Inverse Correlation at the Hip Between Areal Bone Mineral Density Measured by Dual-Energy X-ray Absorptiometry and Cortical Volumetric Bone Mineral Density Measured by Quantitative Computed Tomography

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health, vol. -, no. -, 1e8, 2015 Ó Copyright 2015 by The International Society for Clinical Densitometry 1094-6950/-:1e8/$36.00 http://dx.doi.org/10.1016/j.jocd.2015.01.002

Original Article

Inverse Correlation at the Hip Between Areal Bone Mineral Density Measured by Dual-Energy X-ray Absorptiometry and Cortical Volumetric Bone Mineral Density Measured by Quantitative Computed Tomography Anne Kristine Amstrup,* Niels Frederik Breum Jakobsen, Søren Lomholt, Tanja Sikjaer, Leif Mosekilde, and Lars Rejnmark Osteoporosis Clinic, Department of Endocrinology and Internal Medicine (MEA), THG Aarhus University Hospital, Aarhus, Denmark

Abstract Quantitative computed tomography (QCT) is considered to measure true volumetric bone mineral density (vBMD; mg/cm3) and enables differentiation between cortical and trabecular bone. We aimed to determine the value of QCT by correlating areal BMD (aBMD) by dual-energy X-ray absorptiometry (DXA) with vBMD when using a fixed threshold to delineate cortical from trabecular bone. In a cross-sectional study, 98 postmenopausal women had their hip scanned by DXA and by QCT. At the total hip and the trabecular bone compartment, aBMD correlated significantly with vBMD (r 5 0.74 and r 5 0.63; p ! 0.01, respectively). A significant inverse correlation was found between aBMD and cortical vBMD (r 5 0.57; p ! 0.01). Total hip volume by QCT did not change with aBMD. However, increased aBMD was associated with a decreased trabecular bone volume (r 5 0.36; p ! 0.01) and an increased cortical volume (r 5 0.69; p ! 0.01). Changing the threshold used to delineate cortical from trabecular bone from default 350 mg/cm3 to either 300 or 400 mg/cm3 did not affect integral vBMD ( p 5 89) but had marked effects on estimated vBMD at the cortical ( p ! 0.001) and trabecular compartments ( p ! 0.001). Furthermore, increasing the threshold decreased cortical thickness ( p ! 0.001), whereas the strength parameter in terms of buckling ratio increased ( p ! 0.001). Our results show good agreement between aBMD and integral vBMD. However, using a fixed threshold to differentiate cortical from trabecular bone causes an apparent increase in cortical volume with a decrease in cortical density as aBMD increases. This may be caused by the classification of a larger part of the transition zone as cortical bone with increased aBMD. Key Words: QCT; DXA; vBMD; aBMD.

Organization’s definition of osteoporosis based on T-scores, assessed by dual-energy X-ray absorptiometry (DXA) scans, has been the gold standard in diagnosing osteoporosis (1). However, the DXA techniques have limitations as they may overestimate bone mineral density (BMD) in overweight people (2), patients with osteoarthritis (3), scoliosis, aortic calcification, and vertebral fractures (4) and underestimate BMD in individuals with relatively small bones (5). Furthermore, DXA images are 2-dimensional (2D) assessments of the areal BMD (aBMD; g/cm2), which do not account for alterations in microstructure and geometry or the relative distribution of

Introduction Osteoporosis is a disease resulting from a reduced bone mass with altered bone geometry and microstructure leading to an increased risk of fractures. Since 1994, the World Health Received 10/29/14; Revised 01/06/15; Accepted 01/15/15. *Address correspondence to: Anne Kristine Amstrup, MD, Osteoporosis Clinic, Department of Endocrinology and Internal Medicine (MEA), Aarhus University Hospital, Tage-Hansens Gade 2, 8000 Aarhus C, Denmark. E-mail: [email protected]

1

2 cortical and trabecular bone. For this matter, 3-dimensional (3D) imaging by quantitative computed tomography (QCT) may serve as a better alternative as QCT scans are considered to provide a measure of true volumetric BMD (vBMD; mg/ cm3) including the ability to distinguish between cortical and trabecular bone(6,7). Today, there is an increasing interest in the use of 3D CTbased techniques in relation to osteoporosis (8,9). Whether 3D measuring techniques can improve fracture prediction compared with DXA has not been fully evaluated, and the replacement of DXA scan with QCT in diagnosing osteoporosis in daily practice is still challenging as the amount of radiation by QCT far exceeds that of the DXA scan (10). The advantages, however, of a more differentiated measurement outcome provided by QCT scans may give us more information on indices of importance to fracture risk. Furthermore, it may supply more detailed information on effects of diseases or treatments on changes in bone mineral distribution. Despite the fact that clinical studies more frequently use images from both DXA and QCT in their evaluation of bone and bone changes, little information exists regarding the correlation between these 2 techniques. If QCT images shall serve as a true supplement, or even a replacement, to DXA, it is of interest to examine the association between the mentioned indices. Therefore, in the present study, we aimed to investigate the extent to which aBMD at the hip, as assessed by 2dimensional DXA scans, correlates with total, trabecular, and cortical vBMD as assessed by 3D QCT scanning technique and whether the indices are affected by the cortical threshold used to separate cortical and trabecular bone.

Methods Study Population We studied 98 postmenopausal women aged 63 years (range: 56e76 years) who had been recruited from the general background population. Eighty-one healthy women were diagnosed with osteopenia as they had been screened by DXA and included in an ongoing randomized trial (NCT01690000). Data on the women derive from baseline before any study-related action was taken. Seventeen of the included women had been recruited as healthy controls in another study and did not have DXA scans performed before their inclusion, that is, they were not selected or included because of a known low bone mass (11,12). Exclusion criteria for all participants were impaired renal function (plasma creatinine O120 mmol/L), hypercalcemia (plasma ion O1.32 mmol/L), intestinal malabsorption, impaired liver function, medical conditions known to affect bone, including the use of drugs with effects on calcium homeostasis and bone metabolism such as antiresorptives, and bone anabolic agents as well as diuretics and lithium. None of our participants were on treatment with experimental drugs at the time of investigations. All patients had provided an informed consent before conducting the studies. Both studies are approved by the

Amstrup et al. regional ethic committee of Denmark (#M-2010-0296; M#2012-252-12).

Osteodensitometry by DXA We measured aBMD (g/cm2) on the left hip region using Hologic Discovery scanners (Hologic, Inc., Waltham, MA). All scanners were daily crosscalibrated with a reference phantom to read BMD.

Osteodensitometry by QCT We measured vBMD (mg/cm3) at the left hip by QCT using a Philips Brilliance 40 slices multidetector helical CT scanner (Phillips, Eindhoven, the Netherlands). We scanned with a dose modulation tool (Z-DOM; Phillips) at a voltage of 120 kV. Slice thickness and slice spacing was 3 mm. Field of view was 400 mm and collimation was 40  0.625 mm at the hip. The vBMD was determined using CTXA Hip Exam Analysis protocol, QCTPro (version 4.2.3; Mindways Software, Inc., Austin, TX) in conjunction with a solid-state CT calibration phantom (Model 3; Mindways Software), which was scanned simultaneously with the patients. We performed analysis of the left proximal femur by automatic bone segmentation including the total hip and femoral neck (13). In addition to densitometric measures, we also used the Mindways software to estimate cortical thickness and bone strength as assessed by buckling ratio (BR) at the femoral neck. For initial analyses, the separation algorithm for cortical bone was set at default 350 mg/cm3. In addition, in a subgroup of 50 randomly selected study patients, we studied the correlations between aBMD and vBMD by changing the threshold, delineating cortical from trabecular bone, to 300 and 400 mg/cm3. The reproducibility (coefficient of variation [%]) of the analyses by QCTPro was calculated by repeating evaluation analyses of 10 patients’ data showing a total hip vBMD coefficient of variation of 0.8%.

Statistical Analysis We report results as mean  standard deviation or median with interquartile range (IQR: 25%e75%) unless otherwise stated. Groups were compared using a 2-sample t test or 1way analyses of variance. Associations between variables were assessed by bivariate correlations calculating Pearson’s correlation coefficient (r). Furthermore, we studied associations by using linear regression analysis adjusting for body mass index (BMI) and age. Results from these analyses are reported as unstandardized regression coefficient B (b) with 95% confidence interval (CI). p ! 0.05 Was considered statistically significant. We used IBM SPSS Statistics, version 21 (IBM, New York) for the statistical analyses.

Results Descriptive data are listed in Table 1. Mean age of the 98 participants was 63 years (range: 56e76 years). Associations between studied indices at the total hip are shown in Fig. 1.

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

DXA and QCT Correlation

3

Table 1 Descriptive Data for the 98 Participants Characteristics Age (yr) Height (cm) Weight (kg) BMI (kg/m2) Smokers, n (%) aBMD by DXA, n 5 98 Hip, total (g/cm2) Hip, neck (g/cm2) vBMD by QCT, n 5 98 Total hip, integral vBMD (mg/cm3) Total hip, cortical vBMD (mg/cm3) Total hip, trabecular vBMD (mg/cm3) Femoral neck, integral vBMD (mg/cm3) Femoral, cortical vBMD (mg/cm3) Femoral, trabecular vBMD (mg/cm3)

Median (IQR) 63 165.6 66.6 24.5 3

(56e76) (161.8e170.1) (60.3e76.0) (22.0e26.9) (3)

0.786 (0.734e0.830) 0.645 (0.616e0.700)

positively to aBMD (b 5 34; 95% CI: 26e41; p ! 0.01. When adjusting for BMI and age, total hip volume by QCT was inversely correlated to aBMD (b 5 46.6; 95% CI: 82 to 10.8; p ! 0.05; Table 2). At the femoral neck, total volume (by QCT) did not change with aBMD (b 5 4.0; 95% CI: 10.0 to 2.0; p O 0.05), but an increased aBMD was associated with an increase in cortical volume (b 5 6.5; 95% CI: 5.1e7.9; p ! 0.01) and a decrease in trabecular volume (b 5 10.5; 95% CI: 16.7 to 4.3; p ! 0.01; Fig. 1 [lower panel]).

The Effect of Height, BMI, and Age

993.8 (906.7e1094.0)

Table 3 summarizes the effect of height, BMI, and age on the bone parameters. No significant correlation was observed for age. BMI was positively associated with aBMD at the total hip, whereas no associations were found between BMI and vBMD. However, BMI correlated weakly to integral and trabecular volume at the total hip and femoral neck. Bone size (as estimated by body height) correlated positively with bone volume and inversely with vBMD at the integral and trabecular bone sites at the total hip and femoral neck. Moreover, BR increased significantly with BMI and borderline significantly with body height ( p 5 0.06; data not shown).

142.4 (127.1e153.4)

Cortical Thresholds

253.2 (233.6e282.1) 909.2 (877.7e938.2) 132.4 (120.4e141.6) 259.7 (230.8e288.3)

Note: Median with 25%e75% interquartile range, except for age, where total range is used. Abbr: aBMD, areal bone mineral density; BMI, body mass index; DXA, dual-energy X-ray absorptiometry; IQR, interquartile range; QCT, quantitative computed tomography; vBMD, volumetric bone mineral density.

vBMD vs Total Hip aBMD The strongest correlation was positive and seen between vBMD of the total hip (integral vBMD) and total hip aBMD (r 5 0.74; b, 355.1; 95% CI: 280.0e421.1; p ! 0.01). Similarly, trabecular vBMD correlated positively with aBMD of the total hip (r 5 0.63; b, 132.9; 95% CI: 100.0e165.9; p ! 0.01). In contrast, cortical vBMD correlated inversely with aBMD at the total hip (r 5 0.57; b, 567.0; 95% CI: 731.3 to 402.7; p ! 0.01; Fig. 1 [upper panel]). At the femoral neck, aBMD by DXA also showed a significantly inverse relation to cortical hip vBMD (r 5 0.49; b 5 1358.3; 95% CI: 1843.4 to 873.1; p ! 0.01) along with a positive correlation to trabecular vBMD (r 5 0.56; b 5 119.1; 95% CI: 72.0e166.1; p ! 0.01) and integral vBMD (r 5 0.69; b 5 426.1; 95% CI: 335.0e517.2; p ! 0.01; data not shown). Adjusting for BMI and age did not change the relationships of the aforementioned parameters (Table 2).

Hip Volume vs aBMD As shown in Fig. 1, total hip volume by QCT did not change as a function of aBMD; however, trabecular volume was negatively correlated to aBMD (b 5 66; 95% CI: 100 to 31; p ! 0.01), whereas cortical volume correlated

The effects of changing the threshold for delineating cortical from trabecular bone from the default value of 350 mg/cm3 to 300 mg/cm3 and 400 mg/cm3 are summarized in Table 4. Changing the threshold did not affect integral vBMD or volume at the total hip or femoral neck. However, a marked effect was evident on the cortical and trabecular bone compartments. Mean vBMD increased significantly at both cortical and trabecular bone with increased threshold ( p ! 0.001). Concomitantly, cortical volume decreased, whereas trabecular volume increased with increasing threshold. Furthermore, with increased threshold, average cortical thickness decreased significantly from 0.27  0.1 at 300 mg/cm3 to 0.19  0.1 at 350 mg/cm3 and to 0.16  0.0 at 400 mg/cm3. This was not changed by adjustment for age, and within these analyses, age was not an independent determinant of cortical thickness at the femoral neck ( p 5 0.92; data not shown). Analyses on changes in vBMD as a function of changes in aBMD at different cortical thresholds are shown in Fig. 2 indicating that the inverse association increased with an increase in the threshold value used to differentiate cortical from trabecular bone (r 5 0.23, not significant, at 300 mg/cm3, r 5 0.57; p ! 0.01 at 350 mg/cm3, and r 5 0.64; p ! 0.01 at 400 mg/cm3). On the other hand, no major effects of threshold values were found on the correlations between aBMD and integral (r 5 0.71e0.74) or trabecular vBMD (r 5 0.63e0.69). Similar results were found after adjustment for age and BMI (not shown). Correlations between hip volumes by QCT and aBMD did not change according to threshold as integral hip varied from r 5 0.05 to 0.18 (NS), cortical bone r 5 0.60e0.69 ( p ! 0.01), and trabecular bone r 5 0.20 to 0.36 ( p ! 0.05; Fig. 3).

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

4

Amstrup et al.

Fig. 1. Regression analyses between vBMD by QCT and aBMD by DXA. Regression coefficient (b) and correlation coefficient (r) are shown on each figure. aBMD, areal bone mineral density; DXA, dual-energy X-ray absorptiometry; QCT, quantitative computed tomography; vBMD, volumetric bone mineral density.

Strength Parameter

Discussion

A significant inverse association was present between BR and aBMD at the total hip (r 5 0.44; p ! 0.01) and femoral neck (r 5 0.53; p ! 0.01; data not shown). However, estimated BR was highly dependent on threshold as BR increased with increased threshold values (Table 4).

In this study, we investigated the value of QCT by comparing aBMD and vBMD as measured by DXA and QCT, respectively, at the total hip and femoral neck in 98 postmenopausal women. QCT measures of integral and

Table 2 Multivariate Regression Analyses Between aBMD by DXA and vBMD and Volume by QCT When Adjusted for BMI and Age aBMD by DXA aBMD hip, total QCT vBMD Total hip, integral Total hip, cortical Total hip, trabecular Femoral neck, integral Femoral neck, cortical Femoral neck, trabecular Volume Total hip, integral Total hip, cortical Total hip, trabecular Femoral neck, integral Femoral neck, cortical Femoral neck, trabecular

b

r 0.79** 0.60** 0.67** 0.72** 0.59** 0.49**

387.9** 587.6** 140.0** 400.4** 1436.9** 108.0**

0.37** 0.70** 0.52** 0.33* 0.69** 0.48**

46.6* 35.2** 81.8** 6.1* 6.0** 12.1**

aBMD femoral neck b

r

(325.5 to 450.3) ( 755.0 to 420.2) (106.9 to 173.1) (318.4 to 482.4) ( 1862.0 to 1011.8) (64.6 to 151.5)

0.65** 0.41** 0.58** 0.74** 0.54** 0.50**

344.2** 417.0** 130.8** 446.9** 1440.4** 122.8**

( 82.5 to 10.8) (27.8 to 42.5) ( 114.5 to 49.1) ( 11.5 to 0.7) (4.7 to 7.3) ( 17.5 to 6.7)

0.30* 0.60** 0.38** 0.30* 0.72** 0.45**

21.5 32.7** 54.2** 5.0 6.8** 11.8

(260.0 to 428.0) ( 623.9 to 210.4) (91.6 to 170.1) (360.6 to 533.3) ( 1917.6 to 963.2) (76.2 to 169.3) ( 61.5 to 18.5) (23.7 to 41.7) ( 92.5 to 15.9) ( 11.0 to 0.9) (5.4 to 8.1) ( 17.8 to 5.8)

Note: Correlation coefficients (r) and regression coefficients (b) with 95% confidence interval. Abbr: aBMD, areal bone mineral density; DXA, dual-energy X-ray absorptiometry; QCT, quantitative computed tomography. *p ! 0.05; **p ! 0.01. Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

DXA and QCT Correlation

5

Table 3 Partial Correlation Coefficients (r) by Pearson Between Height, Body Mass Index (BMI), and Age and Bone Parameters as Assessed by QCT (vBMD, Cortical Thickness, Buckling Ratio, and Volume), and by DXA (aBMD) Parameter vBMD Total hip, integral Total hip, cortical Total hip, trabecular Femoral neck, integral Femoral neck, cortical Femoral neck, trabecular Cortical thickness, cm Buckling ratio Volume Total hip, integral Total hip, cortical Total hip, trabecular Femoral neck, integral Femoral neck, cortical Femoral neck, trabecular aBMD Total hip Femoral neck

Height

BMI

Age

0.34** 0.17 0.33** 0.27**

0.08 0.10 0.07 0.15

0.07 0.11 0.14 0.12

0.13

0.15

0.07

0.33**

0.02

0.19

0.12 0.21

0.20 0.24*

0.06 0.14

0.43** 0.04 0.44** 0.36**

0.26** 0.66 0.25* 0.25*

0.11 0.03 0.12 0.01

0.03

0.11

0.04

0.35**

0.26**

0.05

0.19 0.07

0.22* 0.13

0.09 0.03

Abbr: aBMD, areal bone mineral density; DXA, dual-energy Xray absorptiometry; QCT, quantitative computed tomography; vBMD, volumetric bone mineral density. *p ! 0.05; **p ! 0.01.

trabecular vBMD correlated positively and significantly to aBMD. However, cortical vBMD as assessed by QCT showed a significant inverse relation to aBMD. Total hip volume by QCT did not change with aBMD, but an increase in aBMD was associated with a decreased trabecular bone volume and an increased cortical volume. The associations were independent of cortical threshold. Our findings suggest that QCT of the hip is an appropriate measure of integral and trabecular vBMD. This is in accordance with previous studies as Liu et al (14) demonstrated an association similar to ours between integral vBMD and total hip aBMD (r 5 77 vs r 5 74 in our study) in a study with 69 premenopausal women. Furthermore, in a study by Cohen et al (15), investigating the association between aBMD and vBMD at the total hip and femoral neck, showed significant correlation coefficients up to r 5 84. Other studies have described even higher correlations at the hip varying from

r 5 0.82 to r 5 0.97 (16e18) depending on the study population. Although correlations between total hip aBMD and integral vBMD are described, little data have been published on the correlations of trabecular and cortical vBMD. Cohen et al (15) demonstrated a correlation between femoral neck aBMD and trabecular vBMD higher than ours (r 5 0.75e0.85 vs r 5 0.56). Their correlations, however, were only significant in cases (women with osteoporosis or low-trauma fracture) as no significant correlations were observed among the normal controls. Furthermore, the study showed weak nonsignificant positive correlation to cortical vBMD at the femoral neck. This finding is in contrast to our results as we demonstrated a significant inverse correlation between cortical vBMD and aBMD at the total hip and femoral neck. We believe that the most likely explanation for our findings is that the density in the transition zone between trabecular and cortical bone increases with increased aBMD. If a fixed threshold is used to separate cortical from trabecular bone, an increase in bone density results in a relatively larger part of the transition zone being identified as cortical bone and accordingly an increase in cortical volume with a concomitantly lower trabecular bone volume, resulting in an apparently reduced cortical vBMD (Fig. 2). This is further supported by our measures at varying thresholds as a higher cortical threshold is more strongly correlated (inversely) to aBMD owing to the lesser amount of the transition zone being included. In addition, cortical vBMD shows a higher density by increasing threshold along with decreased volume. Consistent with these results, trabecular compartment proofed to increase in density along with increasing volume. The separation between cortical and trabecular bone is considered to be one of the greatest advantages to QCT compared with DXA. These advantages are, however, dependent on several factors including scanner type, operating personnel, and scanner software. Segmentation of cortical bone in general is an ongoing problem for CT techniques (19e22). Furthermore, Engelke et al (23) described how the spatial resolutions of the CT images are often less than the cortical thickness leading to inaccuracy of, for example, density and thickness due to large segmentations. The consequence may be increased cortical volume of interest owing to the included subcortical bone adjacent to the area of interest. BR, a parameter of strength by QCT, is related to cortical instability. It reflects the ability of the bone to resist bending forces at tensile surfaces, and a higher BR is associated with a higher risk of fractures. In agreement with this concept, we demonstrated an inverse correlation to aBMD as a high aBMD is associated with a reduced fracture risk. Furthermore, our data showed that the higher the cortical threshold, the lower the BR appeared as the bone became denser. In accordance with the well-known effect of body size on aBMD (24e26), our analyses showed an increase in aBMD with increased BMI. We also analyzed the associations between body height and bone volume and vBMD. With

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

6

Amstrup et al. Table 4 Effects of Changing the Fixed Thresholds Used to Delineate Cortical From Trabecular Bone (Mean  SD)

QCT Hip vBMD by QCT, mg/cm3 Total hip Integral Cortical Trabecular Femoral neck Integral Cortical Trabecular Cortical thickness, cm Bulking ratio Hip volume by QCT, cm3 Total hip Integral Cortical Trabecular Femoral neck Integral Cortical Trabecular

300 mg/cm3

350 mg/cm3

400 mg/cm3

p value

258  36 773  31 119  15

255  35 916  44 134  15

258  36 1127  93 148  17

0.89 !0.001 !0.001

274 830 120 0.27 6.4

    

47 62 16 0.1 1.5

260 1016 141 0.19 9.1

    

44 138 17 0.1 2.5

269 1384 154 0.16 11.1

    

46 344 20 0.0 3.2

0.29 !0.001 !0.001 !0.001 !0.001

89  13 19  4 70  12

92  14 14  4 78  13

89  14 10  3 79  13

0.29 !0.001 0.001

12  2 2.7  0.6 10  2

12  2 1.8  0.6 11  2

12  2 1.3  0.6 11  2

0.28 !0.001 0.01

Abbr: QCT, quantitative computed tomography; vBMD, volumetric bone mineral density.

increased body height, bone size is expected to increase, which was supported by our analyses, showing a positive correlation between bone volume and height. A priori, we

expected vBMD to be unaffected by bone size. However, in contrast to our expectations, our analyses showed an inverse association between height and vBMD. These findings do,

Fig. 2. Linear regression correlations between vBMD, volume, and aBMD at the hip depending on cortical threshold (V, 300 mg/cm3; C, 350 mg/cm3; , 400 mg/cm3). aBMD, areal bone mineral density; DXA, dual-energy X-ray absorptiometry; QCT, quantitative computed tomography; vBMD, volumetric bone mineral density. Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

DXA and QCT Correlation

7 In conclusion, there is a good and positive agreement between integral and trabecular vBMD and aBMD at the hip. Cortical vBMD, however, is negatively associated to aBMD. If a fixed threshold value is used to differentiate cortical bone from trabecular bone by QCT scans, the density of cortical bone is underestimated as a part of the transition zone may be classified as cortical bone. Independent of the chosen threshold, this causes an apparent decrease in cortical density as overall bone density increases.

References Fig. 3. Illustration of the influence of areal bone mineral density (aBMD) on separation between cortical and trabecular bone using a fixed cutoff level. Cortical, transitional, and trabecular bone is marked on the abscissa illustrating a cross-sectional bone. aBMD is shown at the ordinate. The area under the curve (AUC) is equal to bone volume. As aBMD increases, bone volume allocated as cortical bone increases (AUC) because of an increased amount of transitional bone being allocated as cortical bone.

however, provide a possible explanation for the findings from several observational studies showing an increased risk of hip fractures with increased height (27e29). Accordingly, the mineralization of bone may decrease with increased height causing an increased susceptibility to fractures, which is also supported by our findings of an increased BR with increased body height. As our data suggest, integral and trabecular vBMD are good indicators of aBMD at the hip. Cortical vBMD, however, must be dealt with caution when using a fixed threshold as the separation weakness will lead to inaccurate measures of cortical vBMD. There is no joint agreement regarding the threshold of the cortical compartments, and values between 275 mg/cm3 and 350 mg/cm3 have previously been described (16,30). It is recommended from the manufacturer that a consistent use of a fixed and specific cortical threshold value is preferred when assessing the bone compartments (30). We believe, however, that the current methodology does not allow for separate assessments of cortical bone density. There are several strengths to our study. First, this is to our knowledge the largest and only study of its kind investigating the relationship between DXA aBMD and QCT vBMD divided into integral, trabecular, and cortical bone density and volume among postmenopausal women. Second, our population consists of normal, osteopenic, and osteoporotic women resulting in a very wide spectrum of BMD values which may have increased our power to assess the associations. Furthermore, different cortical thresholds have been used to substantiate the results. There is, however, also a limitation to the study; although 3D vBMD is often referred to as a ‘‘true’’ BMD, our data do not allow for such evaluation as we did not have accesses to bone biopsies allowing for determination of the actual mineral content of bone.

1. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Report of a WHO Study Group. 1994 World Health Organ Tech Rep Ser 843:1e129. 2. Javed F, Yu W, Thornton J, Colt E. 2009 Effect of fat on measurement of bone mineral density. Int J Body Compos Res 7(1):37e40. 3. Okano K, Ito M, Aoyagi K, et al. 2013 Discrepancy in bone mineral densities at different skeletal sites in hip osteoarthritis patients. Mod Rheumatol, http://dx.doi.org/10.3109/14397595. 2013.854078. [Epub ahead of print]. 4. Genant HK, Engelke K, Fuerst T, et al. 1996 Noninvasive assessment of bone mineral and structure: state of the art. J Bone Miner Res 11(6):707e730. 5. Adams J, Bishop N. 2008 DXA in Adults and Children. Primer on the Metabolic Bone Diseases and Disorders of Mineral Metabolism. Washington, DC: American Society for bone and Mineral Research, 152e158. 6. Gupta R, Cheung AC, Bartling SH, et al. 2008 Flat-panel volume CT: fundamental principles, technology, and applications. Radiographics 28(7):2009e2022. 7. Genant HK, Block JE, Steiger P, et al. 1987 Quantitative computed tomography in assessment of osteoporosis. Semin Nucl Med 17(4):316e333. 8. Link TM, Lang TF. 2014 Axial QCT: clinical applications and new developments. J Clin Densitom 17(4):438e448. 9. Imai K. 2014 Recent methods for assessing osteoporosis and fracture risk. Endocr Metab Immune Drug Discov 8(1):48e59. 10. Adams JE. 2009 Quantitative computed tomography. Eur J Radiol 71(3):415e424. 11. Jakobsen NF, Rolighed L, Moser E, et al. 2014 Increased trabecular volumetric bone mass density in familial hypocalciuric hypercalcemia (FHH) type 1: a cross-sectional study. Calcif Tissue Int 95(2):141e152. 12. Jakobsen NF, Rolighed L, Nissen PH, et al. 2013 Muscle function and quality of life are not impaired in familial hypocalciuric hypercalcemia: a cross-sectional study on physiological effects of inactivating variants in the calcium-sensing receptor gene (CASR). Eur J Endocrinol 169(3):349e357. 13. Khoo BC, Brown K, Cann C, et al. 2009 Comparison of QCTderived and DXA-derived areal bone mineral density and T scores. Osteoporos Int 20(9):1539e1545. 14. Liu XS, Cohen A, Shane E, et al. 2010 Bone density, geometry, microstructure, and stiffness: relationships between peripheral and central skeletal sites assessed by DXA, HR-pQCT, and cQCT in premenopausal women. J Bone Miner Res 25(10):2229e2238. 15. Cohen A, Lang TF, McMahon DJ, et al. 2012 Central QCT reveals lower volumetric BMD and stiffness in premenopausal women with idiopathic osteoporosis, regardless of fracture history. J Clin Endocrinol Metab 97(11):4244e4252. 16. Cann CE, Adams JE, Brown JK, Brett AD. 2014 CTXA hipdan extension of classical DXA measurements using quantitative CT. PLoS One 9(3):e91904.

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015

8 17. Cheng X, Wang L, Wang Q, et al. 2014 Validation of quantitative computed tomography-derived areal bone mineral density with dual energy X-ray absorptiometry in an elderly Chinese population. Chin Med J (Engl ) 127(8):1445e1449. 18. Pickhardt P, Bodeen G, Brett A, et al. 2014 Comparison of femoral neck BMD evaluation obtained using Lunar DXA and QCT with asynchronous calibration from CT colonography. J Clin Densitom 18(1):5e12. 19. Treece GM, Poole KE, Gee AH. 2012 Imaging the femoral cortex: thickness, density and mass from clinical CT. Med Image Anal 16(5):952e965. 20. Kang Y, Engelke K, Kalender WA. 2003 A new accurate and precise 3-D segmentation method for skeletal structures in volumetric CT data. IEEE Trans Med Imaging 22(5):586e598. 21. Zebaze R, Ghasem-Zadeh A, Mbala A, Seeman E. 2013 A new method of segmentation of compact-appearing, transitional and trabecular compartments and quantification of cortical porosity from high resolution peripheral quantitative computed tomographic images. Bone 54(1):8e20. 22. Valentinitsch A, Patsch JM, Deutschmann J, et al. 2012 Automated threshold-independent cortex segmentation by 3D-texture analysis of HR-pQCT scans. Bone 51(3):480e487. 23. Engelke K, Adams JE, Armbrecht G, et al. 2008 Clinical use of quantitative computed tomography and peripheral quantitative computed tomography in the management of osteoporosis in

Amstrup et al.

24.

25. 26. 27.

28. 29. 30.

adults: the 2007 ISCD Official Positions. J Clin Densitom 11(1):123e162. Felson DT, Zhang Y, Hannan MT, Anderson JJ. 1993 Effects of weight and body mass index on bone mineral density in men and women: the Framingham study. J Bone Miner Res 8(5): 567e573. Albala C, Yanez M, Devoto E, et al. 1996 Obesity as a protective factor for postmenopausal osteoporosis. Int J Obes Relat Metab Disord 20(11):1027e1032. Looker AC, Flegal KM, Melton LJ III. 2007 Impact of increased overweight on the projected prevalence of osteoporosis in older women. Osteoporos Int 18(3):307e313. Cummings SR, Nevitt MC, Browner WS, et al. 1995 Risk factors for hip fracture in white women. Study of Osteoporotic Fractures Research Group. N Engl J Med 332(12): 767e773. Hemenway D, Feskanich D, Colditz GA. 1995 Body height and hip fracture: a cohort study of 90,000 women. Int J Epidemiol 24(4):783e786. Meyer HE, Tverdal A, Falch JA. 1993 Risk factors for hip fracture in middle-aged Norwegian women and men. Am J Epidemiol 137(11):1203e1211. Mindways. QCT pro 3D Volumetric QCT Pine and Hip Bone Mineral Densitometry and Bone Biomechanical Integrity Assesment. [Version 4.2.3]. 2008.

Journal of Clinical Densitometry: Assessment & Management of Musculoskeletal Health

Volume

-,

2015