Journal of Bone and Mineral Metabolism

, Volume 29, Issue 1, pp 44–53

Comparisons of trabecular and cortical bone in late adolescent black and white females

Authors

    • Department of Pediatrics, Georgia Prevention InstituteMedical College of Georgia
  • Emma M. Laing
    • Department of Foods and NutritionUniversity of Georgia
  • Ruth G. Taylor
    • Department of Foods and NutritionUniversity of Georgia
  • Clifton A. Baile
    • Department of Foods and NutritionUniversity of Georgia
  • Mark W. Hamrick
    • Department of Cellular Biology and AnatomyMedical College of Georgia
  • Daniel B. Hall
    • Department of StatisticsUniversity of Georgia
  • Richard D. Lewis
    • Department of Foods and NutritionUniversity of Georgia
Original Article

DOI: 10.1007/s00774-010-0186-z

Cite this article as:
Pollock, N.K., Laing, E.M., Taylor, R.G. et al. J Bone Miner Metab (2011) 29: 44. doi:10.1007/s00774-010-0186-z

Abstract

Few childhood studies have investigated racial differences in volumetric bone mineral density (vBMD), bone geometry, and bone strength indices measured by three-dimensional bone imaging. The purpose of this study was to compare trabecular and cortical bone parameters at the radius and tibia between late adolescent white and black females using peripheral quantitative computed tomography (QCT). White (n = 25) and black females (n = 25), 18–19 years of age, were pair-matched for age, height, and fat-free soft tissue mass. Peripheral QCT scans were obtained at the 4% (trabecular bone), 20% (cortical bone), and 66% [muscle cross-sectional area (CSA)] sites from the distal metaphyses. Bone strength was determined from vBMD and bone geometry to calculate bone strength index (BSI; trabecular site) and polar strength–strain index (SSI; cortical site). Radial SSI was not different between groups; however, blacks had greater radial BSI (P = 0.02) than whites. After adjustment for the longer forearm in blacks, the greater radial BSI in blacks no longer remained. At the tibia, blacks versus whites had greater bone strength at the trabecular and cortical bone sites (BSI, P = 0.03; SSI, P = 0.04, respectively). When controlling for differences in tibial length and muscle CSA, the higher estimates of bone strength persisted in blacks versus whites (BSI, P = 0.01; SSI, P = 0.02). Our data suggest that when differences in body size are considered, late adolescent black versus white females have a stronger bone profile, due to greater bone geometry and vBMD, at the trabecular and cortical regions of the tibia but not at the radius.

Keywords

RaceAfrican AmericanBone structurepQCT

Introduction

In the United States, osteoporotic fractures occur at twice the rate in white women compared to black women [1]. These racial differences in fragility fractures are thought to be the result of greater areal bone mineral density (aBMD) and/or bone mineral content (BMC) observed in blacks than in whites that seem to be present in childhood and track into adulthood [28]. However, many of these racial differences in aBMD and BMC are attenuated when taking into account differences in age, gender, sexual maturation, and body size [26, 9, 10]. As a result, controversy lies in whether racial differences in fracture risk can be attributed to racial differences in bone strength [11, 12]. Ideally, predicting bone strength or ultimately, a bone’s failure load, requires information of both the material (e.g., mineral density) and geometric (e.g., size and shape) properties of trabecular and cortical bone [13, 14]. Widening the focus beyond bone mass/density to consider bone geometry will undoubtedly advance the knowledge of racial differences in fracture risk.

Dual-energy X-ray absorptiometry (DXA) can be used to assess material surrogates of bone strength (BMC and aBMD); however, because of its two-dimensional methodology, DXA lacks the technical means to measure the geometric parameters related to bone strength and the ability to differentiate between trabecular and cortical bone. DXA has been a useful tool for assessment of BMC and aBMD; however, these two-dimensional bone outcomes are confounded by body size. This limitation of DXA is particularly relevant when studying pediatric bone [15]. Newer three-dimensional bone imaging techniques, such as quantitative computed tomography (QCT) and peripheral QCT (pQCT), may provide greater accuracy in predicting bone strength because they have the capability to measure both volumetric BMD (vBMD) and bone geometry while also distinguishing trabecular and cortical compartments. Thus, QCT/pQCT studies are needed to characterize childhood racial differences in both vBMD and bone geometry as determining how and when racial differences in bone strength manifest can be important for establishing optimal prevention strategies.

To date, QCT/pQCT data on racial differences in trabecular and/or cortical bone between blacks and whites during growth have been limited to four studies [8, 1618]. Gilsanz and colleagues [8] used QCT to assess lumbar spine vBMD in black and white females between the ages of 2 and 20 years, matched for age and sexual maturation. In the overall analyses, groups did not differ in lumbar spine vBMD; however, among those in late puberty (Tanner stage 5), the black females had 23% greater lumbar spine vBMD than the white females. In another QCT study [16], the same group of investigators matched black and white males and females, aged 8–18 years, for age, sex, weight, height, and sexual maturation to assess racial differences in vBMD and cross-sectional area (CSA) at the lumbar spine and femoral midshaft. At the lumbar spine, blacks had higher vBMD than whites, but a similar CSA. Conversely, at the femur, no differences were found in vBMD; however, the black females had greater CSA than their white counterparts, which the authors attributed to their longer femoral bone length [16]. Using pQCT, Wetzsteon et al. [17] compared trabecular and cortical bone parameters at the radius and tibia between 21 white and 23 black children, aged 9–12 years. At the radius and tibia, the investigators found that the black children had significantly stronger bones than the white children as a consequence of greater vBMD and cortical bone CSA [17]. In the fourth study, Leonard et al. [18] compared pQCT-derived cortical vBMD and cortical bone geometry parameters between black and white children, adolescents, and young adults, ranging in age from 5 to 35 years. They found that cortical vBMD and cortical bone geometry parameters at the tibia were significantly greater in blacks compared with whites in Tanner stages 1 through 4; however, in Tanner stage 5 the racial differences in bone outcomes were attenuated. These studies provide important insight into trabecular and cortical vBMD and geometry comparisons between blacks and whites during growth; however, three of the QCT/pQCT studies [8, 16, 18] lack data at non-weight-bearing skeletal sites and the other pQCT study [17] was limited to data from pre- and early pubertal black and white children. To our knowledge, no QCT/pQCT studies have assessed racial differences in trabecular and cortical vBMD and geometry at a non-weight-bearing skeletal site (e.g., radius) between black and white females entering adulthood.

In this pQCT study, we sought to determine racial differences in bone parameters of bone strength at trabecular and cortical sites of the radius and tibia between late adolescent white and black females. Estimates of bone strength were integrated from vBMD and bone geometry parameters at the radius and tibia to calculate the bone strength index (BSI) (trabecular site) and the polar strength–strain index (SSI) (cortical site). Because the functional model of bone development implies that a bone must continually adapt its mass and geometry to withstand loads from increases in bone length, muscle mass, and forces [19, 20], it has been suggested that a bone’s integrity at any age should be analyzed from a functional approach (e.g., relative to muscle size and/or strength and bone length) [2022]. Therefore, in our bone analyses between racial groups, we considered not only the absolute bone measurements but also those measures relative to differences in site-specific muscle CSA and/or limb length.

Materials and methods

Study participants

This study included the baseline data from late adolescent white (n = 25) and black (n = 25) females, aged 18–19 years, recruited from the University of Georgia. These subjects are a subset of a larger cohort study of factors relating to body composition in college-aged females [23]. Because of the importance of height and total fat-free soft-tissue (FFST) mass on bone outcomes [22], we pair-matched a white female with a black female from the larger cohort in the order of age (±2%), height (±1%), and FFST mass (±2%). Inclusion criteria were white or black/African American female, 18–25 years of age, and normal menstruation (e.g., ≥4 menstrual periods in the last 6 months). Participants were excluded if they reported significant weight loss or gain in the past 6 months (±10% initial body weight), participation in competitive high school (grade 9–12) or college athletics, diagnosis of eating disorders, present illness or chronic disease, and/or use of oral contraceptives, dietary supplements, or medications known to affect body metabolism. Procedures were approved by the Institutional Review Board for Human Subjects at the University of Georgia, and all participants provided written consent.

Anthropometry and body composition

Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer (Novel Products, Rockton, IL, USA). Body weight was measured to the nearest 0.1 kg using an electronic scale (Seca Bella 840, Columbia, MD, USA). Limb lengths were measured with anthropometric tape (Rosscraft, Canada) to the nearest 0.10 mm at the forearm (distance between the ulnar styloid process and olecranon) and tibia (the distal edge of the medial malleolus to the tibial plateau).

Body composition outcomes [fat mass (kg), FFST mass (kg), and percentage body fat (%fat)] of the total body were measured using DXA (Delphi A; S/N 70467; Hologic, Bedford, MA. USA). The same technician analyzed all scans using Hologic software, version 11.2 (Whole Body). Quality assurance for DXA was performed via calibration against the manufacturer’s three-step soft tissue wedge (model TBAR; SN 2275) and anthropomorphic phantom (model DPA/QDR-1; SN 9374) composed of different thickness levels of aluminum and lucite, calibrated against stearic acid (100% fat) and water (8.6% fat). Using a one-way random effects model, single measure intraclass correlation coefficients (ICCs) were calculated in females 18–30 years of age (n = 10) scanned twice in our lab during a 7-day period for total body fat mass, FFST mass, and %fat (all R ≥ 0.87).

Peripheral QCT

Peripheral QCT (XCT-2000; Stratec Medizintechnic, Pforzheim, Germany) measurements were taken of the nondominant radius and tibia. Radial and tibial measures were taken at the 4% and 20% sites for bone outcomes, proximal to the articular surface of the distal end of the tibia and radius. The 4% and 20% sites represent areas high in trabecular and cortical bone, respectively; and these specific bone sites along the radius and tibia have been validated against failure load techniques [24, 25]. Each scan was acquired with a 0.4-mm voxel size and a slice thickness of 2.4 mm. The anatomic reference line (for determination of the distal end of the radius or tibia) was identified by acquisition of a 30-mm planar scout view of the joint line and automatically set by the software at 4% or 20% sites. Image processing and calculation of the various bone measures were determined using the Stratec software (version 5.50d), which allows modal options for the following: Contour Mode for detecting the outer bone edge, Peel Mode for defining the way subcortical and trabecular bone are separated, and Cort Mode for separating cortical bone from trabecular bone.

At the trabecular site of the radius and tibia, total and trabecular vBMD (mg/cm3) and total CSA (mm2) were measured using Contour Mode 2 (200 mg/cm3), Peel Mode 2 (400 mg/cm3), and Cort Mode 3 (169 mg/cm3). Using total vBMD and total CSA, BSI (mg2/mm4, Eq. 1) was calculated, which reflects the strength of bone against compression [26, 27]. The following bone variables were assessed at the radial and tibial cortical sites: cortical vBMD (mg/cm3), cortical CSA (mm2), total CSA, cortical BMC (mg/mm), cortical thickness (mm), and polar SSI (mm3). Cortical bone variables for the 20% site were assessed using Contour Mode 1 (710 mg/cm3), Peel Mode 2 (540 mg/cm2), and Cort Mode 1 (480 mg/cm3). The polar SSI (mm3), an estimate of torsional bone strength, is calculated as the integrated product of the section modulus and the density of cortical bone (Eq. 2). Section modulus is calculated as (a × d2)/dmax, where a is the CSA of a voxel, d is the distance of the voxel from the center of gravity, and dmax is the maximum distance of one voxel to the center of gravity. The ratio of cortical vBMD and normal physiological density (ND = 1,200 mg/cm3) provides an estimate of the modulus of elasticity [28].
$$ {\text{Bone}}\,{\text{strength}}\,{\text{index}}\left( {{\text{BSI}},{\text{mg}}^{ 2} /{\text{mm}}^{ 4} } \right) = {\text{total}}\,{\text{CSA}} \times {\text{total}}\,{\text{vBMD}}^{ 2} $$
(1)
$$ \hbox{Polar strength}{-}\hbox{strain index}\left( {{\text{SSI}},{\text{mm}}^{ 3} } \right) = \sum\limits_{i} {{\frac{{\left[ {\left( {a_{i} \times d_{i}^{ 2} } \right)\left( {{\text{cortical}}\,{\text{vBMD}}/{\text{ND}}} \right)} \right.}}{{d_{ \max } }}}}.$$
(2)

The maximal muscle CSA (mm2) of the forearm and lower leg has been accepted as a reasonable surrogate marker of muscle force [29, 30]. Accordingly, a third measurement was taken at the 66% site of both the radius and tibia to evaluate the contribution of muscle force. Contour Mode 1 with a threshold of 34 mg/cm3 and Peel mode 1 were used to obtain the “area of muscle plus bone” (i.e., muscle + radius + ulna or muscle + tibia + fibula) from the “area of subcutaneous fat.” Next, the analysis was performed with Contour Mode 1 and the threshold set at 710 mg/cm3 to determine the “area of bone” (i.e., radius + ulna or tibia + fibula). Muscle CSA is determined by subtracting the “area of bone” from the “area of muscle plus bone.”

All pQCT measures were assessed and analyzed by the same trained operator. The pQCT operator scanned the phantom daily to maintain quality assurance. Test–retest measurements were performed on five females, aged 18–24 years, to determine reliability of the pQCT in our laboratory. The one-way random effects model, ICCs for all pQCT measurements were calculated to be R ≥ 0.97.

Dietary intake

Three-day diet records were used to estimate average intakes of daily energy, macronutrient, calcium, and vitamin D intake per day. The 3 days included 2 weekdays and 1 weekend day. The 3-day diet records were analyzed by Food Processor for Windows version 8.0 (ESHA Research, Salem, OR, USA). In our laboratory, the reliability of diet records was investigated in a previous study of girls 6–10 years of age (n = 10) who completed 3-day diet records twice over a 2-week period. In that investigation, for the one-way random effects model, ICCs were computed for 3-day energy intake and 3-day calcium intake and found to be R = 0.47 and calcium R = 0.71, respectively.

Physical activity

Information on physical activity for the past week was collected using the interviewer-administered 7-day recall questionnaire [31], which has been validated in females within this age group [32]. Participants reported the amount of hours per day spent sleeping as well as the minutes per day spent performing moderate, hard, and very hard activities during the previous week.

Statistical analyses

Data were analyzed using SPSS version 17.0.2 (Chicago, IL, USA) for the Mac OS X. Normal distribution and homogeneity of variances were confirmed by Shapiro–Wilks W and Levene’s tests, respectively. Racial group differences for anthropometric, body composition, physical activity, dietary intake, and unadjusted bone outcome variables were determined using a paired-samples t test (two-tailed). Descriptive statistics for raw variables are presented as mean ± SD. Pearson’s bivariate correlations were conducted separately for the white and black female groups to determine the associations of skeletal site-specific limb length and muscle CSA with bone parameters at the radius and tibia. An F test was performed to test the assumption of homogeneity of regression slopes with regard to the interaction between the independent variables (i.e., racial groups) and the covariates (i.e., forearm length, tibial length, and tibial muscle CSA). Because there were no interactions, analysis of covariance was used to compare the differences in bone response variables between white and black female adolescents after adjusting for muscle CSA and/or limb length differences. Estimated means of bone variables in the adjusted analyses are reported as mean ± SE. P < 0.05 was considered statistically significant in all the analyses. The magnitude of difference between groups was calculated using the Cohen’s d (d) effect size formula: (mean1 – mean2)/pooled SD [33]. Values of 0.2, 0.5, and 0.8 for Cohen’s d indicate small, medium, and large effects, respectively.

Results

Descriptive characteristics

Mean age, weight, height, BMI-for-age, menarche, FFST and fat mass, %fat, forearm and tibial length and muscle CSA, and dietary intakes of the participants are provided in Table 1. By design, no differences were found in mean age, height, weight, and FFST mass values between white and black females. There were also no racial differences in BMI-for-age, menstrual start age, fat mass, and %fat; however, forearm and tibial lengths were greater in the black versus the white group (both P < 0.01). The muscle CSA at the tibia (P < 0.01), but not at the radius, was significantly greater in the white compared with the black females. Mean intakes for all macronutrients and dietary calcium and vitamin D were not different between groups (Table 1).
Table 1

Descriptive characteristics of late adolescent white and black females

 

White females (n = 25)

Black females (n = 25)

Age (years)

18.2 ± 0.4

18.5 ± 0.5

Weight (kg)

63.5 ± 8.8

63.7 ± 9.1

Height (cm)

163.8 ± 6.8

162.9 ± 6.7

BMI (kg/m2)

23.7 ± 3.2

24.1 ± 3.7

Menarche (years)

12.9 ± 1.4

12.3 ± 1.6

FFST mass (kg)

43.5 ± 4.8

44.0 ± 5.4

Fat mass (kg)

18.9 ± 5.4

19.8 ± 6.0

Fat mass (%)

28.9 ± 4.8

29.5 ± 5.5

Forearm length (mm)

256.6 ± 14.8

267.5 ± 12.8*

Tibial length (mm)

372.0 ± 21.5

387.2 ± 26.2*

Forearm muscle CSA (mm2)

2743 ± 360

2728 ± 513

Tibial muscle CSA (mm2)

7636 ± 1487

6292 ± 1082*

Energy intake (kcal/day)

1668 ± 396

1778 ± 351

Carbohydrate (g/day)

201 ± 48

233 ± 59

Protein (g/day)

67 ± 25

69 ± 15

Fat (g/day)

68 ± 25

65 ± 16

Calcium (mg/day)

589 ± 227

656 ± 242

Vitamin D (IU/day)

59 ± 58

56 ± 39

BMI body mass index, FFST fat-free soft-tissue, CSA cross-sectional area

Values are mean ± SD

* P < 0.05; significantly different from white females

No significant differences were observed between the white versus black females in reported hours per day of sleep (7.4 ± 0.6 vs. 7.3 ± 1.2, P = 0.73) or in minutes per day of hard (6 ± 22 vs. 14 ± 21, P = 0.176) and very hard (11 ± 11 vs. 10 ± 13, P = 0.679) activities. However, black versus white females reported more minutes per day of moderate activities (42 ± 19 vs. 11 ± 20, respectively, P = 0.01).

Bivariate correlations of skeletal site-specific limb length and muscle CSA with pQCT bone parameters

Simple relationships of site-specific limb length and muscle CSA (MCSA) with bone parameters are shown separately for the white and black females in Table 2. In the group of white females, tibial length was positively associated with total CSA at trabecular and cortical sites of the tibia (both P < 0.01). Forearm MCSA was positively correlated with total vBMD, BSI, cortical CSA, total CSA (20% site only), cortical BMC, cortical thickness, and SSI (all P < 0.05). At the tibia, positive relationships were found between tibial muscle CSA and the following bone parameters: cortical CSA, cortical BMC, cortical thickness, and SSI (all P < 0.05).
Table 2

Bivariate correlation coefficients of skeletal site-specific limb length and muscle CSA with peripheral quantitative computed tomography (pQCT) bone parameters at the radius and tibia

 

White females (n = 25)

Black females (n = 25)

Limb lengtha

Muscle CSAb

Limb lengtha

Muscle CSAb

4%, trabecular site

 Total vBMD

  Radius

−0.076

0.635*

−0.047

0.234

  Tibia

−0.130

0.221

0.122

0.144

 Trabecular vBMD

  Radius

−0.147

0.323

−0.137

0.061

  Tibia

0.048

0.109

0.164

0.088

 Total CSA

  Radius

0.378

0.089

0.229

0.234

  Tibia

0.690*

0.130

0.061

0.144

 BSI

  Radius

0.102

0.657*

0.110

0.469*

  Tibia

0.228

0.270

0.299

0.418*

20%, cortical site

 Cortical vBMD

  Radius

−0.015

0.039

−0.228

−0.304

  Tibia

−0.006

−0.272

−0.330

−0.219

 Cortical CSA

  Radius

0.326

0.587*

0.146

0.615*

  Tibia

0.113

0.528*

0.419*

0.511*

 Total CSA

  Radius

0.394*

0.401*

0.322

0.640*

  Tibia

0.506*

0.354

0.468*

0.493*

 Cortical BMC

  Radius

0.329

0.596*

0.065

0.501*

  Tibia

0.114

0.501*

0.383

0.499*

 Cortical thickness

  Radius

0.067

0.559*

−0.159

0.076

  Tibia

−0.238

0.411*

0.152

0.246

 SSI

  Radius

0.383

0.432*

0.184

0.580*

  Tibia

0.377

0.401*

0.421*

0.553*

vBMD volumetric bone mineral density, CSA cross-sectional area, BSI bone strength index, BMC bone mineral content, SSI polar strength–strain index

* P < 0.05

aLimb length corresponds to skeletal site-specific relationships (i.e., forearm length and radial bone parameters; tibial length and tibial bone parameters)

bMuscle CSA corresponds to skeletal site-specific relationships (i.e., forearm muscle CSA and radial bone parameters; leg muscle CSA and tibial bone parameters)

In the group of black females, tibial length was positively related with cortical CSA, total CSA (20% site only), and SSI (all P < 0.04). Forearm MCSA was positively associated with BSI, cortical CSA, total CSA (20% site only), cortical BMC, and SSI (all P < 0.01). At the tibia, positive relationships were found between tibial muscle CSA and the following bone parameters: BSI, cortical CSA, total CSA (20% site only), cortical BMC, and SSI (all P < 0.04).

pQCT bone parameter comparisons between late adolescent white and black females

Radius

Group comparisons of the radial bone outcomes (unadjusted and adjusted, controlling for differences in forearm length) at the trabecular and cortical sites are reported in Table 3. At the trabecular site, blacks versus whites had significantly higher values for total vBMD (15.6%, d = 1.09) and BSI (18.5%, d = 0.60), and lower values for total CSA (10.9%, d = 0.59). Although the black females had 7.3% greater trabecular vBMD than whites, this difference was not statistically significant (d = 0.31). Adjusting for the longer forearm length in the black females did not alter the unadjusted bone outcomes for total vBMD (blacks > whites, d = 1.05), trabecular vBMD (black > whites, d = 0.41), and total CSA (whites > blacks, d = 0.97). However, no significant group differences were found for BSI after correcting for forearm length (d = 0.41).
Table 3

Radial pQCT bone outcomes at trabecular and cortical sites in late adolescent white and black females

Radial bone variables

Unadjusteda

Adjustedb

White (n = 25)

Black (n = 25)

P valuec

White (n = 25)

Black (n = 25)

P valued

4%, trabecular site

 Total vBMD (mg/cm3)

343.6 ± 11.8

407.3 ± 14.9

0.001

342.0 ± 14.1

408.9 ± 14.1

0.002

 Trabecular vBMD (mg/cm3)

208.8 ± 7.7

225.2 ± 6.4

0.115

206.8 ± 7.4

227.2 ± 7.4

0.065

 Total CSA (mm2)

285.3 ± 9.4

254.1 ± 9.4

0.014

291.0 ± 9.4

264.4 ± 9.4

0.003

 BSI (mg2/mm4)

341.6 ± 22.7

419.3 ± 24.1

0.023

346.5 ± 24.4

414.5 ± 24.4

0.064

20%, cortical site

 Cortical vBMD (mg/cm3)

1195 ± 3.6

1174 ± 9.0

0.031

1193 ± 7.1

1176 ± 7.1

0.107

 Cortical CSA (mm2)

71.4 ± 1.9

73.4 ± 1.8

0.419

72.3 ± 2.0

72.5 ± 2.0

0.950

 Total CSA (mm2)

98.6 ± 3.2

106.5 ± 3.8

0.037

101.1 ± 3.5

104.1 ± 3.5

0.557

 Cortical BMC (mg/mm)

85.3 ± 2.4

86.2 ± 2.2

0.773

86.2 ± 2.3

85.2 ± 2.3

0.774

 Cortical thickness (mm)

2.68 ± 0.06

2.63 ± 0.05

0.541

2.68 ± 0.07

2.63 ± 0.07

0.648

 SSI (mm3)

232.0 ± 10.1

243.1 ± 11.4

0.336

238.0 ± 10.9

237.1 ± 10.9

0.953

vBMD volumetric BMD, CSA cross-sectional area, BSI bone strength index, BMC bone mineral content, SSI polar strength–strain index

aValues are mean ± SE

bValues are mean ± SE adjusted for forearm length

cTests of significance between groups are based on paired-samples t tests

dTests of significance between groups are based on group main effect with analysis of covariance (ANCOVA)

For the cortical site, no significant group differences were observed for cortical CSA (d = 0.07), cortical BMC (d = 0.02), cortical thickness (d = 0.05), and SSI (d = 0.06). The white females had 1.8% more cortical vBMD (d = 0.52) than blacks, although the black females had a 7.4% greater total CSA (d = 0.48) than whites. However, after adjusting for differences in forearm length, these significant racial differences at the cortical vBMD (d = 0.31) and total CSA (d = 0.04) were no longer observed.

Tibia

Bone measures of the tibia and estimates, controlling for muscle CSA and tibial length differences between groups, at trabecular and cortical sites are summarized in Table 4. At the trabecular bone site, blacks versus whites had significantly higher values for total vBMD at the tibia (20.4%, d = 1.92) and BSI (13.6%, d = 0.54), and lower values for trabecular vBMD at the tibia (10.2%, d = 0.67) and total CSA at the tibia (22.7%, d = 1.86). Adjusting the bone outcomes for the larger tibial muscle CSA in whites and the longer tibial length in blacks did not alter the unadjusted bone findings (total vBMD, blacks > whites, d = 1.77; trabecular vBMD, whites > blacks, d = 0.49; total CSA, whites > blacks, d = 1.65; and BSI, blacks > whites, d = 0.73).
Table 4

Tibial pQCT bone outcomes at trabecular and cortical sites in late adolescent white and black females

Tibial bone variables

Unadjusteda

P valuec

Adjustedb

P valued

White (n = 25)

Black (n = 25)

White (n = 25)

Black (n = 25)

4%, trabecular site

 Total vBMD (mg/cm3)

313.9 ± 7.7

394.1 ± 14.5

0.001

309.9 ± 12.8

398.1 ± 12.8

0.001

 Trabecular vBMD (mg/cm3)

254.5 ± 5.3

228.6 ± 8.8

0.023

254.3 ± 8.0

228.8 ± 8.0

0.043

 Total CSA (mm2)

947.0 ± 23.9

732.4 ± 38.1

0.001

950.8 ± 33.9

728.6 ± 33.9

0.001

 BSI (mg2/mm4)

941.6 ± 47.2

1089.6 ± 46.8

0.031

921.9 ± 470

1109.0 ± 47.0

0.014

20%, cortical site

 Cortical vBMD (mg/cm3)

1171 ± 3.6

1188 ± 4.3

0.016

1172 ± 4.2

1187 ± 4.2

0.023

 Cortical area (mm2)

196.9 ± 5.6

215.2 ± 5.5

0.005

191.9 ± 5.1

220.2 ± 5.1

0.001

 Total CSA (mm2)

346.3 ± 9.2

360.0 ± 10.7

0.213

343.7 ± 8.7

362.2 ± 8.7

0.172

 Cortical BMC (mg/mm)

230.5 ± 6.5

255.7 ± 6.3

0.002

225.0 ± 5.9

261.2 ± 5.9

0.001

 Cortical thickness (mm)

3.63 ± 0.09

3.95 ± 0.10

0.012

3.54 ± 0.10

4.04 ± 0.10

0.002

 SSI (mm3)

1335 ± 51.8

1460 ± 60.1

0.035

1304 ± 50.2

1491 ± 50.2

0.019

vBMD volumetric BMD, CSA cross-sectional area, BSI bone strength index, BMC bone mineral content, SSI polar strength–strain index

aValues are mean ± SE

bValues are mean ± SE adjusted for muscle CSA and tibial length

cTests of significance between groups are based on paired-samples t tests

dTests of significance between groups are based on group main effect using ANCOVA

At the cortical site, blacks had significantly greater values than whites for cortical vBMD (1.4%, d = 0.98), cortical CSA (8.5%, d = 0.60), cortical BMC (9.9%, d = 0.81), cortical thickness (8.1%, d = 0.62), and SSI (8.6%, d = 0.50). There were no group differences at the tibia in total CSA (d = 0.10). The estimated bone outcomes were similar to the unadjusted bone results (cortical vBMD, d = 0.62; cortical CSA, d = 1.31; total CSA, d = 0.23; cortical BMC, d = 1.49; cortical thickness, d = 1.12; and SSI, d = 0.66).

Discussion

To the best of our knowledge, this is the first study to investigate racial differences in vBMD, geometry, and indices of bone strength, assessed by pQCT, at weight-bearing and non-weight-bearing skeletal sites in late adolescent black and white females. A key finding of this study was that when body size is taken into consideration, late adolescent black versus white females have a stronger bone strength profile at the trabecular and cortical sites of the tibia, but not at the radius. In our analyses at the cortical site of the radius, whites had greater cortical vBMD, whereas blacks had greater total bone CSA. A higher cortical vBMD at the radius in whites compared to blacks could either be explained by reduced cortical porosity or increased mineralization; however, radial cortical BMC, which reflects mineralization, was not different between the groups. Further analyses revealed that differences in radial forearm lengths between groups explained the greater cortical vBMD in whites and the larger total bone CSA at the cortical site in blacks. This finding was not unexpected, as with increasing length, bone is exposed to greater mechanical forces as a result of increasing lever arms [22, 34]. Radial forearm length differences did not explain the group differences in total vBMD and total CSA at the trabecular site, suggesting that adaptation to mechanical forces from longer lever arms may affect cortical bone to a greater degree than trabecular bone. It is plausible that the large magnitude of difference found between racial groups in total vBMD at the trabecular region could explain the lower rate of radial fractures observed in elderly blacks over whites [35, 36]. However, our integrated radial bone analyses from bone geometry parameters and vBMD at the trabecular and cortical sites suggest that, for a given forearm length, no differences exist between whites and blacks in estimated bone strength.

Previous childhood studies have shown racial differences at the radius, although most were limited to DXA-derived bone evaluations. In these investigations, aBMD and/or BMC values at the midradius, a site of primarily cortical bone, were shown to be higher in blacks than whites [4, 5, 37]. However, many of these differences in aBMD and BMC no longer remained when statistical adjustments were made for differences in age, sexual maturation, and body size. In a recent pQCT study, Wetzsteon et al. [17] compared trabecular (8% site) and cortical (50% site) bone parameters of the radius and tibia between 21 white and 23 black children, aged 9–12 years. In contrast to the present study, Wetzsteon et al. [17] found black children to have a greater radial BSI and polar SSI than white children, after controlling for differences in age, sex, and site-specific forearm length and muscle CSA. The disparate pQCT findings at the radius can likely be attributed to differences in study design (sexual maturation stage, inclusion of males and females, measurement site along radius, sample size); however, further work is needed in a larger biracial sample of male and female participants at various stages of puberty.

Our data at the trabecular and cortical sites of tibia suggest that blacks have a stronger bone than whites at a weight-bearing skeletal site, which correspond with pQCT-derived tibial bone data in early- to midpubertal children [17, 18] and with QCT-derived femoral bone data in young adults [38]. Consistent with these data, Gilsanz and colleagues [8, 16] found that late adolescent blacks have greater QCT-derived total vBMD at the lumbar spine and a larger total CSA at the femoral midshaft than whites [8, 16]. Our study found greater cortical vBMD at the tibia in blacks versus whites, which differs from findings by Gilsanz et al. [16] who did not detect racial differences in cortical vBMD at the femur. These contrasted cortical vBMD findings at the tibia and femur may be a consequence of the complex nature of mechanical stimuli at various skeletal sites. For instance, animal studies have observed considerable heterogeneity in the response of bone to mechanical stimuli, not only among different skeletal sites but also among different regions of the same bone [39, 40]. Further work is necessary to ascertain various types of loading at different skeletal sites.

Because bones perceive loading-induced strains by adapting their structure to increases in overall strength [19, 20, 41], the use of site-specific surrogates of muscle force (i.e., muscle CSA) and limb length have been suggested when comparing bone parameters in populations of different body sizes [22, 34]. In this study, we sought to determine if racial differences in bone strength could be explained by differences in muscle CSA and/or limb length at each respective site. Our findings suggest that factors other than differences in muscle size and limb length are important with respect to vBMD and bone geometry at weight-bearing skeletal sites. Factors such as diet and physical activity have been shown to be significant determinants of bone development in our age range of participants [42]. In this study, we did not find significant dietary intake differences between groups; however, the black females, on average, reported greater time spent in moderate physical activities compared to white females. Including moderate (or hard) physical activity as a covariate in our analyses did not change any of our bone outcomes between groups (data not shown). Because bone formation is affected by systemic and/or local growth regulators (e.g., sex steroids, growth hormone, insulin-like growth factor 1, glucocorticoids, thyroid hormone, calcium-regulating hormones, and adipocyte-derived hormones), it is possible that racial differences in the initiation and exposure to these systemic and/or local factors could explain the differences in bone measures at loaded skeletal sites between whites and blacks [4346]. Estrogen, for instance, is thought to reduce the bone remodeling threshold on the endocortical surface, thereby sensitizing bone next to marrow to the effect of weight-bearing loads, resulting in greater storage of bone mineral and an increase in bone area [47]. As the black females reported an average menstrual start age of 12.3 years compared to 12.9 years in the white females, it is possible that the stronger cortical bone observed in blacks versus whites at the tibia was influenced by greater estrogen exposure caused by earlier sexual maturation. However, when we statistically controlled for menstrual start age (in addition to forearm length at the radius and tibial length and muscle CSA at the tibia), there was no significant effect of menstrual start age on the adjusted bone outcomes (data not shown). Because our study was not designed to adequately explore whether racial differences in bone outcomes could be attributed to hormonal factors, future work in this area is warranted.

An intriguing result of the study was that blacks, despite having longer forearms and tibias than whites, had smaller total CSA at the trabecular site of the radius and tibia compared to the whites. Total CSA is the area of the entire bone cross section, comprising both the cortical bone and marrow cavity, and can be considered a measure of the bone diameter [48]. It is known that when long bones increase in length, they tend to increase in outer bone size as new bone is added to the outer periosteal surface [49, 50]. In addition to the finding at the trabecular site, there were no significant differences between groups in total CSA at the cortical site. The mechanisms for these findings are unknown and may be the result of the complex interactions of genetic and hormonal factors that could influence bone growth along the shaft. The proportion of trabecular and cortical bone can differ considerably along the length of long bones [48]. For example, trabecular vBMD and bone area vary along the length of the metaphysis, and the cortical vBMD, bone area, and cortical thickness vary along the length of the shaft [51]. As a consequence, measurements obtained at different locations along the shaft of the bone are not theoretically comparable between studies. Because there is no universal consensus to date for the optimal measurement site along the radius and tibia, it is therefore critical to report bone measurement site, along with scan acquisition mode, analysis parameters, and technical quality. Further research is needed to determine the optimal trabecular and cortical bone measurement location with least variability that yields the most clinically relevant information.

The results reported in this investigation should not be extrapolated to the general population because of its cross-sectional design and relatively small sample. Although most of our subjects should have reached their final height, many of them may not have reached their peak bone mass at this age; hence, we might be missing maximum bone mineral gain. The concept of peak bone mass is related to age; however, it has been suggested that bone integrity at any age should be analyzed in a more functional way (e.g., relative to muscle size and/or strength and bone length), as bones should only be strong enough for the demands put on them by normal activities [2022]. Therefore, in our bone analyses between racial groups (individually matched for age, height, and FFST mass), we considered not only the absolute bone measurements but also those measures relative to differences in site-specific muscle CSA and/or limb length. It is important to note that the analysis of muscle loading effects is complex and that the use of muscle CSA does not reflect the functional status of the entire muscle system, including muscle length, contraction velocity, structure, and coordination [52]. Whether the data generated using this technique can be used to predict bone health and risk of skeletal fractures must be validated by subsequent prospective studies. Last, although we did not collect biochemical measures of bone-related hormones, we specifically chose to evaluate postpubertal females (age 18–19 years) rather than those at earlier stages of sexual maturation, which are associated with greater hormonal variation [43]. Despite our efforts to minimize hormonal factors via study design, future investigations should consider systemic and/or local factors at different skeletal sites, as our data at loaded and unloaded skeletal sites provide evidence for local control of bone development.

In conclusion, the differences in bone parameters assessed by pQCT between late adolescent white and black females appear to be site specific (radius or tibia) and dependent upon the type of bone (trabecular or cortical) assessed. Specifically, our integrated radial bone analyses at the trabecular and cortical sites indicate that, for a given forearm length, there are no significant racial differences in radial bone strength. At the weight-bearing tibia, however, our data reflect a stronger bone in black versus white females, independent of muscle CSA and tibial length. Further research is warranted to determine other factors responsible for the racial differences at the weight-bearing, but not at the non-weight-bearing, skeletal sites.

Acknowledgments

The authors thank the study subjects for their participation. We also thank Jessica Principe, Veronique Bredas, Ashley Ferira, and Maria Breen for coordinating the project. N.K.P., R.D.L., E.M.L., C.A.B., M.W.H., and D.B.H. were responsible for the study concept and design; N.K.P., E.M.L., R.G.F., and R.D.L. were responsible for the acquisition of the data. N.K.P. and D.B.H. conducted the statistical analyses. N.K.P., R.D.L., E.M.L., R.G.F., and D.B.H. were responsible for the interpretation of the data and drafting the manuscript. All authors contributed to the revision of the manuscript. Funding was received from The University of Georgia Research Foundation and College of Family and Consumer Sciences.

Conflict of interest statement

None of the authors had any personal or financial conflicts of interest.

Copyright information

© The Japanese Society for Bone and Mineral Research and Springer 2010