| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Osteoporosis Prevention and Treatment Center, the Bone and Mineral Metabolism Laboratory, and the Departments of Physical Medicine & Rehabilitation, Medicine, and Nutrition, Davis Medical Research Center (VM, NEB-S, JDL, E-JH, and AC) and the Departments of Statistics (PKG and BL) and Veterinary Biosciences (LAN), The Ohio State University, Columbus; the School of Allied Health, University of Connecticut, Storrs (JZI); the Department of Endocrinology, Cleveland Clinic, Cleveland (MS); the Nutrition Care Directorate, Walter Reed Army Medical Center, Washington, DC (SLM); and the Bone Imaging Laboratory, Wright State University, Dayton, OH (TNH)
2 Supported by grant RO1 AR40736-01A1 from the National Institutes of Health, grant M01-RR00034 from the Clinical Research Center of the National Institutes of Health, grant 37200-7586 from the National Research Initiative Competitive Grants Program of the US Department of Agriculture, and by Procter & Gamble Company, the National Dairy Council, Ross Products Division Abbott Laboratories, and General Mills Company. 3 Address reprint requests to V Matkovic, Osteoporosis Prevention and Treatment Center & the Bone and Mineral Metabolism Laboratory, Davis Medical Research Center, The Ohio State University, 480 West 9th Avenue, Columbus, OH 43210. E-mail: matkovic.1{at}osu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: This study evaluated the long-term effects of calcium supplementation on bone accretion among females from childhood to young adulthood.
Design: A 4-y randomized clinical trial recruited 354 females in pubertal stage 2 and optionally was extended for an additional 3 y. The mean dietary calcium intake of the participants over 7 y was 
830 mg/d; calcium-supplemented persons received an additional 670 mg/d. Primary outcome variables were distal and proximal radius bone mineral density (BMD), total-body BMD (TBBMD), and metacarpal cortical indexes.
Results: Multivariate analyses of the primary outcomes indicated that calcium-supplementation effects vary over time. Follow-up univariate analyses indicated that all primary outcomes were significantly larger in the supplemented group than in the placebo group at the year 4 endpoint. However, at the year 7 endpoint, this effect vanished for TBBMD and distal radius BMD. Longitudinal models for TBBMD and proximal radius BMD, according to the time since menarche, showed a highly significant effect of supplementation during the pubertal growth spurt and a diminishing effect thereafter. Post hoc stratifications by compliance-adjusted total calcium intake and by final stature or metacarpal total cross-sectional area showed that calcium effects depend on compliance and body frame.
Conclusions: Calcium supplementation significantly influenced bone accretion in young females during the pubertal growth spurt. By young adulthood, significant effects remained at metacarpals and at the forearm of tall persons, which indicated that the calcium requirement for growth is associated with skeletal size. These results may be important for both primary prevention of osteoporosis and prevention of bone fragility fractures during growth.
Key Words: Calcium • growth • skeletal development • peak bone mass • osteoporosis • females
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
= 0.05, with power 
0.8. An invitation form, a food-frequency questionnaire, and a pubertal stage self-report form were mailed to 15 000 female students aged 8–13 y (11). The inclusion criteria were white race, normal physical and mental health, pubertal stage 2, and calcium intake (determined by food frequency) below the threshold level of 1480 mg/d (12). The exclusion criteria were a history of metabolic bone, kidney, liver, or celiac disease; use of oral cortisone, hormones, diuretics, or antiseizure medications; other current systemic, chronic disease; and the presence of clinically significant abnormal laboratory data on screening. Response rates of 8–20% across schools generated initial data on 2000 girls who were willing to participate in the study. From this screening sample, 354 persons satisfied the inclusion criteria and were recruited for the study (Figure 1
).
|
Study protocol
This study was originally designed as a randomized, double-blind, 4-y controlled clinical trial to assess the effect of calcium citrate-malate supplementation (1000 mg/d, given in 4 pills) on the BMD of the total body, the radius, and the metacarpal radiogrammetric measurements of adolescent females during the pubertal growth spurt. The study was subsequently extended for 3 y, into late adolescence, in the subjects who agreed to continue; double-blind status was preserved. On the basis of our previous cross-sectional study (13), participants were first separated into 4 strata with the use of baseline total-body BMD (TBBMD) and body mass index (BMI; in kg/m2), each at 2 levels (above and below the average value). The characteristics of the subjects were TBBMD < 0.879 and BMI < 18.5 in stratum A (n = 122), TBBMD < 0.879 and BMI 18.5 in stratum B (n = 45), TBBMD 0.879 and BMI < 18.5 in stratum C (n = 79), and TBBMD 0.879 and BMI 18.5 in stratum D (n = 108). The purpose of stratification was to allow for the equal distribution of bone measurements between the placebo and supplemented groups. Within each stratum, subjects were randomly assigned to either group. A list of consecutive random assignments to the calcium supplementation and placebo groups within each stratum, prepared by a statistician, led to 177 subjects within each group. A simple coding system linked the drug packages to the randomization list. The Proctor & Gamble Company provided calcium citrate-malate and placebo (microcrystalline cellulose) pills whose palatability and appearance were equal. Subjects were given a 6-mo supply of pills at each visit; additional pills were mailed if a return appointment had to be postponed. Subjects were instructed to take 2 pills in the morning and 2 pills in the evening. Compliance was monitored by pill counts and assessed by fecal calcium density, serum parathyroid hormone (PTH) concentrations, and 24-h urinary calcium excretion.
Outcome measures and confounding factors
Primary outcome measures in this study were bone mineral areal density of the whole body and radius at 2 sites [proximal (33% of the radius length) and distal (10% of the radius length)], and metacarpal cortical index. Metacarpal radiogrammetry was previously used in the study of nutrition, peak bone mass, and hip fracture rates in adults (1) and was shown to distinguish children with upper-limb fractures from their control subjects (14). Secondary outcome measures in the current study were stature, bone width, bone area, and bone turnover markers. Measures to control for confounding factors included assessment of nutritional status (ie, dietary calcium, protein, energy intake, and 24-h urinary sodium), energy expenditure, body weight, skeletal age, and pubertal development. Medical history, physical examination, and completed dietary and physical activity questionnaires and weight, height, and bone mass measurements of the whole body and the forearm were obtained at baseline and every 6 mo; blood (drawn between 0800 and 1700), 24-h urine, and stool samples (no baseline data) were obtained at baseline and annually. Hand X-rays for skeletal age and radiogrammetry of the metacarpal bones were obtained at baseline and at years 4 and 7. The subject's weight was measured to the nearest 0.1 kg while the subject was wearing normal indoor clothing but no shoes. Standing height to the nearest 0.1 cm was recorded on a wall-mounted stadiometer (nearest 0.1 cm) while the subject was without shoes and with the mandible plane parallel to the floor (11). The subjects self-assessed their pubertal stage by marking on a chart the appropriate figures of sexual development, and the onset of menarche was documented within 6 mo of the event (15). Body composition (ie, lean body mass and body fat) and BMD were measured by using dual-energy X-ray absorptiometry (DXA; GE-Lunar DPX-L) with DPX-L software (version 1.3q; GE-Lunar). The precision errors (%CV) for the whole body and radius shaft BMD measurements were 0.5% and 0.8%, respectively (15). The two-person interobserver error for DXA analysis was < 0.1%. Daily phantom measurements on the DXA indicated a steady but extremely slow machine drift (a total decrease of 1.8% over 5 y); BMD was adjusted accordingly. Skeletal age was determined from radiographs of the nondominant hand by using the FELS method (16). Radiogrammetry of the metacarpals was performed by using the automated X-Posure System (Pronosco A/S, Vedboek, Denmark; 17). The CVs for the static and repositioned measurements were 0.0% and 0.2% for bone width and 0.5% and 0.6% for cortical thickness, respectively (18). Metacarpal cortical area (CA) and total cross-sectional area (TA) and the metacarpal cortical index, the ratio of CA to TA, were calculated. Nutritional status was assessed from 3-d dietary food records by using NUTRITIONIST III software (version 8.5l; Hearst Corp, San Bruno, CA). Weight-corrected energy expenditure was estimated by recording activity in 15-min intervals for 2 d (11, 19). Serum urine and stool specimens were stored at –80 °C and –20 °C, respectively, and analyzed in multiple batches at different times throughout the study. Basic blood and urine chemistry measurements were made by using a Hitachi 717 Chemistry Analyzer (Boehringer-Mannheim, Indianapolis), and serum calcium was adjusted per total protein. Stool samples were collected before visits and aliquot used for fecal calcium density (mg calcium/g dry feces) measurement as previously described (20). Urine and fecal calcium were measured by atomic absorption spectrophotometry. Urinary N-telopeptides (NTX; nmol · L bone collagen equivalent/24 h) were measured by a competitive inhibition enzyme-linked immunosorbent assay (Osteomark; Ostex International, Seattle). Serum 25-hydroxyvitamin D3 [25(OH)D3] was measured in each sample (excluding baseline) by using a radioimmunoassay (RIA) with 125I-labeled tracer (21). Serum osteocalcin was measured by immunoradiometric assays, and PTH was measured by Allegro immunoassay kits (both: Nichols Institute, San Juan Capistrano, CA). The stability of the PTH was shown by 99.1% agreement between repeated assays of a subsample (n = 10) in 1994 and 2000.
Statistical analysis
This long-term study allowed us to evaluate the effectiveness of calcium supplementation on bone mineral accretion during the period when most of the bone mass is accumulated. The distributions of confounding factors within the calcium-supplemented and placebo groups were compared by using the two-sample Kolmogorov-Smirnov tests for equality of distributions of average cumulative response over all visits for the 4- and 7-y cohorts. For the 7-y cohort, the distributions of average cumulative response over all visits during the first 4 y were also compared.
The primary outcome variables were analyzed by using repeated-measures multivariate analysis of variance (MANOVA) (visits 1–9 or 1–15) consisting of main effects and interactions for treatment and visits, as well as a more general MANOVA (visits 2–9 or 2–15), in which the observations on the same person were treated as response variables with the use of an arbitrary variance-covariance matrix with baseline values as covariates. The gains from baseline were also analyzed by using repeated-measures MANOVA with baseline values as covariates. If the repeated-measures MANOVA indicated a significant interaction effect or if the general MANOVA indicated significant differences between the mean responses, the follow-up univariate analyses [t test for gain and analysis of covariance (ANCOVA), with baseline values as covariates for the placebo and supplemented groups] at the endpoints of the original 4-y and the extended 7-y clinical trial are also presented. All these tests were made at = 0.05. Even though the subjects made 
15 visits, the mean (± SD) data are shown only for the baseline and the 4-y and 7-y endpoints.
An intent-to-treat analysis was performed to measure the programmatic effectiveness of calcium supplementation while ignoring the lack of full compliance (22, 23). Because the distributions of the confounding factors over time did not differ significantly, we did not include any of these variables in the analysis of primary outcomes. Furthermore, because the baseline measurements for cohorts in the 4-y and 7-y analyses did not differ significantly, no adjustments for differential dropouts were made.
Significant interactions in the repeated-measures MANOVA, as well as the results from the general MANOVA and the follow-up univariate analyses, suggested a possible catch-up phenomenon in bone mass acquisition in the placebo group at some of the skeletal sites. To obtain an insight into the skeletal physiology of bone growth and the timing of the catch-up phenomenon, longitudinal analyses based on the linear mixed-effect model (LME; 24, 25), with years since menarche (YSM) as the time line, group effects as fixed, and subject effects as random, were carried out by using data from subjects for whom there were 2 observations and with known onset of menarche (n = 252) for some of these skeletal sites.
The YSM is the best determinant of bone physiology in pubertal females. It serves as a biological clock for the events associated with menarche (26-28), and it minimizes the effect of heterogeneity on subjects' bone biology within each visit. The fixed group effects for bone mass accretion patterns in our LME model are represented by regression splines, ie, piece-wise cubic polynomials; constraints are such that the overall function is continuous and smooth at the joining points of these polynomials (knots). We used a natural-spline model (25), with fixed specified knots that further assumes linear trends in segments outside the lower and upper boundary knots. The large-sample chi-square tests for the significance of fixed effects in LME are based on likelihood ratio statistics (25). We also obtained 95% confidence bands for the difference in bone mineral areal density patterns for the corresponding groups. The band provides the estimated range of YSM values during which the patterns being compared differ significantly. Conservatively, when this confidence band contains zero, the difference at that time point is not significant at = 0.05. In addition, because the chi-square test is geared to test multiple hypotheses, it is possible that the LME model P value indicates highly significant fixed effects, but the confidence band may show insignificant differences at the full range of YSM time points. This phenomenon is similar to the F test in ANOVA, in which the F test shows significance, but the pairwise comparisons may not differ significantly.
This modeling strategy allows flexible shapes for regression functions in the analysis of longitudinal data (29-32). Our model assumes that a subject-specific bone-mass accretion pattern is a random effect that differs from its treatment group by an intercept term. It is similar to the model reported earlier in a bone-mass modeling study (32). This model was used to estimate patterns for the primary and secondary outcome variables and calcium metabolic measurements. On the basis of bone biology, 4 knots at –2, 0, +2, and +4 YSM were used in bone outcome models. For the bone turnover markers, serum calcium, serum PTH, and fecal and urinary calcium concentrations, which were collected yearly, 3 knots at –2, 0, and +2 YSM were used. Because the randomization to the 2 treatment groups was performed within each of the 4 strata, our initial model for the 2 x 2 x 2 factorial experiment included distinct regression splines for each subgroup (25). However, the factor "baseline BMI" was dropped from the model because it showed no significant effect on the bone accretion profiles. Furthermore, the factor "baseline TBBMD" had only a significant main effect (± 0.038, P < 0.001) in the TBBMD accretion patterns without any interaction with the pill or placebo factor. The models were fitted for different combinations of boundary knots, all of which provided qualitatively similar results. The results based on boundary knots at –3 YSM (baseline average – SD) and at +6 YSM (last visit average + SD) are presented here.
The biologic efficacy (22, 23) of calcium intake was evaluated by using LME analysis of data after post hoc stratification of subjects based on the average total cumulative calcium intake over time (ie, above or below the habitual dietary calcium intake of 830 mg/d), irrespective of assigned group. Total calcium intake for subjects in the supplemented group included dietary calcium plus pill calcium after adjustment for compliance.
Finally, post hoc stratifications according to the final height or the TA above and below the median (ie, tall or short and larger or smaller bones, respectively) were also performed in subgroup analyses to establish whether body or bone size affects calcium requirement as previously indicated (33, 34). The two-factor (ie, group and size) ANCOVA for BMD at proximal radius and cortical bone mass (CA), with the baseline value of the response variable used as a covariate, was performed at both the year 4 and the year 7 endpoints to assess the overall differences among the 4 subgroups by using the F test. The contrast between the means of the placebo and calcium-supplemented groups for persons within each size subgroup was tested by using t tests to assess their interactions. We used S-PLUS 2000 for WINDOWS software (Professional Release 3; Insightful Corporation, Seattle) for all statistical analyses (35).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Baseline characteristics of the study population are presented in Table 1. There were no significant differences at baseline in any of the values between the placebo and calcium-supplemented groups. Moreover, the cohorts with 2 visits, those who completed the 4-y and 7-y intervention, and those in the high or low calcium intake subgroups did not differ significantly in baseline characteristics between the 2 groups. In addition, the distributions of the confounding factors, including pubertal stage, skeletal age, dietary proteins, energy intake and expenditure, and urinary sodium, did not differ significantly between the 2 groups throughout the study (range of P values: 0.30–0.83). The average serum calcidiol concentration over time was 27.9 ± 10.2 ng/mL and 27.5 ± 10.9 ng/mL in the supplemented and placebo groups, respectively (range of P values: 0.33–0.55). The placebo group had a slightly higher proportion of ever-users of contraceptives (23%) than did the supplemented group (13%). However at age 18 y, the two-way (ie, group and contraceptive use indicator) ANOVA for the bone variables found no interaction between the factors TBBMD (P = 0.642), proximal radius BMD (P = 0.758), and CA:TA (P = 0.527).
|
). The average total calcium intake in the supplemented group (1498 ± 318 mg/d) was close to the calcium intake threshold for adolescents (12). Total calcium intake in the supplemented group declined by 10% from age 15.5 y to age 18 y, primarily because of a decrease in compliance with pill taking (Table 2). In addition, several placebo group subjects had an average cumulative calcium intake comparable to the total calcium intake in the supplemented group.
|
) than did the placebo group. PTH concentrations in the placebo group reached their peak by the onset of menarche and declined thereafter, subsequent to changes in serum calcium. Fecal calcium density and urinary calcium were lower in both groups during the premenarcheal period than after menarche. The trend was not clinically significant, but the fact that the pattern changes at menarche is clinically significant.
|
), and urinary NTX excretion (nmol · L bone collagen equivalent/d; P = 0.31; Figure 3) between supplemented and placebo groups.
|
). Nor were differences in gains in stature, metacarpal TA, and bone mineral areas confirmed after 7 y of intervention (Table 4). In addition, a repeated-measure MANOVA for height, metacarpal TA, and bone mineral areas, in which baseline values were used as covariates, showed no significant group x time interactions or group main effects.
|
|
shows the data for these variables at baseline and at the year 4 endpoint and their gains from the baseline, along with the follow-up univariate analyses for BMD at the 3 sites and for metacarpal CA and CA:TA. The follow-up univariate analyses of these data at the year 4 endpoint show that TBBMD, proximal and distal radius BMD, metacarpal CA:TA, and the gains in TBBMD and metacarpal CA:TA, when baseline measurements were used as covariates, were significantly higher in the calcium-supplemented group than in the placebo group.
For the 7-y cohort, the general MANOVA for comparing the mean responses on primary outcome variables (visits 2–15), in which their baseline values were used as covariates, found a highly significant treatment effect on TBBMD (supplemented group: n = 60; placebo group: n = 67; P < 0.0004), proximal radius BMD (supplemented group: n = 56; placebo group: n = 63; P < 0.024), and distal radius BMD (supplemented group: n = 57; placebo group: n = 64; P < 0.028). The repeated-measures MANOVA (visits 1–15), in which baseline values were used as covariates, found highly significant treatment x visits interaction for all 3 variables (P < 0.0006). The repeated-measures MANOVA of the gains from the baseline for this cohort, in which baseline measurements were used as covariates, found highly significant treatment x visit interactions in TBBMD (P < 0.0097) and proximal radius BMD (P < 0.001). A subanalysis of gain from the baseline by using general MANOVA on the follow-up data from the first 4 y only (visits 2–9) found that the supplemented and placebo groups were significantly different (TBBMD: P < 0.0004; proximal radius BMD: P < 0.004). However, the general MANOVA for the gains from baseline for the same cohort during the 3-y study extension (visits 10–15) did not find a difference between the groups (TBBMD: P < 0.220; proximal radius BMD: P < 0.346). The repeated-measures MANOVA for metacarpal CA and CA:TA, in which baseline values were used as covariates (visits 1, 9, 15; supplemented group: n = 76; placebo group: n = 96), found that treatment x visit interactions were significant for these variables (CA: P < 0.024; ratio of CA to TA: P < 0.0034). However, for the gains from baseline, when baseline measurements were used as covariates, the interaction was significant for CA:TA (P < 0.0000) but not for CA (P > 0.11). Table 4 provides data for these variables at baseline, at the year 4 and year 7 endpoints, and their gains from the baseline, along with the results of follow-up univariate analyses for BMD at the metacarpal CA and CA:TA, adjusted for the baseline values. The differences between the supplemented and placebo groups in TBBMD and distal BMD, which were significant at the year 4 endpoint, were no longer significant at the year 7 endpoint. In addition, the metacarpal CA and CA:TA and their gains from baseline were significantly higher in the supplemented group than in the placebo group at both these endpoints (Table 4). Collectively, these analyses clearly indicated a need for further examination of the longitudinal behavior of bone mass accretion in the supplemented and placebo groups.
The longitudinal analysis based on the LME model revealed a significant influence of calcium supplementation on TBBMD (P < 0.001) and proximal radius BMD (P < 0.001) (Figure 4). The calcium-supplemented group showed a faster rate of bone mass accretion from the beginning of the study; the maximum difference between the BMD of the 2 groups at both of these sites occurred during the interval from –1 YSM to +1 YSM. After +1 YSM, the differences between the groups started to diminish, and they became insignificant after +3 and +5 YSM for TBBMD and proximal radius BMD, respectively (Figure 4).
|
). The high-calcium-intake subgroup (average intake: 1353 ± 342 mg/d; n = 180) has persistently higher BMD of the proximal radius from puberty to young adulthood than does the low-calcium-intake subgroup (average intake: 668 ± 118 mg/d; n = 89).
|
). Unlike the tall subgroup, the short subgroup did not show an effect of calcium supplementation on the proximal radius BMD (Figure 6).
|
830 mg/d was adequate for persons with smaller metacarpals but insufficient for those genetically predetermined to develop larger bones.
Bone fracture
Twenty girls from the placebo group and 9 girls from the calcium-supplemented group reported having a bone fracture because of a moderate trauma during the trial. There were 3 forearm fractures in the supplemented group (33%) and 11 in the placebo group (48%). The average timing of the fracture was +1.2 ± 0.4 YSM, which coincides with the bone-modeling phase in skeletal development and overlaps with the timing of the maximal effect of calcium supplementation on BMD among young females in this study (Figure 4).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
670 mg/d) beyond a habitual dietary calcium intake of 830 mg/d (increasing total calcium intake up to 1500 mg/d) positively influences bone mass acquisition throughout the bone-modeling phase of the pubertal growth spurt. This effect diminishes during the skeletal consolidation of late adolescence that is due to the catch-up phenomenon in bone mineral accretion. At the beginning of young adulthood, the positive effect of calcium supplementation was evident at all skeletal regions of interest; however, the only differences that remained significant were those at the metacarpals and at the proximal radius among the subgroup with high calcium intake and in tall persons.
The explanation for the catch-up phenomenon lies within the skeletal physiology of growth, which dictates calcium requirement (4, 12). Calcium requirement is the highest during the pubertal growth spurt, when most of the retained calcium contributes to skeletal build-up to accommodate longitudinal bone growth and periosteal bone expansion (4, 13). During this interval, 37% of the entire adult skeletal mass is accumulated (13). Therefore, inadequate calcium intake during this period compromises the bone mineral accretion rate (3). During the pubertal growth spurt, there was an obvious discrepancy between optimal and actual calcium intakes among the placebo group in this study, which was possibly reflected in a higher fracture rate, even though that was not a research outcome variable. However, when bone modeling slows down (by epiphyseal closure that results from increased estradiol secretion after menarche), the demand for calcium declines, and the calcium intake threshold decreases from 1480 to 957 mg/d (12). The habitual calcium intake in the young women in the current study is very close to this reduced requirement and is sufficient to accommodate bone consolidation, which allows BMD to catch up slowly.
These conclusions are supported by the blood, stool, and urine chemistry measurements. During bone modeling of the pubertal growth spurt, the PTH concentrations in the placebo group reached a maximum and declined thereafter. The increase in PTH could increase intracortical remodeling and porosity as indicated previously (36) and is reflected in a slightly increased urinary NTX excretion. The pubertal growth spurt is the time when serum calcitriol concentration is at its maximum and is increasing calcium absorption to accommodate skeletal needs (37), which results in a lower fecal calcium excretion (4), and presumably in a lower fecal calcium density. Calcium absorption decreases during bone consolidation in late adolescence (4), and fecal calcium density increases concomitantly. Calcium supplementation during the pubertal growth spurt partially blunts PTH secretion, which suggests that calcium intake is adequate, and bone resorption decreases concomitantly, as previously documented by calcium kinetics (38). The changes in serum PTH reflect the fluctuations in serum calcium concentration during growth. Serum calcium is lower during the bone-modeling phase of the pubertal growth spurt than it is either before that phase or later, during the bone-consolidation phase. Urinary calcium follows those changes closely (4). Urinary calcium excretion is lower during the bone-modeling phase than during skeletal consolidation. The difference in urinary calcium excretion between the 2 groups was much smaller during the pubertal growth spurt (7%) than it was in late adolescence (18%), which implies that urinary calcium is less dependent on intake during bone modeling, as documented by balance studies (3, 4). Calcium supplementation increased urinary calcium more during bone consolidation, despite a drop in calcium intake (by 10%), which suggests that the calcium supply is adequate.
The catch-up phenomenon in bone mass acquisition suggests the existence of a reversible mineral deficit that is acquired during the pubertal growth spurt (36) and that can be repaired during bone consolidation. The catch-up phenomenon was apparent, in particular at the total body. However, at the metacarpals and at the proximal forearm of persons accustomed to different calcium intakes, the difference created during bone modeling was maintained into young adulthood, which implies incomplete catch-up and a lower peak bone mass. The lower the habitual calcium intake in the population, the higher the possibility of a low peak bone mass at skeletal maturity. The results of ecologic studies (1, 39) conducted among populations accustomed to very low lifetime calcium intakes point in this direction. Metacarpal radiogrammetric measurements (ie, CA and CA:TA) were lower in women from a low-calcium-intake area of Croatia, who were accustomed to calcium intakes of 400 mg/d over a lifetime and who had a concomitantly high rate of hip fracture (1). Similar findings with regard to peak bone mass were obtained in animal experimentations by manipulating calcium in the diet early in life (40, 41).
In addition to BMD, the current study evaluated the effect of calcium supplementation on longitudinal (height) and periosteal (width) bone expansion and on bone mineral area. The results show that calcium supplementation in addition to the habitual calcium intake of 830 mg/d had no effect on bone geometry measurements.
To ascertain whether BMD is related to calcium supplementation among persons of different body frames, its interactions with final height and periosteal expansion (TA) were examined (33, 34). Among subjects destined to be taller or to have larger bones, those with higher calcium intakes had significantly higher BMD than did those with lower calcium intakes. These results strongly support the notion that dietary calcium requirement for skeletal development is size dependent.
All clinical trials with calcium or dairy product supplementation in children and adolescents that have been completed to date (3, 5-10, 42-44) showed a positive effect of intervention on bone mass, but they were all too short (1–3 y) to address the question of whether it is the adaptation of bone tissue to nutritional challenge that leads to peak bone mass. The increase in bone mass observed in those short-term studies could be explained to a large extent by the phenomenon of the bone-remodeling transient (45). In some of the studies reported earlier, the difference in bone mass between the groups diminished after calcium intervention was discontinued (46, 47), which indicated that the bone accretion gained in the first transient was lost as a result of the second (45). In other studies, the effects of intervention were maintained 1–3 y after discontinuation of treatment (44, 48, 49); this result may be specific to the calcium source, to the amount of habitual dietary calcium intake, or both.
In the current study, however, calcium supplementation continued without interruption for 7 y, which allowed for the adaptation in bone behavior during the bone-modeling and bone-consolidation phases in skeletal development and extended long after the first bone-remodeling transient effect ended (12 mo after the beginning of intervention). According to calcium balance studies (4), the calcium requirement of the participants during the 7-y follow-up was changing as well. The effect of the second transient, following the discontinuation of calcium supplementation after 7 y, therefore is not expected because dietary calcium requirements have been reduced to an amount closer to habitual dietary calcium intake. Thus, if we assume that these women will maintain their dietary habits, it is unlikely that the differences in bone mass measurements observed at these skeletal regions of interest at the beginning of young adulthood will disappear by the cessation of intervention. A follow-up study may be necessary for confirmation.
Some studies indicate that calcium intake may play a role in the prevention of fractures due to bone fragility during growth (50, 51). The fracture rate was not a primary research outcome, but the reported fracture incidence in our study points in this direction, although the sample size may be too small for adequate interpretation. Because the peak incidence of fractures due to bone fragility coincides with the pubertal growth spurt (52, 53), our results indicate that calcium intakes at threshold level may reduce the risk of fracture during this stage of skeletal development, irrespective of the catch-up phenomenon in bone mass acquisition that occurs thereafter. The possibility that calcium intakes may reduce the risk of fracture is particularly important given the large number of childhood forearm fractures and the rising incidence over the last 30 y (54). A long-term intervention study of calcium supplementation in children that evaluates the forearm fracture as the main research outcome should be conducted to resolve this issue.
It is of great interest to pediatricians to tailor the nutritional recommendations for adolescents to their individual calcium needs, which are maximal at the pubertal growth spurt. However, such specification may be difficult to implement in practice because a person's rates of bone growth and skeletal maturity cannot be predicted precisely. The results of the current study also have implications for dietary calcium intake standards for children and adolescents worldwide; the standards for one ethnic group might not be suitable for another. Each country should develop its own standards that are specific to the people living in the region. Factors such as ethnicity, dietary habits (including salt intake), sunlight exposure, and activity level all play a role, but stature and body frame must also be considered.
In summary, this study documents that calcium supplementation in excess of a habitual calcium intake of 830 mg/d affects BMD during the pubertal growth spurt, but there is a diminishing effect thereafter that is due to the catch-up phenomenon in bone mineral accretion. By young adulthood, significant effects of calcium supplementation were present at metacarpals and at the proximal forearm in subjects who had better calcium compliance and in subjects who developed larger body frames. These results imply that standards for dietary calcium intake in adolescence should be based on growth rate and body and bone size development. The results of this study may be important for the prevention of bone fragility fractures during growth, as well as for the primary prevention of osteoporosis.
|
ACKNOWLEDGMENTS |
|---|
VM, TNH, and PKG were involved in the study design and study conduction; NEBS, JDL, JZI, MS, LAN, SLM, and EJH were involved in various phases of data collection; PKG and BL performed the statistical analysis; TNH was involved in designing and implementing quality control for the DXA measurements; NEBS, JZI, and JDL were involved in database development and maintenance; VM, PKG, BL, AC, JDL, and JZI were involved in data analysis; VM and PKG had a major role in writing the manuscript; and all authors contributed to the writing of the manuscript. All authors have declared that no conflicts of interest exist.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. M. Hill, M. Braun, M. Kern, B. R. Martin, J. W. Navalta, D. A. Sedlock, L. McCabe, G. P. McCabe, M. Peacock, and C. M. Weaver Predictors of Calcium Retention in Adolescent Boys J. Clin. Endocrinol. Metab., December 1, 2008; 93(12): 4743 - 4748. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Maalouf, M. Nabulsi, R. Vieth, S. Kimball, R. El-Rassi, Z. Mahfoud, and G. El-Hajj Fuleihan Short- and Long-Term Safety of Weekly High-Dose Vitamin D3 Supplementation in School Children J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2693 - 2701. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Atkinson, G. P. McCabe, C. M. Weaver, S. A. Abrams, and K. O. O'Brien Are Current Calcium Recommendations for Adolescents Higher than Needed to Achieve Optimal Peak Bone Mass? The Controversy J. Nutr., June 1, 2008; 138(6): 1182 - 1186. [Full Text] [PDF]
|
|
|
|
 |
|
 V. Lilleby Review: Bone status in juvenile systemic lupus erythematosus Lupus, August 1, 2007; 16(8): 580 - 586. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Braun, C. Palacios, K. Wigertz, L. A Jackman, R. J Bryant, L. D McCabe, B. R Martin, G. P McCabe, M. Peacock, and C. M Weaver Racial differences in skeletal calcium retention in adolescent girls with varied controlled calcium intakes Am. J. Clinical Nutrition, June 1, 2007; 85(6): 1657 - 1663. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Straub Calcium Supplementation in Clinical Practice: A Review of Forms, Doses, and Indications Nutr Clin Pract, June 1, 2007; 22(3): 286 - 296. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Abrams and G. J. Strewler Adolescence: How Do We Increase Intestinal Calcium Absorption to Allow for Bone Mineral Mass Accumulation? IBMS BoneKEy, May 1, 2007; 4(5): 147 - 157. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. I. Gafni and J. Baron Childhood Bone Mass Acquisition and Peak Bone Mass May Not Be Important Determinants of Bone Mass in Late Adulthood Pediatrics, March 1, 2007; 119(Supplement_2): S131 - S136. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Lentle and D. Worsley Osteoporosis Redux J. Nucl. Med., December 1, 2006; 47(12): 1945 - 1959. [Full Text] [PDF]
|
|
|
|
 |
|
 L. L. DeBar, C. Ritenbaugh, M. Aickin, E. Orwoll, D. Elliot, J. Dickerson, N. Vuckovic, V. J. Stevens, E. Moe, and L. M. Irving YOUTH: A Health Plan-Based Lifestyle Intervention Increases Bone Mineral Density in Adolescent Girls Arch Pediatr Adolesc Med, December 1, 2006; 160(12): 1269 - 1276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. M. Fiorito, D. C. Mitchell, H. Smiciklas-Wright, and L. L. Birch Girls' Calcium Intake Is Associated with Bone Mineral Content During Middle Childhood J. Nutr., May 1, 2006; 136(5): 1281 - 1286. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Zhu, Q. Zhang, L. H. Foo, A. Trube, G. Ma, X. Hu, X. Du, C. T Cowell, D. R Fraser, and H. Greenfield Growth, bone mass, and vitamin D status of Chinese adolescent girls 3 y after withdrawal of milk supplementation Am. J. Clinical Nutrition, March 1, 2006; 83(3): 714 - 721. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. El-Hajj Fuleihan, M. Nabulsi, H. Tamim, J. Maalouf, M. Salamoun, H. Khalife, M. Choucair, A. Arabi, and R. Vieth Effect of Vitamin D Replacement on Musculoskeletal Parameters in School Children: A Randomized Controlled Trial J. Clin. Endocrinol. Metab., February 1, 2006; 91(2): 405 - 412. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. P. Heaney and C. M. Weaver Newer Perspectives on Calcium Nutrition and Bone Quality J. Am. Coll. Nutr., December 1, 2005; 24(suppl_6): 574S - 581S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Cheng, A. Lyytikainen, H. Kroger, C. Lamberg-Allardt, M. Alen, A. Koistinen, Q. J. Wang, M. Suuriniemi, H. Suominen, A. Mahonen, et al. Effects of calcium, dairy product, and vitamin D supplementation on bone mass accrual and body composition in 10-12-y-old girls: a 2-y randomized trial Am. J. Clinical Nutrition, November 1, 2005; 82(5): 1115 - 1126. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Abrams, I. J. Griffin, K. M. Hawthorne, and L. Liang Height and Height Z-Score Are Related to Calcium Absorption in Five- to Fifteen-Year-Old Girls J. Clin. Endocrinol. Metab., September 1, 2005; 90(9): 5077 - 5081. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. K. Bachrach Calcium and Peak Bone Mass: How Much Is Needed and When? IBMS BoneKEy, April 1, 2005; 2(4): 11 - 14. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
