Calcium-Enriched Foods and Bone Mass Growth in Prepubertal Girls
J. Clin. Invest.
© The American Society for Clinical Investigation, Inc.
Volume 99, Number 6, March 1997, 1287–1294
Calcium-enriched Foods and Bone Mass Growth in Prepubertal Girls:
A Randomized, Double-blind, Placebo-controlled Trial
Jean-Philippe Bonjour,* Anne-Lise Carrie,
and Rene Rizzoli*
Division of Bone Diseases, formerly Division of Clinical Pathophysiology, WHO Collaborating Center for Osteoporosis and Bone
Disease, Department of Internal Medicine,
Division of Nuclear Medicine, Department of Radiology,
and Reproduction, Department of Pediatrics, University Hospital, Geneva; and
Serge Ferrari,* Helene Clavien,* Daniel Slosman,
Division of Biology of Growth
Nestle Research Center, Vaud, Switzerland
High calcium intake during childhood has been suggested
to increase bone mass accrual, potentially resulting in a
greater peak bone mass. Whether the effects of calcium sup-
plementation on bone mass accrual vary from one skeletal
region to another, and to what extent the level of spontaneous
calcium intake may affect the magnitude of the response
has, however, not yet been clearly established. In a double-
blind, placebo-controlled study, 149 healthy prepubertal
girls aged 7.9
0.1 yr (mean
two food products containing 850 mg of calcium (Ca-suppl.)
or not (placebo) on a daily basis for 1 yr. Areal bone mineral
density (BMD), bone mineral content (BMC), and bone size
were determined at six sites by dual-energy x-ray absorpti-
ometry. The difference in BMD gain between calcium-sup-
plemented (Ca-suppl.) and placebo was greater at radial
(metaphysis and diaphysis) and femoral (neck, trochanter,
and diaphyses) sites (7–12 mg/cm
bar spine (2 mg/cm
per yr). The difference in BMD gains
between Ca-suppl. and placebo was greatest in girls with a
spontaneous calcium intake below the median of 880 mg/d.
The increase in mean BMD of the 6 sites in the low-calcium
consumers was accompanied by increased gains in mean
BMC, bone size, and statural height. These results suggest a
possible positive effect of calcium supplementation on skel-
etal growth at that age. In conclusion, calcium-enriched
foods significantly increased bone mass accrual in prepu-
bertal girls, with a preferential effect in the appendicular
skeleton, and greater benefit at lower spontaneous calcium
J. Clin. Invest.
1997. 99:1287–1294.) Key words:
bone mineral density
SEM) were either allocated
per yr) than in the lum-
The bone mass accrual occurring during childhood and adoles-
cence is a major determinant of peak bone mass, and thereby
of the risk of osteoporotic fractures occurring in advanced age
(1–3). At the end of the growth period, a large variance in
bone mineral density (BMD)
and content (BMC), in either
the axial or the appendicular skeleton, is observed both in
healthy females and males (4–8). Many genetic and environ-
mental factors have been suggested to influence bone mass ac-
cumulation during this period (9–15).
It is usually accepted that increasing calcium intake during
childhood and adolescence can promote a greater increase in
bone mass, and thereby a higher peak bone mass (16–21). An-
alyzing the relationship between bone mass and spontaneous
calcium intake, however, indicates that not all studies (22–27)
have found a positive correlation between these two variables.
Only a few prospective randomized double-blind interven-
tion trials have examined the effects of calcium supplements in
children and adolescents (12, 28–30). Although the results of
these studies suggest that calcium supplementation can posi-
tively influence bone mass gain in this population, the magni-
tude of the effects appears to be different in the axial and ap-
pendicular skeleton as well as at the metaphyseal or diaphyseal
levels (12, 28–30). Significantly, the mean bone mass gain in re-
sponse to calcium supplements has been found to be only
modest at the lumbar spine and proximal femur, two major
sites of subsequent osteoporotic fracture (12, 28, 30). The diffi-
culty of consistently demonstrating a significant effect of cal-
cium in the various parts of the skeleton may be related to dif-
ferences in the osteodensitometric method used (for example,
dual photon absorptiometry [DPA] [12, 28] versus the more
precise dual-energy x-ray absorptiometry [DXA] ), to dif-
ferences in the stage of pubertal maturation at the time of cal-
cium supplementation (28, 29), or to differences in the level of
spontaneous calcium intake, as observed in adult subjects (31).
We conducted a randomized, double-blind, placebo-con-
trolled study in a homogenous cohort of prepubertal girls to
assess the effects of calcium supplementation on bone mass
gain at various sites of the skeleton. In the context of a large-
scale program for nutritional prevention during childhood of
osteoporosis, as well as to anticipate the risk of low long-term
compliance to pharmaceutical calcium supplements at a popu-
lation level, we administered palatable calcium-enriched foods
commonly taken at breakfast or as snacks to children living in
an affluent western society.
Subjects and study design.
Committee of the Department of Pediatrics of the University Hospi-
The protocol was approved by the Ethical
Address correspondence to Prof. Jean-Philippe Bonjour, M.D., Divi-
sion of Bone Diseases, formerly Division of Clinical Pathophysiology,
Department of Internal Medicine, University Hospital, CH-1211
Geneva 14, Switzerland. Phone: 41-22-3729950; FAX: 41-22-3829973.
Received for publication 23 July 1996 and accepted in revised form
7 January 1997.
density; BMC, bone mineral content; BMD, bone mineral density;
Ca-suppl., calcium-supplemented; DXA, dual energy x-ray absorpti-
ometry; ROI, region of interest.
Abbreviations used in this paper:
BMAD, bone mineral apparent
Bonjour et al.
tal of Geneva. Informed consent was obtained from the parents and
their children. Healthy prepubertal caucasian girls with a mean age
SEM) of 7.93
0.04 yr (range: 6.6–9.4 yr) were recruited through
the Public Health Youth Service of the Geneva district over an 8-mo
period, from April to November, 1993. The following exclusion crite-
ria were applied: no parental approval, ratio weight/height less than
the 3rd or greater than the 97th percentile according to Geneva refer-
ence values, presence of physical signs of puberty, chronic disease,
gastro-intestinal disease capable of inducing malabsorption, congeni-
tal or acquired bone disease, or regular use of medication. 149 sub-
jects were randomized into two groups stratified according to their
level of spontaneous calcium intake to avoid an uneven distribution
of the low- or high-calcium consumers. A limit of 900 mg/d, as as-
sessed by an initial food frequency questionnaire, was used to deter-
mine the cutoff point for each group. For unknown reasons, the drop-
out rate after 4 mo of enrollment was larger (15 vs. 4) in one of the
two experimental groups. Since for technical reasons production of
the specially prepared foods could not be pursued beyond a given
time, 10 out of the 149 enrolled subjects were deliberately assigned
for the rest of the enrollment period to the group having shown the
greater dropout rate during the first phase of the study. This explains
the difference in the number of subjects enrolled between the two
groups of the intention-to-treat cohort (see results). At the end of the
intervention period, opening of the blinded code revealed that the
larger dropout rate during the first 4 mo was in the calcium-supple-
Calcium supplement (Ca-suppl.).
used to fortify several food products: cakes (two kinds), biscuits, fruit
juices, powdered drinking chocolate, chocolate bars, and yogurts. The
calcium contents (mg/serving) of calcium-enriched vs. placebo food
products was as follows: chocolate cakes, 516 vs. 33; caramel cakes,
512 vs. 41; biscuits, 548 vs. 8; fruit juices, 383 vs. 11; powdered drink-
ing chocolate, 530 vs. 4; chocolate bars, 429 vs. 80; yogurts, 478 vs.
174. The corresponding phosphorus content (mg/serving) were: choc-
olate cakes, 285 vs. 52; caramel cakes, 274 vs. 46; biscuits, 311 vs. 37;
fruit juices, 73 vs. 21; powdered drinking chocolate, 277 vs. 26; choco-
late bars, 252 vs. 98; yogurt, 304 vs. 146. The corresponding energy
contents (kcal/serving) were: chocolate cakes, 185 vs. 187; caramel
cakes, 186 vs. 188; biscuits, 197 vs. 203; fruit juices, 120 vs. 119; pow-
dered drinking chocolate, 64 vs. 60; chocolate bars, 270 vs. 274; yo-
gurt, 182 vs. 180. These foods had previously been tested for palat-
ability and long-term tolerance in a pilot study involving a small
cohort of adolescent girls. On average, the intake of two calcium-
enriched products per day provided a calcium supplement of about
850 mg. The control (placebo) group was given daily two similar
products in terms of energy, protein, lipid and mineral content, but
without added calcium. The maximal difference between the calcium-
enriched and placebo foods was as follows: energy, 6 kcal; proteins,
0.2 g; lipids, 0.3 g; carbohydrates (essentially lactose), 0.9 g. Over the
entire intervention period (48 wk) each subject received 12 sets of 56
food products (1 set every 4 wk). Parents were instructed to ensure
consumption of the two food products every day in place of similar
foods taken for breakfast or snacks. Compliance was recorded by the
parents and verified through regular phone calls and interviews by a
trained dietician (H. Clavien). Clinical examinations were conducted
by a pediatrician (G. Theintz) at the beginning and at the end of the
intervention period, with recording of anthropometric variables and
assessment of pubertal status.
Calcium intake assessment.
An estimate of the spontaneous cal-
cium intake during the intervention period was obtained by averaging
the results of three frequency questionnaires made at 0, 24, and 48 wk.
Measurement of bone variables.
and areal bone mineral density (BMD, g/cm
mass were determined at the beginning and the end of the study pe-
riod by dual energy x-ray absorptiometry (DXA) using a Hologic
QDR-2000 instrument (Waltham, MA). Six skeletal sites were as-
sessed: distal metaphysis of the radius, diaphysis of the radius (desig-
nated respectively as the ultradistal region and as the one-third re-
Calcium from milk extract was
Bone mineral content (BMC, g)
) as estimates of bone
gion of the radial shaft according to the manufacturer’s software),
femoral neck, femoral trochanter, femoral diaphysis, and L2–L4 ver-
tebrae in antero-posterior view as previously reported (8, 32). The co-
efficient of variation of repeated measurements at these sites (as de-
termined in young healthy adults) varied between 1.0 and 1.6% from
BMD, and 0.3 to 3.0% for BMC and bone area. The possible influ-
ence of the calcium-enriched foods on skeletal size was assessed by
analyzing the changes in scanned bone area (mm
of interest described above. At the level of the femoral neck, the
height (i.e., the dimension parallel to the hip axis) of the region of in-
terest (ROI) rectangular box was maintained constant from one ex-
amination to the other. Only the width (i.e., the dimension perpendic-
ular to the hip axis) was adjusted to the growth of the bone to scan a
similar proportion of soft tissue on each side of the femoral neck. For
the trochanter, the ROI box was maintained constant. In addition, at
the level of L2–L4, the height (mm) of the region of interest was de-
termined to be the number of lines from the lower edge of L4 to the
upper edge of L2 multiplied by a conversion factor (1 line
mm). An estimate of the mean vertebral width was derived by divid-
ing the area by the height of L2–L4. At the level of the midfemoral
shaft, the external diameter of the diaphysis was estimated by an addi-
tional analysis. A region of interest was defined at baseline as having
a constant length (45 lines
midfemoral shaft (halfway between the greater trochanter and the
upper edge of the patella). To assess to what extent calcium-induced
increase in areal BMD would be due to an effect on bone size as com-
pared to volumetric mineral density, an estimate of this latter variable
(bone mineral apparent density, BMAD) was calculated as previously
described for lumbar spine, midradius, and femoral neck (33–34).
Follow-up after treatment discontinuation.
ment cohort, 100 girls (45 from the placebo and 55 from the calcium-
supplemented group) underwent a third examination 1 yr after the
end of the intervention phase. Bone and anthropometric variables
were measured, and calcium intake, as assessed by a fourth frequency
questionnaire, was also recorded.
Expression of the results and statistical analysis.
tervention lasted 48 wk, the changes in the various anthropometric
and osteodensitometric variables are expressed in units/yr without
correction for the difference between 1 yr and 48 wk. The differences
in the anthropometric and osteodensitometric variables in the cal-
cium-enriched and placebo groups were analyzed both in terms of ac-
tive-treatment cohort, (which included the subjects who consumed
the study foods during 48 wk) and an intention-to-treat cohort (ac-
counting for all subjects who entered the study and had measure-
ments redone at 48 wk). Bone mass gains were determined separately
for each skeletal site; the average BMD changes at the six studied
sites were also calculated. Results were expressed as either absolute
per yr) or relative terms (%/yr). Mean change in the bone area
of the six regions of interest was also computed to assess potential
changes in bone size. All results are given as mean
ences between the Ca-suppl. and placebo groups were evaluated us-
ing a two-tailed Student’s
-test for unpaired values. In addition, for
the subjects on whom three measurements (at 0, 48, and 96 wk) were
available, an ANOVA test for repeated measures was applied to
evaluate whether any calcium effect observed at the end of the inter-
vention period would be maintained one year after treatment discon-
tinuation. Finally, a comparative analysis of the effect of Ca-suppl. on
areal (BMD) vs. volumetric bone mineral density (BMAD) in radial
and femoral diaphysis was made by calculating the changes in Z score
from the mean values recorded at baseline in the intention-to-treat
) of the six regions
45.14 mm) starting from the
Within the active treat-
Although the in-
SEM. The differ-
Cohorts studied and compliance.
initially enrolled, 144 had their BMD measured at baseline and
48 wk later (intention-to-treat cohort). Among the five subjects
Of the 149 prepubertal girls
Calcium-Enriched Foods and Bone Mass Growth in Prepubertal Girls
who declined to be examined at the end of the 1-yr interval,
three belonged to the Ca-suppl. group, and two to the placebo
group. At baseline no difference between the two groups was
found with respect to age, statural height, body weight, or
BMD at any skeletal site (Table I). From the initial cohort, 108
girls remained compliant (active-treatment cohort) to the
study food products over the whole study period. Among the
41 subjects (28%) who dropped out from the study, 40 re-
ported lassitude with the consumption of two imposed food
products per day, and one left the Geneva district. Dropout
times after the onset of the intervention phase ranged from 4
to 32 wk (mean, 15.3
16) in the Ca-suppl. and placebo groups, re-
spectively. In the active-treatment cohort, 89.9
53) of the study foods were consumed
25) and 4 to 28 wk (mean,
throughout the intervention period in the Ca-suppl. and pla-
cebo groups, respectively, as estimated by regular interviews of
the subjects’ mothers. After 48 wk, none of the examined sub-
jects displayed any sign of puberty.
Spontaneous calcium intake, as determined
by three frequency questionnaires at 16-wk intervals, were
similar in both groups (916
42, and 879
and placebo groups, respectively). Taking into account the in-
dividual compliance over the 48-wk intervention period, the
mean daily amount of supplemented calcium ingested with the
enriched foods was 807
10 mg. Thus, the calcium intake in-
creased from 916
42 to 1723
44 mg/d in the Ca-suppl. group.
Anthropometric and bone mass changes.
anthropometric and osteodensitometric values recorded in
the active-treatment groups after a 1-y interval. At baseline,
the mean values in the Ca-suppl. and the placebo groups did
not differ between the two groups, and were similar to those of
the intention-to-treat groups presented in Table I. The in-
creases in statural height, body weight, and BMD at the vari-
ous skeletal sites in the placebo group were within the ex-
pected range for prepubertal girls of this age, and were all
statistically significant (
0.001 by two-sided paired sample
analysis). The corresponding BMD gains calculated in either
per yr) or relative (%/yr) terms are pre-
sented in Table II and Fig. 1, respectively. At all sites, the
mean increment in bone mass was greater in the Ca-suppl.
than in the placebo group. However, the difference in BMD
was smaller in the lumbar spine than in the radial and femoral
sites. The magnitude of the calcium effect, further estimated as
the ratio of the mean bone mass gain in the Ca-suppl. group
over that observed in the placebo group (
cebo) (Table II), ranged from 1.09 to 1.78 (mean of six mea-
sured skeletal sites, 1.52). Interestingly, the
cebo ratio for lumbar spine BMD was quite similar to that
calculated for height gain. In contrast to all osteodensitometric
variables and to statural height, the mean change in either
body weight or BMI was not greater in the Ca-suppl. group
than in the placebo group (Table II).
36 mg/d in Ca-suppl.
Table II shows the
Table I. Baseline Characteristics of Prepubertal Girls Assigned
to Consume Foods Enriched in Calcium (Intention-to-Treat
Cohort) or Not (Placebo)
No. of subjects
Radial metaphysis BMD (g/cm
Radial diaphysis BMD (g/cm
Femoral neck BMD (g/cm
Femoral trochanter BMD (g/cm
Femoral diaphysis BMD (g/cm
Lumbar spine (L2–L4) BMD (g/cm
All values are given as mean
None of the differences between means recorded in the calcium-supple-
mented and the placebo groups were found to be statistically significant
SEM. Areal BMD values are in mg/cm
Table II. Anthropometric and Osteodensitometric Values Determined at Baseline and 48 Wk Later in Prepubertal Girls
Consuming Foods Enriched or Not in Calcium (Active-Treatment Cohort)
Baseline48 wkGainBaseline 48 wkGain
Radial metaphysis BMD (g/cm
Radial diaphysis BMD (g/cm
Femoral neck BMD (g/cm2)
Femoral trochanter BMD (g/cm2)
Femoral diaphysis BMD (g/cm2)
Lumbar spine (L2–L4) BMD (g/cm2)
All values are given as mean?SEM. Areal BMD values are in mg/cm2 . All values recorded at 48 wk were statistically significant as compared to base-
line values at P ? 0.001 by two-sided paired Student’s t-test analysis. ? Ca suppl. / ? Placebo values correspond to the ratio of the gains in the calcium
supplemented over that in the placebo group as calculated from the means of the individual differences recorded at 48 wk and at baseline. *P ? 0.08,
‡P ? 0.05, §P? 0.02, ?P ? 0.01 as compared to the corresponding change in the placebo group.
Bonjour et al.
The mean increment in BMD for the six measured skeletal
sites, as expressed in both absolute and relative terms, is illus-
trated in Fig. 2. The difference between the Ca-suppl. and pla-
cebo groups was highly significant in both the active-treatment
and intention-to-treat cohorts. The effect on mean BMD gain
remained highly significant (P ? 0.02), even after adjustment
for age, height, and weight at entry (data not shown). The cor-
responding mean increases in both BMC and bone area were
also greater in the Ca-suppl. group than in the placebo group,
but did not reach statistical significance (data not shown). In
the Ca-suppl. subjects belonging to the intention-to-treat co-
hort, the change in mean BMD of the six skeletal sites was pos-
itively correlated (r ? 0.24, P ? 0.038, n ? 77) with the cumu-
lative amount of calcium ingested from the enriched foods.
Effects of calcium supplements in relation to spontaneous
calcium intake. To evaluate the effects of calcium-enriched
foods on bone mass gain in relation to spontaneous calcium in-
take (as assessed by averaging the results from the food fre-
quency questionnaires made at 0, 24, and 48 wk), the Ca-suppl.
and placebo groups were divided in two subgroups according
to the median of their spontaneous calcium intake (855 and
880 mg/d in the intention-to-treat and active-treatment co-
horts, respectively). At baseline, there was no statistically sig-
nificant difference between the Ca-suppl. and placebo groups
Figure 1. BMD gains in prepubertal girls consuming food products
enriched or not in calcium during 48 wk (active-treatment cohort) or
for various lengths of time (intention-to-treat cohort). Bars represent
means?SEM. The number of subjects in the active-treatment cohort
was as follows: placebo (open bars) ? 53, Calcium-Supplemented
(dashed bars) ? 55. In the intention-to-treat cohort, numbers were:
placebo (open bars) ? 67, calcium-supplemented (dashed bars) ? 77.
#P ? 0.08; *P ? 0.05; **P ? 0.02; ***P ? 0.01.
Figure 2. Mean BMD gains in prepubertal girls consuming food
products enriched or not in calcium during 48 wk (active-treatment
cohort) or for various lengths of time (intention-to-treat cohort). For
each subjects the yearly gains recorded at the six studied skeletal sites
presented in Fig. 1 (relative values), and Table III (absolute values)
were averaged. Bars represent means?SEM. The number of subjects
in the active-treatment cohort was as follows: placebo (open bars) ?
53, calcium-supplemented (dashed bars) ? 55. In the intention-
to-treat cohort, numbers were: placebo (open bars) ? 67, calcium-
supplemented (dashed bars) ? 77. ***P ? 0.001.
Table III. Changes in Anthropometric Variables, BMD, BMC and Bone Size in Response to Calcium-Enriched Foods in
Prepubertal Girls Distributed According to their Spontaneous Calcium Intake
Intention-to-treat cohort Active-treatment cohort
PlaceboCa suppl. PlaceboCa suppl.PlaceboCa suppl.PlaceboCa suppl.
n ? 36 n ? 36n ? 31n ? 41 n ? 29n ? 25n ? 24n ? 30
? Bone area
4.7?0.25.2?0.2* 5.3?0.25.1?0.24.8?0.25.4?0.3* 5.3?0.2 5.3?0.2
Values are given as mean?SEM. aThe median of the spontaneous dairy calcium intake was 855 and 880 in the intention-to-treat and active-treatment
cohorts, respectively. BMD, BMC, and bone area values are the means of the six skeletal sites presented in Table IV and averaged in each subject.
The total calcium intake corresponds to the sum of the dairy calcium taken spontaneously and of calcium in the enriched food products. *P ? 0.08,
‡P ? 0.05, §P ? 0.02, ?P ? 0.01 as compared to the corresponding placebo group.
Calcium-Enriched Foods and Bone Mass Growth in Prepubertal Girls
in terms of BMD at any site (data not shown); this was ob-
served in both the spontaneously low- and high-calcium intake
groups, which had mean calcium intakes of 650?33 (n ? 72)
and 1143?33 mg/d (n ? 72), respectively. Both statural
height and body weight, however, were significantly greater
among the spontaneously high- versus low-calcium consumers
(129.1?0.7 vs. 126.5?0.7 cm, P ? 0.001; 27.5?0.5 vs. 25.7?0.5
kg, P ? 0.001, respectively). The mean BMD gain in response
to calcium supplements, as expressed in absolute (Table III) or
relative (Fig. 3) terms, was more pronounced among the low-
calcium consumers (? Ca-suppl. /? placebo: 1.42–1.58 g/cm2,
depending on the site examined). Indeed, linear regression
analysis showed that the mean BMD gain at the six skeletal
sites was significantly correlated to the total calcium consump-
tion in the spontaneously low (r ? 0.32, P ? 0.007), but not in
the high-calcium intake subgroups of the intention-to-treat co-
hort (r ? 0.02, NS).
Effect of calcium on BMC, skeletal size, and statural height.
As shown in Table III, there were also significant increments
in both mean BMC (? Ca-suppl /? placebo ratio, 1.21–1.23)
and mean bone area of the six investigated skeletal sites (? Ca-
suppl. /? placebo ratio, 1.15–1.16) in the subgroup of low-
calcium consumers. The change in mean bone size was ac-
companied by a trend toward greater statural height gain (?
Ca-suppl. /? placebo ratio, 1.11–1.13) which nearly achieved
statistical significance. Indeed, the reduced gain in bone area
and statural height observed in the spontaneously low- versus
high-calcium consumers appeared to be corrected by the cal-
cium-enriched foods (Table III).
A similar pattern was observed in the projected area of the
lumbar spine, and further morphometric analysis indicated
that in the spontaneously low-calcium intake subgroup, the
vertebral height gain at L2–L4 was significantly greater in the
Ca-suppl. than in the placebo group (Table IV). At the level of
the femoral midshaft, a significant difference in width change
was also observed between the Ca-suppl. and the placebo groups
in the spontaneously low-calcium intake subgroup (Table 4).
To evaluate whether the calcium effect on bone mass was
associated with a commensurate increase in volumetric min-
eral density, gain in Z scores from baseline values was calcu-
lated for both BMD and BMAD at the level of the radial and
femoral diaphyses, two sites where the response was particu-
larly significant. In the low-calcium consumers of the active
treatment cohort presented in Table III, the difference in
BMD Z score gain between the two groups was greater (Ca-
suppl. [n ? 29] minus placebo [n ? 25] ? ? 0.28 and ? 0.18,
P ? 0.05, in radial and femoral diaphysis, respectively) than
that in BMAD Z score (Ca-suppl. minus placebo ? ?0.05 and
?0.08, P ? 0.1, in radial and femoral diaphysis, respectively).
Follow-up 1 yr after treatment termination. 100 girls (Ca-
suppl., n ? 55; placebo, n ? 45) out of the 108 in the active
treatment cohort were followed up to 1 yr after they stopped
Figure 3. Changes in BMD in response to calcium-enriched foods in
prepubertal girls distributed according to their spontaneous calcium
intake. Bars represent means?SEM. BMD values are the mean of
the 6 skeletal sites. The corresponding absolute changes in BMD are
presented in Table III. The median spontaneous calcium intake is in-
dicated above the bars. The mean spontaneous calcium intake was
650?16 and 1143?33 mg/d in the intention-to-treat cohort, and
675?17 and 1185?39 mg/d in the active-treatment cohort. ***P ?
0.01 as compared to the corresponding placebo group.
Table IV. Changes at the Level of the Lumbar Spine and Midfemoral Shaft at the End of the Intervention (? 2-1) and 1 yr After
Discontinuation of the Treatment (? 3-1) in Girls Having a Spontaneously Low Calcium Intake
PlaceboCa suppl. PlaceboCa suppl.
n ? 24n ? 25 n ? 24n ? 25
? Bone area (mm2/yr)
? Width (mm/yr)
? Height (mm/yr)
? BMC (mg/yr)
? BMD (mg/cm2/yr)
? Bone area (mm2/yr)
? Width (mm/yr)
? BMC (mg/yr)
? BMD (mg/cm2 per yr)
Values are mean?SEM recorded in girls having had a third examination 1 yr after treatment discontinuation (active-treatment cohort). All gains
were statistically significant by paired t-test analysis. *P ? 0.08, ‡P ? 0.05, §P ? 0.02 as compared to the placebo group using an ANOVA test for re-
Bonjour et al.
the consumption of the study foods. Most of the absolute dif-
ferences in mean bone mass and size gains recorded at the end
of the intervention period was still detectable 1 yr after termi-
nation of the dietary intervention (mean BMD, 58?3 vs. 53?3
mg/cm2, P ? 0.05; mean BMC, 1265?66 vs. 1136?56 mg, P ?
0.101; mean bone area, 143?6 vs. 131?5 mm2, P ? 0.08; statu-
ral height, 11.9?0.3 vs. 11.2?0.3 cm, P ? 0.09, in the Ca-suppl.
and placebo groups, respectively, ANOVA test for repeated
measurement). Among the 100 girls who underwent a third ex-
amination, 49 belonged to the group of spontaneously low-cal-
cium consumers in whom the effects of the calcium-enriched
foods on BMD, BMC, scanned bone area and statural height
had been particularly significant. At the levels of the lumbar
spine and femoral midshaft, the differences in height and
width gain, observed at the end of the intervention period
were still statistically significant 1 yr after treatment discontin-
uation (Table IV).
A recently held National Institutes of Health Consensus De-
velopment Conference concluded that a calcium intake above
the Recommended Dietary Allowance (RDA) of 800 mg/d
may lead to an increased rate of bone accumulation in children
aged 6–10 yr (35). This report underlined the necessity to con-
sider the effects of calcium on regional changes in bone mass,
and recommended the improvement of strategies to achieve
and maintain optimal dietary intake of calcium by both nutri-
tional and supplemental means (35). It is appropriate to dis-
cuss the present study in relation to these considerations.
Compliance rate. In the present study, the compliance rate
was about 75% after 1 yr. A higher rate could have been antic-
ipated, considering that calcium was supplemented using en-
riched palatable foods usually taken at either breakfast or as
snacks, and known to be well accepted by children. The psy-
chological constraint on the families involved in controlled tri-
als, however, and the fact that they were blinded with respect
to the study product, played an important role in this phenom-
Changes in BMD gains. Overall, our results indicate that
increasing calcium intake from approximately 900 to 1750 mg/d
led to a substantial increment, 3.5 to 5.0%/yr (Fig. 2), in the
rate of bone mass accumulation in prepubertal girls. The ap-
pendicular sites appeared to be more responsive than the axial
skeleton. Indeed, the effect was found to be minimal in the
lumbar spine, as assessed by measuring spinal BMD in the
classical antero-posterior view. Among the appendicular re-
gions, the most obvious positive effects were observed at sites
essentially consisting of cortical bone (e.g., radial and femoral
diaphysis). Nevertheless, BMD gain was also positively influ-
enced by calcium supplementation at the trochanter and femo-
ral neck levels as well as at the radial metaphysis. So far, only
one placebo-controlled study has compared the effects of cal-
cium supplementation at different sites of the skeleton in 22
twin pairs of prepubertal female and male children (28). The
most prominent effects were also observed in the appendicular
skeleton at the level of the midshaft radius, with weaker re-
sponses at the levels of lumbar spine (? 1.0%/yr) and proximal
femur. In this region, the calcium effect did not achieve statisti-
cal significance, despite a longer (3 yr) duration of the inter-
vention. Similarly, in a study undertaken in prepubertal Chi-
nese girls, calcium supplements were associated with a higher
radial BMD gain (30). Taken together, these results (as well as
other studies in experimental animals  and human subjects
[37, 38]) suggest that the appendicular skeleton, (particularly
regions predominantly composed of compact bone) appear to
be more sensitive than the axial skeleton to the effect of cal-
cium supplementation above RDA. Importantly, however, cal-
cium supplementation appears also to have an effect on BMC
and bone size at the level of the lumbar spine. Therefore, the
use of only areal BMD in the classical DPA/DXA antero-pos-
terior view may well underestimate the effect of calcium sup-
plementation on the axial skeleton during growth.
Spontaneous calcium intake and response to calcium sup-
plements. A thorough meta-analysis of calcium balance data
recently indicated that there is a threshold level of calcium in-
take in humans (39), as previously demonstrated in animal ex-
periments (40). In children and adolescents, the level above
which a rise in calcium intake is no longer associated with an
increase in calcium retention appears to be set between 1200
to 1500 mg/d (39). Accordingly, our study clearly indicates that
the response to calcium supplementation in terms of bone
mass gain was markedly dependent on the level of spontane-
ous calcium intake. Indeed, the benefit was 3.5-fold (2.1 vs.
0.6%/yr) greater in the spontaneously low- (e.g., less than 850
mg/d) versus high-calcium consumers. Our data suggest that a
calcium intake below 800–900 mg/d may not be sufficient for
optimal bone mass accrual in 7–9 yr-old prepubertal girls. On
the other hand, the effect of calcium supplementation appears
to be rather mild at that age when the spontaneous consump-
tion is larger than 1100–1200 mg/d. To our knowledge, this is
the first evidence from a placebo-controlled study showing
that there is a calcium intake threshold below which supple-
mentation can be particularly beneficial to prepubertal chil-
dren. A review of controlled clinical trials testing the effect of
calcium supplementation on BMD in postmenopausal women
also indicated a greater benefit in low-calcium consumers (41).
Our results suggest that the use of a simple frequency ques-
tionnaire could be useful for identifying children who would
particularly benefit from an increased calcium intake.
Mechanisms of the effects of calcium supplements. The posi-
tive effects of calcium previously reported in double-blind, pla-
cebo-controlled clinical trials carried out in prepubertal chil-
dren (28, 30) have essentially been ascribed to a reduction in
bone remodeling (42). In favor of this notion was the fact that
the level of serum osteocalcin, a biochemical marker of bone
remodeling in adults, was significantly reduced in the calcium-
supplemented prepubertal group (28). Such an explanation is
compatible with the currently favored mechanism proposed to
account for the inhibitory effect of calcium on age-related
bone loss (31, 41, 43–46).
In the two aforementioned studies, the positive bone mass
effect of calcium (given as either citrate malate or carbonate
salts) was not associated with any change in either bone area
or statural height (28, 30). Our study was not aimed at eluci-
dating the mechanism of calcium effect on bone mass. Conse-
quently, and to maintain the compliance level as high as possi-
ble, the protocol did not include any blood or urine sampling.
Nevertheless, careful examination of the changes in scanned
bone area and in statural height suggests that calcium supple-
mentation, as provided in our study, could affect bone model-
ing. Indeed, in the group of spontaneous low-calcium consum-
ers, calcium-enriched foods enhanced the gain of both mean
scanned bone area and statural height to the level achieved in
Calcium-Enriched Foods and Bone Mass Growth in Prepubertal Girls
the spontaneously high-calcium consumers. Morphometric
analysis of the changes observed in the lumbar spine and in
femoral diaphysis suggests that calcium could enhance both
the longitudinal and the cross-sectional growth of the bones.
As reviewed by Nordin (4), the possibility that milk calcium
may positively influence longitudinal growth has been sus-
pected for several decades. In many studies it was difficult to
discriminate between the effects of calcium from those of milk-
contained proteins or energy. In the present study, both cal-
cium-enriched and placebo food products provided identical
amounts of proteins and energy. Thus, the possibility that milk
calcium may affect not only the remodeling, but also the mod-
eling of the skeleton, deserves to be further investigated. As
previously indicated, the tested foods were enriched with cal-
cium from milk extract. With the daily consumption of two
tested foods, the mean difference in the phosphorus intake was
386 mg/d as compared to 870 mg/d mean difference of calcium.
Based on previous dietary diary recording in prepubertal girls
of that age (47), one can estimate that the enriched foods in-
creased the intake of phosphorus by 30–35% as compared to a
100% increase for calcium intake. To what extent this increase
in the phosphorus intake played a contributing role in the stim-
ulation of bone mass accrual reported in this study remains to
An important issue is whether the positive effect on bone
mass gain will be sustained after discontinuation of calcium
supplementation, and thus be ultimately translated into higher
peak bone mass. As mentioned above, our data suggest that
milk calcium supplementation may increase bone mass not
only by inhibiting the process of remodeling, but also by stimu-
lating bone modeling. According to this hypothesis one may
expect that the contribution of the modeling effect on bone
mass will not completely disappear after discontinuation of the
calcium supplementation. Indeed, our data suggest that part of
the gain in bone mass and bone size was not lost 1 yr after the
end of the calcium supplementation. Further prospective long-
term studies will eventually provide a definite answer to the
question of whether or not a transient rise in the intake of cal-
cium during growth can permanently increase the amount of
bone tissue through a modeling effect in certain parts of the
In summary, calcium-enriched food given to prepubertal
girls with a spontaneous calcium intake below RDA signifi-
cantly increased bone mass gain at several sites of the appen-
dicular skeleton, including the proximal femur. Morphometric
analysis provided evidence for a calcium effect on bone model-
ing which was associated with an increase in statural growth.
We are indebted to Dr. P. Hazeghi, M.D., and to the Geneva Public
Youth Service for the recruitment of the subjects, to the team of the
bone densitometry unit, and to Mrs. S. Gardiol, C. Ang and M.-C.
Brandt for their help in preparing the manuscript.
This work was supported by Nestec Ltd., Lausanne, Switzerland,
and by the Swiss National Science Foundation (grants 32-32415.91
1. Riggs, B.L., and L.J. Melton III. 1988. Osteoporosis. Etiology, Diagnosis,
and Management. Raven Press, New York.
2. Hui, S.L., C.W. Slemenda, and C.C. Johnston, Jr. 1990. The contribution
of bone loss to postmenopausal osteoporosis. Osteoporosis Int. 1:30–34.
3. Johnston, C.C., Jr., and C. Longcope. 1990. Premenopausal bone loss - a
risk factor for osteoporosis. N. Engl. J. Med. 323:1271–1272.
4. Nordin, B.E.C. 1976. Calcium, phosphate, and magnesium metabolism.
Clinical Physiology and Diagnostic Procedures. Churchill Livingstone, Edin-
burgh, London, and New York.
5. Hui, S.L., C.C. Johnston, Jr., and R.B. Mazess. 1985. Bone mass in normal
children and young adults. Growth. 49:34–43.
6. Pollitzer, W.S., and J.J.B. Anderson. 1989. Ethnic and genetic differences
in bone mass: a review with a hereditary vs. environmental perspective. Am. J.
Clin. Nutr. 50:1244–1259.
7. Ott, S.M. 1990. Attainment of peak bone mass. J. Clin. Endocrinol.
8. Bonjour, J.P., G. Theintz, B. Buchs, D. Slosman, and R. Rizzoli. 1991.
Critical years and stages of puberty for spinal and femoral bone mass accumula-
tion during adolescence. J. Clin. Endocrinol. & Metab. 73:555–563.
9. Pocock, N.A., J.A. Eisman, J.L. Hopper, M.G. Yeates, P.N. Sambrook,
and S. Eberl. 1987. Genetic determinants of bone mass in adults - a twin study.
J. Clin. Invest. 80:706–710.
10. Seeman, E., J.L. Hopper, L.A. Bach, M.E. Cooper, E. Parkinson, J.
McKay, and G. Jerums. 1989. Reduced bone mass in daughters of women with
osteoporosis. N. Engl. J. Med. 320:554–558.
11. Tylavsky, F.A., A.D. Bortz, R.L. Hancock, and J.J.B. Anderson. 1989.
Familial resemblance of radial bone mass between premenopausal mothers and
their college-age daughters. Calcif. Tissue Int. 45:265–272.
12. Matkovic, V., D. Fontana, C. Tominac, P. Goel, and C.H. Chesnut III.
1990. Factors that influence peak bone mass formation: A study of calcium bal-
ance and the inheritance of bone mass in adolescent females. Am. J. Clin. Nutr.
13. Lutz, J., and R. Tesar. 1990. Mother-daughter pairs: spinal and femoral
bone densities and dietary intakes. Am. J. Clin. Nutr. 52:872–877.
14. Bachrach, L.K., P.A. Guido, D. Katzman, I.F. Litt, and R. Marcus. 1990.
Decreased bone density in adolescent girls with anorexia nervosa. Pediatrics.
15. Bell, N.H., J. Shary, J. Stevens, M. Garza, L. Gordon, and J. Edwards.
1991. Demonstration that bone mass is greater in black than in white children.
J. Bone Miner. Res. 6:719–723.
16. Sandler, R.B., C.W. Slemenda, R.E. LaPorte, J.A. Cauley, M.M.
Schramm, M.L. Barresi, and A.M. Kriska. 1985. Postmenopausal bone density
and milk consumption in childhood and adolescence. Am. J. Clin. Nutr. 42:270–
17. Chan, G. 1991. Dietary calcium and bone mineral status of children and
adolescents. Am. J. Dis. Child. 145:631–634.
18. Sentipal, J.M., G.M. Wardlaw, J. Mahan, and V. Matkovic. 1991. Influ-
ence of calcium intake and growth indexes on vertebral bone mineral density in
young females. Am. J. Clin. Nutr. 54:425–428.
19. Tylavsky, F.A., J.J.B. Anderson, R.V. Talmage, and T.N. Taft. 1992.
Are calcium intakes and physical activity patterns during adolescence related to
radial bone mass of white college-age females? Osteoporosis Int. 2:232–240.
20. Lloyd, T., N. Rollings, M.B. Andon, L.M. Demers, D.F. Eggli, K. Kisel-
horst, H. Kulin, J.R. Landis, J.K. Martel, G. Orr, and P. Smith. 1992. Determi-
nants of bone density in young women. I. Relationships among pubertal devel-
opment, total body bone mass, and total body bone density in premenarchal
females. J. Clin. Endocrinol. & Metab. 75:383–387.
21. Ruiz, J.C., C. Mandel, and M. Garabedian. 1995. Influence of spontane-
ous calcium intake and physical exercise on the vertebral and femoral bone
mineral density of children and adolescents. J. Bone Miner. Res. 10:675–682.
22. Katzman, D.K., L.K. Bachrach, D.R. Carter, and R. Marcus. 1991. Clin-
ical and anthropometric correlates of bone mineral acquisition in healthy ado-
lescent girls. J. Clin. Endocrinol. Metab. 73:1332–1339.
23. Kröger, H., A. Kotaniemi, P. Vainio, and E. Alhava. 1992. Bone densit-
ometry of the Spine and Femur in Children by Dual-Energy X-ray Absorptio-
metry. Bone Miner. 17:75–85.
24. Grimston, S.K., K. Morrison, J.A. Harder, and D.A. Hanley. 1992. Bone
mineral density during puberty in Western Canadian children. Bone Miner. 19:
25. Kröger, H., A. Kotaniemi, L. Kröger, and E. Alhava. 1993. Develop-
ment of bone mass and bone density of the spine and femoral neck - a prospec-
tive study of 65 children and adolescents. Bone Miner. 23:171–182.
26. Young, D., J.L. Hopper, C.A. Nowson, R.M. Green, A.J. Sherwin, B.
Kaymakci, M. Smid, C.S. Guest, R.G. Larkins, and J.D. Wark. 1995. Determi-
nants of bone mass in 10- to 26-year-old females: A twin study. J. Bone Miner.
27. Carrié, A. L., and J. P. Bonjour. 1995. Osteoporosis as a pediatric prob-
lem. Pediatr. Clin. North Am. 42:811–824.
28. Johnston, C.C., J.Z. Miller, C.W. Slemenda, T.K, Reister, S. Hui, J.C.
Christian, and M. Peacock. 1992. Calcium supplementation and increases in
bone mineral density in children. N. Engl. J. Med. 327:82–87.
29. Lloyd, T., M.B. Andon, N. Rollings, J.K. Martel, J.R. Landis, L.M. De-
mers, D.F. Eggli, K. Kieselhorst, and H.E. Kulin. 1993. Calcium supplementa-
tion and bone mineral density in adolescent girls. JAMA. 270:841–844.
Bonjour et al.
30. Lee, W.T.K., S.S.F. Leung, S.H. Wang, Y.C. Xu, W.P. Zeng, J. Lau, S.J.
Oppenheimer, and J.C.Y. Cheng. 1994. Double-blind, controlled calcium sup-
plementation and bone mineral accretion in children accustomed to a low-cal-
cium diet. Am. J. Clin. Nutr. 60:744–750.
31. Dawson-Hughes, B., G. E. Dallal, E. A. Krall, L. Sadowski, N. Sahyoun,
and S. Tannenbaum. 1990. A controlled trial of the effect of calcium supple-
mentation on bone density in postmenopausal women. N. Engl. J. Med. 323:
32. Theintz, G., B. Buchs, R. Rizzoli, D. Slosman, H. Clavien, P.C. Si-
zonenko, and J.P. Bonjour. 1992. Longitudinal monitoring of bone mass accu-
mulation in healthy adolescents: Evidence for a marked reduction after 16
years of age at the levels of lumbar spine and femoral neck in female subjects. J.
Clin. Endocrinol. Metab. 75:1060–1065.
33. Katzman D.K., L.K. Bachrach, D.R. Carter, and R. Marcus. 1991. Clini-
cal and anthropometric correlates of bone mineral acquisition in healthy ado-
lescent girls. J. Clin. Endocrinol. Metab. 73:1332–1339.
34. Carter D.R., M.L. Bouxsein, and R. Marcus. 1992. New approaches for
interpreting projected bone densitometric data. J. Bone Miner. Res. 7:137–145.
35. National Institutes of Health. 1994. Optimal calcium intake. NIH Con-
sensus statement 12:1–31.
36. Gilsanz, V., T.F. Roe, J. Antunez, M. Carlson, M.L. Duarte, and W.G.
Goodman. 1991. Effect of dietary calcium on bone density in growing rabbits.
Am. J. Physiol. 260:471–476.
37. Holbrook T.L., E. Barrett-Connor, D.L. Wingard. 1988. Dietary cal-
cium and risk of hip fracture: 14-year prospective population study. Lancet. ii:
38. Matkovic V., D. Fontana, C. Tominac, P. Goel, C.H. Chesnut. 1990.
Factors that influence peak bone mass: a study of calcium balance and the in-
heritance of bone mass in adolescent females. Am. J. Clin. Nutr. 52:878–888.
39. Matkovic, V., and R. P. Heaney. 1992. Calcium balance during human
growth: evidence for threshold behavior. Am. J. Clin. Nutr. 55:992–996.
40. Forbes, R.M., K.E. Weingartner, H.M. Parker, R.M. Bell, and J.W. Erd-
man, Jr. 1979. Bioavailability to rats of zinc, magnesium and calcium in casein-,
egg- and soy protein-containing diets. J. Nutr. 109:1652–1660.
41. Dawson-Hughes, B. 1991. Calcium supplementation and bone loss: a re-
view of controlled clinical trials. Am. J. Clin. Nutr. 54:(suppl.)274–280.
42. Parfitt, A.M. 1994. The two faces of growth: Benefits and risks to bone
integrity. Osteoporosis Int. 4:382–398.
43. Heaney, R.P. 1993. Nutritional factors in osteoporosis. Annu. Rev. Nutr.
44. Kanis, J.A. 1994. Calcium nutrition and its implications for osteoporosis.
Part I. Children and healthy adults. Eur. J. Clin. Nutr. 48:757–767.
45. Kanis, J.A. 1994. Calcium nutrition and its implications for osteoporosis.
Part II. After menopause. Eur. J. Clin. Nutr. 48:833–841.
46. Chevalley, T., R. Rizzoli, V. Nydegger, D. Slosman, C.H. Rapin, J.P.
Michel, H. Vasey, and J.P. Bonjour. 1994. Effects of calcium supplements on
femoral bone mineral density and vertebral fracture rate in vitamin-D-replete
elderly patients. Osteoporosis Int. 4:245–252.
47. Clavien H., G. Theintz, R. Rizzoli, and J.-P. Bonjour. 1996. Does pu-
berty alter dietary habits in adolescents living in a western society? J. Adolesc.