anthropometric and bone mass measurements, and the Tukey-Kramer honest significant difference test was used to compare individual means if the overall ANOVA was significant at P < .05 (JMP; SAS Institute, Inc, Cary, NC). Repeated-measures ANOVA was used to determine whether there were significant group differences and whether changes over time differed among dietary groups (time by group interaction). Approval was obtained from the Children's Hospital Medical Center Institutional Review Board, and written informed consent was obtained from parents.

RESULTS

Phase I: First 6 Months

Table 3. Characteristics of Infants Completing the Study Through 6 Months of Age [View Table]

Table 4. Absolute Changes (Mean ± SD) in Anthropometric Measurements and Bone Mass Between 1 and 6 Months of Age by Feeding Group During the First 6 Months of Age [View Table]

Because BMC is highly correlated with body size, additional analyses were performed to determine whether group differences in bone mass remained significant when weight, length, and bone area were included as covariates.¹⁵ When including these variables in the regression models, both the low-mineral formula- and human milk-fed groups had similar BMC, which was lower than the that in the moderate-mineral group at 3 and 6 months of age. Adjusted mean BMC values for the moderate-mineral formula group compared with the low-mineral- and human milk-fed groups were 127.8 ± 1.5 (SEM) g vs 119.2 ± 1.5 and 122.1 ± 1.4 g, respectively, at 3 months of age (ANOVA, P < .001) and 168.7 ± 2.5 g vs 157.6 ± 2.5 and 158.7 ± 2.4 g, respectively, at 6 months of age (ANOVA, P < .001). Although there were slight gender differences in BMC at 3 and 6 months of age, with male infants having higher BMC than female infants, these differences were explained by gender differences in body weight. Therefore, it was not necessary to include gender as a covariate.

To determine whether among human milk-fed infants supplemental feedings had an effect on BMC, the relationships between BMC and the ounces of formula consumed were investigated at 3 and 6 months of age. At 1 month of age, all but three infants were exclusively fed human milk. The mean ounces of formula consumed among the infants fed human milk were 16 ± 14 (SD) and 23 ± 14 oz/d at 3 and 6 months, respectively. Weight, length, and BMC at any age were not associated with ounces of formula consumed. In addition, there were no significant differences in weight, length, and BMC at 3 or 6 months of age between human milk-fed infants receiving less than 4 oz/d and those receiving more.

BMC at 6 months of age, among the formula-fed infants, was correlated with both average dietary phosphorus intake (r = .592; P < .001) and the average dietary calcium intake (r = .620; P < .001) during the first 6 months. The relationships between BMC and these minerals remained significant even after controlling for caloric intake (P = .004 for phosphorus; P = .02 for calcium). It was not possible to determine the independent effects of dietary calcium and phosphorus on BMC because of the strong correlation of these minerals with each other (r = .871; P < .001).

Phase II: Second 6 Months

Table 5. Characteristics of Infants After Randomization at 6 Months of Age (Includes Only Infants Completing 12 Months) [View Table]

Table 6. Changes (Mean ± SD) in Anthropometric Measurements and Bone Mass Between 6 and 12 Months of Age by Feeding Group During the Second 6 Months of Age [View Table]

Although the increases in anthropometric measurements between 6 and 12 months of age were not related to the second 6-month feeding group, there were significant differences during the second 6 months by the first 6-month feeding groups. Infants who received human milk during the first 6 months gained slightly less weight during the second 6 months than infants fed either low- or moderate-mineral formula (1.92 ± 0.45, 2.24 ± 0.55, and 2.21 ± 0.60 g, respectively; P = .05). Mean changes in BMC between 6 and 12 months, adjusted for changes in weight, length, and bone area, differed significantly among the early feeding groups, with human milk-fed infants having a greater gain in BMC than either the low- or moderate-mineral formula groups (P = .04): 81 ± 16 g in human milk-fed infants and 73 ± 15 and 71 ± 15 g in the low- and moderate-mineral formula groups, respectively. Infants fed whole cow milk during the second 6 months were excluded from this analysis because of the small number of infants completing the study. By 12 months of age there were no differences in BMC in either the early or late feeding groups (Fig 2).

DISCUSSION

We found that by 1 month of age there already were differences among the feeding groups. These findings indicate either a dramatic effect of early feeding on bone metabolism (because the infants have been consuming the formula for only 2 to 4 weeks) or baseline differences between infants fed human milk or formula. Because the early differences at 1 month of age may be spurious, the changes in anthropometric and bone measurements from baseline (1 month of age) among the different diet groups were investigated. Differences in total body bone mass accretion among feeding groups persisted, even after statistically adjusting for body size, as suggested by others.¹⁵

We observed a significant effect of infant mineral intake on bone mass accretion during the first 6 months of life, but this effect was not evident during the second 6 months. Despite a significant difference in mineral content of formulas consumed during the second 6 months, bone mass accretion did not differ. It is possible that the mineral content of both the moderate- and high-mineral formulas were high enough to have reached a threshold effect of these minerals on bone mass accretion already. A difference in the response during the second 6 months might possibly have been observed had a lower-mineral formula been used. However, the formulas used represent commercially available formulas that are marketed to the 6- to 12-month-old infant.

The design of the trial enabled us to evaluate short- and long-term effects of diet during the first 6 months of life on changes in bone mass accretion throughout the first year. We were able to determine whether bone mass accretion during the second 6 months differed according to the mineral intake during the first 6 months. A greater bone mass accretion was demonstrated during the second 6 months of life among infants who received human milk early in life compared with infants who received either low- or moderate-mineral formula. This difference seems to indicate, using the human milk-fed infant as the standard of reference, that a slowdown occurs during the second 6 months among infants fed formula in the first 6 months of life. This slowdown resulted in similar bone mass at 1 year of age among the different feeding groups.

The results of a randomized trial of calcium supplementation in young children found that supplementation increased bone mass accretion among prepubertal children.²⁰ However, 1 year after cessation of the trial, bone mass measurements were similar among both groups.²⁰ Our findings, coupled with those of other investigators, indicate that the effect of mineral intake on bone mass accretion during growth seems to be transient and not long term.

In conclusion, we found that the mineral intake of an infant's diet is significantly associated with bone mass accretion during the first 6 months of life. In addition, infants fed human milk during the first 6 months of life had a greater gain in bone mass during the second 6 months of life when they were provided a moderate- to high-mineral formula than infants previously fed low- or moderate-mineral formula. These early effects of diet on bone mass accretion resulted in similar total body bone mass at 1 year of age.