Differences in bone microarchitecture between postmenopausal Chinese-American and white women


  • For a Commentary on this article, please see Bouxsein (J Bone Miner Res. 2011;26:1389–1391. DOI: 10.1002/jbmr.442).


Chinese-American women have lower rates of hip and forearm fracture than white women despite lower areal bone density (aBMD) by dual X-ray absorptiometry (DXA). We recently reported higher trabecular (Dtrab) and cortical (Dcomp) bone density as well as greater trabecular (Tb.Th) and cortical thickness (C.Th) but smaller bone area (CSA), as measured by high-resolution peripheral quantitative computed tomography (HR-pQCT), in premenopausal Chinese-American compared with white women. These findings may help to account for the lower fracture rate among Chinese-American women but were limited to measurements in premenopausal women. This study was designed to extend these investigations to postmenopausal Chinese-American (n = 29) and white (n = 68) women. Radius CSA was 10% smaller in the Chinese-American versus the white group (p = .008), whereas their C.Th and Dcomp values were 18% and 6% greater (p < .001 for both). Tibial HR-pQCT results for cortical bone were similar to the radius, but Tb.Th was 11% greater in Chinese-American versus white women (p = .007). Tibial trabecular number and spacing were 17% lower and 20% greater, respectively, in Chinese-American women (p < .0001 for both). There were no differences in trabecular or whole-bone stiffness estimated by microstructural finite-element analysis, but Chinese-American women had a greater percentage of load carried by the cortical bone compartment at the distal radius and tibia. There was no difference in load distribution at the proximal radius or tibia. Whole-bone finite-element analysis may indicate that the thicker, more dense cortical bone and thicker trabeculae in postmenopausal Chinese-American women compensate for fewer trabeculae and smaller bone size. © 2011 American Society for Bone and Mineral Research.


Asian women generally have lower areal bone mineral (aBMD) density as measured by dual-energy X-ray absorptiometry (DXA) when compared with white women and other racial groups.1–5 Despite lower aBMD, numerous studies also report lower rates of hip and wrist fractures in Asian women than in white women.1, 6–8 These seemingly incongruent findings remain poorly understood. Using high-resolution peripheral quantitative computed tomography (HR-pQCT), we and others9, 10 recently reported that Chinese premenopausal women have smaller bone area at the radius and tibia, greater cortical bone density and thickness, and greater trabecular thickness than white premenopausal women. Additionally, we found greater trabecular bone density among Chinese-American women at both sites compared with white women.9 Although these findings may account for the lower fracture rates among Chinese-American women, it is unclear whether the differences observed in premenopausal women are relevant to postmenopausal Chinese-American and white women.

Recent advances in quantitative computational technologies allow for assessment of skeletal mechanical competence by using high-resolution, image-based microstructural finite-element analysis (µFEA). Because high-resolution images can differentiate cortical and trabecular bone, the mechanical load distribution between the two compartments also can be quantified. Validation studies have demonstrated the accuracy of mechanical measurements based on HR-pQCT and µFEA techniques at the distal radius11 and distal tibia.12 In addition, HR-pQCT-based µFEA has been used in several clinical studies to elucidate differences in mechanical competence between subjects with and without osteoporosis13 and those with and without a history of fracture.14–19

The purpose of this investigation was twofold: (1) to extend the analyses of bone density and microarchitecture by HR-pQCT to postmenopausal white and Chinese-American women and (2) to investigate whether differences in HR-pQCT variables result in differences in mechanical competence, as determined by µFEA of HRpQCT images.

Material and Methods


Ninety-seven postmenopausal women (68 white and 29 Chinese American) were studied. We chose to study Chinese-American women because, according to one study, Chinese Americans have the lowest rates of hip fracture among any Asian-American subgroups.6 Participants were recruited by newspaper and Internet advertisements, flyers, and directly at primary-care physician offices. Inclusion criteria were self-reported Chinese or white ancestry, postmenopausal status, and age between 58 and 69 years. This age range was selected in order to study women who were past the perimenopausal transition period but who have not reached the age where comorbid conditions/medications are likely to affect bone metabolism. Women were screened by history and biochemical evaluation [ie, creatinine, phosphorus, calcium, and parathyroid hormone (PTH) determinations and liver function tests] for conditions or medications known to affect bone metabolism. Exclusion criteria included untreated hyperthyroidism [thyroid-stimulating hormone (TSH) < 0.5], renal dysfunction [estimated glomerular filtration rate (eGFR ≤ 60 mL/min) calculated using the Modification of Diet in Renal Disease (MDRD) equation], liver dysfunction (aspartate transaminase or alanine transaminase 2 times the upper limit of normal), intestinal malabsorption owing to any cause, history of malignancy other than nonmelanomatous skin cancer, metabolic bone diseases such as primary or secondary hyperparathyroidism, HIV disease, organ transplantation, fragility fracture, and drug exposures affecting bone metabolism (ie, current or past use of glucocorticoids, tacrolimus, cyclosporine, methotrexate, teriparatide, calcitonin, or aromatase inhibitors; current use of hormone-replacement therapy or raloxifene; and current or more than 1 year of use of bisphosphonates). All patients gave written informed consent. Patients were compensated for study participation and travel expenses within the guidelines of the Columbia University Institutional Review Board, which approved this study.

Clinical evaluation

Information regarding past medical as well as surgical history and medications was collected. Weight and height were measured by balance beam and a wall-mounted, calibrated Harpenden stadiometer (Holtain Ltd., Crymych, UK), respectively.

Bone densitometry and microarchitecture

Areal bone mineral density (aBMD) by DXA was measured at the lumbar spine (L1–L4; LS), total hip (TH), femoral neck (FN), and one-third radius (⅓ rad) using a QDR 4500A machine (Hologic, Waltham, MA, USA). Participants were measured on the same densitometer using the same software, scan speed, and technologist certified by the International Society of Clinical Densitometry. In vivo precision, determined according to the standard method, at this facility is 1.28% at the lumbar spine, 1.36% at the hip, and 0.70% at the distal radius (one-third site).20

Volumetric bone mineral density (vBMD) and microarchitecture were measured at the nondominant forearm and tibia using an HR-pQCT instrument (XtremeCT, Scanco Medical AG, Bassersdorf, Switzerland), as described previously.9, 13, 14, 19, 21, 22 The following variables are reported at the radius and tibia: mean cross-sectional area (CSA), mean cortical thickness (Ct.Th), cortical and trabecular vBMD (Dcomp, and Dtrab), trabecular number (Tb.N) and separation (Tb.Sp), and trabecular thickness (Tb.Th). In vivo reproducibility values for HR-pQCT variables at our center are (coefficient of variability) 0.55% to 1.06% for density measures, 0.16% to 1.25% for area measures, 0.91% to 1.53% for cortical thickness, and 3.65% to 5.22% for trabecular microarchitecture variables.

Finite-element analyses for HR-pQCT images of the distal radius and tibia

Each thresholded HR-pQCT image of the distal radius or distal tibia was converted to a µFE model by directly converting bone voxels to eight-node elastic brick elements with an element size of 82 × 82 × 82 µm3.13, 19 Bone tissue was modeled as an isotropic linear elastic material with a Young's modulus (Es) of 15 GPa and a Poisson's ratio of 0.3.23 A uniaxial displacement equaling 1% of the bone segment height was applied perpendicularly to the distal surface of the radius or tibia, whereas the proximal surface was imposed with zero displacement along the same direction. Both ends of the tibia were allowed to expand freely in the transverse plane. The total reaction force was calculated from the linear µFE analysis, and the axial stiffness was calculated as the reaction force divided by the imposed displacement. Similarly, trabecular bone stiffness was calculated for the trabecular bone compartment after removing the cortex. Based on the whole-bone µFE model, the percentage of load carried by the cortical bone at the distal and proximal surfaces of the whole-bone segment was calculated as the reaction force at the cortical bone surface divided by the total reaction force at the bone surface. All µFE analyses were performed by using a customized element-by-element preconditioned conjugate gradient solver24 and implemented on a Dell XPS PC workstation (Dell, Inc., Round Rock, TX, USA).

Statistical analysis

Data are expressed as mean ± SD. Comparisons of group characteristics between the Chinese-American and white groups were evaluated by independent two-sided t test. Criterion values were adjusted for unequal variances where appropriate. Bone density, microarchitecture, and FEA variables for each site were first compared between the two racial groups without adjustment using two-sided t tests and then compared again after adjustment for differences in age alone and both age and body mass index (BMI) using generalized linear models. Data were not adjusted for years since menopause because it strongly correlated with age, and age was more strongly associated with bone density variables. Analyses were adjusted for BMI rather than weight and height because the latter two variables are highly correlated with each other. Both unadjusted and adjusted p values are reported to show the influence of covariates on comparisons. For all analyses, a two-tailed p ≤ .05 was considered to indicate statistical significance. Statistical analysis was performed using SAS, Version 9.2 (SAS Institute, Cary, NC, USA).


As shown in Table 1, Chinese-American women were slightly younger, shorter, weighed less, and had lower BMI values than white women. There was no difference in current smoking status (0% versus 1.5%, p = .32). White women were more likely to consume alcohol (25% versus 7%, p = .04), but average alcohol intake in both groups was low (≤1 drink/day). aBMD by DXA at the LS, TH, FN, and ⅓ radius did not differ significantly between the two groups (Table 2) before or after adjusting for age alone or age and BMI.

Table 1. Study Group Characteristics
 Chinese-American mean ± SD (n = 29)White mean ± SD (n = 68)p Value Chinese-American versus white
Age (years)61.1 ± 2.163.5 ± 3.0<.0001
Height (inches)61.7 ± 2.064.0 ± 2.3<.0001
Weight (pounds)128 ± 17147 ± 27.0001
BMI (m/kg2)23.6 ± 2.625.3 ± 4.9.03
Years since menopause (years)10 ± 412 ± 5.06
Table 2. aBMD by DXA in Chinese-American and White Women
SiteChinese-American mean ± SDWhite mean ± SDUnadjusted p valuep Value adjusted for agep Value adjusted for age and BMI
Lumbar spine (g/cm2)0.901 ± 0.130.921 ±
Total hip (g/cm2)0.829 ± 0.090.819 ±
Femoral neck (g/cm2)0.690 ± 0.070.689 ±
⅓ Radius (g/cm2)0.618 ± 0.060.621 ±

There were major racial differences in vBMD and structure, as measured by HR-pQCT (Table 3, Fig. 1). Total CSA was 10% lower at the radius and 8.5% lower at the tibia in Chinese-American than in white women. C.Th was 18% and 16% greater in Chinese-American than in white women at the radius and tibia, respectively. Dcomp was 6% higher at the radius and 5% higher at the tibia in Chinese-American than in white women. There was no difference in Dtrab at the radius or tibia.

Table 3. vBMD and Microarchitecture by HR-pQCT in Chinese-American and White Women
 Chinese-American mean ± SDWhite mean ± SDUnadjusted p valuep Value adjusted for agep Value adjusted for age and BMI
 CSA (mm2)201 ± 33224 ±
 C.Th (mm)0.86 ± 0.140.73 ± 0.20.0004.03.003
 Dcomp (mg HA/ccm)916 ± 54862 ± 77.0002.02.005
 Dtrab (mg HA/ccm)118 ± 36125 ±
 Tb.N (1/mm)1.60 ± 0.341.73 ±
 Tb.Th (mm)0.061 ± 0.0130.060 ±
 Tb.Sp (mm)0.60 ± 0.170.59 ±
 CSA (mm2)613 ± 83670 ±
 C.Th (mm)0.99 ± 0.200.85 ±
 Dcomp (mg HA/ccm)828 ± 57789 ±
 Dtrab (mg HA/ccm)138 ± 29148 ±
 Tb.N (1/mm)1.5 ± 0.251.8 ± 0.33<.0001<.0001<.0001
 Tb.Th (mm)0.078 ± 0.0150.070 ±
 Tb.Sp (mm)0.61 ± 0.110.51 ± 0.11<.0001<.0001<.0001
Figure 1.

Percentage difference in HR-pQCT and FEA measurements at the (A) distal radius and (B) distal tibia in Chinese-American and white women. Asterisks denote significance of comparison (*p < .05).

There were several between-group site-specific differences in trabecular microarchitecture. Tb.N was 16% lower at the tibia and there was a trend toward lower Tb.N (7.5%, p = .08) at the radius in Chinese-American compared with white women. Tb.Sp was not different at the radius but was 20% greater at the tibia in Chinese-American than in white women. Tb.Th did not differ at the radius, whereas it was 11% greater in Chinese-American women at the tibia. Representative HR-pQCT images are shown in Fig. 2. As shown in Table 3, after adjustment for age or age and BMI, differences at the radius remained significant. At the tibia, the differences in CSA, Tb.N, Tb.Th, and Tb.Sp remained significant after adjusting for age as well as age and BMI, whereas the difference in C.Th was attenuated after adjusting for age alone.

Figure 2.

Smaller bone size and thicker cortex of (left) Chinese-American compared with (right) white women illustrated by representative 3D HR-pQCT images of the radius (top) and tibia (bottom): (A) Radius, Chinese-American; (B) radius, white; (C) tibia, Chinese-American; and (D) tibia, white.

As shown in Table 4 and Fig. 1, there was no difference in FEA-estimated trabecular or whole-bone stiffness at the radius or tibia before or after adjusting for age alone or age and BMI, but Chinese-American women had a greater percentage of load carried by the cortical bone compartment at the distal surface of the radius and tibia than white women that remained significant after adjustment for covariates.

Table 4. HR-pQCT-Based Finite Element Analysis in Chinese-American and White Women
 Chinese-American mean ± SDWhite mean ± SDUnadjusted p valuep Value adjusted for agep Value adjusted for age and BMI
 Whole-bone stiffness (kN/mm)71.6 ± 12.568.4 ±
 Trabecular bone stiffness (kN/mm)7.2 ± 6.98.7 ±
 Distal cortical load distribution (%)52 ± 1244 ± 9.0006.005.005
 Proximal cortical load distribution (%)93 ± 692 ±
 Whole-bone stiffness (kN/mm)207 ± 32206 ±
 Trabecular bone stiffness (kN/mm)75 ± 2682 ±
 Distal cortical load distribution (%)33 ± 827 ±
 Proximal cortical load distribution (%)72 ± 869 ±

In order to further assess the influence of differences in age between the groups on bone density, structure, and stiffness, women limited to ages 59 to 66 years (the age range containing 75% of participants) were compared between the racial groups. Results did not differ from those of the entire group (data not shown).


We found that postmenopausal Chinese-American women have smaller bone size than white women but thicker and denser cortices. At the tibia, Chinese-American women have thicker trabeculae than white women as well but fewer trabeculae, leading to greater trabecular spacing. It is unclear why there are different patterns at the radius versus the tibia, but they could be secondary to the effects of greater weight bearing at the tibia in white women (although results did not change with adjustment for BMI). These results generally are in agreement with our earlier report9 and that of Wang and colleagues10 in premenopausal women. In our prior study, however, premenopausal Chinese-American women not only had thicker, denser cortices, and thicker trabeculae than white women, but there also was no difference in trabecular number or spacing, leading to higher trabecular bone density in Chinese than in young white women. These data may suggest that the thicker, denser cortices observed in Chinese-American compared with white premenopausal women persist after menopause, whereas higher trabecular bone density does not. However, given the cross-sectional design and small sample sizes of these studies, we cannot rule out the possibility that cohort effects between our pre- and postmenopausal groups account, at least in part, for these differences. Because our study was not longitudinal, we did not formally compare age-related differences between pre- and postmenopausal white women with those in pre- and postmenopausal Chinese-American women. Prospective studies examining racial differences in age-related bone microarchitectural changes will be an interesting area of future investigation because our results could suggest that Chinese-American women may be better able to preserve bone microarchitecture with aging than white women.

FEA, which incorporates both trabecular and cortical elements of bone quality into an approximation of mechanical competence, was used to estimate whole-bone stiffness. Whole-bone FEA demonstrating similar mechanical competence despite these structural differences suggests that the thicker and denser cortical bone and thicker trabeculae in postmenopausal Chinese-American women compensate for fewer trabeculae and smaller bone area. The lack of difference in stiffness could translate into an advantage for Chinese women with regard to fracture risk. Lower weight and shorter height would be expected to produce less force during a fall from a standing height on a bone of similar stiffness, and therefore, Chinese women may be less likely to fracture. Owing to the strong correlation between stiffness at the proximal femur and distal tibia,25 these results may be relevant to the clinical observation that Chinese-American women have lower rates of hip fracture than white women. Perhaps even more pertinent to this discussion is the analysis of biomechanical load sharing between cortical and trabecular compartments because there is a growing body of evidence that cortical bone is the main determinant of bone strength in the femoral neck.26–28 We found that Chinese-American women have a higher percentage of load carried by cortical bone than white women at the proximal but not the distal radius and tibia. The greater load carried by cortical bone proximally is consistent with the greater amount of cortical bone present in Chinese women. The greater cortical density observed in Chinese women could be secondary to lower cortical porosity, greater mineralization, or both—both areas for future investigation. The reasons for differences in load sharing between the proximal and distal sites were not specifically addressed by our study but may be related to differences in the relative amounts of trabecular versus cortical bone at these sites.

Our study has several limitations. The cohort studied was a relatively small, convenience sample of women, and thus the results could have been influenced by selection bias. However, our earlier results for premenopausal women are similar to the larger cohort published in Australia.10 While we attempted to exclude women with conditions or taking medications known to influence bone quality both by history and biochemical evaluation, we cannot be certain that the observed differences or lack thereof are not due to the inclusion of some women with such factors. The groups differed in age, and it is possible that some of the observed differences could be due to this factor. The mean difference in age, however, was small (2 years), and we adjusted for this difference statistically. Further, we repeated analyses limiting the groups to the age range of overlap to help ensure that the observed differences were not due to a disparity in age.

Other limitations include those imposed by the methodology. Direct measurements of trabecular microstructure by HR-pQCT are limited because the resolution of the XtremeCT is at the upper end of the width of individual trabeculae. Therefore, trabecular structure is assessed using an algorithm with model assumptions,29, 30 and results for trabecular microarchitecture could differ if obtained by direct analysis of this index by bone biopsies. Additionally, as noted by Seeman and colleagues,10 shorter limb length in Chinese-American women and the fixed distance of HR-pQCT measurement from the reference line may mean that the region of interest in Chinese-American women is more proximal, thus containing proportionately more cortical bone than in white women. Since we did not measure limb length in our study groups, we cannot exclude the possibility that this could account in part for the differences observed in our study. However, Seeman and colleagues found similar results even when adjusting the site of measurement for differences in limb length.10

Our study did not investigate the environmental or genetic reasons for differences in bone size, density, and structure between the racial groups (other than anthropometric measures) and did not examine nonskeletal factors that might influence differences in fracture rates. Wang and colleagues suggested that earlier menarche (estrogen exposure) in Chinese versus white women might result in narrower and shorter bones, whereas less resorptive activity (as indicated by lower bone turnover markers) may lead to the thicker trabeculae observed in Chinese women.10 It will be important to conduct future studies examining the physiologic basis for the observed differences. Despite these limitations, this study provides the first data quantitating differences in vBMD, microarchitecture, and mechanical competence in postmenopausal white and Chinese-American women. They help to account for the relative protection from forearm and hip fracture that has been observed among Chinese-American women in comparison to white women.


All the authors state that they have no conflicts of interest.


MD Walker and XS Liu contributed equally to this work.

This work was supported by NIH Grants K23 AR053507, R01 AR051376, and AR058004, and UL1 RR024156, a National Osteoporosis Foundation grant, and the Mary and David Hoar Fellowship Program of the New York Community Trust and the New York Academy of Medicine. We are indebted to Dr Clyde Wu, whose vision and support were instrumental in the design and implementation of this study.