| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
INSERM Research Unit 403 (S.B., F.M., P.D.D.), and Claude Bernard University of Lyon, 69437 Lyon, France; and Orthopedic Biomechanics Laboratory (M.L.B.), Beth Israel Deaconess Medical Center, Boston, Massachusetts 02215
Address all correspondence and requests for reprints to: Prof. Pierre D. Delmas, INSERM Unit 403, Hôpital Edouard Herriot, Pavillon F, 69437 Lyon Cedex 03, France. E-mail: delmas{at}lyon.inserm.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective and Patients: We determined the short-term reproducibility of this device by measuring 15 healthy volunteers three times each. We compared HR-pQCT measurements in 108 healthy premenopausal, 113 postmenopausal osteopenic, and 35 postmenopausal osteoporotic women. Furthermore, we compared values in postmenopausal osteopenic women with (n = 35) and without previous fracture history (n = 78).
Design and Setting: We conducted a cross-sectional study in a private clinical research center.
Intervention and Main Outcome Measure: We took HR-pQCT measurements of the radius and tibia. Femoral neck and spine BMD were measured in postmenopausal women by dual-energy x-ray absorptiometry.
Results: Precision of HR-pQCT measurements was 0.7–1.5% for total, trabecular, and cortical densities and 2.5–4.4% for trabecular architecture. Postmenopausal women had lower density, trabecular number, and cortical thickness than premenopausal women (P < 0.001) at both radius and tibia. Osteoporotic women had lower density, cortical thickness, and increased trabecular separation than osteopenic women (P < 0.01) at both sites. Furthermore, although spine and hip BMD were similar, fractured osteopenic women had lower trabecular density and more heterogeneous trabecular distribution (P < 0.02) at the radius compared with unfractured osteopenic women.
Conclusion: HR-pQCT appears promising to assess bone density and microarchitecture at peripheral sites in terms of reproducibility and ability to detect age- and disease-related changes.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In vivo three-dimensional quantification of trabecular structure by either high-resolution peripheral quantitative computed tomography (HR-pQCT) or micro-magnetic resonance imaging (µMRI) was first introduced over 10 yr ago (7, 8, 9). In these initial imaging systems, the nominal isotropic resolution was 170 µm for HR-pQCT, and 80–200 µm in plane with a slice thickness of 400–700 µm for µMRI. Since then, technical advances have improved the spatial resolution of µMRI markedly, with recent studies achieving nominal resolutions of 137 x 137 µm in plane with a slice thickness of 350 µm (10). Additional advances have led to the introduction of a new HR-pQCT system that permits in vivo assessment of trabecular architecture and volumetric BMD at the distal radius and tibia with a nominal isotropic voxel size of 82 µm (11).
The overall aim of our study was to evaluate the clinical performance of this new HR-pQCT system. Specifically, we determined the short-term reproducibility of density and architecture measurements, assessed the correlation among the parameters, and compared density and architectural parameters from healthy premenopausal, osteopenic postmenopausal, and osteoporotic postmenopausal women. We also compared parameters in osteopenic postmenopausal women with and without a history of fracture.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
To assess short-term reproducibility, 15 healthy women (aged 21–47 yr) underwent three separate HR-pQCT scans of the distal radius and tibia within a 1-month period (12).
To determine differences in bone density and microarchitecture between healthy individuals and those with osteoporosis, we evaluated three groups of women: premenopausal, postmenopausal osteopenic, and postmenopausal osteoporotic. The premenopausal group comprised 108 healthy individuals (aged 19–45 yr) who had no conditions known to affect bone or mineral metabolism. The postmenopausal group comprised 148 individuals who had not been menstruating for at least 1 yr. Patients were excluded if they had known current metabolic disorders other than osteoporosis. Based on the lower of their BMD measurements at the lumbar spine or proximal femur, these women were classified as osteopenic (–1 < T-score < –2.5; n = 113 women, aged 52–88 yr) or osteoporotic (T-score 
–2.5; n = 35 women, aged 61–82 yr). Measurements at the wrist and at distal tibia were excluded in nine and six women, respectively, because of patient motion. This resulted in a final sample of 106 osteopenic women with HR-pQCT measurements at the distal radius, 109 osteopenic women with HR-pQCT scans at the distal tibia, and 33 osteoporotic women with HR-pQCT scans at the distal radius and at the distal tibia. All measurements from the premenopausal group were included in the final analyses. Measurements were usually performed at the nondominant limb unless there was a history of fracture on that limb, in which case the nonfractured limb was measured.
Personal history of fractures was recorded by a questionnaire. Because of the low number of subjects having had fractures in the premenopausal and postmenopausal osteoporotic groups, differences between individuals with and without a personal history of fracture after the age of 45 yr, excluding those of the face, skull, fingers, and toes, were analyzed in the postmenopausal osteopenic group only. Thus, among the 106 and 109 osteopenic women with valid HR-pQCT scans at the distal radius and distal tibia, respectively, 35 reported at least one previous fracture (44 fractures in total). Fracture types included wrist and distal forearm (n = 18), forearm (n = 5), humerus (n = 4), metacarpal (n = 1), femur (n = 1), patella (n = 1), lower leg (n = 5), ankle (n = 4), metatarsal (n = 3), and ribs (n = 2).
The protocol was approved by an independent Ethics Committee, and all patients gave written informed consent before participation.
Measurement of BMD and bone microarchitecture
BMD (g/cm2) at the lumbar spine and femoral neck was measured in postmenopausal women using DXA (Hologic QDR4500; Hologic, Bedford, MA).
Volumetric BMD and microarchitecture were measured at the distal radius and distal tibia using a three-dimensional HR-pQCT system (XtremeCT; Scanco Medical AG, Bassersdorf, Switzerland). This system uses a two-dimensional detector array in combination with a 0.08-mm point-focus x-ray tube, enabling the simultaneous acquisition of a stack of parallel CT slices with a nominal resolution (voxel size) of 82 µm. The following settings were used: effective energy of 60 kVp, x-ray tube current of 95 mA, and matrix size of 1536 x 1536.
At each skeletal site, 110 CT slices were obtained, thus delivering a three-dimensional representation of approximately 9 mm in the axial direction. The arm or leg of the patient was immobilized during the examination in an anatomically formed carbon fiber shell. An anteroposterior scout view was used to define the measurement region. Briefly, a reference line was manually placed at the endplate of the radius and tibia, as shown in Fig. 1A
. The first CT slice was 9.5 mm and 22.5 mm proximal to the reference line for the distal radius and distal tibia, respectively. The effective dose was less than 3 µSv per measurement (11) with a measurement time of 2.8 min. Attenuation data were converted to equivalent hydroxyapatite (HA) densities. Quality control, based on Shewart rules, was monitored by daily scans of a phantom containing rods of HA (densities of 0, 100, 200, 400, and 800 mg HA/cm3) embedded in a soft-tissue equivalent resin (QRM, Moehrendorf, Germany).
|
For follow-up measurements, an algorithm automatically uses the cross-sectional area (CSA, mm2) within the periosteal boundary of the radius or tibia to match the volumes of interest on the baseline and follow-up scans, and thus only the bone volume common to the previous scan(s) is used to assess density and microarchitectural measurements (19). Hence, of the initial 110 slices, on average 103 (range, 91–108) were analyzed in the follow-up scans.
The outcome variables used in our analyses included volumetric bone density (g HA/cm3) for entire (Dtot), trabecular (Dtrab), and cortical (Dcort) regions; cortical thickness (CTh, µm); and trabecular bone volume fraction (BV/TV, %), thickness (TbTh, µm), number (TbN*, mm–1), separation (TbSp, µm), and intra-individual distribution of separation (TbSp SD, µm).
Statistical analysis
For each subject in the reproducibility study, a coefficient of variation (CV) calculated as the SD of the three repeated measurements divided by the subject mean. Furthermore, the short-term precision errors were then calculated as root-mean-square (RMS) average of the precision errors for each of the subjects. The relationships between density and architecture were studied using Pearson or Spearman correlation, depending on the distribution of variables. The differences among premenopausal, postmenopausal osteopenic, and postmenopausal osteoporotic individuals were assessed by ANOVA or by a Kruskal-Wallis test (two-tailed tests), followed, when significant, by Student’s t test or Wilcoxon signed rank test for pair-wise comparisons (one-side tests). Finally, differences between the postmenopausal osteopenic women with and without fracture history were assessed by Student’s t test or Wilcoxon signed rank test (one-side tests). For the correlation between HR-pQCT parameters, the Bonferroni correction led us to consider statistical significance at P < 0.02. All statistical analyses were performed on SPSS software (version 12.0).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Reproducibility (RMS-CV) of density (total, trabecular, and cortical) measurements ranged from 0.7–1.5% (Table 1). In comparison, the reproducibility of structural parameters (trabecular number, thickness, separation, and distribution and cortical thickness) was slightly lower, with RMS-CVs ranging from 0.9–4.4% (Table 1). Measurement reproducibility was similar at the distal radius and distal tibia.
|
The relationships among age, density, and architecture parameters are shown in Table 2. We observed a significant age-related decline in total, trabecular, and cortical bone density at the distal radius (r = –0.50 to –0.67) and distal tibia (r = –0.46 to –0.79). Architectural parameters were also significantly correlated with age (|r| = 0.35–0.65 and |r| = 0.28–0.65, respectively, at the radius and tibia). When the analysis was restricted to the population-based sample of 108 premenopausal women, aged 19–45 yr, there was no significant correlation with age for any parameters (data not shown). As expected, total density was strongly correlated to both trabecular and cortical density (r = 0.80–0.83 for both). In contrast, trabecular and cortical densities were only modestly related to each other. Trabecular density was strongly correlated to trabecular architectural measurements at both the distal radius (|r| = 0.78–0.91) and tibia (|r| = 0.71–0.75). Cortical density was highly correlated with cortical thickness but weakly correlated with trabecular architecture. Height and weight were moderately correlated with the cross-sectional bone area at both sites but were only poorly or not significantly correlated with the other parameters.
|
Mean values and statistical comparisons among premenopausal, postmenopausal osteopenic, and postmenopausal osteoporotic women are shown in Table 3, and representative three-dimensional images in Fig. 1
, B–E. The results of ANOVA and Kruskal-Wallis test showed differences between the three groups for all HR-pQCT parameters except for mean cross-sectional area (P < 0.001). Postmenopausal osteopenic and osteoporotic women did not differ in age, height, or weight. At both the radius and tibia, all parameters, except for mean cross-sectional area, were significantly different between premenopausal and postmenopausal women. Specifically, trabecular and cortical density were 19–37% and 11–18%, respectively, lower in postmenopausal than premenopausal women (P < 0.001 for all). Postmenopausal women also had lower trabecular number (–12 to –23%; P < 0.001) and thickness (–8 to –19%; P < 0.001), decreased cortical thickness (–28 to –41%; P < 0.001), and increased trabecular separation (21–38%; P < 0.001) and intra-individual distribution of separation (53–83%, P < 0.001) compared with premenopausal women.
|
). Compared with those classified as osteopenic, osteoporotic women had lower bone density (–5 to –18%), decreased trabecular number (–10% at the tibia) and thickness (–10% at the radius), increased trabecular separation (+9 to +12%) and intra-individual distribution of separation (+19% at the tibia), and decreased cortical thickness (–15 to –18%) (P < 0.02 for all). Discrimination of osteopenic women with and without history of fracture
Osteopenic women with and without a history of fracture did not differ with regard to age, height, weight, years since menopause, BMD of the lumbar spine and femoral neck, and HR-pQCT measurements at the distal tibia (Table 4). In contrast, at the distal radius, total density (–10.0%; P = 0.015), cortical density (–3.4%; P = 0.042), trabecular density (–12.3%; P = 0.018), and trabecular number (–8.5%; P = 0.037) were significantly lower and trabecular separation (+12.8%; P = 0.027) and intra-individual distribution of separation (+25.6%; P = 0.011) were significantly higher in women with a history of fracture compared with those with no previous fracture (Fig. 2). After the Bonferroni correction, differences were borderline or no more significant for the following parameters: cortical density, trabecular number, and trabecular separation.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
We found that the short-term reproducibility of HR-pQCT parameters was quite similar at the distal radius and distal tibia, with CVs less than 1.5% for bone density and from 2.5–4.4% for trabecular architecture. It should be noted that our precision data were generated in healthy individuals, and the values may be higher among osteoporotic subjects. Nonetheless, the reproducibility for density measurements by HR-pQCT is similar to that reported for distal and ultradistal radius BMD by DXA (20, 21). The reproducibility of trabecular architecture measurements is also similar to or better than previous studies using a lower-resolution in vivo HR-pQCT system (19, 22) or high-resolution MRI (23, 24). Reproducibility of trabecular structure measurement by MRI has been reported to be 2–9%, depending on the specific parameter and analysis algorithm (23, 25, 26). Patient movement and difficulty in matching the analysis volume in repeat studies have been cited as primary error sources for follow-up measurements (23). Patient movement was also a concern in our study, because nine women (3.5%) at the distal radius and six women (2.3%) at the distal tibia were eliminated because of obvious movement artifacts that rendered the scan unsuitable for analysis. Given that several studies have shown that MRI is capable of detecting changes in trabecular structure caused by disease or treatment (24, 26, 27, 28, 29, 30), our findings of comparable or better reproducibility in HR-pQCT measurements are encouraging in this regard.
We observed significant age-related changes in density, trabecular structure, and cortical thickness, as evidenced both by correlations with age (Table 2) and by comparing pre- with postmenopausal women (Table 3). At both the distal radius and distal tibia, the correlations between density and age were similar to those recently reported in a large population-based study with a broader age range than our subjects (31). We confirm a number of previous in vitro and in vivo studies that have reported the decline in trabecular bone density (i.e. BV/TV) to be accompanied by declines in trabecular number and thickness (32, 33, 34). In our study, the decline in trabecular number was slightly more pronounced than the decline in thickness. Possible explanations for this observation include the possibility that thinner trabeculae are resorbed first, thereby increasing the mean thickness of those remaining and/or by the possibility that the trabeculae that remain adapt to the increased mechanical demand by increasing their size. Similar findings have been noted at the calcaneus, where the rate of age-related declines in trabecular number was reported to be approximately twice that of the decline in trabecular thickness (27). We also observed the expected age-related declines in cortical thickness. Thus, the HR-pQCT measurements appear useful for gaining insight into structural mechanisms underlying various causes of and treatments for skeletal fragility, because they are capable of assessing architecture and density changes in both the cortical and trabecular bone compartments. For example, this may be particularly useful in delineating the subtle compartment-specific effects of mild hyperparathyroidism or intermittent PTH administration (35). Moreover, by assessing individual bone compartments with excellent precision, HR-pQCT may improve the ability to monitor the response to antiresorptive therapy, a critical clinical issue (4).
An interesting finding of our study was the ability of HR-pQCT measurements of trabecular bone density and intra-individual distribution of separation at the distal radius to discriminate between osteopenic women with and without a history of fracture, whereas spine and hip BMD measurements did not (Table 4). Admittedly, these findings must be interpreted with caution because our study had a relatively small sample size with various types of fractures, and thus, our results need to be confirmed in a larger prospective study. Whereas several previous studies have shown that in vivo measurements of trabecular structure can be used to identify osteoporotic women with and without fractures and that trabecular structure may improve predictions of fracture risk based on BMD alone, generally the fracture cases in these studies also have lower BMD than controls (34, 36, 37, 38). An exception to this is the study by Link et al. (39) who demonstrated that MRI-based measurements of trabecular structure differed among cardiac transplant patients with and without vertebral fractures, whereas BMD did not. Moreover, Laib et al. (18) have reported that the distribution of trabecular separation (TbSp SD) was the MRI index that was most sensitive for discrimination of individuals with and without spine fractures. Placed in the context of current clinical dilemmas, our preliminary results are intriguing, because over half of osteoporotic fractures occur in women with osteopenia by WHO criteria, and identification of factors that predispose these osteopenic women to fracture is needed (1, 3).
Regarding the personal history of fracture, there were 2-fold more fractures at the arm than at the leg, which may explain why only measurements at the distal radius were able to discriminate among fractured vs. unfractured subjects. Moreover, as the radius is not weight bearing, one may speculate that there are fewer mechanical stimuli to maintain structural integrity in the face of estrogen deficiency, and perhaps trabecular architectural changes are more rapid and/or more dramatic at the radius than the distal tibia. Additional studies are needed to better understand the utility of each of these skeletal sites with regard to fracture risk prediction.
It should be noted that our study design had several limitations. First, the study was cross-sectional, and therefore the ability of the HR-pQCT measurements to predict fracture risk prospectively and to monitor age-, disease-, and treatment-related changes in bone density or architecture requires additional validation. Second, as mentioned previously, the sample size was relatively small; the premenopausal group comprised 108 individuals, thereby precluding establishment of normal reference values. Moreover, we did not have the ideal population-based sample with equal representation across the lifespan to accurately assess age-related patterns.
There are several technical limitations that should be considered as well. We noted, as have others, moderate to strong correlations between trabecular density and trabecular architecture. These observations must be viewed carefully because in the current study, trabecular density and trabecular number are independently measured, whereas BV/TV, trabecular thickness, and trabecular separation are derived. Additional studies are required to discern the combination of density and architectural features that will be most useful clinically. Moreover, we found that cortical thickness and cortical density were very strongly correlated (Table 2). This high correlation suggests that partial volume effects limit the reliability of cortical density measurements. In support of this view, we note that cortical thickness and density were more strongly correlated at the distal radius (r2 = 0.90), where average cortical thickness ranged from 190-1170 µm, than they were at the distal tibia (r2 = 0.76), where average cortical thickness is higher (280–1790 µm). Removal of cortical surface voxels before computation of cortical density may improve the reliability of the cortical density measurements, which at present should be viewed with caution. Even if the reliability is improved, cortical density measurements will still reflect the combined influence of porosity and the degree of matrix mineralization, because a resolution on the order of 10 µm is required to directly assess cortical porosity in humans (40).
Despite these limitations, the current study has provided new information regarding the potential clinical utility of HR-pQCT measurements with 82-µm isotropic voxel size. Additional studies are required to confirm and extend these preliminary results. Altogether, however, our findings indicate that HR-pQCT appears promising as a technique to assess bone density and trabecular microarchitecture at peripheral skeletal sites, both in terms of reproducibility and ability to detect age- and disease-related changes.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online September 27, 2005
Abbreviations: BMD, Bone mineral density; BV, bone volume; CSA, cross-sectional area; µCT, microcomputed tomography; CTh, cortical thickness; CV, coefficient(s) of variation; D, volumetric bone density; DXA, dual-energy x-ray absorptiometry; HA, hydroxyapatite; HR-pQCT, high-resolution peripheral quantitative computed tomography; µMRI, micro-magnetic resonance imaging; RMS, root mean square; TbN*, trabecular number; TbSp, trabecular separation; TBSp SD, intra-individual distribution of separation; TbTh, trabecular thickness; TV, trabecular volume.
Received June 3, 2005.
Accepted September 15, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  T. Chevalley, J.-P. Bonjour, S. Ferrari, and R. Rizzoli Influence of Age at Menarche on Forearm Bone Microstructure in Healthy Young Women J. Clin. Endocrinol. Metab., July 1, 2008; 93(7): 2594 - 2601. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. K. Genant, K. Engelke, and S. Prevrhal Advanced CT bone imaging in osteoporosis Rheumatology, July 1, 2008; 47(suppl_4): iv9 - iv16. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. B. Kalpakcioglu, S. Morshed, K. Engelke, and H. K. Genant Advanced Imaging of Bone Macrostructure and Microstructure in Bone Fragility and Fracture Repair J. Bone Joint Surg. Am., February 1, 2008; 90(Supplement_1): 68 - 78. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Lee, V. Gilsanz, and T. A. L. Wren Limitations of Peripheral Quantitative Computed Tomography Metaphyseal Bone Density Measurements J. Clin. Endocrinol. Metab., November 1, 2007; 92(11): 4248 - 4253. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Lentle and D. Worsley Osteoporosis Redux J. Nucl. Med., December 1, 2006; 47(12): 1945 - 1959. [Full Text] [PDF]
|
|
|
|
 |
|
 E. Seeman and P. D. Delmas Bone quality--the material and structural basis of bone strength and fragility. N. Engl. J. Med., May 25, 2006; 354(21): 2250 - 2261. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
