JOURNAL OF BONE AND MINERAL RESEARCH Volume 12, Number 5, 1997 Blackwell Science, Inc. q 1997 American Society for Bone and Mineral Research

Comparisons of Noninvasive Bone Mineral Measurements in Assessing Age-Related Loss, Fracture Discrimination, and Diagnostic Classification* STEPHAN GRAMPP, HARRY K. GENANT, ASHWINI MATHUR, PHILIPP LANG, MICHAEL JERGAS, ¨ ER, YING LU, and MONICA CHAVEZ MASAHIKO TAKADA, CLAUS-C. GLU

ABSTRACT The purpose of this study was to examine the commonly available methods of noninvasively assessing bone mineral status across three defined female populations to examine their interrelationships, compare their respective abilities to reflect age- and menopause-related bone loss, discriminate osteoporotic fractures, and classify patients diagnostically. A total of 47 healthy premenopausal (age 33 6 7 years), 41 healthy postmenopausal (age 64 6 9 years), and 36 osteoporotic postmenopausal (age 70 6 6 years) women were examined with the following techniques: (1) quantitative computed tomography of the L1–L4 lumbar spine for trabecular (QCT TRAB BMD) and integral (QCT INTG BMD) bone mineral density (BMD); (2) dual X-ray absorptiometry of the L1–L4 posterior-anterior (DXA PA BMD) and L2–L4 lateral (DXA LAT BMD) lumbar spine, of the femoral neck (DXA NECK BMD) and trochanter (DXA TROC BMD), and of the ultradistal radius (DXA UD BMD) for integral BMD; (3) peripheral QCT of the distal radius for trabecular BMD (pQCT TRAB BMD) and cortical bone mineral content (BMC) (pQCT CORT BMC); (4) two radiographic absorptiometric techniques of the metacarpal (RA METC BMD) and phalanges (RA PHAL BMD) for integral BMD; and (5) two quantitative ultrasound devices (QUS) of the calcaneus for speed of sound (SOS CALC) and broadband ultrasound attenuation (BUA CALC). In general, correlations ranged from (r 5 0.10–0.93) among different sites and techniques. We found that pQCT TRAB BMD correlated poorly (r 0.46) with all other measurements except DXA UD BMD (r 5 0.62, p 0.0001) and RA PHAL BMD (r 5 0.52, p 0.0001). The strongest correlation across techniques was between QCT INT BMD and DXA LAT BMD (r 5 0.87, p 0.0001), and the weakest correlation within a technique was between pQCT TRAB BMD and pQCT CORT BMC (r 5 0.25, p 0.05). Techniques showing the highest correlations with age in the healthy groups also showed the greatest differences among groups. They also showed the best discrimination (as measured by the odds ratios) for the distinction between healthy postmenopausal and osteoporotic postmenopausal groups based on age-adjusted logistic regression analysis. For each anatomic site, the techniques providing the best results were: (1) spine, QCT TRAB BMD (annual loss, 21.2% [healthy premenopausal and healthy postmenopausal]); Student’s t-value [not the T score], 5.4 [healthy postmenopausal vs. osteoporotic postmenopausal]; odds ratio, 4.3 [age-adjusted logistic regression for healthy postmenopausal vs. osteoporotic postmenopausal]); (2) hip, DXA TROC BMD (20.46; 3.5; 2.2); (3) radius, DXA UD BMD (20.44; 3.3; 1.9) and pQCT, CORT BMC (20.72; 2.9; 1.7); (4) hand, RA PHAL (20.51; 3.6; 2.0); and (5) calcaneus, SOS (20.09; 3.4; 2.1) and BUA (20.52; 2.6; 1.7). Despite these performance trends, the differences among sites and techniques were statistically insignificant ( p > 0.05) using age-adjusted receiver operating characteristic (ROC) curve analysis. Nevertheless, kappa score analysis (using 22.0 T score as the cut-off value for osteopenia and 22.5 T score for osteoporosis) showed that in general the diagnostic agreement among these measurements in classifying women as osteopenic or osteoporotic was poor, with kappa scores averaging about 0.4 (exceptions were QCT TRAB/ INTG BMD, DXA LAT BMD, and RA PHAL BMD, with kappa scores ranging from 0.63 to 0.89). Often different patients were estimated at risk by using different measurement sites or techniques. (J Bone Miner Res 1997;12:697–711)

*Presented in part by H.K.G. at the XIth International Workshop on Bone Densitometry, Gleneden Beach, Oregon, U.S.A., September 24 –28, 1995.(28)

Musculoskeletal Section and Osteoporosis Research Group, Department of Radiology, University of California, San Francisco, California, U.S.A.

697

698

GRAMPP ET AL.

INTRODUCTION STEOPOROSIS is the most common generalized disease of the skeleton, causing reduction of bone mass and changes in bone structure. The loss of integrity of the trabecular network and the cortical shell result in a reduction in bone strength and in fragility fractures. Current methods for evaluating skeletal status, assessing osteoporosis, and determining fracture risk rely mostly on the noninvasive measurement of bone mineral content (BMC) and bone mineral density (BMD).(1–5) Of these methods, the most common is dual X-ray absorptiometry (DXA), typically used at the posterior-anterior and lateral lumbar spine, proximal femur, and forearm. Quantitative computed tomography (QCT) overcomes the projectional limitations of DXA and allows a selective measurement of trabecular bone in the spine. The validity of QCT of the lumbar spine is generally accepted and the method is regarded as a sensitive tool for noninvasive measurement of the BMD.(6 –11) Recently, peripheral QCT (pQCT) instruments have become available leading to renewed interest in measurements of the radius.(11–16) Similarly, new techniques to measure BMD of the phalanges(17–19) and metacarpals(20 –23) using radiographic absorptiometry (RA) have shown promising results. The only established techniques for noninvasive assessment of bone status that do not require radiation are ultrasound-based techniques measuring ultrasound velocity (speed of sound, SOS) and broadband ultrasound attenuation (BUA) at the calcaneus and other sites.(24 –27) The purpose of this study(28) was to examine the abovementioned methods of noninvasively assessing bone mineral status across three defined female populations (healthy premenopausal, healthy postmenopausal, and osteoporotic postmenopausal women) to examine their interrelationships, compare their respective abilities to reflect age- and menopause-related bone loss, discriminate women with osteoporotic vertebral fractures, and classify women diagnostically as osteopenic or osteoporotic.

O

MATERIALS AND METHODS We studied 47 healthy premenopausal (PRE, age 33 6 7 years), 41 healthy postmenopausal (POST, age 64 6 9 years), and 36 osteoporotic postmenopausal (OSTEO, age 70 6 6 years) women. Exclusion criteria were history of generalized disease of bone, malignant disease, or trauma at the measurement sites, or any drug treatment that could influence bone metabolism. The diagnosis of osteoporosis was defined as the presence of at least one atraumatic vertebral fracture. A fracture was determined by a semiquantitative assessment of morphologic changes of the thoracic and lumbar spine on lateral conventional radiographs.(29) This assessment was based on the qualitative appearance of each vertebra, with a fracture defined by altered morphology and a decrease in vertebral height of approximately 20% or more at the anterior, middle, or posterior aspect of the vertebral body. Nearly all of these women (Table 1) were examined with each of the following techniques. (1) QCT with a GE 9800

computerized tomography (CT) scanner (General Electric, Milwaukee, WI, U.S.A.). The standard protocol for spinal bone mineral analysis(6,30,31) was used. It uses a 10 mm thick axial slice in the midsection of four vertebral bodies from L1 to L4. Measurements were performed with a single energy technique at 80 kVp. Calibration of the CT image was achieved by a simultaneous scanning of a calibration phantom (Image Analysis, Columbia, KY, U.S.A.) containing various inserts of hydroxyapatite-equivalent material. The images were evaluated on an off-line image processing workstation with an automated quantitative image evaluation technique.(30) Two different regions of interest (ROIs) in each vertebra were evaluated for BMD.(30) These were the elliptical ROIs, (trabecular bone of the vertebral body) and the integral ROIs (a portion of the trabecular and cortical bone of the entire vertebra representing a crosssection through the midvertebra and portions of the posterior elements but no transverse processes or vertebral endplates). BMD was determined for each ROI (trabecular [QCT TRAB BMD] and integral [QCT INTG BMD]) and expressed in hydroxyapatite-equivalent units (mg/cm3).(6,30 –33) (2) DXA with a Hologic QDR-2000 scanner (Hologic Inc., Waltham, MA, U.S.A.) using pencil beam, 140/70 kVp, 2.0 mA and the manufacturer’s recommended standard analysis procedures for the posterior-anterior lumbar spine at L1–L4 (DXA PA BMD), the lateral lumbar spine at L2–L4 (DXA LAT BMD), the proximal femur at the femoral neck (DXA NECK BMD), trochanter (DXA TROC BMD), intertrochanteric region (DXA INTER BMD), and Ward’s triangle (DXA WARD BMD), and the distal radius at the ultradistal (DXA UD BMD), 1/3 (DXA 1/3 BMD), total (DXA TOT BMD), and midradius (DXA MID BMD). The software allowed anatomic separation of the different ROIs. Spine scans were performed with the patient in the supine position. Lateral scans were performed following a PA scan in which the vertebral levels were identified with the subject in supine position. Hip scans were performed with the subject in supine position. Distal radius scans were performed with the subject’s arm placed palm down on the scanning table, scanning distal to the radial styloid and covering the ultradistal, the distal third radius, and the ulna. The projectional BMD results for all ROIs were given in grams per square centimeter. (3) Peripheral QCT measurements of the nondominant radius were obtained with a Stratec XCT-960 scanner (Stratec GmbH, Pforzheim, Germany), using a single energy technique (47 kVp, 0.3 mA). One axial slice (2.5 mm thickness) was obtained at a region defined to be at 4% of the ulnar length proximal to the most proximal site of the radial articular surface. The axial slice was automatically evaluated to define trabecular, cortical, and total ROIs. To calculate the total ROI, the software used a thresholding algorithm to define the outer boundary of the bone. The trabecular ROI was generated by reducing the area of the total ROI by 55%. This was done by trimming pixels from the periphery to define a core area, which consists exclusively of trabecular bone. The cortical was defined through a second thresholding algorithm identifying bone with high density in the outer rim. Constant threshold levels were used for all subjects (0.5 mg/cm3 for total bone; 0.7 mg/cm3

NONINVASIVE BONE MINERAL MEASUREMENTS for cortical bone). Based on this approach, pQCT TOT BMD, pQCT TRAB BMD, pQCT CORT BMD and so on were calculated. From area, slice thickness, and BMD (mg/ cm3), the BMC (g) (for example pQCT CORT BMC) of each region was calculated. (4) Two computerized RA techniques were performed on conventional radiographs of the nondominant hand. Two manufacturer-specific aluminum wedges were used as reference phantoms: for the second metacarpal (Chugai Pharmaceutical Co. Ltd, Tokyo, Japan) and for the middle phalanges (Computed Osteosystems, Yellow Springs, OH, U.S.A.). Lateral radiographs were taken for each subject by direct exposure. The hands were radiographed twice with each phantom in accordance with the procedures provided by the manufacturers (50 kVp, 400 mA, 0.8 s, and 60 kVp, 300 mA, 0.5 s; the mean of the double measurement was used in the data analysis). The focus-to-film distance was 105 cm. An aluminum alloy reference wedge was placed parallel to the middle phalanx of the index finger (phalanx measurement) or parallel to the main axis of the entire hand (metacarpal measurement). Developed films were sent to both manufacturers for analysis. In order to do this, each film was digitized by a high resolution video camera. For the phalanx measurements, ROIs were defined encompassing the wedge and the middle phalanges; ROIs for the metacarpal measurements were encompassing the wedge and a 6- to 8-mm-long portion (10% of the length) of the second metacarpal. Through a curve fitting process applied to the wedge image data, a set of parameters was calculated defining the relationship between absorber density and optical density for each individual film. These parameters were applied to the bone image data to compute bone mineral mass. The integral BMDs (RA CH METC BMD and RA CO PHAL BMD) for the respective bones were determined with both techniques and given as aluminumequivalent units in grams per square centimeter. (5) Two quantitative ultrasound devices made by Walker Sonix (WS, Hologic Inc.) and the Lunar Achilles (LA, Lunar Corp., Madison, WI, U.S.A.) were used for measurements of the calcaneus of the nondominant side. Both systems used an emitting and receiving transducer either side of the heel, with water (heated to 358C) as the coupling agent. The systems used a transmitting transducer with a central frequency of 0.5 MHz, which was electronically exited to produce a broadband spectra. Ultrasonic waves with a predominant bipolar appearance were transmitted through the heel, detected by receiving transducers, and digitized for analysis. SOS (m/s) and BUA (dB/MHz) were determined (SOS WS CALC, BUA WS CALC, SOS LA CALC, and BUS LA CALC) with both machines according to the manufacturers’ recommended standard procedures. The means and standard deviations of the measurements for the three subject groups were calculated. The Student’s t-test (t values and p values) and percent decrement were used for comparing the different measurements for reflecting intergroup differences. Annual changes were expressed as percent changes relative to the predicted values at age 30 and as fractional standard deviations relative to the standard deviation (SD) of PRE. Correlations with age along with p values are also reported. Odds ratios (for 1 SD

699 decrease in the measured parameter) and 95% confidence limits based on the age-adjusted logistic regression were calculated to measure the discriminative ability (for discriminating between OSTEO and POST group) and the risk of osteoporotic fracture associated with the measured parameter. The pairwise comparisons of the discriminative abilities were tested using age-adjusted receiver operating characteristic (ROC) curve analysis. Pairwise comparisons of all techniques were obtained by pooling all subjects (PRE, POST, OSTEO) and using Pearson’s correlation coefficients (r), percent standard errors of the estimate (CV), and p values for testing significance of the correlations. To compare techniques for their diagnostic ability, a kappa score analysis was done on the postmenopausal women (POST and OSTEO) and osteoporotic postmenopausal women (OSTEO). This was done by classifying every women from the postmenopausal groups as osteopenic if her T score with respect to the reference group (PRE) was less than 22.0. Similarly every subject from the osteoporotic group was classified as osteoporotic if her T score compared with the reference group (PRE) was less than 22.5. The T score for an individual woman and a particular measurement is defined as the measurement minus the mean measurement for the young normals (PRE) divided by the SD of the measurement in the PRE group. Note that the T score is measuring the position of an individual woman with respect to the PRE group and is different from the Student’s t value.

RESULTS Results showed that all parameters (except pQCT TRAB BMD) between the PRE and POST groups had statistically significant differences ( p # 0.05) (Table 1). Except for pQCT TRAB BMD, pQCT CORT BMD, pQCT TRAB BMC, SOS WS CALC, and BUA LA CALC, other measurements showed statistically significant differences between POST and OSTEO groups ( p # 0.05; Table 1). The Student’s t values and percent decrements demonstrate the different levels of significance (Table 1 and Fig. 1). Linear regressions showed that in healthy women (PRE AND POST) the only parameter not significantly correlated with age was pQCT TRAB BMD. All other parameters were significantly correlated with age ( p # 0.05; Table 1 and Fig. 2). The highest correlation with age was found at the spine with r 5 20.78 for QCT TRAB BMD and QCT INTG BMD and r 5 20.72 for DXA LAT BMD. The highest annual bone loss was measured by QCT TRAB BMD (1.18%), DXA WARD BMD (1.0%), and QCT INTG BMD (0.80%). The smallest annual changes were registered for both measurements of the ultrasound parameter SOS (SOS WS CALC 5 20.02; SOS LA CALC 5 20.09). However, when expressed in terms of SD rather than percentages (relative to PRE) the annual changes for ultrasound SOS were comparable to most other techniques. Age-adjusted odds ratios for discriminating the OSTEO from POST group were significant for QCT TRAB BMD (4.3, 95% confidence limits 1.8, 10.0) and QCT INTG BMD

700

GRAMPP ET AL. TABLE 1. COMPARISON

OF

MEANS, AGE-RELATED LOSSES,

Mean SD Method parameter (n)

PRE

POST

OSTEO

178 33

106 30

72 22

QCT TRAB BMD mg/cm3 (123) QCT INTG BMD mg/cm3 (123) DXA PA BMD g/cm2 (124)

289 36

208 31

176 29

1.02 0.12

0.90 0.12

0.79 0.15

DXA LAT BMD g/cm2 (123)

0.82 0.08

0.65 0.08

0.58 0.07

DXA FEM NECK BMD g/cm2 (124) DXA FEM TROC BMD g/cm2 (124) DXA FEM INTER BMD g/cm2 (124) DXA FEM WARD BMD g/cm2 (124) DXA FEM TOT BMD g/cm2 (124) DXA RAD UD BMD g/cm2 (122) DXA RAD 1/3 BMD g/cm2 (122) DXA RAD MID BMD g/cm2 (122) DXA RAD TOT BMD g/cm2 (122) pQCT RAD TRAB BMD mg/cm3 (106) pQCT RAD CORT BMD mg/cm3 (106) pQCT RAD TOT BMD mg/cm3 (106) pQCT RAD TRAB BMC mg (106)

0.82 0.12

0.67 0.12

0.62 0.10

0.69 0.10

0.58 0.08

0.52 0.09

1.09 0.14

0.90 0.13

0.84 0.15

0.76 0.15

0.50 0.14

0.43 0.11

0.94 0.12

0.77 0.11

0.71 0.12

0.44 0.05

0.38 0.04

0.34 0.06

0.68 0.05

0.60 0.07

0.56 0.07

0.59 0.05

0.53 0.06

0.48 0.07

0.57 0.04

0.50 0.05

0.46 0.06

192 30

175 53

161 41

692 70

650 85

618 69

361 49

323 62

290 53

68 13

59 17

55 16

% decr. t value p value PRE vs. POST

% decr. t value p value POST vs. OSTEO

40.7 10.7 0.0001 27.8 11.2 0.0001 11.7 4.6 0.0001 20.8 9.7 0.0001 18.0 5.8 0.0001 15.8 5.7 0.0001 17.4 6.4 0.0001 34.2 8.1 0.0001 18.1 6.5 0.0001 14.0 6.1 0.0001 11.8 6.3 0.0000 10.2 5.8 0.0000 12.3 6.5 0.0000 9.0 1.8 0.0829 6.1 2.3 0.0200 10.5 2.9 0.0050 13.2 2.8 0.0070

31.5 5.4 0.000 15.5 4.7 0.000 12.0 3.6 0.001 11.4 4.2 0.000 7.7 2.0 0.045 11.5 3.5 0.001 6.7 2.1 0.040 14.0 2.5 0.020 7.8 2.4 0.020 10.4 3.3 0.002 6.7 2.9 0.006 9.4 3.2 0.002 8.0 3.3 0.001 8.1 1.2 0.230 4.9 1.7 0.090 10.2 2.4 0.020 6.8 1.0 0.300

AND

FRACTURE ASSOCIATIONS

BY

TECHNIQUE

Annual loss (%)/SD PRE 1 POST

Cor (2) w. age p value PRE 1 POST

Odds ratio LCL/UCL POST vs. OSTEO

ROC AREA SE POST vs. OSTEO

1.18 0.064

0.78 0.0001

4.3 1.8/10.1

0.82 0.05

0.80 0.064

0.78 0.0001

3.0 1.5/6.1

0.80 0.05

0.30 0.026

0.41 0.0001

2.4 1.4/4.2

0.78 0.05

0.58 0.058

0.72 0.0001

2.6 1.4/5.0

0.79 0.05

0.53 0.036

0.56 0.0001

1.5 0.9/2.5

0.71 0.06

0.46 0.033

0.54 0.0001

2.2 1.3/3.8

0.77 0.05

0.47 0.030

0.55 0.0001

1.5 0.9/2.5

0.71 0.06

1.00 0.040

0.70 0.0001

1.6 0.9/2.9

0.72 0.06

0.48 0.030

0.57 0.0001

1.6 0.9/2.7

0.73 0.06

0.44 0.039

0.62 0.0001

1.9 1.0/3.5

0.74 0.06

0.38 0.052

0.64 0.0000

1.6 0.9/2.7

0.72 0.06

0.35 0.042

0.63 0.0000

2.1 1.0/4.6

0.73 0.06

0.46 0.053

0.66 0.0000

2.2 1.1/4.4

0.74 0.06

0.22 0.014

0.18 0.1000

1.4 0.9/2.4

0.70 0.06

0.19 0.019

0.31 0.0070

1.3 0.7/2.3

0.71 0.06

0.31 0.023

0.35 0.0020

1.5 0.9/2.8

0.73 0.06

0.38 0.019

0.30 0.0070

1.2 0.7/2.1

0.69 0.06

NONINVASIVE BONE MINERAL MEASUREMENTS

701

TABLE 1. CONTINUED

Mean SD Method parameter (n) pQCT RAD CORT BMC mg (106) pQCT RAD TOT BMC mg (106) RA CH METC BMD A1 eqival. (103) RA CO PHAL BMD A1 eqival. (103) BUA WS CALC db/MHz (123) SOS WS CALC m/s (121) BUA LA CALC db/MHz (115) SOS LA CALC m/s (115)

PRE

POST

OSTEO

145 40

111 38

87 30

283 43

241 32

221 35

2.48 0.30

2.13 0.28

1.91 0.29

107 9

88 11

79 10

89 18

74 16

64 18

1517 11

1505 8

1504 27

117 13

111 12

106 15

1580 38

1532 30

1510 26

% decr. t value p value PRE vs. POST

% decr. t value p value POST vs. OSTEO

23.2 3.7 0.0004 14.8 4.9 0.0000 14.2 4.9 0.0001 18.6 8.2 0.0001 16.5 4.1 0.0001 0.8 5.6 0.0001 5.1 2.2 0.0310 3.0 6.3 0.0001

21.9 2.9 0.006 8.3 2.4 0.020 10.3 3.3 0.002 10.0 3.6 0.001 13.5 2.6 0.010 0.1 0.32 0.750 5.2 1.9 0.068 1.5 3.4 0.001

(3.0, 1.5, 6.1), for DXA PA BMD (2.4, 1.4, 4.2), DXA LAT BMD (2.6, 1.4, 5.0), DXA TROC BMD (2.2, 1.3, 3.8), DXA UD BMD (1.9, 1.0, 3.5), DXA MID BMD (2.1, 1.0, 4.6), DXA RAD TOT BMD (2.2, 1.1, 4.4), for RA CO PHAL BMD (2.0, 1.1, 3.6), and for BUA WS CALC (1.7, 1.0, 3.0) and SOS LA CALC (2.1, 1.2, 3.9) (Table 1 and Fig. 3). The area under the age adjusted ROC curves for the different measurements were not significantly different from each other ( p . 0.05). For further comparisons between and among techniques, we selected from each technique those parameters that are most widely used. We also included the parameters that showed the best performance if they were not the most commonly used parameters. (See Tables 2 and 3). Correlations, percentage standard error of estimate (CV), and p values based on all women and on linear regressions for the selected parameters are given in Table 2. Correlations ranged from 0.11 to 0.94 among different sites and techniques. We found that pQCT TRAB BMD correlated poorly, with correlations less than 0.46 with other measurements (except DXA UD BMD, r 5 0.62, p # 0.0001 and RA CO PHAL BMD, r 5 0.52, p # 0.0001). Comparing all techniques, the strongest correlation across techniques was between QCT INTG BMD and DXA LAT BMD (r 5 0.87, p # 0.0001) and the weakest within a

Annual loss (%)/SD PRE 1 POST

Cor (2) w. age p value PRE 1 POST

Odds ratio LCL/UCL POST vs. OSTEO

ROC AREA SE POST vs. OSTEO

0.72 0.027

0.46 0.0001

1.7 0.9/3.2

0.73 0.06

0.44 0.029

0.54 0.0000

1.4 0.8/2.7

0.69 0.06

0.47 0.038

0.63 0.0001

1.7 0.9/3.3

0.73 0.06

0.51 0.062

0.71 0.0001

2.0 1.1/3.6

0.77 0.06

0.52 0.026

0.45 0.0001

1.7 1.0/3.0

0.73 0.06

0.02 0.034

0.57 0.0001

1.0 0.6/1.5

0.69 0.06

0.18 0.016

0.30 0.0050

1.4 0.8/2.4

0.70 0.06

0.09 0.039

0.63 0.0001

2.1 1.2/3.9

0.78 0.05

technique was between pQCT TRAB BMD and pQCT CORT BMC (r 5 0.25, p # 0.01). The mean correlation and the associated SD for each technique was calculated (Fig. 4). On the average, the QCT techniques correlated best while the pQCT and QUS techniques showed weaker correlations. By kappa analysis the diagnostic agreement among the different measurements was generally poor in classifying postmenopausal women (POST and OSTEO) as osteopenic (T score , 2.0) or in classifying women with vertebral fractures (OSTEO) as osteoporotic (T score , 2.5) (Table 3 and Figs. 5 and 6). Exceptions were the kappa analysis comparisons among QCT TRAB, QCT INTG, DXA LAT, and RA PHAL, which gave kappa scores between 0.65 and 0.89 for the classification as osteopenic and between 0.63 and 0.87 for the classification as osteoporotic. Even in these cases, different patients were sometimes estimated at risk by different measurement approaches (Figs. 5 and 6).

DISCUSSION Measurements of different bone parameters such as BMD, BMC, SOS, or BUA at various body sites have been found useful in assessing skeletal status, assessing osteopo-

702

GRAMPP ET AL.

FIG. 1. (A) QCT TRAB BMD, (B) DXA LAT BMD, (C) DXA FEM TROC BMD, (D) DXA RAD UD BMD, (E) pQCT CORT RAD BMC, (F) RA CO PHAL BMD, (G) BUA WS CALC, and (H) SOS LA CALC in the discrimination of study groups. Means with standard deviation lines are given.

rosis, and determining osteoporotic fracture risk. QCT of the spine has been used to assess the skeletal status and detect age- and disease-related changes,(6 –11,30 –33) but it has been less widely used than DXA due to its cost and complexity. More recently, QUS, pQCT, and computerized

RA have been introduced as reliable methods for assessing skeletal status. Most of these techniques have been evaluated and compared in limited combinations in different studies. Since the study populations and/or the definition of osteoporosis dif-

NONINVASIVE BONE MINERAL MEASUREMENTS

FIG. 2.

703

Linear correlations with age (r values) shown as a bar chart.

FIG. 3.

Age-adjusted odds ratios are shown as a bar chart.

fered, the results of the different studies cannot be easily compared. In addition, all techniques reported in this manuscript have not yet been used in one single population and the results compared. Ensuring that all the patients were measured by each technique enabled comparability of the different techniques.

Detection of age- and menopause-related bone loss In our study, we found that all the techniques except pQCT RAD TRAB BMD readily differentiated between healthy premenopausal and healthy postmenopausal subjects and reflected age- and menopause-related bone loss. In differentiating between healthy premenopausal and

healthy postmenopausal women, the best abilities (based on percentage decrement and Student’s t values) were shown by QCT, followed by DXA LAT BMD of the spine. Similar results were also obtained by Steiger et al.(31) who found a decrement of 42% in the comparison of healthy premenopausal and healthy postmenopausal women for QCT TRAB BMD and 28% for QCT INTG BMD. DXA of the spine showed better abilities in the lateral (DXA LAT BMD) measurements compared with the PA (DXA PA BMD) measurements. This agrees with the results obtained by Guglielmi et al.,(7) who found the highest sensitivity and specificity for the detection of age- and osteoporosis-related changes for QCT TRAB BMD followed by DXA LAT BMD and by DXA PA BMD. The RA CO PHAL BMD

TABLE 2. CORRELATION COEFFICIENTS (R) AND %SEE (CV) FOR ALL WOMEN (PRE, POST AND OSTEO) P VALUE 5 NS, NOT SIGNIFICANT AT 0.05; 21 SIGNIFICANCE AT 0.05; 22 AT 0.01; 23 AT 0.001; 24 AT 0.0001

r CV p QCT TRAB BMD QCT INTG BMD DXA PA BMD DXA LAT BMD DXA FEM NECK BMD DXA FEM TROC BMD DXA UD RAD BMD pQCT TRAB RAD BMD pQCT CORT RAD BMC RA CH METC BMD RA CO PHAL BMD BUA WS CALC (db/MHz) SOS WS CALC (m/s) BUA LA CALC (db/MHz) SOS LA CALC (m/s)

DXA QCT QCT DXA DXA FEM TRAB INTG PA LAT NECK BMD BMD BMD BMD BMD 1

0.94 0.72 8.8 12.1 24 24 0.94 1 0.83 15.1 9.6 24 24 0.72 0.83 1 30.2 14.0 24 24 0.87 0.88 0.74 21.0 11.9 11.7 24 24 24 0.71 0.77 0.76 30.6 16.1 11.3 24 24 24

0.87 9.2 24 0.88 8.9 24 0.74 12.8 24 1

0.71 14.2 24 0.77 12.8 24 0.76 13.1 24 0.71 14.2 24 0.71 1 13.3 24

0.77 0.83 0.76 0.81 0.84 27.5 14.2 11.2 11.2 11.0 24 24 24 24 24

BUA BUA SOS LA SOS WS RA DXA DXA pQCT pQCT RA CO CALC WS CALC LA FEM UD TRAB CORT CH TROC RAD RAD RAD METC PHAL (db/ CALC (db/ CALC BMD BMD BMD BMC BMD BMD MHz) (m/s) MHz) (m/s) 0.77 12.1 24 0.83 10.7 24 0.76 12.3 24 0.81 11.3 24 0.84 10.4 24 1

0.76 0.76 0.62 0.67 0.67 0.65 27.9 16.3 13.5 13.8 14.8 14.4 24 24 24 24 24 24

0.76 11.0 24 0.76 11.0 24 0.62 13.3 24 0.67 12.5 24 0.67 12.5 24

0.40 23.0 24 0.41 22.8 24 0.36 23.2 24 0.40 22.9 24 0.40 22.9 24

0.64 28.7 24 0.62 29.4 24 0.48 32.8 24 0.55 31.2 24 0.53 31.8 24

0.71 12.1 24 0.72 11.9 24 0.57 14.0 24 0.66 12.8 24 0.60 13.7 24

0.81 9.8 24 0.80 10.1 24 0.61 13.5 24 0.75 11.3 24 0.62 13.3 24

0.64 19.9 24 0.60 20.7 24 0.50 22.4 24 0.61 20.5 24 0.54 21.8 24

0.40 1.1 24 0.37 1.1 24 0.31 1.1 23 0.41 1.1 24 0.37 1.1 24

0.40 11.5 24 0.41 11.4 24 0.34 11.8 23 0.35 11.7 24 0.34 11.8 23

0.79 1.7 24 0.76 1.9 24 0.58 2.3 24 0.74 1.9 24 0.67 2.1 24

0.65 0.45 0.49 0.60 0.69 0.55 0.34 0.29 0.72 12.8 22.2 32.6 13.7 12.3 21.6 1.1 12.0 2.0 24 24 24 24 24 24 23 22 24 1

0.40 0.41 0.36 0.40 0.40 0.45 0.62 40.3 23.2 16.0 17.2 19.1 17.5 13.3 24 24 24 24 24 24 24

0.62 0.69 0.74 0.77 0.56 0.42 0.24 0.66 19.7 25.9 11.4 10.8 21.3 1.1 11.1 2.1 24 24 24 24 24 24 21 24 1

0.64 0.62 0.48 0.55 0.53 0.49 0.69 0.25 33.8 20.0 15.1 15.6 17.7 17.1 12.1 24.2 24 24 24 24 24 24 24 21

0.25 0.38 0.52 0.41 0.29 0.11 0.46 36.3 15.5 14.5 23.3 1.2 11.8 2.6 21 24 24 24 22 NS 24 1

0.71 0.72 0.57 0.66 0.60 0.60 0.74 0.38 0.62 32.7 18.5 14.9 14.7 16.5 16.1 12.2 23.5 30.3 24 24 24 24 24 24 24 24 24

0.62 0.63 0.42 0.27 0.30 0.42 13.2 13.2 23.2 1.2 11.3 2.6 24 24 24 22 22 24 1

0.81 0.80 0.61 0.75 0.62 0.69 0.77 0.52 0.63 0.73 27.1 16.0 14.5 13.1 16.2 14.6 11.6 21.8 29.9 11.8 24 24 24 24 24 24 24 24 24 24

0.73 0.51 0.28 0.26 0.60 11.6 23.7 1.2 12.0 2.4 24 24 22 21 24 1

0.64 0.60 0.50 0.61 0.54 0.55 0.56 0.41 0.42 0.51 0.62 33.3 20.2 15.0 15.0 16.9 15.9 14.0 22.8 33.9 14.8 13.2 24 24 24 24 24 24 24 24 24 24 24

0.62 0.45 0.38 0.75 21.5 1.1 11.4 2.0 24 24 24 24 1

0.40 0.37 0.31 0.41 0.37 0.34 0.42 0.29 0.27 0.28 0.45 0.52 39.2 23.3 16.3 17.2 18.6 18.1 15.1 23.9 36.0 16.2 14.9 22.2 24 24 22 24 24 23 24 22 22 22 24 24

0.52 0.41 0.83 1.0 11.4 1.6 24 24 24 1

0.40 0.41 0.34 0.35 0.34 0.29 0.24 0.11 0.30 0.26 0.38 0.41 0.24 38.9 23.2 16.3 17.7 19.1 18.5 16.3 25.2 35.1 16.3 15.4 23.8 1.1 24 24 23 24 23 22 21 NS 22 21 24 24 22

0.24 0.56 12.1 2.4 22 24 1

0.79 0.76 0.58 0.74 0.67 0.72 0.66 0.46 0.42 0.60 0.75 0.83 0.56 0.42 26.1 16.6 14.2 12.7 15.1 13.4 12.6 22.5 33.4 13.6 10.9 14.5 1.0 11.4 24 24 24 24 24 24 24 24 24 24 24 24 24 24

0.42 2.6 24 1

NONINVASIVE BONE MINERAL MEASUREMENTS

705

TABLE 3. KAPPA SCORES FOR THE CLASSIFICATION OF POSTMENOPAUSAL WOMEN (POST AND OSTEO) AS OSTEOPENIC (T SCORE , 22.0) AND OSTEOPOROTIC POSTMENOPAUSAL (OSTEO) AS OSTEOPOROTIC (T SCORE , 22.5)

k t,22.0 t,22.5 QCT TRAB BMD QCT INTG BMD DXA PA BMD DXA LAT BMD DXA FEM NECK BMD DXA FEM TROC BMD DXA UD RAD BMD pQCT TRAB RAD BMD pQCT CORT RAD BMC RA CH METC BMD RA CO PHAL BMD BUA WS CALC (db/MHz) SOS WS CALC (m/s) BUA LA CALC (db/MHz) SOS LA CALC (m/s)

DXA QCT QCT DXA DXA FEM TRAB INTG PA LAT NECK BMD BMD BMD BMD BMD

BUA BUA SOS LA WS SOS RA DXA DXA pQCT pQCT RA CO CALC WS CALC LA FEM UD TRAB CORT CH TROC RAD RAD RAD METC PHAL (db/ CALC (db/ CALC BMD BMD BMD BMD BMD BMD MHz) (m/s) MHz) (m/s)

1.00

0.89

0.41

0.75

0.34 0.47

0.41

0.25

0.18

0.32 0.68

0.20

0.29

0.04

0.38

1.00

0.87

0.54

0.81

0.16 0.30

0.31

0.17

0.22

0.33 0.75

0.12

0.12

0.03

0.28

0.89

1.00

0.46

0.75

0.32 0.50

0.43

0.29

0.21

0.32 0.71

0.24

0.33

0.05

0.41

0.87

1.00

0.60

0.86

0.18 0.35

0.43

0.15

0.12

0.28 0.78

0.14

0.14

0.05

0.33

0.41

0.46

1.00

0.37

0.43 0.66

0.40

0.16

0.19

0.35 0.31

0.30

0.33

0.06

0.27

0.54

0.60

1.00

0.55

0.15 0.35

0.22

0.07

0.34

0.50 0.44

0.17

0.31

0.04

0.09

0.75

0.75

0.37

1.00

0.26 0.44

0.33

0.26

0.21

0.34 0.65

0.27

0.38

0.06

0.47

0.81

0.86

0.55

1.00

0.21 0.40

0.35

0.21

0.08

0.44 0.63

0.16

0.17

0.05

0.34

0.34

0.32

0.43

0.26

1.00 0.56

0.55

0.21

0.10

0.39 0.19

0.30

0.50

0.13

0.37

0.16

0.18

0.15

0.21

1.00 0.47

0.07 20.04 20.04 20.08 0.06 20.04

0.25

0.17

0.28

0.47

0.50

0.66

0.44

0.56 1.00

0.33

0.35

0.20

0.33 0.36

0.25

0.36

0.04

0.42

0.30

0.35

0.35

0.40

0.47 1.00

0.38

0.64

0.35

0.53 0.31

0.33

0.33

0.07

0.37

0.41

0.43

0.40

0.33

0.55 0.33

1.00

0.44

0.31

0.44 0.36

0.36

0.44

0.07

0.38

0.31

0.43

0.22

0.35

0.07 0.38

1.00

0.58

0.29

0.41 0.44

0.24 20.04

0.07

0.58

0.25

0.21

0.16

0.26

0.21 0.35

0.44

1.00

0.22

0.33 0.35

0.22

0.05

0.29

0.17

0.15

0.07

0.21 20.04 0.64

0.58

1.00

0.41

0.57 0.32

0.47 20.04. 20.06

0.40

0.18

0.21

0.19

0.20

0.10 0.20

0.31

0.22

1.00

0.29 0.17

0.16

0.25

0.13

0.23

0.22

0.12

0.34 20.08 20.04 0.35

0.29

0.41

1.00

0.26 0.19

0.17

0.25

0.24 20.07

0.32

0.32

0.35

0.34

0.38 0.33

0.44

0.33

0.29

1.00 0.30

0.27

0.23

0.11

0.24

0.33

0.28

0.49

0.44

0.08 0.53

0.41

0.57

0.26

1.00 0.43

0.48

0.33

0.08

0.19

0.68

0.71

0.30

0.65

0.19 0.36

0.36

0.35

0.17

0.30 1.00

0.33

0.23

0.03

0.50

0.75

0.78

0.44

0.63

0.06 0.31

0.44

0.32

0.19

0.43 1.00

0.23

0.10

0.07

0.35

0.20

0.24

0.30

0.27

0.30 0.25

0.36

0.22

0.16

0.27 0.33

1.00

0.55

0.18

0.50

0.12

0.14

0.17

0.16 20.04 0.33

0.24

0.47

0.17

0.48 0.23

1.00

0.31 20.05

0.37

0.29

0.33

0.33

0.38

0.50 0.36

0.44

0.16

0.25

0.23 0.23

0.55

1.00

0.63

0.12

0.14

0.31

0.17

0.25 0.33 20.04 20.04

0.25

0.33 0.10

0.31

1.00 20.04 20.04

0.04

0.05

0.06

0.63

0.13 0.03

0.07

0.13

0.11 0.03

0.18

0.03

1.00

0.07

0.03

0.05

0.04

0.05

0.17 0.07

0.07 20.06

0.24 20.08 0.07 20.05 20.04

1.00

0.12

0.38

0.41

0.27

0.47

0.37 0.43

0.38

0.29

0.23

0.24 0.50

0.50

0.63

0.07

1.00

0.28

0.33

0.09

0.34

0.28 0.37

0.58

0.40 20.08

0.19 0.35

0.37 20.04

0.12

1.00

0.05

0.16

0.03

706

GRAMPP ET AL.

FIG. 4. Mean intersite/technique correlations (average r values by technique) shown as a bar chart.

technique showed good distinction between healthy premenopausal and healthy postmenopausal women and better sensitivity to age-related changes than pQCT, QUS, and DXA of either femur or radius. A similar trend was reported by Matsumoto et al.,(23) who found that RA of the metacarpals provided a sensitivity for discrimination between age groups similar to that of DXA PA BMD. In addition, we found that for DXA of the proximal femur, the best parameter was the DXA FEM WARD BMD; for DXA of the radius, DXA RAD TOT BMD; for pQCT of the radius, pQCT RAD CORT BMC; for RA of the hand, RA CO PHAL BMD; and for QUS of the calcaneus, SOS LA CALC. Similarly, when investigating age- and menopause-related changes by linear regression analysis, we found a strong inverse relationship between bone parameters and age for most sites and techniques. Our results, in agreement with many studies,(7,10,31,34,35) suggest that the sensitivity to detect age-related changes, expressed as percentage annual loss, was higher at the spine measured with QCT TRAB BMD than with QCT INTG BMD, DXA LAT BMD, or DXA PA BMD. We found that DXA PA BMD showed comparatively smaller age-related changes compared with other axial and many appendicular measures except pQCT and QUS. This finding agrees with that of Steiger et al.,(36) who found smaller changes with age in DXA PA BMD than in DXA at the femur or at the calcaneus. The authors suggested that this may be influenced by the fact that DXA PA spine measurements are confounded by degenerative changes of the vertebrae, a suggestion strongly supported by Ito et al.(9) and also suggested in our study population. We also found a comparatively high annual change of around 21% for DXA FEM WARD. Similar results for DXA of the femur were reported by Greenspan et al.,(37) who found that in women aged 65 and over, the highest annual BMD changes (20.86%) occurred at the Ward’s triangle. However, results given by Steiger et al.(36) for

age-related changes in women aged 65 and over showed larger changes for DXA FEM NECK BMD and DXA FEM TROC BMD (20.82% and 21.30%, respectively). Those numbers differ from ours principally because our measure of annual loss substantially reflects the menopause while Steiger’s reflect age-related effects among the very elderly. For QUS measurements, we found that BUA WS CALC, which largely covers a volume of trabecular bone, showed a comparatively high annual loss of 20.52% (the SOS percent decrement would generally be low compared with other parameters due to a large denominator). This supports the concept that skeletal sites consisting predominantly of trabecular bone respond faster to metabolic stimuli than cortical bone.(38,39) This measure, however, was the only ultrasound parameter in our study which showed such a substantial annual decrement. Similar results were reported by Schott et al.(40) who found annual changes in a female population of 20.29% by BUA LA CALC and of 20.07% by SOS LA CALC. Moris et al.(41) found age correlations in women of r 5 20.62 for SOS and r 5 20.59 for BUA, similar to our observed correlations. In our study, the measurements at the radius most strongly influenced by age and menopause were DXA RAD 1/3 BMD (20.46%/year) and pQCT RAD COR BMC (20.38%). Age-related changes in cortical BMC at the radius appear to be largely due to endosteal thinning of the cortical bone since both cortical area and cortical BMC diminish proportionally. The importance of changes in the peripheral cortical bone was recently emphasized in several studies at the forearm using DXA(36,42– 44) as well as pQCT.(13,15) However, some authors have reported significant losses in radial trabecular BMD of 20.5 to 21.1%/ year in women, especially shortly after the menopause.(14,45) The differences between our results and these studies might be caused by differences between study populations, especially due to different criteria used in the definition of healthy subjects. Moreover, in contrast to

NONINVASIVE BONE MINERAL MEASUREMENTS

707

FIG. 5. Mean (1SD) kappa scores for the classification of postmenopausal women (POST and OSTEO) as osteopenic (T score , 22.0 relative to PRE) shown as a bar chart.

FIG. 6. Mean (1SD) kappa scores for the classification of women with vertebral fractures (OSTEO) as osteoporotic (T score , 2.5 relative to PRE) shown as a bar chart.

these pQCT studies, which separated cortical bone and soft tissue by an iterative contour-finding algorithm(14) or an individual threshold setting,(45) a constant threshold was used for all subjects in our study. Although this consistency of threshold guaranteed a high measurement precision,(15) it under-represented the total amount of bone in cases of low overall density or thin cortical rims. Differences in the data evaluation process were therefore most pronounced in elderly subjects and might be responsible for the inconsistency in the trabecular BMD results. With respect to RA techniques, our data showed reasonable age-related changes at both the phalanges and the

metacarpals. The age-related annual losses were comparable to other peripheral measurements used in our study. This result is consistent with the results published for RA at both sites of the hand. In this context, Fukunaga et al.(21) demonstrated that RA CH METC BMD showed age-related losses comparable to DXA RAD BMD and better then DXA PA BMD of the lumbar spine. Matsumoto et al.(23) reported annual losses in women by RA of the metacarpals which ranged between 20.8% (in the eighth decade) and 21.5% (in the sixth decade). One reason our numbers differ could be that in calculating annual bone loss we used all the women with ages ranging from 21 to 79,

708 while Matsumoto et al.(23) calculated losses for women in each decade.

Discrimination of osteoporotic women Of equal interest to the comparisons of the two healthy groups (reflecting age- and menopause-related changes) are the comparisons of the measurements in their ability to discriminate between healthy postmenopausal and osteoporotic postmenopausal women. We used the Student’s t-test, odds ratios, and area under the ROC curve as statistical approaches to quantify the ability of the measurements to discriminate the osteoporotic postmenopausal group from the healthy postmenopausal group. Since most of the measurements are influenced by age, we adjusted for age in the logistic regression and ROC analysis. The results of odds ratios analysis reflected closely the trends shown by the Student’s t-test. In our study, QCT TRAB BMD and QCT INTG BMD gave the best results based on age-adjusted odds ratios, followed by DXA LAT BMD and DXA PA BMD. The same trend, with QCT demonstrating the highest sensitivity for distinction between normal and osteoporotic women, was found in studies by Guglielmi et al.(7) and by Jergas et al.,(35) who compared QCT with DXA LAT BMD and DXA PA BMD, by Pacifici et al.,(10) and by Van Berkum et al.,(46) who compared QCT with DXA PA BMD. Some appendicular measurements made in our study also gave a significant odds ratio but were generally lower compared with measurements at the spine. The parameters giving the highest odds ratio for each of these techniques were: for DXA of the femur, FEM TROC BMD; for DXA of radius, RAD TOT BMD; for pQCT of the radius, RAD CORT BMC; for RA of the hand, CO PHAL BMD; and for QUS of the calcaneus, SOS LA CALC. The trend of spine measurements showing better discrimination for vertebral fractures has been reported in many cross-sectional studies comparing the abilities of measurements of the spine with those of peripheral sites.(34,47–50) In contrast, it has been reported in both cross-sectional and longitudinal studies that peripheral measurements may be equal to spinal measurements when the latter are obtained by projectional techniques such as DXA. Cummings et al.(51) and Davis et al.(52) found that calcaneal DXA measurements were better than the DXA PA BMD for assessing hip fracture risk and age-related bone loss. Matsumoto et al.,(23) in a comparison of RA CH METC BMD and DXA PA BMD, found the methods were equivalent in their abilities to discriminate an osteoporotic group from a healthy group. Grampp et al., Bjarnson et al., and Overgaard et al. found that DXA of the radius had a discriminative ability comparable to spinal DXA.(15,53,54) Even with the trends shown in our study, the discriminative abilities of various measurements based on the ageadjusted ROC analysis were not significantly different. Due to the relatively small sample size of our study, the power for detecting small differences is not very great. Therefore, the comparison of odds ratios was only able to depict trends in our study and could become significant in larger populations. So with the small sample sizes in mind and the trends shown above, it can be suggested that the ability of

GRAMPP ET AL. QCT TRAB BMD for the discrimination of osteoporotic subjects is strongest, followed by QCT INTG BMD. DXA at the spine as well as other measurements at hip, radius, hand, and calcaneus also demonstrate these discriminatory abilities, similar to each other but weaker than those of QCT.

Correlations among measurement techniques Correlations between the different measurements ranged from weak to strong depending on the anatomic site and technique used. Average correlations among all the techniques/sites were only modest (Fig. 7) and were similar to those given in the recent literature.(7,10,11,15,17,19 –22,34 –36,45,47,55,56) Generally, correlations between measurements at the lumbar spine including QCT and DXA were modest to strong. Correlations between the various measurement ROIs at the femur using DXA were modest, and between the various calcaneus measurements using QUS weak to moderate. At the radius, measurements performed with pQCT and DXA did not show strong correlations, and trabecular and cortical bone parameters by pQCT were only weakly correlated. We found modest correlations when comparing axial and peripheral measurement sites. One remarkable point may be that in our study the comparison between X-ray based methods and ultrasound measurements were only marginally lower than those between X-ray based methods alone. This agrees with results given by other authors for measurements of BUA(40,41,50,57,58) and SOS.(41,59) The only study that reported substantially different results from ours for pQCT was from Rico et al.(60) In their study of radial pQCT, correlation coefficients for comparisons between total BMD and cortical BMD, total BMD and trabecular BMD, and trabecular BMD and cortical BMD were r 5 0.95, 0.62, and 0.43, respectively. This could be due to different study populations, a different definition of osteoporosis, and/or a different algorithm to differentiate cortical bone from soft tissue.

Diagnostic agreement among measurement techniques The selection of a specific threshold level for t score is rather arbitrary, and at any given point the prevalence of “disease” differs substantially by site and technique. In this exercise we selected a T score of 22 for osteopenia to achieve a reasonable balance between “disease” and “nondisease” in this elderly female population. The selection of T score of 22.5 for osteoporosis is compatible with WHO proposed criteria and perhaps appropriate for the vertebral fracture group. Considering the relatively modest correlation coefficients found between the different parameters, techniques, and sites, it is not surprising that by kappa analysis the diagnostic agreement among these measures in classifying women as osteopenic or osteoporotic was generally poor when using specific thresholds. The exception to this was the agreement of QCT TRAB/INTG BMD with DXA LAT BMD and with RA CO PHAL BMD. Often different patients were estimated at risk by different measurement

NONINVASIVE BONE MINERAL MEASUREMENTS approaches, (as indicated by the rather low kappa scores). These levels of diagnostic disagreement may result from a number of factors. All of the techniques have their own specific error sources as well as fundamentally different methodologies and units of expression that undermine their respective abilities to define true biological relationships within or between anatomic sites or between individuals.(1,61) Beyond technical considerations, genetic and environmental influences and pathophysiologic forces may impact the cortical and trabecular bone envelopes or the appendicular and axial skeleton in a different manner in different individuals,(36,49,52,62) and therefore some of the measured discrepancies reflect true anatomic variations that result in unlike diagnostic classifications. These observations have major clinical ramifications. The choice of the same T score (22.0 for osteopenia and 22.5 for osteoporosis in this study) for different sites and different techniques may not be advisable. More importantly, use of a single measure and a specific cut-off to diagnose and treat an individual patient is problematic and perhaps unwise. There are limitations in the interpretation of our study results. First, due to our small population size, many of the differences do not reach statistical significance and some apparent differences may not be real. Second, this study is cross-sectional in design and can provide inferences about vertebral fracture discrimination but no information specifically about fracture risk prediction or the ability to monitor disease or treatment-induced changes. Third, our results may be influenced by our definition for osteoporosis, requiring the presence of a vertebral fracture.(63) This diagnostic criterion may favor axial measurements over peripheral measurements, whereas selecting hip fracture or other long-bone fracture could provide different ranking of the methods. All of the techniques examined were capable of assessing age- and menopause-related changes, and of defining the skeletal status in these women. Spinal QCT appeared to provide the most robust performance overall. The other measurements at the axial or appendicular sites were relatively similar to each other. The results derived from the various techniques were only moderately correlative, which precluded the prediction of one parameter from another parameter in the individual woman and caused disagreement in diagnostic classification in many women.

709

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

ACKNOWLEDGMENTS 16.

We thank Cynthia Hayashi-Clark ARRT, Susan Schoen ARRT, and Vesta March ARRT for performing the measurements and analyses of the patient scans.

17.

REFERENCES 1. Genant HK, Engelke K, Fuerst T, Glu ¨er CC, Grampp S, Harris ST, Jergas M, Lang T, Lu Y, Majumdar S, Mathur A, Takada M 1996 Noninvasive assessment of bone mineral and structure: State of the art. J Bone Miner Res 11:707–730. 2. Black D, Cummings SR, Genant HK, Nevitt MC, Palermo L,

18.

19.

Browner W 1992 Axial and appendicular bone density predict fractures in older women. J Bone Miner Res 7:633– 638. Mosekilde L, Bentzen SM, Ørtoft G, Jorgensen J 1989 The predictive value of quantitative computed tomography for vertebral body compressive strength and ash density. Bone 10:465– 470. Spector TD, McCloskey EV, Doyle DV, Kanis JA 1993 Prevalence of vertebral fracture in women and the relationship with bone density and symptoms: The Chingford Study. J Bone Miner Res 8:817– 822. Grampp S, Jergas M, Glu ¨er CC, Lang P, Brastow P, Genant HK 1993 Radiological diagnosis of osteoporosis: Current methods and perspectives. Radiol Clin N Am 31:1133–1145. Genant HK, Cann CE, Ettinger B, Gordan GS 1982 Quantitative computed tomography of vertebral spongiosa: A sensitive method for detecting early bone loss after oophorectomy. Ann Intern Med 97:699 –705. Guglielmi G, Grimston SK, Fischer KC, Pacifici R 1994 Osteoporosis: Diagnosis with lateral and posteroanterior dual x-ray absorptiometry compared with quantitative CT. Radiology 192:845– 850. Genant HK, Glu ¨er CC, Steiger P, Faulkner KG 1992 Quantitative computed tomography for the assessment of osteoporosis. In: Moss AA, Gamsu G, Genant HK (eds.) Tomography of the Body, Vol. 2. W.B. Saunders, Philadelphia, PA, U.S.A., pp. 523–549. Ito M, Hayashi K, Yamada M, Uetani M, Nakamura T 1993 Relationship of osteophytes to bone mineral density and spinal fracture in men. Radiology 189:497–502. Pacifici R, Rupich R, Griffin M, Chines A, Susman N, Avioli LV 1990 Dual energy radiography versus quantitative computer tomography for the diagnosis of osteoporosis. J Clin Endocrinol Metab 70:705–710. Grampp S, Jergas M, Lang P, Steiner E, Fuerst T, Glu ¨er CC, Mathur A, Genant HK 1996 Quantitative CT assessment of the lumbar spine and radius in patients with osteoporosis. AJR Am J Roentgenol 167:133–140. Ru ¨egsegger P, Durand E, Dambacher MA 1991 Localization of regional forearm bone loss from high resolution computed tomographic images. Osteoporos Int 1:76 – 80. Ru ¨egsegger P, Durand EP, Dambacher MA 1991 Differential effects of aging and disease on trabecular and compact bone density of the radius. Bone 12:99 –105. Schneider P, Bo ¨rner W 1991 Periphere quantitative Computertomographie zur Knochenmineralmessung mit einem neuen speziellen QCT-Scanner. Fortschr Ro ¨ntgenstr 154:292–299. Grampp S, Lang P, Jergas M, Glu ¨er CC, Mathur A, Engelke K, Genant HK 1995 Assessment of the skeletal status by peripheral quantitative computed tomography of the forearm: Shortterm precision in-vivo and comparison to dual x-ray absorptiometry. J Bone Miner Res 10:1566 –1576. Grampp S, Majumdar S, Jergas M, Newitt D, Lang P, Genant HK 1996 Distal radius: In vivo assessment with quantitative MR imaging, peripheral quantitative CT, and dual x-ray absorptiometry. Radiology 198:213–218. Kleerekoper M, Nelson DA, Flynn MJ, Pawluszka AS, Jacobsen G, Peterson EL 1994 Comparison of radiographic absorptiometry with dual-energy x-ray absorptiometry and quantitative computed tomography in normal older white and black women. J Bone Miner Res 9:1745–1749. Yang S-O, Hagiwara S, Engelke K, Dhillon MS, Guglielmi G, Bendavid EJ, Soejima O, Nelson D, Genant HK 1994 Radiographic absorptiometry for bone mineral measurement of the phalanges: Precision and accuracy study. Radiology 192:857– 859. Cosman F, Herrington B, Himmelstein S, Lindsay R 1991

710

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

GRAMPP ET AL. Radiographic absorptiometry: A simple method for determination of bone mass. Osteoporos Int 2:34 –38. Derisquebourg T, Dubois P, Devogelaer JP, Meys E, Duquesnoy B, Nagant de Deuxchaisnes C, Delcambre B, Marchandise X 1994 Automated computerized radiogrammetry of the second metacarpal and its correlation with absorptiometry of the forearm and spine. Calcif Tissue Int 54:461– 465. Fukunaga M, Tomomitsu T, Otsuka N, Imai H, Morita R, Nishii Y 1990 Indexes of bone mineral content on second metacarpal bone roentgenogram analyzed by digital image processing: A comparison with other bone mass quantifying methods. Radiat Med 8:230 –235. Hayashi Y, Yamamoto K, Fukunaga M, Ishibashi T, Takahashi K, Nishii Y 1990 Assessment of bone mass by image analysis of metacarpal bone roentgenograms: A quantitative digital image processing (DIP) method. Radiat Med 8:173–178. Matsumoto C, Kushida K, Yamazaki K, Imose K, Inoue T 1994 Metacarpal bone mass in normal and osteoporotic Japanese women using computed x-ray densitometry. Calcif Tissue Int 55:324 –329. Baran DT, Kelly AM, Karellas A, Gionet M, Price M, Leahy D, Steuterman S, McSherry B, Roche J 1988 Ultrasound attenuation of the os calcis in women with osteoporosis and hip fractures. Calcif Tissue Int 43:138 –142. Heaney RP, Avioli LV, Chestnut CH, Lappe J, Recker RR, Brandburger GH 1989 Osteoporotic bone fragility: Detection by ultrasound transmission velocity. JAMA 261:2986 –2990. Agren M, Karellas A, Leahey D, Marks S, Baran D 1991 Ultrasound attenuation of the calcaneus: A sensitive and specific discriminator of osteopenia in postmenopausal women. Calcif Tissue Int 48:240 –244. Glu ¨er CC, Wu CY, Jergas M, Goldstein SA, Genant HK 1994 Three quantitative ultrasound parameters reflect bone structure. Calcif Tissue Int 55:46 –52. Genant HK, Grampp S, Mathur A, Glu ¨er CC, Jergas M, Takada M, Lang P, Lu Y, Chavez M 1996 Comparisons of non-invasive bone mineral measurements in assessing agerelated loss, fracture discrimination, and diagnostic classification. Osteoporos Int 6:76 –110. Genant HK, Wu CY, van Kuijk C, Nevitt M 1993 Vertebral fracture assessment using a semi-quantitative technique. J Bone Miner Res 8:1137–1148. Genant HK, Cann CE, Pozzi-Mucelli RS, Kanter AS 1983 Vertebral mineral determination by quantitative computed tomography: Clinical feasibility and normative data. J Comput Assist Tomogr 7:554 –559. Steiger P, Block JE, Steiger S, Heuck A, Friedlander A, Ettinger B, Harris ST, Glu ¨er CC, Genant HK 1990 Spinal bone mineral density by quantitative computed tomography: Effect of region of interest, vertebral level, and technique. Radiology 175:537–543. Cann CE, Genant HK 1980 Precise measurement of vertebral mineral content using computed tomography. J Comp Assist Tomogr 4:493–500. Cann CE, Genant HK, Kolb FO, Ettinger B 1985 Quantitative computed tomography for prediction of vertebral fracture risk. Bone 6:1–7. Reinbold W, Genant H, Reiser U, Harris S, Ettinger B 1986 Bone mineral content in early-postmenopausal and postmenopausal osteoporotic women: comparison of measurement method. Radiology 160:469 – 478. Jergas M, Breitenseher M, Glu ¨er CC, Yu W, Genant HK 1995 Estimates of volumetric bone density from projectional measurements improve the discriminatory capability of dual x-ray absorptiometry. J Bone Miner Res 10:1101–1110. Steiger P, Cummings SR, Black DM, Spencer NE, Genant HK

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

1992 Age-related decrements in bone mineral density in women over 65. J Bone Miner Res 7:625– 632. Greenspan SL, Maitland LA, Myers ER, Krasnow MB, Kido TH 1994 Femoral bone loss progresses with age: A longitudinal study in women over age 65. J Bone Miner Res 9:1959 –1965. Jones CD, Laval-Jeantet AM, Laval-Jeantet MH, Genant HK 1987 Importance of measurement of spongious vertebral bone mineral density in the assessment of osteoporosis. Bone 8:201–206. Genant H, Cann C, Ettinger B, Gordan G 1982 Determination of bone mineral loss in the axial skeleton of oophorectomized women using quantitative computed tomography. J Comput Assist Tomogr 6:217–218. Schott AM, Hans D, Sornay-Rendu E, Delmas PD, Meunier PJ 1993 Ultrasound measurements on os calcis: Precision and age-related changes in a normal female population. Osteoporos Int 3:249 –254. Moris M, Peretz A, Tjeka R, Negeban N, Wouters M, Bergmann P 1995 Quantitative ultrasound bone measurements: Normal values and comparison with bone mineral density by dual x-ray absorptiometry. Calcif Tissue Int 57:6 –10. Mazess RB, Peppler WW, Chesney RW, Lange TA, Lindgren U, Smith E 1983 Does bone measurement on the radius indicate skeletal status? Concise communication. J Nucl Med 25:281–288. Ga¨rdsell P, Johnell O, Nilsson BE, Gullberg B 1993 Predicting various fragility fractures in women by forearm bone densitometry: a follow-up study. Calcif Tissue Int 52:348 –353. Riis BJ, Christiansen C 1988 Measurement of spinal or peripheral bone mass to estimate early postmenopausal bone loss? Am J Med 84:646 – 653. Butz S, Wu ¨ster C, Scheidt-Nave C, Go ¨tz M, Ziegler R 1994 Forearm BMD as measured by peripheral quantitative computed tomography (pQCT) in a German reference population. Osteoporos Int 4:179 –184. Van Berkum FN, Birkenha¨ger JC, Van Veen LC, Zeelenberg J, Birkenha¨ger-Frenkel DH, Trouerbach WT, Stijnen T, Pols HA 1989 Noninvasive axial and peripheral assessment of bone mineral content: A comparison between osteoporotic women and normal subjects. J Bone Miner Res 4:679 – 685. Heuck AF, Block J, Glu ¨er CC, Steiger P, Genant HK 1989 Mild versus definite osteoporosis: Comparison of bone densitometry techniques using different statistical models. J Bone Miner Res 4:891–900. Bagur A, Vega E, Mautalen C 1994 Age dependence of the normal/abnormal difference of bone mineral density in osteoporotic women. Bone Miner 26:209 –218. Riggs BL, Wahner HW, Dunn WL, Mazess RB, Offord KP, Melton III LJ 1981 Differential changes in bone mineral density of the appendicular and axial skeleton with aging. J Clin Invest 67:328 –335. Herd RJ, Blake GM, Miller CG, Parker JC, Fogelman I 1994 The ultrasonic assessment of osteopenia as defined by dual x-ray absorptiometry. Br J Radiol 67:631– 635. Cummings SR, Black DM, Nevitt MC, Browner W, Cauley J, Ensrud K, Genant HK, Palermo L, Scott J, Vogt TM 1993 Bone density at various sites for prediction of hip fractures: the study of osteoporotic fractures. Lancet 341:72–75. Davis JW, Ross PD, Wasnich RD 1994 Evidence for both generalized and regional low bone mass among elderly women. J Bone Miner Res 9:305–309. Overgaard K, Hansen MA, Riis BJ, Christiansen C 1992 Discriminatory ability of bone mass measurements (SPA and DEXA) for fractures in elderly postmenopausal women. Calcif Tissue Int 50:30 –35. Bjarnson K, Nilas L, Hassager C, Christiansen C 1995 Dual

NONINVASIVE BONE MINERAL MEASUREMENTS

55.

56.

57.

58.

59.

60.

energy x-ray absorptiometry of the spine: Decubitus lateral versus anteroposterior projection in osteoporotic women: Comparison to single energy x-ray absorptiometry of the forearm. Bone 16:255–260. Yu W, Glu ¨er CC, Grampp S, Jergas M, Fuerst T, Wu CY, Lu Y, Fan B, Genant HK 1995 Spinal bone mineral assessment in postmenopausal women: A comparison between dual x-ray absorptiometry and quantitative computed tomography. Osteoporos Int 5:433– 439. Overton TR, Wheeler GD 1992 Bone mass measurements in the distal forearm using dual-energy x-ray absorptiometry and gamma-ray computed tomography: A longitudinal, in vivo comparative study. J Bone Miner Res 7:375–381. Massie A, Reid DM, Porter RW 1993 Screening for osteoporosis: Comparison between dual energy x-ray absorptiometry and broadband ultrasound attenuation in 1000 perimenopausal women. Osteoporos Int 3:107–110. Young H, Howey S, Purdie DW 1993 Broadband ultrasound attenuation compared with dual-energy x-ray absorptiometry in screening for postmenopausal low bone density. Osteoporos Int 3:160 –164. Faulkner KG, McClung MR, Coleman LJ, Kingston-Sandahl E 1994 Quantitative ultrasound of the heel: Correlation with densitometric measurements at different skeletal sites. Osteoporos Int 4:42– 47. Rico H, Gonzalez-Riola J, Revilla M, Villa LF, Gomez-Cas-

711 tresana F, Escribano J 1994 Cortical versus trabecular bone mass: Influence of activity on both bone components. Calcif Tissue Int 54:470 – 472. 61. Genant HK, Lu Y, Mathur AK, Fuerst TP, Cummings SR 1996 Classification based on DXA measurements for assessing the risk of hip fractures. J Bone Miner Res 11(suppl 1): S120. 62. Hannan M, Felson D, Anderson JJ 1992 Bone mineral density in elderly men and women: Results from the Framingham osteoporosis study. J Bone Miner Res 7:547–553. 63. Consensus Development Conference 1993 Diagnosis, prophylaxis, and treatment of osteoporosis. Am J Med 94:646 – 650.

Address correspondence and reprint requests to: Harry K. Genant, M.D. Professor of Radiology, Medicine, and Orthopedic Surgery Executive Director, Osteoporosis Research Group Department of Radiology Musculoskeletal Section University of California–San Francisco 505 Parnassus Avenue Box 0628 San Francisco, CA 94143 U.S.A. Received in original form August 14, 1996; in revised form November 1, 1996; accepted December 5, 1996.

Comparisons of noninvasive bone mineral ... - Wiley Online Library

techniques: (1) quantitative computed tomography of the L1–L4 lumbar .... used in the data analysis). ... were applied to the bone image data to compute bone.

2MB Sizes 6 Downloads 193 Views

Recommend Documents

Comparisons of noninvasive bone mineral ... - Wiley Online Library
0.40. 0.37. 0.31. 0.41. 0.37. 0.34. 0.42. 0.29. 0.27. 0.28. 0.45. 0.52. 1. 0.24. 0.56. WS. 39.2. 23.3. 16.3. 17.2. 18.6. 18.1. 15.1. 23.9. 36.0. 16.2. 14.9. 22.2. 12.1. 2.4.

Comparisons of noninvasive bone mineral measurements
(QUS) of the calcaneus for speed of sound (SOS CALC) and broadband ultrasound attenuation (BUA CALC). In ... techniques providing the best results were: (1) spine, QCT TRAB BMD (annual loss, 1.2% ... The only established techniques for noninvasive ..

Paired comparisons of carbon exchange ... - Wiley Online Library
Jun 10, 2002 - Abstract. The effects of harvest on European forest net ecosystem exchange (NEE) of carbon and its photosynthetic and respiratory components (GPP (gross primary production) and. TER (total ecosystem respiration)) were examined by compa

ELTGOL - Wiley Online Library
ABSTRACT. Background and objective: Exacerbations of COPD are often characterized by increased mucus production that is difficult to treat and worsens patients' outcome. This study evaluated the efficacy of a chest physio- therapy technique (expirati

The Metaphysics of Emergence - Wiley Online Library
University College London and Budapest University of. Technology and Economics. I. Mental Causation: The Current State of Play. The following framework of ...

Competing paradigms of Amazonian ... - Wiley Online Library
September 2014, immediately after the accepted version of this manuscript was sent to the authors on 18 September. 2014. doi:10.1111/jbi.12448. Competing ..... species are there on earth and in the ocean? PLoS Biology, 9, e1001127. Moritz, C., Patton

Principles of periodontology - Wiley Online Library
genetic make-up, and modulated by the presence or ab- sence of ... can sense and destroy intruders. ..... observation that apparently healthy Japanese sub-.

poly(styrene - Wiley Online Library
Dec 27, 2007 - (4VP) but immiscible with PS4VP-30 (where the number following the hyphen refers to the percentage 4VP in the polymer) and PSMA-20 (where the number following the hyphen refers to the percentage methacrylic acid in the polymer) over th

Recurvirostra avosetta - Wiley Online Library
broodrearing capacity. Proceedings of the Royal Society B: Biological. Sciences, 263, 1719–1724. Hills, S. (1983) Incubation capacity as a limiting factor of shorebird clutch size. MS thesis, University of Washington, Seattle, Washington. Hötker,

Kitaev Transformation - Wiley Online Library
Jul 1, 2015 - Quantum chemistry is an important area of application for quantum computation. In particular, quantum algorithms applied to the electronic ...

Hormonal regulation of appetite - Wiley Online Library
E-mail: [email protected]. Hormonal regulation of appetite. S. Bloom. Department of Metabolic Medicine, Imperial. College London, London, UK. Keywords: ...

PDF(3102K) - Wiley Online Library
Rutgers University. 1. Perceptual Knowledge. Imagine yourself sitting on your front porch, sipping your morning coffee and admiring the scene before you.

Standard PDF - Wiley Online Library
This article is protected by copyright. All rights reserved. Received Date : 05-Apr-2016. Revised Date : 03-Aug-2016. Accepted Date : 29-Aug-2016. Article type ...

Authentic inquiry - Wiley Online Library
By authentic inquiry, we mean the activities that scientists engage in while conduct- ing their research (Dunbar, 1995; Latour & Woolgar, 1986). Chinn and Malhotra present an analysis of key features of authentic inquiry, and show that most of these

TARGETED ADVERTISING - Wiley Online Library
the characteristics of subscribers and raises advertisers' willingness to ... IN THIS PAPER I INVESTIGATE WHETHER MEDIA TARGETING can raise the value of.

Verbal Report - Wiley Online Library
Nyhus, S. E. (1994). Attitudes of non-native speakers of English toward the use of verbal report to elicit their reading comprehension strategies. Unpublished Plan B Paper, Department of English as a Second Language, University of Minnesota, Minneapo

PDF(270K) - Wiley Online Library
tested using 1000 permutations, and F-statistics (FCT for microsatellites and ... letting the program determine the best-supported combina- tion without any a ...

Phylogenetic Systematics - Wiley Online Library
American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024. Accepted June 1, 2000. De Queiroz and Gauthier, in a serial paper, argue that state of biological taxonomy—arguing that the unan- nointed harbor “wide

PDF(270K) - Wiley Online Library
ducted using the Web of Science (Thomson Reuters), with ... to ensure that sites throughout the ranges of both species were represented (see Table S1). As the ...

Standard PDF - Wiley Online Library
Ecology and Evolutionary Biology, University of Tennessee, Knoxville, TN 37996, USA,. 3Department of Forestry and Natural. Resources, Purdue University ...