Girls' Calcium Intake Is Associated with Bone Mineral Content During Middle Childhood

Abstract

We examined longitudinally the association between calcium intake and total body bone mineral content (TBBMC) in 151 non-Hispanic white girls. Intakes of dairy, energy, and calcium were assessed using three 24-h dietary recalls in girls at ages 5, 7, 9, and 11 y. We assessed their total-body bone mineral content with dual-energy X-ray absorptiometry at ages 9 and11 y. Dairy foods comprised the major contributor (70%) to calcium intake over the 6-y period; 28% of calcium came from other foods, and 2% from supplements. By age 9 and 11 y, the majority of girls did not meet calcium recommendations. Higher calcium intake at ages 7 and 9 y was associated with higher TBBMC at age 11 y. Calcium intake at age 9 y was also positively associated with TBBMC gained from age 9 to 11 y. Calcium intake at age 11 y was not correlated with TBBMC at the same age. Relations between calcium intake and TBBMC did not differ for total calcium and for calcium from dairy sources, likely reflecting the fact that dairy products were the major source of calcium in this sample. Results from the present study provide new longitudinal evidence that calcium intake, especially calcium from dairy foods, can have a favorable effect on girls' TBBMC during middle childhood.

Adequate amounts of calcium are essential throughout the life cycle to promote bone and overall health, and to help reduce the risk of osteoporosis (1). Osteoporosis is considered a major public health threat for an estimated 44 million U.S. women and men ≥50 y old, causing an annual health care cost of $17 billion in 2001 (2). During childhood and adolescence, the increases in bone mass that occur with growth require adequate intakes of calcium and other nutrients provided by milk and milk products. A low dietary intake of calcium during childhood and adolescence may jeopardize achievement of genetically determined peak bone mass (3). Dairy foods are the most important source of dietary calcium, providing more that half of total intake (4). Unfortunately, children and adolescents, especially girls, do not consume the recommended 2−3 servings of milk and milk products each day (5). As a result of declining milk consumption in recent years, 70% of girls ages 6−11 y do not meet current calcium recommendations (6).

Numerous randomized controlled trials conducted in children and adolescents showed that higher calcium intakes, whether provided through supplements, fortified foods, or dairy products, increased bone mass during the intervention period compared with unsupplemented controls (7-22). Some follow-up studies showed that the positive effect of calcium supplementation on bone mineral status disappears after the supplementation period ceases (23,24). However, Bonjour et al. (8) demonstrated that the benefits of increased calcium intake persisted after cessation of the trial.

Conflicting findings were obtained from observational studies assessing relations between usual calcium intakes and bone health using self-reported measures of dietary intake (20, 25-38). Most of the observational studies were cross sectional, and longitudinal observational studies investigating the long-term relations between usual calcium intakes and bone health in children are limited (20,33,35,37,38). To our knowledge, no other observational studies have investigated the relation between calcium intake over 6 y and total-body bone mineral content (TBBMC)³ and change in TBBMC across middle childhood. In addition, there is no clear evidence concerning the differential influence of dairy and nondairy sources of calcium on bone health.

There is disagreement in the scientific community concerning the appropriate adjustment for bone mineral content (BMC), especially in studies focused on growing children. Researchers differ concerning whether BMC should be adjusted for growth and size, and which adjustments are most appropriate (39,40). Bone health is influenced by multiple factors (e.g., hormonal status), in addition to calcium intake, and the effects of calcium may be obscured during periods of rapid change by other factors affecting bone health, such as during puberty or menopause (41). In the present study, we focused on assessing the effects of calcium intake on bone accretion during a period of rapid growth; for this reason, we chose to adjust for differences in pubertal growth using height velocity. This is particularly problematic when studying a sample that is going through puberty and in which large differences in pubertal status exist (42).

The objective of this study was to examine longitudinally the association between girls' reported usual calcium intake and TBBMC. We hypothesized that at similar height velocities, girls' calcium intake from age 5 to 11 y would be associated with TBBMC at age 11 y. Given that calcium retention starts to increase at age 9 y, we also hypothesized that at similar height velocities, girls' calcium intake at age 9 y would be associated with a change in TBBMC from age 9 to 11 y. Dietary sources of girls' calcium intake were also evaluated.

SUBJECTS AND METHODS

Subjects

At entry into the study, participants included 197 5-y-old girls (mean age, 5.4 ± 0.4 y) and their parents; 192 of these families were reassessed 2 y later when the girls were 7 y old (mean age, 7.3 ± 0.3 y). A 3rd assessment with 183 families was conducted 2 y later when the girls were 9 y old (mean age, 9.34 ± 0.3 y), followed by a 4th assessment with 177 families they were age 11 y (mean age, 11.34 ± 0.3 y). This sample includes 151 families with complete data on all measures pertaining to the study. The families who were excluded due to missing data did not differ from those included in the analyses in family income, mothers' education level, fathers' education level, or parents' and daughters' mean BMI.

The eligibility criteria for participation at the time of recruitment included living with both biological parents, the absence of severe food allergies or chronic medical problems affecting food intake, and the absence of dietary restrictions involving animal products. Families were recruited for participation in the study using flyers and newspaper advertisements. In addition, families with age-eligible female children within a 5-county radius received mailings and follow-up phone calls (Metromail). Parents were generally in their mid 30s at the time of recruitment (mothers 35.4 ± 4.8 y; fathers 37.4 ± 5.4 y). Participating families were non-Hispanic white, predominately middle income, with a mean of 15 ± 2 y of education for fathers and mothers. Approximately equal numbers of families reported incomes in the following ranges $20,000−$35,000, $35,000−$50,000, and >$50,000 when the girls were 5 y old. Parents were slightly overweight at the time of the first measurement with a mean BMI of 26.4 ± 6.05 kg/m² for mothers, and 28.0 ± 4.35 kg/m² for fathers. The Pennsylvania State University Institutional Review Board approved all study procedures, and parents provided consent for their family's participation before the study began.

Measures

24-Hour dietary recall

Three recalls were obtained per respondent at each time of measurement; 2 weekdays and 1 weekend day during the summer and fall months were randomly selected over a 2-wk period. Mothers were the primary reporters of girls' intake at each age; girls were asked to be present during all interviews to facilitate the recall process. Interviews were conducted by trained staff at The Pennsylvania State University Diet Assessment Center using the computer-assisted Nutrition Data System for research (NDS-R, Nutrition Coordinating Center, University of Minnesota). At each wave, data were analyzed using the most current version of the NDS database. When the girls were 5 y old, NDS Version 2.91, Nutrient database version 26, food database 11a (1996) was used, whereas the Nutrient Data System for Research (NDS-R) was used when the girls were 7, 9, and 11 y old. Version 4.01_30 (2000), version 4.02_31 (2001), and version 4.06_34 (2003) were used at age 7, 9, and 11 y, respectively. Food portion posters (2D Food Portion Visual, Nutrition Consulting Enterprises) were used to assist in the estimation of food amounts.

Dietary supplement intake was assessed by additional questions during the 24-h recall. Nutrient data were averaged over 3 d to obtain an estimate of dairy, energy, and calcium intakes. Mean calcium intakes were compared with Adequate Intake recommendations (43). Girls' calcium intakes at ages 5 and 7 y were compared with the recommendations for 4- to 8-y-old children (800 mg/d); at ages 9 and 11 y, intakes were compared with the recommendations for 9- to 13-yold girls (1300 mg/d). Mean dairy intakes were compared with the dietary guidelines (5), which recommend 3 servings (3 cups) of dairy group foods daily for children who require 6694 kJ/d or more, and 2 servings (2 cups) daily for those with lower energy needs (1 cup equivalent is 1 cup low-fat or fat-free milk or yogurt; 1½ oz (42.5 g) of low-fat or fat-free natural cheese; 2 oz (57 g) of low-fat or fat-free processed cheese) (5). Dairy, energy, and calcium intakes estimates were based on foods consumed. Calcium intake at each age included intake from multivitamin-mineral supplements, and sources of calcium included dairy foods and other sources, including supplements.

We created an Energy Estimate Requirements (EER) range for girls' 4−8 and 9−18 y, in which we used their mean age, weight, height, and physical activity coefficients (sedentary, low active, active, and very active). The range of values in EER reflects possible differences in the physical activity coefficients of participants (44). For girls aged 4−8 and 9−18 y, the range of EER was 5159−8954 and 6276−11297 kJ/d, respectively.

Body mass index

Height and weight were measured by a trained staff member following procedures described by Lohman et al. (45). Children were dressed in light clothing and measured without shoes. Height was measured in triplicate to the nearest 0.1 cm using a Shorr Productions stadiometer (Irwin Shorr). Weight was measured in trip-licate to the nearest 0.1 kg using a Seca Electronic Scale. BMI-for-age percentiles were calculated using growth charts from the CDC (46). Overweight is defined as a BMI-for-age percentile ≥85%.

Height velocity

Girls' height velocity was calculated by subtracting height at age 11 y from height at age 9 y. Height velocity is an indicator of the pubertal growth spurt and reflects growth in stature; among girls in this age group, it can serve as an indicator of pubertal growth (47).

Total-body bone mineral content

Girls' TBBMC was assessed at age 9 and 11 y using dual energy X-ray absorptiometry. A trained technician obtained measurements with children in a supine position, in light clothing without shoes. Whole-body scans were obtained using a Hologic QDR 4500W (S/N 47261) instrument in the array scan mode. Scans were analyzed using the whole-body software, QDR4500 Whole Body Analysis. The TBBMC was expressed in grams.

In this study, BMC and not bone mineral density (BMD) was used as the outcome measure to assess the relation between calcium intake and bone mass. Because BMD is BMC divided by bone area, density is related to mass and is not a sensitive measure of bone accumulation associated with growth and increase in skeletal size (39,48).

Data analysis

Data were analyzed using SAS (version 8.02); a P-value < 0.05 was used to indicate significant effects. Repeated-measures ANOVA was conducted to evaluate the time effects of girls' height, weight, TBBMC, energy, calcium, and dairy intake. Pairwise comparisons of significant effects were computed using contrast statements. Because the contrast statements are equivalent to multiple t tests, a Bonferroni correction was used to control the overall error rate at P < 0.05 (i.e., individual contrasts were considered significant at P < 0.0127).

The relations between calcium intake across the ages from 5 to 11 y and TBBMC, and between calcium intake and change (Δ) in TBBMC were assessed using correlational analysis. Spearman correlations were used because variables were not normally distributed, and partial correlations were employed to adjust for height velocity. Partial correlations allowed us to determine the association between calcium intake and TBBMC by removing the effects of height velocity from both calcium intake and TBBMC. During the adolescence growth spurt, calcium retention is greater; therefore, dependence on calcium intake would be greater during this time (3). Additionally, height velocity is one of the determinants of BMC; gain in mass is very rapid during adolescence and up to 25% of peak bone mass is acquired during the 2-y period encompassing the peak growth in height (49).

RESULTS

Weight status, bone mineral content, and nutrient intake

BMI increased significantly at each time period; mean BMI-for-age percentiles were higher at ages 9 and 11 y than at ages 5 and 7 y (Table 1). The percentage of girls classified as being overweight increased significantly from age 5 to 11 y. At age 9 y, on average, girls' mean TBBMC (Table 1) was consistent with data presented by Faulkner et al. (50) (907 ± 236 g); however, at age 11 y, girls' mean TBBMC was higher than the normative BMC (1151 ± 296 g) (50). There was a significant increase in TBBMC from age 9 to 11 y (348 ± 105 g), reflecting growth during this period. Girls' mean height velocity from age 9 to 11 y was 13 ± 2.4 cm/2 y.

TABLE 1.

Girls' height, weight, BMI, and TBBMC from ages 5 to 11 y¹,²

Age, y	5	7	9	11
Height, cm	111.1 ± 4.7a	123.7 ± 5.4b	135.8 ± 6.3c	149.0 ± 7.3d
	(99.3 − 122.6)	(109.0 − 136.5)	(117.7 − 152.8)	(126.5 − 169.6)
Weight, kg	19.6 ± 3.1a	25.4 ± 5.2b	34.3 ± 8.5c	44.7 ± 11.3d
	(14.2 − 35.2)	(17.7 − 45.3)	(22.1 − 58.7)	(26.1 − 74.7)
BMI-for-age percentile	59.1 ± 26.4a	58.3 ± 27.0a	64.0 ± 26.7b	64.1 ± 27.2b
	(1.7 − 99.4)	(0.3 − 98.8)	(3.2 − 99.2)	(2.0 − 99.1)
Overweight, %	19 ± 40.0a	20 ± 40.4a	31 ± 46.3b	28 ± 45.1b
TBBMC, g	NA2	NA	946 ± 148a	1293 ± 218b
			(642 − 1417)	(809 − 1967)

¹Values are means ± SD (range), n = 151. Means in a row without a common letter differ, P < 0.0001.

²NA, not applicable.

Girls' mean daily energy intake fell within the range of estimated energy requirements for girls' 4−8 and 9−18 y old (Table 2) (44). At age 5 y, on average, girls met the recommended 2 daily servings of dairy; however, by age 7, 9, and 11 y, girls consumed fewer than the 3 daily servings of dairy recommended by the USDA dietary guidelines (51). Dairy foods were the major contributor (70%) to calcium intake over the 6-y period, whereas 28% of calcium came from other sources of food, and 2% from supplements. That girls' dairy intake remained stable and did not increase to 3 servings/d by age 7, 9, and 11 y may explain in part why 67 and 73% of the girls failed to meet the recommended 1300 mg/d for calcium by 9 and 11 y, respectively. The percentage of girls using calcium from supplements did not vary from ages 5 to 11 y (P = 0.07), with values of 24, 17, 20, and 15% at ages 5, 7, 9, and 11 y, respectively.

TABLE 2.

Girls' energy, and dairy and calcium intakes from ages 5 to 11 y, based on 3-d recalls at each time point¹

Age, y	5	7	9	11
Energy intake, kJ/d	6335 ± 1356a	7054 ± 1360b	7632 ± 1469c	7795 ± 1904c
	(3582 − 9912)	(3502 − 11548)	(4280 − 12452)	(4385 − 1463)
Dairy, servings/d	2.7 ± 1.3a	2.8 ± 1.3b	2.8 ± 1.3a	2.8 ± 1.4a
	(0.4 − 8.5)	(0.4 − 7.0)	(7.0 − 7.5)	(0.1 − 8.3)
Total calcium,2 mg	849 ± 340a	866 ± 297a	925 ± 317b	946 ± 415b
	(277 − 2032)	(268 − 1905)	(347 − 2001)	(286 − 2572)
Calcium from dairy, mg	609 ± 326a	620 ± 291a	632 ± 300a	629 ± 342a
	(58 − 1783)	(121 − 1701)	(102 − 1811)	(48 − 2164)
Calcium from nondairy foods, mg	219 ± 87a	235 ± 73a	282 ± 121b	289 ± 162b
	(75 − 543)	(81 − 457)	(87 − 804)	(105 − 1493)
Calcium from supplements, mg	23 ± 54a	15 ± 40a	17 ± 38a	15 ± 41a
	(0 − 450)	(0 − 2000)	(0 − 155)	(0 − 200)

¹Values are means ± SD (range), n = 151. Means in a row without a common letter differ, P < 0.0001, except for total calcium and energy (P < 0.01).

²Includes calcium from supplements.

TBBMC at age 11 y and change in TBBMC from age 9 to 11 y

The relation between girls' calcium intake with TBBMC at age 11 y and change in TBBMC from ages 9 to 11 y was assessed after adjusting height velocity to correct for differences in pubertal growth among individuals (Table 3). Data are presented showing these relations for total calcium and for calcium from dairy foods, nondairy sources, and from supplements. It is unlikely that total calcium intake at age 5 y was associated with TBBMC at age 11. Higher total calcium intake and calcium intake from dairy foods at 7 and 9 y were associated with higher TBBMC at age 11. In contrast, there was no association between calcium intake from nondairy sources at 7 and 9 y and TBBMC at age 11 y. There was no association between total calcium intake at age 11 y and TBBMC at age 11 y. Finally, total calcium intake at 9 y and calcium intake from dairy foods at 7 and 9 y were associated with increases in TBBMC from age 9 to 11 y. There was no evidence of an association between calcium intake from nondairy sources and change in TBBMC from age 9 to 11 y. Analyses were also conducted without adjusting for height velocity. Total calcium intake at age 9 y was associated with TBBMC at age 11 y. There was no evidence of other significant associations.

TABLE 3.

Adjusted and unadjusted Spearman correlation coefficients between calcium intake and TBBMC in 151 girls at age 11 y, and Δ TBBMC from age 9 to 11 y¹

	TBBMC at age 11 y		Δ TBBMC from age 9−11 y
Age, y	Adjusted2r	Unadjusted r	Adjusted r	Unadjusted r
Total Ca intake,3 mg/d
5	0.13	0.08	0.12	0.00
7	0.23**	0.13	0.15	0.02
9	0.19*	0.17*	0.19*	0.12
11	0.01	0.00	NA4
Ca intake from dairy, mg/d
5	0.10	0.04	0.11	0.02
7	0.24**	0.15	0.18*	0.00
9	0.16*	0.12	0.16*	0.06
11	0.01	0.05	NA
Nondairy calcium intake,2 mg/d
5	0.13	0.15	0.10	0.05
7	0.01	0.00	0.13	0.08
9	0.13	0.15	0.13	0.15
11	0.11	0.10	NA

¹Asterisks indicate significant correlations

^*P < 0.05

^**P < 0.01.

²Adjusted for height velocity.

³Includes calcium from supplements.

⁴NA, not applicable.

DISCUSSION

This prospective study supports the growing body of evidence that calcium intake, especially calcium from dairy foods, can favorably affect TBBMC during childhood. Findings from this research are consistent with other observational studies in children (20,33,37). Lee et al. (33) studied the relation between long-term calcium intakes during the first 5 y of life and TBBMC. In agreement with our study, Lee et al. (33) found that previous calcium intake was significantly associated with TBBMC, whereas the cross-sectional association between current calcium intake and TBBMC was not established. This observation confirms that given the nature of bone remodeling, the effect of calcium intake on TBBMC takes several months or even a year to become manifest (52).

Few observational studies examined the effects of calcium intake on TBBMC accretion, and the results are controversial (20,35,37,38,53). In agreement with our data, Barr et al. (37) found that calcium intake was associated with a change in TBBMC over 2 y in pubertal girls (ages 9−12 y). Another 2-y prospective investigation demonstrated that when beverages low in nutrient density replaced milk, adolescent girls had reduced TBBMC accretion (53). Conversely, data reported by Lloyd et al. (35) showed no association between mean calcium intake over 8 y and bone gain during adolescence. Another 1-y prospective study among school-aged children (aged 5−19 y) in Copenhagen found that size-adjusted BMC accretion was not related to calcium intake (38).

It is possible that our ability to detect an association between calcium intake and TBBMC and change in TBBMC, and the absence of such an association in other studies, resulted from differences in the adjustments applied. To date, researchers differ concerning whether BMC should be adjusted for growth and size (39,40). Unfortunately, results among studies may vary depending on the adjustments applied, thus making it difficult to compare results. According to Heaney (48), using a BMC unadjusted for growth-related variables (i.e., bone area, height, and weight) is the correct approach. If the hypothesis is that calcium influences the amassing of tissue and the increase in its size, adjusting increases in mass for increases in bone area or height would factor out the effects of growth and thus create a variable similar to BMD (48). Because height velocity is an indicator of pubertal status, the most powerful factor influencing growth during this period, and is also one of the determinants of BMC in this study, we adjusted for height velocity.

The low correlations between calcium intake and TBBMC in this study underscore the multifactorial nature of bone health, with calcium intake as only one of those factors. For example, puberty has a powerful effect on bone mass; therefore, it can hide the effects of calcium on bone mass. In addition, calcium is a threshold nutrient (41); thus, the importance of calcium intake is especially evident at intakes below the threshold. Girls who have intakes above the threshold level contribute nothing to the correlation except statistical noise (41). It is possible that the lack of association observed between nondairy calcium and TBBMC was due to the finding that the range in intakes from nondairy calcium was much narrower than the range in dairy calcium.

Consistent with national data (Continuing Survey of Food Intakes by Individuals and NHANES III) (54,55), in this study, dairy foods were the major food source of calcium (70%), indicating that consumption of dairy foods is critical to meet recommendations for this nutrient. In this study, calcium intake from supplements was included, and the data show that the mean contribution of calcium supplements was very small because a relatively small percentage of the girls used supplements. Our data show that the percentage of girls using calcium supplements was consistent with the percentage of young adults in the United States ages 18−24 y using calcium supplements (56).

Although the present study provides the longitudinal data required to address several important questions, the study also has several limitations. The sample is a homogeneous one; girls were prepubertal, non-lactose intolerant non-Hispanic white, and our findings cannot be generalized to other racial or ethnic populations or boys.

To our knowledge, no other observational studies reported the relation between calcium intake over 6 y and TBBMC and change in TBBMC across middle childhood. Most of the data on TBBMC accretion comes from clinical trials assessing the effect of calcium supplementation and fortification, and few long-term observational studies exist that assessed the relation between calcium and TBBMC accretion within the range of usual intakes. In addition, there is no clear evidence on dairy and nondairy sources of calcium and their differential influence on bone health. Furthermore, the longitudinal design of this investigation offers advantages over previous cross-sectional studies by assessing calcium intake by 24-h dietary recall repeatedly over 6 y. The results of cross-sectional studies using 1 or 2 assessments may be inaccurate, given that calcium intake at one point in time may not reflect lifetime calcium intake, and bone mass at one point in time is the result of long-term confounding variables that are not assessed (43).

In conclusion, this research provides confirmatory evidence of the importance of habitual calcium intake, especially calcium from dairy sources, in contributing to bone mass and bone mineral accretion, and attainment of peak bone mass during growth, which is assumed to be a critical factor osteoporosis in later life. The results of the current study suggest that increasing calcium intake among children should continue to be a major focus of interventions. These findings, which showed that ∼70% of total calcium is obtained from dairy sources, suggest that increasing dairy intake among girls could be an especially effective strategy to increase total calcium intake. The development of dietary habits that include frequent milk intake is likely to lead to higher calcium intake in later years. For example, Fisher and colleagues (34), in a mother-daughter study, concluded that milk availability to the daughters at meals and snacks was associated with meeting calcium recommendations.