Predictors of Bone Mass by Peripheral Quantitative Computed Tomography in Early Adolescent Girls

Journal of Clinical Densitometry, vol. 4, no. 4, 313–323, Winter 2001 © Copyright 2001 by Humana Press Inc. All rights of any nature whatsoever reserved. 1094-6950/01/4:313–323/$12.75

Original Article

Predictors of Bone Mass by Peripheral Quantitative Computed Tomography in Early Adolescent Girls Laurie Moyer-Mileur, PHD, RD,1 Bin Xie, MS,2 Shauna Ball, MS, RD,1 Cynthia Bainbridge, PHD,2 Diane Stadler, PHD, RD,2 and Webster S. S. Jee, PHD3 1Center

for Pediatric Nutrition Research, Department of Pediatrics, School of Medicine, 2Division of Food and Nutrition, College of Health; and 3Division of Radiobiology, School of Medicine, University of Utah, Salt Lake City, UT

Abstract This cross-sectional study used peripheral quantitative computed tomography (pQCT) to evaluate the influences of age, body size, puberty, calcium intake, and physical activity on bone acquisition in healthy early adolescent girls. The pQCT technique provides analyses of volumetric bone mineral density (vBMD) (mg/cm3) for total as well as cortical and trabecular bone compartments and bone strength expressed as polar strength strain index (mm2). Bone mass of the nondominant distal and midshaft tibia by pQCT and lumbar spine and hip by dual X-ray absorptiometry (DXA) were measured in 84 girls ages 11–14 yr. Pubertal stage, menarche status, anthropometrics, and 3-d food intake and physical activity records were collected. Total and cortical bone mineral content and vBMD measurements by pQCT were significantly related to lumbar spine and femoral neck BMD measurements by DXA. We did not note any significant determinants or predictors for trabecular bone mass. Body weight was the most important predictor and determinant of total and cortical bone density and strength in healthy adolescent girls. Menarche, calcium intake, height, body mass index, and weight-bearing physical activity and age were also identified as minor but significant predictors and determinants of bone density and strength. Bone measurements by the pQCT technique provide information on bone acquisition, architecture, and strength during rapid periods of growth and development. Broader cross-sectional studies using the pQCT technique to evaluate the influence of age, gender, ethnicity, puberty, body size, and lifestyle factors on bone acquisition and strength are needed. Key Words: Adolescent; bone mineral density; puberty; peripheral quantitative computed tomography.

increased risk of fracture (1). Primary prevention of osteoporosis consists of maximizing peak bone mass during childhood, adolescence, and young adulthood, and minimizing bone loss later in life, especially after menopause for women (2). The majority of bone mass is accumulated by age 18, with about 45% deposited during the adolescent growth spurt (3). There are a number of recent investigations of determinants and predictors of bone mass in children and adolescents (4–11). Numerous factors, including age, body size, puberty, calcium

Introduction Osteoporosis is a disease characterized by low bone mass and microarchitectural deterioration of bone tissue leading to enhanced bone fragility and

Received 02/02/01; Revised 03/13/01; Accepted 04/10/01. Address correspondence to Laurie Moyer-Mileur, Center for Pediatric Nutrition Research, Room 2A100, Department of Pediatrics, School of Medicine, University of Utah, Salt Lake City, UT 84132. E-mail: [email protected]

313

314 intake, and physical activity, have been identified as important determinants and predictors of bone mass. However, in all these studies, bone mass in terms of bone mineral density (BMD) was assessed with single photon absorptiometry, dual photon absorptiometry, or dual X-ray absorptiometry (DXA), which provide “areal density” expressed as grams/square centimeter. This restriction of the inherent planar nature of the measurement makes a true geometric assessment of a bone and estimation of bone strength impossible (12,13), and estimation of bone strength grossly approximate. As a result, it is difficult to predict or evaluate with certainty the importance of these determinants to the development of volumetric BMD (vBMD) or bone strength during adolescence. Peripheral quantitative computed tomography (pQCT) has been recently introduced to the clinical users and investigators with the purpose of solving this problem. pQCT provides a three-dimensional display of the data, an integrated measurement of combined cortical and trabecular bone and a separate measurement for cortical and trabecular bone (12). In addition, based on the architectural and BMD information provided by pQCT, the strength strain index (SSI) can be calculated to reflect bone strength. No previous study using pQCT to evaluate the determinants and predictors of bone development in preadolescent and adolescent girls has been reported. Limited data are available on the normal variation in bone mass of the peripheral skeleton measured by pQCT in adolescents (4,6,7). The purpose for the present cross-sectional study was to evaluate the influences of age, body size, puberty, calcium intake, and physical activity on bone mass measured by pQCT in healthy preadolescent and adolescent females.

Materials and Methods Subjects and Protocol The protocol was approved by the Review Committee for Research with Human Subjects Institutional Review Board, University of Utah, and informed consent was obtained from each subject and her parent. Eighty-four healthy girls ages 11–14 from Salt Lake City, UT, were recruited. Participants were selected based on the evaluation of questionnairereported medical and menstrual histories and med-

Journal of Clinical Densitometry

Moyer-Mileur et al. ication use. The inclusion criteria included the following: generally healthy as determined by a medical history, no use of medications that might affect calcium metabolism or bone growth, no tobacco or alcohol use, absence of diseases known to affect bone development, eating disorders, or pregnancy.

Questionnaires Subjects completed a modified past-year, leisuretime activity questionnaire (14,15), which was designed to determine energy expenditure in leisure activities in the last year. From a list of physical activities, subjects were required to chose those in which they participated during the last year and to indicate the frequency and duration of participation in each activity. Subjects were also encouraged to complete an open-ended question about participation in activities not included in the list. Some types of activities such as cycling and swimming were considered as non-weight-bearing physical activities that were defined as not being supported by the body mass. Only weight-bearing physical activities was used in the analysis. Total year leisure weight-bearing activity level in total past-year hours per week was determined as the summed product of the frequency and duration of weight-bearing activity. The hours per week of all weight-bearing activity were also converted to the estimated metabolic cost and expressed as MET-hours per week by multiplying hours per week for each specific activity by the estimated MET value. An MET is a multiple of the resting metabolic rate and represents the millimeters of O2 consumed (kilogram of body mass · minute) (15). A 3-d food record of all foods and beverages consumed over two weekdays and one weekend day was obtained. All subjects were given detailed instructions on how to complete the record and were given specially prepared sheets on portion size and how to record food intake. The computer program Food Processor I (ESHA Research, Salem, OR) was used to analyze the food records and calculate the average daily calcium intake. Percentage of recommend dietary allowance (RDA) was also calculated according to the RDA for calcium revised in 1989 (16).

Anthropometric Measurements All subjects’ body height, weight, and pubertal status were determined by a trained research nurse.

Volume 4, 2001

Bone Mass by pQCT in Early Adolescent Girls

315

Fig. 1. Examples of (A) scout scan for identification of distal tibia reference line and cross-sectional measurements of (B) distal and (C) midshaft tibia by pQCT.

Height in centimeters was measured with the subject standing without shoes. Body weight in kilograms was determined on a balance scale with the subject in light clothing and without shoes. Body mass index (BMI) (weight/height2) was calculated. Pubertal status was evaluated by assessing breast and pubic hair and was expressed in Tanner stage (17).

Bone Mass Assessment All subjects’ bone mass was measured on the nondominant leg at the distal and midshaft tibia from the 10 and 66% length of the tibia from the distal end by pQCT (XCT 2000; Norland, Fort Atkinson, WI). These sites were selected because bone composition at the distal site is predominantly trabecular bone while the midshaft site is composed primarily of cortical bone. The distal site scan is accomplished by performing a planar scout view over the joint line of interest formed to determine the anatomic reference line at the distal end of the tibia (Fig. 1A). A tomographic scan is then performed at the 10% distal site from the distal end of the tibia (Fig. 1B). The midshaft tibia site is determined by external measurement of the extreme ends of the tibia using the distal end of the medial mealeolus and the internal point of articulation of the knee as landmarks. The 66% site of the total length is determined and marked. The sub-

Journal of Clinical Densitometry

ject is seated on the scanner chair with the extended nondominant leg resting inside the concentric acrylic cylinder (14 cm diameter) at the central gantry. The gantry and computed tomography positioning laser are manually placed over the 66% site mark and the scan is performed (Fig. 1C). A constant voxel size of 0.6 mm and translational speed of 30 mm/s was used for both sites in all subjects. To analyze trabecular bone, a contour mode with a threshold of 269 mg/cm3 is used to separate soft tissue and bone. By default setting of 45% inner core of bone area as trabecular bone, 55% of the outer bone area is considered as cortical and subcortical bone and are peeled off. Based on that data, the trabecular vBMD (mg/cm3) is calculated. A constant default threshold of 710 mg/cm3 is used to analyze the cortical bone. In this way, trabecular pixels with the attenuation coefficient lower than the defined threshold are removed from the scan, leaving a cortical shell, from which cortical vBMD and cortical thickness are calculated. In addition, SSI was determined at both the distal and midshaft sites based on the calculation of the cross-sectional moment of inertia divided by the maximum distance of any voxel from the center of gravity. This index represents the geometric and material properties of bone and is supposed to reflect the bone strength with respect to bending or torsion.

Volume 4, 2001

316 The polar SSI predicts the torsion bone strength, and the axial SSI predicts the bending bone strength with respect to the x or y-axis. Since the axial values are affected by the different rotational positions of the leg during the measurement and the polar SSI is independent of rotation, polar SSI is used for bone strength analysis. A threshold of 740 mg/cm3 is used to define the bone tissue assumed to contribute mainly to bone strength. In our laboratory, the shortterm coefficients of variation (CVs) in 15 subjects including repositioning are 1.0 and 1.9% for cortical and trabecular vBMD, respectively. Measurements of the lumbar spine (L2–L4 vertebrae) and hip (femoral neck) by DXA (XR-26; Norland) were obtained for comparison of predictors of areal BMD (aBMD) vs vBMD. Precision studies with 15 repositioned subjects showed a CV of < 1.0% for lumbar spine and femoral neck aBMD. The subjects’ radiation exposure was 23 mrem for pQCT measurements and 10 mrem for DXA measurements, for a total of 33 mrem.

Statistical Analyses Data were analyzed using the SPSS program version 10.0 (SPSS, Chicago, IL). The mean, standard deviation (SD), and range were given as descriptive statistics. Pearson’s r correlation was used to determine the associations between each related variable. Partial correlation was used to determine the relationship between bone variables and pubertal stage after control of age and body size as covariates. Scatter plot and univariate regression were used to define the independent relationship between bone mass and the presence of menarche. Stepwise multiple linear regression models were fit to estimate the effect of predictors (i.e., age, pubertal stage, body size, weight-bearing physical activity, calcium intake) on aBMD and vBMD, cortical thickness, and polar SSI. Bone mass between subjects in the absence or presence of menarche were compared using two-sided student’s t-test. A p value ≤ 0.05 was considered significant for all statistical analyses.

Results The mean and SDs for age, anthropometry, weight-bearing physical activity, and calcium intake in terms of average daily intake and percentage of

Journal of Clinical Densitometry

Moyer-Mileur et al. Table 1 Results of Descriptive Analysis of Age, Anthropometry, Weight-Bearing Physical Activity, and Calcium Intake (n = 84)

Age (yr) Pubertal stage Weight (kg) Height (cm) BMI (kg/m2) Weight-bearing activity (MET-h/wk) Calcium intake (mg/d) RDA of calcium (%)

Mean

SD

Range

12.8 2.5 50.1 158.5 19.8 36.8

0.8 0.5 12.2 8.1 3.9 25.7

11–14 1–4 30.0–102.7 140–180 14.7–35.3 5–107

935 77.9

364 34.0

261–2219 22–185

RDA are presented in Table 1. Pubertal stage distribution by Tanner stage was as follows: stage 1, 1.2%; stage 2, 41.6%; stage 3, 56%; stage 4, 1.2%. Forty-one girls (49%) reported the onset of menarche. Total and cortical and trabecular bone bone mineral content ([BMC], mg), bone area (cm2), vBMD, cortical thickness, and polar SSI by pQCT and aBMD for the lumbar spine and femoral neck are summarized in Table 2. As expected, values for the distal tibia site reflect the predominance of trabecular bone while the midshaft tibia values reflect the predominance of cortical bone. Values for lumbar spine and femoral neck BMD were similar. Summaries of the correlational (bivariate) relationship among (1) vBMD aBMD, cortical thickness, and polar SSI and the most theoretically important independent variables and (2) bone variables measured by pQCT vs DXA are reported in Tables 3 and 4. Pubertal stage, menarche, weight, and height were significantly correlated with distal tibia total cortical thickness (p < 0.05) while menarche, weight, height, and BMI were correlated with polar SSI (p < 0.05). Distal tibia total vBMD was associated with pubertal stage (p < 0.05) and menarche (p < 0.01). No correlation was found between distal tibia trabecular vBMD and related factors. Weak associations were found at the midshaft tibia for menarche, BMI and weightbearing activity and total vBMD, age and cortical vBMD, and weight-bearing activity and polar SSI (p < 0.05). Midshaft cortical thickness correlated with

Volume 4, 2001

Bone Mass by pQCT in Early Adolescent Girls Table 2 Results of Descriptive Analysis of Bone Variables (n = 84) Mean Distal Tibia (10%) Total BMC (mg) 212.3 Bone area (cm2) 578.3 vBMD (mg/cm3) 370.4 Cortical BMC (mg) 89.0 Bone area (cm2) 94.2 vBMD (mg/cm3) 906.8 Trabecular BMC (mg) 102.9 Bone area (cm2) 427.7 vBMD (mg/cm3) 240.4 Cortical thickness (mm) 1.2 Polar SSI (mm2) 750.3 MidShaft Tibia (66%) Total BMC (mg) 322.0 Bone area (cm2) 553.2 vBMD (mg/cm3) 589.8 Cortical BMC (mg) 248.2 Bone area (cm2) 238.1 vBMD (mg/cm3) 1042.9 Trabecular BMC (mg) 40.7 Bone area (cm2) 235.9 vBMD (mg/cm3) 171.0 Cortical thickness (mm) 3.3 Polar SSI (mm2) 1640.0 Lumbar spine aBMC (mg) 3312.4 aBMD (mg/cm2) 862.4 Femoral neck BMC (mg) 389.1 aBMD (mg/cm2) 860.5

SD

Range

42.0 110.5 54.9

128.1–341.4 396.4–932.4 255.5–511.2

31.8 24.9 65.0

14.3–165.0 28.7–189.4 759.5–1030.5

27.0 88.4 28.9 0.4 293.5

53.3–187.5 246.9–751.3 161.2–319.8 0.2–2.0 89.2–1550.1

48.5 108.2 671

237.1–452.7 373.1–824.9 439.4–713.2

33.4 165.0–316.9 30.3 172.8–304.0 39..8 969.3–1138.7 17.0 76.6 32.7 0.4 358.2

18.0–93.1 120.2–464.4 107.0–269.6 2.4–4.2 972.2–2492.4

887.5 151.8

1752–6050 520.0–1181.2

78.5 142.8

250.0–602.0 590.1–1191.0

weight-bearing activity and daily calcium intake (p < 0.05). Age, pubertal stage, menarche, weight, height, and BMI were strongly associated with lumbar spine aBMD, while femoral neck aBMD correlated with pubertal stage, menarche, weight, height, BMI, and weight-bearing activity (p < 0.05) (Table 3). The distal and midshaft cortical BMC, BA, vBMD, cortical thickness, and polar SSI were signif-

Journal of Clinical Densitometry

317 icantly associated with lumbar spine BMC and aBMD. Tibia total BMC and vBMD and trabecular vBMD at both sites were strongly associated with lumbar spine BMC and aBMD. A weaker negative correlation was found between midshaft tibia BA and lumbar spine BMC and aBMD. Midshaft tibia total BMC and vBMD cortical BMC, BA, and vBMD, cortical thickness, and polar SSI were significantly associated with femoral neck BMC and vBMD. Correlations between the distal tibia and femoral neck were significant for total BMC and vBMD, trabecular vBMD, all cortical bone variables, and polar SSI and femoral neck aBMD. Only trabecular vBMD at the distal tibia site was found to correlate significantly with femoral neck BMC. In stepwise multiple linear regression, midshaft tibia total BMC and cortical vBMD were the most significant predictors of lumbar spine aBMD (R2 = 0.82), and midshaft cortical BMC and total vBMD were the most significant predictors of femoral neck aBMD (R2 = 0.68). Partial correlation, after control for age, body size, and menarche was found to be strongly associated with distal and midshaft tibia total vBMD (r = 0.28 and r = 0.25, respectively; p < 0.05). No relationship was found between distal tibia trabecular vBMD and pubertal stage (r = –0.08; NS). Partial correlation of menarche to lumbar spine and femoral neck aBMD found stronger associations (r = 0.48 and r = 0.50, respectively; p < 0.01). Scatter plots in Fig. 2A–D illustrate the positive association between menarche and vBMD and aBMD. Comparisons of bone variable by menarche status (pre- vs postmenarche) are presented in Table 5. Higher values for distal tibia total vBMD and cortical thickness (p < 0.01), midshaft tibia total vBMD (p = 0.03), and lumbar spine and femoral neck aBMD (p < 0.01) were observed in girls who had begun menstruation. In stepwise multiple linear regression (Table 6), using vBMD and aBMD, cortical thickness, and polar SSI as dependent variables, body weight, menarche, calcium intake, BMI, height, and weightbearing activity were significant predictors. It is estimated that 6.8 to 55.7% of the variance in bone mass and strength could be explained by these factors. Body weight accounted for 19.4–42.8% of the variance in distal and midshaft tibia polar SSI as well as lumbar spine and femoral neck aBMD. The presence of menarche accounted for 10.4–12.0% of the vari-

Volume 4, 2001

318

Moyer-Mileur et al. Table 3 Summary of Correlation Between Bone Variables and Related Factors a

Distal tibia Total vBMD (mg/cm3) Trabecular vBMD (mg/cm3) Cortical thickness (mm) Polar SSI (mm2) MidShaft tibia Total vBMD (mg/cm3) Cortical vBMD (mg/cm3) Cortical thickness (mm) Polar SSI (mm2) Lumbar spine aBMD (mg/cm2) Femoral neck aBMD (mg/cm2)

Weightbearing Calcium activity intake

Age

Pubertal stage

Menarche

Weight

Height

0.23b –0.05 0.34b 0.33b

0.22b –0.04 0.23b 0.63c

0.33b 0.11 0.31b 0.21b

0.37b –0.07 0.26b 0.37c

0.15 0.06 0.23b 0.28b

0.39b 0.01 0.56c 0.30b

0.05 0.07 0.03 0.06

–0.15 0.03 –0.15 –0.06

0.06 0.24b 0.04 0.18

0.10 0.39b 0.08 0.30b

0.24b 0.41c 0.16 0.10

–0.35b –0.07 –0.49c 0.79c

–0.17 0.15 0.19 0.77c

–0.21b 0.11 0.15 0.64c

0.25b 0.10 0.24b 0.33b

0.27 0.13 0.27b –0.01

0.21b

0.47c

0.60c

0.63c

0.51b

0.56b

0.17

0.01

0.51c

0.52c

0.59d

0.35b

0.59b

0.21b

0.03

–0.11

BMI

a

Correlation coefficient; n = 84, two-tailed test. Statistical significance ≤ 0.05 level. c Statistical significance ≤ 0.01 level. b

ance observed in distal and midshaft tibia total vBMD, distal tibia cortical thickness, and lumbar spine and femoral neck aBMD. Height accounted for an additional 4.2% of the variance in distal tibia cortical thickness while calcium intake contributed to 6.8% of the variance in midshaft tibia cortical thickness. BMI contributed to 5.8% of the variance in midshaft tibia total vBMD. In addition, results of correlation among bone variables showed that vBMD and aBMD, cortical thickness, and polar SSI had consistently, highly significant relationships with cortical, trabecular, and total bone area and BMC.

Discussion Our cross-sectional study of 84 healthy early adolescent girls provides new information on determinants of distal tibia bone development measured by pQCT. Both genetic and environmental factors influence bone development and acquisition. It is estimated that 60–80% of the variance in bone mass is explained by genetic factors (18). Pubertal stage and body size are variables that are highly related to genetic factors, while dietary habits such as calcium

Journal of Clinical Densitometry

intake and physical activity level are more likely associated with environment. In our study, subjects’ body weight and onset of menarche accounted for the majority of the variance in total and cortical vBMD and and polar SSI of the tibia and aBMD of the lumbar spine and hip. Among these variables, body weight accounted for 19.4–42.8%, which was the most import determinant of bone density and strength. Trabecular vBMD was not influenced by age, body size, pubertal stage, or lifestyle factors. This finding is supported by Neu et al. (19), who recently studied the distal radius by pQCT in 371 children and adults and found no change in vBMD during adolescence or pubertal maturation. Midshaft total and cortical BMC and vBMD measured by pQCT were significantly related to aBMD of the lumbar spine and femoral neck measured by DXA in our study population. Trabecular BMC or vBMD measurement by pQCT did not correlate to DXA lumbar spine or femoral neck measurements. Both aBMD by DXA and total vBMD by pQCT are a sum of cortical and trabecular compartments. This explains the significant relationship between pQCT measurements of total and cortical BMC and vBMD

Volume 4, 2001

Bone Mass by pQCT in Early Adolescent Girls

319

Table 4 Summary of Correlation Between Bone Variables Measured by pQCT vs DXAa Lumbar spine

Femoral neck

BMC (mg)

BMD (mg/cm2)

BMC (mg)

BMD (mg/cm2)

0.56b 0.01 0.76b

0.57b –0.03 0.78b

0.21 0.04 0.28

0.50b 0.08 0.56b

0.12 0.17 0.50b

0.11 –0.22 0.50b

0.25 0.08 0.37c

0.21 –0.05 0.47b

Cortical 0.66b BMC (mg) 0.59b Bone area (cm) 0.56b vBMD (mg/cm3) Cortical thickness (mm) 0.65b Polar SSI (mm2) 0.40c Midshaft tibia Total BMC (mg) 0.59b Bone area (cm) –0.04 vBMD (mg/cm3) 0.55b Trabecular BMC (mg) –0.04 Bone area (cm) –0.25c vBMD (mg/cm3) 0.38b Cortical BMC (mg) 0.62b Bone area (cm) 0.50b vBMD (mg/cm3) 0.66b Cortical thickness (mm) 0.49b Polar SSI (mm2) 0.42b

0.68b 0.59b 0.58b

0.12 0.10 0.14

0.49b 0.47b 0.36c

0.66b 0.35c

0.13 0.28

0.48b 0.28

0.56b –0.12 0.59b

0.52b 0.08 0.33c

0.60b –0.02 0.49b

Distal tibia Total BMC (mg) Bone area (cm) vBMD (mg/cm3 ) Trabecular BMC (mg) Bone area (cm) vBMD (mg/cm3)

0.01 –0.30c 0.44b 0.60b 0.47b 0.68b 0.50b 0.35c

0.07 –0.20 0.21 0.55b 0.53b 0.31c 0.42b 0.46b

–0.04 –0.06 0.20 0.68b 0.60b 0.55b 0.54b 0.49b

a

Correlation coefficient; n = 84, two-tailed test. Statistical significance ≤ 0.01 level. c Statistical significance ≤ 0.05 level. b

and DXA measurements of lumbar spine and femoral neck aBMD and the absence of association for trabecular BMC or vBMD. The findings of pubertal stage and menarche as important determinants confirm and extend previous observations (6,11,20,21). The positive relationships, as shown in the scatter plots, suggest that at the onset of puberty and menstruation, driven by

Journal of Clinical Densitometry

growth hormone (GH) and sex hormones, bone mass accumulation increases. This finding is also supported by Gilsanz et al. (22), who observed increases in vBMD during puberty in both Caucasian and African American adolescents. Together with GHs, estrogen is assumed to be an important factor that could account for this observation. In our study, subjects who had begun menarche were found to have

Volume 4, 2001

320

Moyer-Mileur et al.

Fig. 2. Scatter plots illustrating positive association between onset of menarche and (A) distal and (B) midshaft tibia total vBMD and (C) lumbar spine and (D) femoral neck aBMD.

significantly higher bone mass accumulation and higher bone strength than those who had not yet started menstruation. In a multiple regression model, pubertal stage, BMI, and age were highly interrelated, although multicollinearity was not found to be significant However, partial correlations, which identified BMI and age as covariates, revealed a

Journal of Clinical Densitometry

strong positive relationship between pubertal stage and bone mass in terms of total and cortical vBMD and bone strength in polar SSI. Significant relationships were also found between age and midshaft cortical vBMD and lumbar spine aBMD. However, chronologic age was not an important determinant of bone density or strength. This

Volume 4, 2001

Bone Mass by pQCT in Early Adolescent Girls

321

Table 5 Comparison of Bone Compartment vBMD, Cortical Thickness, and Bone Strength by pQCT and Lumbar Spine and Hip BMD by DXA to Subjects’ Menarcheal Statusa

Age Tanner stage Weight (kg) Height (cm) BMI (kg/m2) Distal tibia Total vBMD (mg/cm3) Trabecular vBMD (mg/cm3) Cortical thickness (mm) Polar SSI (mm3) Midshaft tibia Total vBMD (mg/cm3 Cortical vBMD (mg/cm3) Cortical thickness (mm) Polar SSI (mm3) Lumbar spine aBMD (mg/cm2) Femoral neck aBMD (mg/cm2) a

Premenarche (n = 43)

Postmenarche (n = 41)

p value

12.6 (0.8) 2.2 (0.5) 46.6 (13.9) 157.0 (8.9) 18.7 (4.3)

13.0 (0.8) 2.9 (0.4) 54.5 (12.6) 161.0 (6.7) 20.9 (3.9)

0.01 0.00 0.01 0.02 0.02

353.5 (54.4) 237.0 (30.2) 1.1(0.4) 695.8 (322.8)

390.9 (52.9) 244.3 (31.9) 1.3 (0.3) 817.7 (244.0)

0.003 0.35 0.004 0.06

573.9 (64.8) 1040.2 (35.4) 3.2 (0.4) 1620.2 (384.6)

606.5 (67.0) 1053.9 (35.8) 3.4 (0.4) 1664.4 (335.3)

0.03 0.08 0.06 0.58

755.1 (125.0)

932.6 (114.4)

0.001

769.2 (113.2)

911.1 (125.9)

0.001

n = 85; two-tailed t-test.

Table 6 Stepwise Multiple Regression Model for pQCT and DXA Bone Variables and Related Factors Regression Distal tibia Total vBMD (mg/cm3) Cortical thickness (mm) Polar SSI (mm2) Midshaft tibia Total vBMD (mg/cm3) Cortical thickness (mm) Polar SSI (mm2) Lumbar spine BMD (mg/cm2) Femoral neck BMD (mg/cm2)

Journal of Clinical Densitometry

Variance

r

r2

p

348.082 + 40.615 menarche –07829 + 0.216 menarche + 0.0118 Height 178.04 + 11.926 Weight

Menarche (11.7%) Menarche (11.2%); height (4.2%) Weight (19.4%)

0.36 0.42

0.13 0.002 0.18 0.001

0.45

0.21 0.000

693.956 – 6.718 BMI + 46.457 menarche 3.025 + 0.00034 calcium intake 936.586 + 14.557 weight 0.476 + 0.00626 Weight + 0.116 menarche 0.470 + 0.00542 Weight + 0.104 menarche + 0.0012 weight-bearing activity

BMI (5.8%); menarche (10.4%) Calcium intake (6.8%) Weight (24.3%) Weight (42.8%); menarche (12.9%) Weight (36.7%); menarche (12.0%); weight-bearing activity (4.1%)

0.43

0.19 0.001

0.29 0.50 0.76

0.08 0.02 0.25 0.000 0.57 0.000

0.74

0.55 0.000

Volume 4, 2001

322 finding may be owing to the narrow age range of our sample (11–14 yr). Our results are consistent with other studies in which the covariance of age with other potentially more important determinants of bone such as puberty and body weight were carefully controlled (4,10). We found that weight and height were significantly correlated with distal tibia cortical thickness and polar SSI and lumbar spine and femoral neck aBMD. BMI was also significantly correlated with distal tibia polar SSI and lumbar spine aBMD. In the multiple regression model, body weight was a significant predictor of tibia bone strength and aBMD. BMI, which may be a better index to reflect body size, was just a weak predictor of total vBMD in the multiple regression model. The influence of body size in terms of height and weight was consistent with previous studies. In a study by Boot et al. (11), height had no significant influence on spine BMD, whereas in studies by Rice et al. (4), Bonjour et al. (20), and Miller et al. (23), height was shown to be significantly correlated with bone mineral in children and adolescents. Several studies have observed that body weight was strongly associated with BMC or BMD (24,25) However, Sentipal et al. (5) found that body weight was not related to bone status. Daily calcium intake, as calculated from a 3-d food record, was 935 ± 364 mg/d, equivalent to 78% of the RDA. This value was slightly higher than the average intake of 771 mg of calcium/d for girls ages 12–19 reported in the USDA Continuing Survey for Food Intakes by Individuals, 1994–96 (26). A small but significant relationship was found between calcium intake and tibia cortical thickness. A significant, positive effect of calcium intake on bone mineralization has been suggested by several crosssectional (5,9) and prospective (23–30) studies; however, other cross-sectional studies have not (20,31,32). Disagreement surrounding the effect of calcium intake on bone development may be owing to the influence of the pubertal stage and hormone activity. Our subjects were studied at the onset of puberty when bone is maximally stimulated by GHs and estrogen, thus obscuring the relatively small (33) but important effect of dietary calcium. In our study, level of weight-bearing physical activity expressed as estimated metabolic cost-hours per week, which might be more accurate to reflect sub-

Journal of Clinical Densitometry

Moyer-Mileur et al. jects’ activity level, was only significantly correlated with total bone area and BMC, and trabecular bone area. However, it was still a weak but significant predictor of midshaft tibia bone density and strength and femoral neck BMD in the multiple regression model. No significant correlation between physical activities and bone mineral has been reported (4,11). However, our results support those of Slemenda et al. (30), who found that the level of weight-bearing physical activity expressed in total hours per week was positively correlated to BMD at the radius and hip in girls ages 5–14 (30). Rubin et al. (6) also found a positive correlation between physical activity and lumbar spine BMD.

Conclusion This cross-sectional study provides reference information on tibia bone mass in healthy female adolescents from a sample of 84 adolescent girls ages 11–14. Bone mass defined as total and cortical vBMD and aBMD increases during puberty. Trabecular bone mass was not influenced by age, puberty, body size, or other lifestyle factors. Body weight was the most important predictor and determinant of bone density and strength in healthy adolescent girls. Menarche, calcium intake, height, BMI, and level of weight-bearing physical activity and age were also identified as minor but significant predictors and determinants of bone density and strength. Bone measurements by pQCT allow the separation of cortical and trabecular bone compartments and provide information on bone acquisition, architecture, and strength during rapid periods of growth and development. Broader cross-sectional studies using the pQCT technique to evaluate the influence of age, gender, ethnicity, puberty, body size, and lifestyle factors on bone acquisition and strength are needed.

Acknowledgments We acknowledge the support of Primary Children’s Medical Center Foundation and the efforts of Leona Hollingsworth, Tricia Pratt, Ryan Roberts, Susan Runyon, and Marnie Shepperd Clay. Preliminary data were presented at the 1st International Workshop on Musculoskeletal Interactions, Santorini, Greece, May 1999.

Volume 4, 2001

Bone Mass by pQCT in Early Adolescent Girls

References 1 Conference Report. 1991 Consensus Development Conference: prevention and treatment of osteoporosis. Am J Med 90:107–110. 2 Matkovic V. 1992 Calcium and peak bone mass. J Intern Med 231:151–160. 3 Matkovic V, Jelic T, Wardlaw GM, et al. 1993 Timing of peak bone mass in Caucasian females and its implication for the prevention of osteoporosis. J Clin Invest 2:799–808. 4 Rice S, Blimkie CJ, Webber CE, et al. 1993 Correlates and determinants of bone mineral content and density in healthy adolescent girls. Can J Physiol Pharmacol 71(12):923–930. 5 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(2):425–428. 6 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(6):863–870. 7 Gunnes M, Lehmann EH. 1996 Physical activity and dietary constituents as predictors of forearm cortical and trabecular bone gain in healthy children and adolescents: a prospective study. Acta Paediatr 85:19–25. 8 De Schepper J, Derde MP, Broeck VD, Piepsz A, Jonckheer MH. 1991 Normative data for lumbar spine bone mineral content in children: influence of age, height, weight, and pubertal stage. J Nucl Med 32(2):216–220. 9 Gunnes M, Lehmann EH. 1995 Dietary calcium, saturated fat, fiber and vitamin C as predictors of forearm cortical and trabecular bone mineral density in healthy children and adolescents. Acta Paediatr 84:388–392. 10 Molgaard C, Thomsen BL, Michaelsen KF. 1998 Influence of weight, age and puberty on bone size and bone mineral content in healthy children and adolescents. Acta Paediatr 87:494–499. 11 Boot AM, Ridder MA, Pols HA, Krenning EP, Muinck KS. 1997 Bone mineral density in children and adolescents: relation to puberty, calcium intake, and physical activity. J Clin Endocrinol Metab 82(1):57–62. 12 Kleerekoper M, Nelson DA. 1997 Which bone density measurement? J Bone Miner Res 12:712–714. 13 Sievanen H, Koskue V, Rauhio A, Kannus P, Heinonen A, Vuori I. 1998 Peripheral quantitative computed tomography in human long bones: evaluation of in vitro and in vivo precision. J Bone Miner Res 13(5):871–882. 14 Aaron DJ, Kriska AM, Dearwater SR, Cauley JA, Metz KF, LaPorte RE. 1995 Reproducibility and validity of an epidemiologic questionnaire to assess past-year physical activity in adolescents. Am J Epidemiol 142(2):191–201. 15 Aaron DJ, Kriska AM, Dearwater SR, et al. 1993 The epidemiology of leisure physical activity in an adolescent population. Med Sci Sports Exerc 25(7):847–853. 16 USRDA. 1989 National Research Council, National Academy of Sciences, Washington, DC, 10th ed. 17 Tanner JM. 1962 Growth at Adolescence, 2nd ed. Oxford: Blackwell. England

Journal of Clinical Densitometry

323 18 Woolf AD, Dixon AS. 1998 Osteoporosis: A Clinical Guide, 2nd ed. London: Martin Dunitz. 19 Neu CM, Manz F, Rauch F, Merkel A, Schonau E. 2001 Bone densities and bone size at the distal radius in healthy children and adolescents: a study using peripheral quantitative computed tomography. Bone 28:227–232. 20 Bonjour JP, Theintz F, 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(3):555–563. 21 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(6):1332–1339. 22 Gilsanz V, Roe TF, Mora S, Costin G, Goodman WG. 1991 Changes in vertebral bone density in black girls and white girls during childhood and puberty. N Engl J Med 325(23):1597–1600. 23 Miller JZ, Slemenda CW, Meaney FJ, Reister TK, Hui S, Johnston CC. 1991 The relationship of bone mineral density and anthropometric variables in healthy male and female children. Bone Miner 14(2):137–152. 24 Lloyd T, Rollings N, Andon MB, et al. 1992 Determinants of bone density in young women. I. Relationships among pubertal development, total body bone mass, and total body bone density in premenarchal females. J Clin Endocrinol Metab 75(2):383–387. 25 Russell-Aulet M, Wang J, Thornton JC, Colt EW, Pierson RN Jr. 1993 Bone mineral density and mass in a cross-sectional study of white and Asian women. J Bone Miner Res 8(5):575–582. 26 U.S. Department of Agriculture, Agricultural Research Service. 1997 Data Tables: Results from USDA’s 1994–96 Continuing Survey of Food Intakes by Individuals and 1994–96 Diet and Health Knowledge Survey. In: 1994–96 Continuing Survey of Food Intakes by Individuals and 1994–96 Diet and Health Knowledge Survey. CD-ROM, NTIS accession no. PB98-500457. 27 Chan GM. 1991 Dietary calcium and bone mineral status of children and adolescents. Am J Dis Child 145:631–634. 28 Lloyd T, Andon MB, Rollings N, et al. 1993 Calcium supplementation and bone mineral density in adolescent girls. JAMA 270(7):841–844. 29 Cadogan J, Eastell R, Jones N, Barker ME. 1997 Milk intake and bone mineral acquisition in adolescent girls: randomized, controlled intervention trial. BMJ 315(7118):1255–1260. 30 Slemenda CW, Miller JZ, Hui SL, Reister TK, Johnston CC Jr. 1991 Role of physical activity in the development of skeletal mass in children. J Bone Miner Res 6(11):1227–1233. 31 Grimston SK, Morrison K, Harder JA, Hanley DA. 1992 Bone mineral density during puberty in western Canadian children. Bone Miner 19(1):85–96. 32 Kroger H, Kotaniemi A, Vainio P, Alhava E. 1992 Bone densitometry of the spine and femur in children by dual-energy X-ray absorptiometry. Bone Miner 17(3):75–85. 33 Clark PA, Rogot AD. 1996 Growth hormones and sex steroid interactions at puberty. Endocrinol Metab Clin North Am 25(3):665–681.

Volume 4, 2001