SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The objective of this study was to investigate the association between dietary calcium intake and radial bone density among young women, over the whole range of intake and at different levels of calcium intake. The study design was a cross-sectional, observational multicenter investigation in six European countries. One thousand one hundred and sixteen healthy Caucasian girls aged 11–15 years and 526 women aged 20–23 years participated, after having been selected from larger population samples to represent a large range in calcium intake. Bone mineral density (BMD) was measured with dual-energy X-ray absorptiometry at the ultradistal and middistal radius. Calcium intake was assessed with 3-day food records. Other potential determinants of BMD were measured by anthropometry or questionnaires. Mean calcium intake among the girls varied between 609 mg/day in Italy and 1267 mg/day in Finland; intakes for women were in a similar range. After adjustment for height, weight, and age at menarche for the women, and adjustment for age, height, weight, Tanner stage, and bone area for the girls, radial BMD at both sites did not significantly vary among quartiles of calcium intakes for both age groups. In multivariate linear regression, calcium was weakly positively associated with BMD at both sites in the girls (per 100 mg of calcium: β = 0.57 mg/cm2, p = 0.03 for middistal BMD and β = 0.56 mg/cm2, p = 0.01 for ultradistal BMD). For middistal BMD, the association was observed predominantly in premenarcheal girls. The associations were no longer statistically significant after full adjustment for all determinants of BMD, except again in pre-menarcheal girls. Radial BMD in the women was not associated with calcium intake, except after full adjustment for determinants of BMD, when ultradistal BMD became inversely associated with calcium intake (per 100 mg β = −1.02, p = 0.03); this finding was due to results in one of the countries and not found in other countries. There was no evidence for a different relation between calcium and BMD at different levels of intake; although there was a positive association at calcium intake levels <600 mg/day, the interaction was not significant and there was no consistent trend over intake categories. These results do not support the hypothesis that dietary calcium is a determinant of peak BMD in European women, for a wide range of intake. This study does not provide evidence that Recommended Dietary Allowances for calcium should be increased.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Most of the skeletal peak bone mass is already attained at the age of 16–18 years,(1,2) although some additional growth may take place during the third and fourth decade, depending on the skeletal site. Among other environmental factors, nutrition, and in particular calcium intake,(3,4) is assumed to influence whether the genetically determined maximal peak bone mass is reached. It has been suggested that there is a level of calcium intake below which accumulation of skeletal mass is a function of intake and above which accumulation is constant, and not dependent on further increases in intake.(5) There is no agreement as yet on the level of this threshold intake. Some suggest it may be higher than the current Recommended Daily Allowance (RDA).5 Matkovic et al.(6) estimated threshold intakes, from a meta-analysis of more than 500 balance studies, to be 1090, 1390, 1480, and 957 mg/day for infants, children, adolescents, and young adults, respectively. However, these high estimates are disputed(7) and, moreover, not strongly supported by data from population studies. Only a few observational studies(8-11) report a relation between habitual calcium intake and bone density during adolescence. However, intervention studies have shown an increased bone mineral density (BMD) in children who received a daily calcium supplement in comparison with children receiving a placebo,(12-16) but this effect appears to vanish after stopping supplementation.(17,18) There have been no supplement studies continuing until peak bone density has been reached, so it is not clear whether a higher peak bone mass can be attained by long-term high calcium intake. Results of cross-sectional(19,20) and prospective studies(21-23) in young adult women are inconsistent and show only weak positive associations at the most.(24-26)

The objective of this study was to investigate the relation between dietary calcium intake and radial bone density among two age groups (adolescent girls aged 11–15 years and women aged 20–23 years). A European multicenter study design was chosen in order to obtain a large range in calcium intake. The relatively large study population allowed the evaluation of the hypothesis of a threshold level.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Population

In each of the six participating countries (Denmark, Finland, France, Italy, the Netherlands, and Poland), subjects with high and low calcium intake were selected from random population samples of about 750 adolescent girls (11–15 years of age) and 375 young adult women (20–23 years of age). In Finland, Denmark, Poland, and the Netherlands, random samples from the local population registries were obtained and subjects invited to participate; response rates varied from 25.4% to 51.5% for girls and from 28.6% to 62% for the women. In France, girls and women were recruited via general practitioners and gynecologists in two geographic areas, Rhone-Alpes and Marseille; in Italy, girls were recruited from all eight secondary schools in the town of Rende, and women from the University of Calabria (response rates 100%). A selection on calcium intake was made with a food frequency questionnaire as a screening instrument, to assure an adequate range of calcium intake within each country. Of those selected, the participation rate varied between 64% and 84% for the girls and between 52% and 91% for the women; data were collected for 1116 girls and 526 women. The study was approved by the local medical-ethical committees and all participants (and their parents when required) gave their informed consent.

Subjects were excluded when indicating a chronic disease in general, diseases related to bone or calcium metabolism in particular, use of corticosteroids, noncaucasian origin, sporting more than 7 h/week, (previous) pregnancy, irregular menstruation (for the women only), vegetarianism or macrobiotism, or any prescribed diet (except energy-restricted diet).

Food intake assessment

Calcium intake was assessed by two methods. As a first screening and selection procedure, to assure a wide range of calcium intake, usual calcium intake during the previous month was assessed by a 20-item food frequency questionnaire (FFQ)(27,28) in the population samples. Second, in the selected study populations, calcium intake was assessed by a 3-day record method.

The FFQ was adapted for each country, to include specific sources of calcium. Frequency of consumption was reported in 10 categories: rarely/never, 1 day/month, 1 day/2 weeks, 1 day/week, and subsequently 2, 3, 4, 5, 6, or 7 days/week. Portion sizes were quantitated by respondents in terms of household measures (slices, spoons, cups, glasses) to which standard weights were assigned. The use of food supplements was asked. Completeness, credibility of reported number of servings and consistency in reported consumption frequencies were checked. Data on mean daily consumption of food products in grams were converted to calcium intake using the local food composition tables. In each country, 50 women from the low end and 50 women from the high end of the calcium intake distribution were invited for the actual study. The girls were stratified on age, in 1-year categories, before selecting persons with low and high intakes (20 low, 20 high, for each category). In France, all girls and women were selected, because of the size of the sample.

The FFQ is an appropriate method to classify subjects within a country, but not for comparison among countries. To estimate calcium intake in a comparable way, a 3-day food record method was used. The selected subjects were asked to record everything they consumed during a consecutive Wednesday, Thursday, and Friday, the week before their visit to the Institute. Time of day, food, quantity, and recipes of composite dishes were recorded. The parent responsible for meal preparation was asked to assist in completing the food records. At the visit to the Institute, the food records were checked by a dietitian for completeness; household measures were verified by comparison with standard measures. Nutrient intake was calculated per day, and mean intake of calcium and energy was calculated as the average over 3 days.

Although a preselection on low and high calcium intake was made with the FFQ, the calcium intake from the food records showed a continuous distribution, including the middle range of intake. This is due to the phenomenon of regression to the mean, and to the differences in methodology.

Bone density measurements

Bone mineral content (BMC) and bone area (BA) at the middistal radius (one-third distal point between the styloid process and the tip of the olecranon of the elbow) and the ultradistal radius of the nondominant arm were measured by dual-energy X-ray absorptiometry (DXA; p-DXA Osteoscan, Nederburgh BV, Bunschoten, Netherlands). The BMC in milligrams was divided by the projected area of the bone to derive the BMD in milligrams per square centimeter, as a means of normalizing results for bone size. The same DXA instrument was used in each country. It was calibrated after transportation to a new location and daily during measurement days, against a reference phantom. To evaluate long-term precision, the daily results of the phantom BMD measurements were plotted against time. There was no overall correlation between time and BMD (r = −0.04). Analysis of variance showed that none of the mean phantom BMD values in subsequent countries was significantly different from the mean value in the first country (326.2 ± 3.2 mg/cm2). Bone mineral measurements were performed by the same two operators in the Netherlands (April–June 1995), Finland (September–October 1995), Denmark (November–December 1995), and Italy (January–February 1996). In France (March–April 1996), one operator was replaced, and in Poland (May–June 1996) another team performed the densitometry. The coefficient of variation for 10 measurements of the same subject, with repositioning, was 1.6–2.7% for BMD ultradistal and 1.1–2.4% for BMD middistal (assessed at different times and by different operators). The variation between operators, expressed as the mean percentage difference between paired measurements, was about 2.5%.

Anthropometry and other measurements

Anthropometry was performed by the Osteoscan operators in each country. Height and weight were measured with the subject wearing light clothing and no shoes. The triceps skinfold, midupper arm circumference, and hand grip-strength were measured at the nondominant arm. Pubertal stage was defined as the Tanner stage of breast development (M1–M5), assessed by comparison with standard photographs.

Subjects completed a self-administered questionnaire on menstrual function and use of oral contraceptives; smoking habits; alcohol use; daylight exposure; height, weight; and education of parents; and physical activity. Physical activity was determined for the previous month. The questionnaire covered activities at school, work, and in leisure time (sports and household activities).(29) For the girls, the questionnaire comprised 58 items, for the women 88. Each activity was assigned a score for energy expenditure and a score for the weight-bearing characteristic. All activities were combined into a summary score for energy expenditure and one for weight-bearing activity. The questionnaire was checked in an interview setting. Daylight exposure was estimated by questions on the hours spent outdoors, separately for fall/winter, spring/summer, and summer holidays, and separately for week and weekend days. The answers to these questions were combined into one score for mean daylight exposure.

Statistical analysis

All procedures described below were performed for girls and women separately. Variables that were not normally distributed (triceps skinfold, daylight exposure, energy expenditure, weight-bearing activity, and calcium intake) were natural log transformed for Pearson correlation analyses. Potential determinants of BMD were identified by means of Pearson's correlation analysis and further evaluated in the pooled dataset with stepwise multivariate regression analysis. It has been noted before that when the relationship between BMC and BA is not directly proportional, part of the variation of BMD within a population will be due to differences in bone size between individuals.(30) The use of BMD could then lead to spurious associations with other variables, such as dietary intake, which are themselves related to bone size through their dependence on overall body size. For the girls, there was indeed a correlation between BMD and BA, for both sites, even after height and weight were included in the model. Therefore, BA was included as a covariable in all multivariate models with BMD as dependent variable, for the girls. Categorical variables (Tanner score [five stages], menarcheal status [pre/post], smoking habits [yes/no], oral contraceptive use [yes/no]) were entered in the models as dummy variables. Next, the relation between calcium intake and radial BMD was investigated within centers and overall. Calcium intake was calculated from the food records. In the multivariate analyses, first we included a fixed set of covariables: height, weight, and age at menarche for the women, age, height, weight, Tanner stage and BA for the girls, both in the analyses of the pooled data and in country-specific analyses. Next, the overall set of predictors of BMD were included, which had been identified with stepwise linear regression analyses. For specific countries, not all the variables mentioned were significant predictors of BMD, but we chose to include the same sets of covariables. Due to missing values for some of the covariables, multivariate analyses were performed for 1109 girls and 516 women. First, BMD was calculated for quartiles of calcium intake, and the differences tested, with analysis of covariance. The linear regression of BMD on individual calcium intake was evaluated, adjusting for the main determinants of BMD. To see if there was a threshold value for a linear relation between calcium and BMD, regression coefficients for calcium were estimated at different levels of calcium intake. The validity of this stratified approach was evaluated by regression diagnostics for the different strata. In the adolescent population, stratified analysis was performed for pre- and postmenarcheal girls. A p value of < 0.05 was considered significant. All analyses were performed with BMDP Statistical Software (version 7.0).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Lifestyle and anthropometric characteristics of the girls and women are listed in Table 1. Table 2 gives the calcium intake estimated from the food records and middistal and ultradistal BMD by country, as well as height and weight. Mean intakes among the girls were similar to intakes among the women. High intakes were observed in Finland, Denmark, and the Netherlands, with mean intakes around 1100–1200 mg/day. Intakes in France and Poland were intermediate (800–900 mg/day) and in Italy relatively low intakes were seen (mean about 600–700 mg/day). Mean BMD among the girls was highest in France, significantly higher than in the Netherlands (p < 0.05) and (middistally) Denmark (p < 0.01); among the women, middistal BMD in Italy (p < 0.01) and ultradistal BMD in Finland (p < 0.01) were significantly lower than in other countries.

Table Table 1..  Lifestyle and Anthropometric Characteristics of Girls Aged 11–15 Years and Women Aged 20–23 Years
Thumbnail image of
Table Table 2..  Mean (±SD) Calcium Intake and BMD per Country
Thumbnail image of

In the pooled data, ultradistal BMD was not correlated with calcium intake (natural-log transformed), neither in girls nor in women. Middistal BMD was weakly positively correlated with calcium intake among the women (r = 0.13, p < 0.01), but not among the girls. However, a number of other parameters were associated with either calcium or BMD, or both, thereby potentially biassing the relation between calcium intake and BMD.

Among girls, calcium intake was significantly correlated with age (r = −0.09, p < 0.01), height (r = 0.20, p < 0.001), grip strength (r = 0.12, p < 0.001), and energy intake (r = 0.39, p < 0.001). Among women, significant positive correlations were observed for calcium and height (r = 0.28, p < 0.001), weight (r = 0.12, p < 0.01), grip strength (r = 0.16, p < 0.001), and energy intake (r = 0.49, p < 0.01). Determinants of middistal and ultradistal BMD were evaluated in the pooled data with stepwise linear regression analysis, starting with a maximum model including all potential determinants identified with univariate correlation analysis: age, height, weight, smoking, daylight exposure, menarche, oral contraceptive use, physical activity (energy expenditure), energy intake, calcium intake, grip strength, triceps skinfold, arm circumference, and in addition, for the girls, BA and Tanner stage. Among girls, a number of growth- related parameters were identified as statistically significant and independent predictors of BMD: age, Tanner score, BA, height, weight, triceps skinfold, arm circumference, grip strength, and menarche (yes/no). For the women, the main predictors of BMD were different for the ultradistal and middistal site. Ultradistal BMD was best predicted by height, weight, triceps skinfold, grip strength, and age at menarche. Middistal BMD values were best predicted by height, weight, menarche, and daylight exposure. In multivariate analyses of the relation between calcium intake and BMD for the girls, we first included age, height, weight, Tanner score, and BA, and additionally the other determinants triceps skinfold, arm circumference, grip strength, menarche, and energy intake. For the women, multivariate analyses include first height, weight, and age at menarche, and additionally triceps skinfold, grip strength, and energy intake for ultradistal BMD and daylight exposure and energy intake for middistal BMD.

For the overall data as well as for the individual countries, we calculated the adjusted BMD per quartile of calcium intake (Tables 3 and 4). Overall, BMD did not significantly vary among the categories of calcium intake. However, after additional adjustment for grip strength, triceps skinfold, and energy intake, ultradistal BMD in the highest calcium quartile for the women was significantly lower than in the second quartile. This finding could be attributed to the relatively low ultradistal BMD in Denmark in the highest calcium quartile. In the Netherlands, middistal BMD in the girls was significantly higher in the highest calcium intake quartile compared with lower quartiles, but this observation disappeared after additional adjustment for the other determinants of BMD in girls. In other countries, BMD did not significantly vary with calcium intake, among girls nor women.

Table Table 3..  BMD* in Quartiles of Calcium Intake Among 11–15 Year Old Girls, Overall and by Country
Thumbnail image of
Table Table 4..  BMD* in Quartiles of Calcium Intake Among 20–23 Year Old Women, Overall and by Country
Thumbnail image of

In linear regression analysis, after adjustment for age, height, weight, Tanner stage, and BA, middistal BMD was weakly associated with calcium intake among the girls (β for calcium [per 100 mg] = 0.57 ± 0.27, p = 0.03); a similar association was observed for ultradistal BMD (β = 0.56 ± 0.22, p = 0.01) (Table 5). Additional adjustment for the other predictors of BMD mentioned above and total daily energy intake lowered the regression coefficients (β = 0.47; p = 0.10 and 0.42; p = 0.06, for mid- and ultradistal BMD, respectively). For the women, there was no association between calcium intake and BMD at either site, when adjusting for height, weight and age at menarche (β [per 100 mg of calcium] = 0.47 ± 0.44, p = 0.29 for middistal BMD; β = −0.32 ± 0.43, p = 0.45 for ultradistal BMD). After additional adjustment, the regression coefficient for calcium in relation to middistal BMD was 0.79 (p = 0.11), while the inverse association between calcium intake and ultradistal BMD became statistically significant (β = −1.02, p = 0.03). Stratified by country, the estimated regression coefficients for calcium varied and in general did not significantly contribute to the prediction of the BMD, with exception of middistal BMD among girls in the Netherlands and ultradistal BMD among women in Denmark, similar to the results from the analysis of covariance (Table 5). The overall inverse association between calcium and ultradistal BMD among women was not consistent over countries.

Table Table 5..  Regression Coefficients for Calcium in Multivariate Linear Regression Models for Mid-distal and Ultradistal BMD, by Country and Overall
Thumbnail image of

Serum 25-hydroxyvitamin D concentrations were available for four countries (not for Finland and Italy). Vitamin D status was not associated with either middistal or ultradistal BMD or calcium intake. Inclusion of vitamin D in the multivariate models did not appreciably change the regression coefficients for calcium.

To investigate a potential nonlinear association between calcium intake and BMD, the population was stratified into four categories of calcium intake, roughly corresponding with the quartiles of the calcium intake distribution in our population: <600 mg/day, 600–900 mg/day, 900–1200 mg/day, and >1200 mg/day. Within each category, the adjusted regression coefficient for calcium was estimated (Table 6). Calcium significantly contributed to the variation in mid-distal BMD at the lowest intake level among girls. Among women, the regression coefficient for calcium at levels <600 mg/day was similar as for girls, but not statistically significant. After additional adjustment for the other determinants of BMD, the regression coefficients for calcium intake <600 mg/day were 6.28 (p = 0.02) for the girls and 7.87 (p = 0.05) for the women. There was no association between calcium and BMD at the highest intake level. However, there was no consistent trend over the four intake categories, and the interaction term was not statistically significant. For ultradistal BMD, the association with calcium intake did not depend on the level of calcium intake. Inclusion of a quadratic term for calcium in the linear regression models did not improve the fit of any of the models.

Table Table 6..  Regression Coefficients for Calcium in Different Strata of Calcium Intake
Thumbnail image of

Finally, it was evaluated whether the relation between calcium and BMD was dependent on pubertal development. Among premenarcheal girls (n = 424), calcium (per 100 mg) was a significant determinant of middistal BMD, adjusted for age, height, weight, Tanner stage, and BA (β = 1.42, p < 0.01), but not among postmenarcheal girls (β = 0.24, p > 0.05). The interaction was statistically significant (p = 0.04). For ultradistal BMD there was no significant interaction with pubertal development. In a further stratification, it was analyzed whether the association of calcium and BMD in different calcium intake categories was different for the pre- and postmenarcheal groups. The results were similar as for the total population of girls: for mid-distal BMD a relatively stronger association at levels <600 mg/day (β = 5.76, p = 0.41 in premenarcheal girls; β = 6.74, p < 0.05 in postmenarcheal girls), no association at intakes >1200 mg/day and no consistent trend in the intermediate intake levels. This observation does not support the notion that the relation between calcium intake and BMD before menarche is different than after menarche.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

In this cross-sectional study among 1116 adolescent girls and 526 young adult women in six European countries, no association was observed in multivariate models between dietary calcium intake and BMD of either the middistal or ultradistal radius in women. Additional adjustment for grip strength, triceps skinfold, and energy intake resulted in a significant inverse relation of ultradistal BMD with calcium intake, seen predominantly in the Danish women. Among the girls there was a weak positive association with BMD at both sites, which was no longer statistically significant after full adjustment for all determinants of BMD, or when comparing across calcium intake quartiles. The relation between calcium intake and middistal BMD among girls was seen predominantly among premenarcheal girls, also after full adjustment. There was no evidence for a different relation between calcium and BMD at different levels of intake; although there was a positive association at calcium intake levels <600 mg/day, the interaction was not significant and there was no consistent trend over intake categories.

We have investigated two age groups within the same populations, with exactly the same methods. Therefore, the comparability of data between age groups and among countries in this study is high. Densitometric and anthropometric data were collected by the same technical staff for all subjects with the same device, and identical questionnaires were used in each center.

Calcium intake was first assessed with a short screening questionnaire in a large random population (about 750 girls and 375 women in each country). For the actual study, girls and women were selected from the low and high ends of the calcium intake distribution in each country. In this study population, calcium (and energy intake) was measured by means of a 3-day diet record method. The calcium intake distribution obtained by this method was continuous within each country, so there was no gap between subjects with low and high intake. This is due to the phenomenon of regression to the mean and to differences in the methods used. It is likely that the food frequency questionnaire tends to overestimate intake in subjects with a high intake, by overestimation of the frequency of consumption; low intakes may have been underestimated in subjects with other sources of calcium than dairy products. The mean intakes assessed in this study are not necessarily representative for the respective countries. To test our hypothesis, it was important to have a large range of calcium intake rather than a representative sample of the population.

Measurements in the different countries were performed at different times of the year. Seasonal variation in BMD has been reported for postmenopausal women(31); bone loss was higher during fall and winter, mainly associated with low vitamin D status. To our knowledge, there is no information about seasonal changes in BMD in the age group we studied. The girls and women in our study are healthy and not a risk group with respect to vitamin D status; we did not observe an association of vitamin D status with BMD. When we included season in the multivariate analysis, assuming that the analyses in Finland, Denmark, and Italy were performed in the winter season (September–February) and in the other countries in spring/summer, the estimates for the regression coefficient for calcium (per 100 g/day) increased: for girls, middistal from 0.57–0.69, ultradistal from 0.56–0.68; for women, middistal from 0.47–0.53, ultradistal from −0.32 to −0.24. However, this change is marginal in relation to mean BMD, and significance levels did not change. Thus, an effect of seasonal variation on BMD cannot be excluded and if there is such an effect, it is likely to attenuate a potential relation between calcium and BMD. However, the additional analyses we performed suggest that such an effect is also likely to be minor.

In girls aged 11–15, accumulation of skeletal mass is an ongoing process, while 20- to 23-year-old women have attained peak bone mass of the radius.(3) In this young adult age group, calcium was not associated with middistal BMD. For the girls, the weak positive association was no longer significant after additional adjustment for menarche, grip strength, triceps skinfold, arm circumference, and energy intake. These are variables that are themselves associated with body size and bone size. Prentice et al.(30) have argumented that, when BMC and BA are not directly proportional, BMD partly varies with bone size. In our girl population, this was the case. If calcium intake also partly varies with body size, this will introduce a biased association between calcium and BMD. Therefore, we adjusted in all analyses for height, weight, and BA, but even then it is possible that the initial association between calcium and BMD was due to residual confounding related to body and bone size.

Girls and women from countries with higher mean usual intakes of calcium did not have higher BMD compared with girls and women in countries with lower calcium consumption. Also within most countries, calcium intake did not significantly contribute to the variation in BMD. The strength of the association between calcium and BMD within countries did not depend on the mean level of intake of calcium.

The observed inverse association between calcium intake and ultradistal BMD, representative of trabecular bone, was mainly due to the results in one country, Denmark. Because it was not seen in other countries, and there is no immediate likely biological explanation of this finding, which has not been reported in other studies, it is attributed to chance. The determinants of BMD in Denmark were not different from those in other countries, but in theory it is possible that high calcium consumption among Danish women is associated with specific dietary or lifestyle habits that are detrimental to trabecular bone density, and that were not assessed in this study.

In a number of observational studies, calcium intake was weakly associated with BMD at different skeletal sites, in children(8,10,11) as well as young adult women.(24-26) In other studies, calcium intake was not a significant determinant of BMD,(19-20,22,32) which could have been due to small variation in calcium intake in relatively homogeneous populations. Welten et al.(33) evaluated the relationship between dietary calcium and bone mass in premenopausal women in a meta-analysis. They concluded that there was overall evidence for a positive association, expressed in an overall partial correlation coefficient of 0.08 (95% confidence interval, 0.05–0.12). The present study shows a similar result, though not statistically significant: the adjusted correlation between calcium intake and middistal BMD among women was 0.07. For ultradistal BMD however, the association with calcium intake was inverse (partial r = −0.10, p < 0.05).

In a French cross-sectional study,(11) calcium was found to be a determinant of vertebral bone density in prepubertal children, not in older children. The present study yields a similar finding. It is a point of discussion whether increased calcium intake during childhood and adolescence leads to higher peak bone mass. One of the calcium intervention trials(13) suggested that additional calcium would affect BMD in prepubertal children only. Other interventions(12,14-16) were also performed in mainly prepubertal populations. In one of these,(16) a greater benefit of supplementation was observed in girls with lower spontaneous calcium intake. The reported intervention studies have followed children during maximal 3 years, after which they had not yet reached their peak bone mass. Follow-up studies indicate that several years after stopping calcium supplementation, an effect on bone density is no longer measurable. In a 15-year longitudinal study in the Netherlands,(21) calcium intake in adolescence was not a determinant of peak bone mass, when physical activity was adjusted for. These data, as well as the present results, suggest that the effect of calcium on BMD in prepubertal children is a transient phenomenon due to changes in bone remodeling, as postulated by Kanis.(7)

Requirements of calcium during growth are based on balance studies, taking into account calcium absorption, skeletal accretion rates, and obligatory losses. However, there is a lack of studies in adolescents, and the RDA is largely extrapolated from adult data. Table 7 gives the RDAs for the countries participating in the CALEUR study. Several authors have shown a positive calcium balance even at relatively high calcium intakes, taking this as evidence that current RDAs, especially for adolescents, may be too low to attain the full potential peak bone mass. Kanis(7) ascribes the observed association between calcium intake and balance to a statistical artefact and states that requirements should not be based on these data.

Table Table 7..  Recommended Dietary Allowances for Calcium in the CALEUR Participating Countries
Thumbnail image of

The present study does not support a role of dietary calcium in the attainment of peak bone mass at the average levels of intake in Europe. Current European RDAs for calcium appear to be adequate for an optimal development of peak bone mass.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The authors thank the following persons for contributing to the study by data collection, questionnaire development, data management, biochemical analyses, and technical and administrative support.

Denmark: Lise Hejl, Donna Arbuckle-Lund, Birthe Weinell, Jørg Andreasen, and Jens Bonde Petersen.

Finland: Eija Orreveteläinen and Pirkko Peronius.

France: Cecilia Iovanna, Jeanne Fabre-Monges, Josianne Casanova, Catherine Gauthier, Lynda Hedreville, Martine Charrel, Bénédicte Guibert, Christine Fourniel, and all the practitioners that participated in the study.

Italy: Daniela Bonofiglio, Anna Giorno, Stefania Catalano, Saveria Apuila, Diego Sisci, Nicola Fico, and Stefanie Marsico.

the Netherlands: Liesbeth Abbink, Geertje Huijbers, Carolien van Tuyl, Hanny Leezer, Henny Brants, Corine Beemster, Evelien Krijger, Wanda Bemelmans, Hanneke den Breeijen, Marco Bouman, Kitty van Loon, Anita Volkers, Marijke van Nispen, Jan Catsburg, Wilma Kreuning, Jaap-Jan Stevenhagen, Annemiek van de Velde, Anne-Claire Verheul, Angelique Prins, and Geert Hoorneman.

Poland: Malgorzata Rogalska-Niedžwiedž, Božena Wajszczyk, Anna Lachowicz, Zofia Chwojnowska, Elžbieta Kaszak, and Zofia Kozak.

The CALEUR study was financially supported by the European Commission DG XII (grant BMH1-CT94–1523), the Dairy Foundation for Nutrition and Health, CERIN, Laboratoire Innothera, the Dutch Ministry of Public Health, Oda og Hans Svenningsens Fond, Mejeriforeningen, Danish Dairy Board, the Polish State Committee of Scientific Research (grant 4P05D02910), Tetra Laval Service GmbH–Warsaw Branch.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Matkovic V, Jelic T, Wardlaw GM, Ilich JZ, Goel PK, Wright JK, Andon MB, Smith KT, Heaney RP 1994 Timing of peak bone mass in Caucasian females and its implications for the prevention of osteoporosis J Clin Invest 93:799808.
  • 2
    Bonjour JP, Theintz G, Buchs B, Slosman D, Rizzoli R 1991 Critical years and stages of puberty for spinal and femoral bone mass accumulation during adolescence J Clin Endocrinol Metab 73:555563.
  • 3
    Anderson JJB, Tylavsky FA, Halioua L, Metz JA 1993 Determinants of peak bone mass in young adult women: A review Osteoporos Int (Suppl) 1:S32S36.
  • 4
    Sowers MR, Galuska DA 1993 Epidemiology of bone mass in premenopausal women Epidemiol Rev 15:374398.
  • 5
    Matkovic V, Ilich JZ 1993 Calcium requirements for growth: Are current recommendations adequate? Nutr Rev 51:171180.
  • 6
    Matkovic V, Heaney RP 1992 Calcium balance during human growth: Evidence for threshold behavior Am J Clin Nutr 55:992996.
  • 7
    Kanis JA 1994 Calcium nutrition and its implications for osteoporosis. Part I. Children and healthy adults Eur J Clin Nutr 48:757767.
  • 8
    Chan GM 1991 Dietary calcium and bone mineral status of children and adolescents Am J Dis Child 145:631634.
  • 9
    Sentipal JM, Wardlaw GM, Mahan J, Matkovic V 1991 Influence of calcium intake and growth indexes on vertebral bone mineral density in young females Am J Clin Nutr 54:425428.
  • 10
    Rubin K, Schirduan V, Gendreau P, Sarfarazi M, Mendola R, Dalsky G 1993 Predictors of axial and peripheral bone mineral density in healthy children and adolescents, with special attention to the role of puberty J Pediatr 123:863870.
  • 11
    Ruiz JC, Mandel C, Garabedian M 1995 Influence of spontaneous calcium intake and physical exercise on the vertebral and femoral bone mineral density of children and adolescents J Bone Miner Res 10:675682.
  • 12
    Lloyd T, Andon MB, Rollings N, Martel JK, Landis JR, Demers LM, Eggli DF, Kieselhorst K, Kulin HE 1993 Calcium supplementation and bone mineral density in adolescent girls JAMA 270:841844.
  • 13
    Johnston CC, Miller JZ, Slemenda CW, Reister TK, Hui S, Christian JC, Peacock M 1992 Calcium supplementation and increases in bone mineral density in children N Engl J Med 327:8287.
  • 14
    Lee WTK, Leung SSF, Wang SH, Xu YC, Zeng WP, Lau J, Oppenheimer SJ, Cheng JCY 1994 Double-blind controlled calcium supplementation and bone mineral accretion in children accustomed to low calcium diet Am J Clin Nutr 60:744750.
  • 15
    Lee WTK, Leung SSF, Leung DMY, Tsang HSY, Lau J, Cheng JCY 1995 A randomized double-blind controlled calcium supplementation trial, and bone and height acquisition in children Br J Nutr 74:125139.
  • 16
    Bonjour JP, Carrie AL, Ferrari S, Clavien H, Slosman D, Theintz G, Rizzoli R 1997 Calcium-enriched foods and bone mass growth in prepubertal girls: A randomized, double-blind, placebo-controlled trial J Clin Invest 99:12871294.
  • 17
    Lee WTK, Leung SSF, Leung DMY, Cheng JCY 1996 A follow-up study on the effects of calcium-withdrawal and puberty on bone acquisition of children Am J Clin Nutr 64:7177.
  • 18
    Slemenda CW, Reister TK, Hui SL, Miller JZ, Christian JC, Johnston CC 1994 Influences on skeletal mineralization in children and adolescents: Evidence for varying effects of sexual maturation and physical activity J Pediatr 125:201207.
  • 19
    Mazess RB, Barden HS 1991 Bone density in premenopausal women: Effects of age, dietary intake, physical activity, smoking, and birth control pills Am J Clin Nutr 53:132142.
  • 20
    Fehily AM, Coles RJ, Evans WD, Elwood PC 1992 Factors affecting bone density in young adults Am J Clin Nutr 56:579586.
  • 21
    Welten DC, Kemper HCG, Post GB, Van Staveren WA 1994 Weight-bearing activity during youth is a more important factor for peak bone mass than calcium intake J Bone Miner Res 9:10891096.
  • 22
    Välimäki MJ, Kärkkäinen M, Lamberg-Allardt C, Laitinen K, Alhava E, Heikkinen J, Impivaara O, Mäkelä P, Palmgren J, Seppänen R, Vuori I, the Cardiovascular Risk in Young Finns Study Group 1994 Exercise, smoking, and calcium intake during adolescence and early adulthood as determinants of peak bone mass BMJ 309:230235.
  • 23
    Recker RR, Davies M, Hinders SM, Heaney RP, Stegman MR, Kimmel DB 1992 Bone gain in young adult women JAMA 268:24032408.
  • 24
    Metz JA, Anderson JJB, Gallagher PN 1993 Intakes of calcium, phosphorus, and protein, and physical activity level are related to radial bone mass in young adult women Am J Clin Nutr 58:537542.
  • 25
    Sowers M, Wallace RB, Lemke LH 1985 Correlates of forearm bone mass among women during maximal bone mineralization Prev Med 14:585596.
  • 26
    Kanders B, Dempster DW, Lindsay R 1998 Interaction of calcium nutrition and physical activity on bone mass in young women J Bone Miner Res 3:145149.
  • 27
    Hulshof KFAM, Van der Heiden-Winkeldermaat HJ, Kistemaker C, Van Beresteijn ECH 1989 De calciuminneming uit zuivelprodukten: Meting via een schriftelijke vragenlijst Voeding 50:302306.
  • 28
    Stafleu A, Burema J, Hulshof KFAM, Van Staveren WA 1989 Reproduceerbaarheid van een vragenlijst naar de calcium- en vitamine D-inneming T Soc Gezondheidsz 67:370374.
  • 29
    Verheul ACM, Prins AN, Kemper HCG, Kardinaal AFM, Van Erp-Baart MAJ 1998 Validation of a weight-bearing physical activity questionnaire in a study of bone density in girls and women Pediatr Exerc Sci 10:3847.
  • 30
    Prentice A, Parsons TJ, Cole TJ 1994 Uncritical use of bone mineral density in absorptiometry may lead to size-related artifacts in the identification of bone mineral determinants Am J Clin Nutr 60:837842.
  • 31
    Rosen CJ, Morrison A, Zhou H, Storm D, Hunter SJ, Musgrave K, Chen T, Wei W, Holick MF 1994 Elderly women in northern New England exhibit seasonal changes in bone mineral density and calciotropic hormones Bone Miner 25:8392.
  • 32
    Katzman DK, Bachrach LK, Carter DR, Marcus R 1991 Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls J Clin Endocrinol Metab 73:13321339.
  • 33
    Welten DC, Kemper HCG, Post GB, Van Staveren WA 1995 A meta-analysis of the effect of calcium intake on bone mass in young and middle aged females and males J Nutr 125:28022813.