- Email: [email protected]
Normative Data for Digital X-Ray Radiogrammetry From a Female and Male German Cohort
Journal of Clinical Densitometry, vol. 9, no. 3, 341e350, 2006 Ó Copyright 2006 by The International Society for Clinical Densitometry 1094-6950/06/9:341e350/$32.00 DOI: 10.1016/j.jocd.2006.05.010
Original Article
Normative Data for Digital X-Ray Radiogrammetry From a Female and Male German Cohort Joachim Bo¨ttcher,*,1 Alexander Pfeil,1 Max L. Scha¨fer,1 Alexander Petrovitch,1 Bettina E. Seidl,1 Hans-Joachim Mentzel,1 Gabriele Lehmann,2 Ansgar Malich,3 Jens-Peter Heyne,1 Gert Hein,2 Gunter Wolf,2 and Werner A. Kaiser1 1
Institute of Diagnostic and Interventional Radiology, Friedrich-Schiller-University Jena, Jena, Germany; 2Department of Rheumatology and Osteology, Clinic of Internal Medicine III, Friedrich-Schiller-University Jena, Jena, Germany; and 3 Sued-Harz Klinikum, Department of Radiology, Nordhausen, Germany
Abstract This study presents German reference data for digital X-ray radiogrammetry (DXR) differentiated by males as well as females, and quantifies for gender-specific and age-related differences including all DXR parameters. This study also documents the effects of different X-ray settings (e.g., radiographs of the wrist or the hand) on DXR measurements. There were 2085 patients who were prospectively enrolled (954 females and 1131 males) from a data pool of 11,915 patients with radiographs of the nondominant hand or wrist. All patients underwent measurements of bone mineral density (BMD), cortical thickness, bone width, and the metacarpal index (MCI) using DXR technology. These data showed a continuous age-related increase of the DXR parameters to the point of peak bone mass, then a continuous decline beyond the peak bone mass with accentuated age-related cortical bone loss in women. Peak bone mass is reached at approximately 30e34 yr for women and 45e49 yr for men. In addition, men had a significantly higher DXR BMD (mean: þ12.8%) compared with woman in all age groups. Regarding the impact of various X-ray settings (e.g., X-raywrist vs. X-rayhand), no significant difference was observed between both groups, men as well as women. The development of digital imaging technology has enabled more precise measurements of several radio-geometric features. The present study estimated normative reference values for DXR in German Caucasian women and men. Based on this reference data, a valid and reliable quantification of disease-related demineralization based on measurements of DXR BMD and MCI is now available for the Caucasian ethnic group. Key Words: Bone mineral density; bone width; digital X-ray radiogrammetry; metacarpal index; normative data.
is a radio-geometrical technique used to estimate the thickness of the cortical bone partition. For more than 40 yr, conventional radiogrammetry has been the primary method for assessment of bone status from hand radiographs (6e8). This method measures total width of a bone and the medullary width at the midpoint of the second metacarpal of the nondominant hand. The measured data are typically used to calculate various indices, such as the metacarpal index (i.e., the ratio of total bone width and cortical thickness), the Barnett and Nordin index (i.e., the percentage of cortical thickness) (6), the Garn index (i.e., the cortical area) (9), and the Exton-Smith index (i.e., related to the cortical area and the surface area) (10), and thus to quantify changes of the cortical bone.
Introduction Digital X-ray radiogrammetry (DXR) is a new diagnostic tool for the measurement of cortical bone mineral density (BMD) on metacarpals using digitized radiographs (1e5). The DXR is based on conventional radiogrammetry, which
Received 01/18/06; Revised 05/22/06; Accepted 06/03/06. *Address correspondence to: Joachim Bo¨ttcher, MD, Institute of Diagnostic and Interventional Radiology, Friedrich-Schiller-University Jena, Erlanger Allee 101, Jena, 07747 Germany. E-mail: [email protected]
341
342 Conventional radiogrammetry and the previously mentioned indices are most sensitive to cortical bone changes (i.e., periosteal apposition and especially endosteal resorption). Despite conventional radiogrammetry use being relatively inexpensive and widely available, it is limited because of its imprecision (from 2.4% up to 11%) due to the difficulty in identification of the endosteal margin and precise marking of the midshaft location by the operator-dependent radiogrammetry measurements (11e13). Digital X-ray radiogrammetry is a new operator-independent diagnostic tool, providing automated measurement of cortical BMD on the metacarpals using digitized radiographs with an excellent intra-radiograph (0.05e0.33%) and interradiograph reproducibility (0.26e1.54%). The high precision is based on refinement, computerization, and the implementation of algorithms for automatic and operator-independent image analysis by the DXR technology (14,15), which is based on a combined computerized radiogrammetry and textural analysis of the 3 middle metacarpal bones. Although DXR BMD estimates were not taken at a point known for high incidence of fracture, in many studies DXR has confirmed its predictive value regarding fracture risk (1,2), has provided evidence for therapeutic effects in the treatment of postmenopausal osteoporosis (3), and has revealed closed associations to dual-energy X-ray absorptiometry (DXA) measurements of the spine, femur, and forearm (5,16). The DXR-technique seems to bridge the gap between radiogrammetry and densitometry by estimating BMD from radiogrammetry measurements in case of primary and secondary osteoporosis. Another promising application of DXR is the quantification of periarticular bone loss in rheumatoid arthritis. In comparison with other osteodensitometric techniques (DXA, peripheral quanitative computer tomography, quantitative ultrasound), DXR could detect and quantify the diseaserelated cortical bone loss caused by the course and severity of rheumatoid arthritis (17e21). A third potential application of DXR technology is the determination of peripheral bone status in childhood (22,23). Children with cystic fibrosis, Turner syndrome, Marfan syndrome, osteogenesis imperfecta, inflammatory bowel diseases, and renal disorders often have disease-related or therapy-induced cortical osteopenia develop, frequently requiring radiographs of the hand for disease monitoring (based on bone age determination because of identification of diseaserelated retardation), which could be used for DXR estimates. The aim of this study was to present German reference data of DXR for Caucasian men and women to quantify the gender-specific as well as age-related differences of DXR BMD, and to evaluate the effects of different X-ray settings on DXR measurements.
Patients and Methods Patients This prospective study enrolled 2085 patients (954 females and 1131 males) between 2001 and 2005 from a data pool of Journal of Clinical Densitometry
Bo¨ttcher et al. 11,915 patients with hand or wrist radiographs, which were performed during admission in the emergency unit of a university in the Eastern part of Germany to exclude fractures due to trauma. As an important part of our study design, a detailed questionnaire was completed by all subjects to record possible exclusion criteria mentioned as follows. The remaining 2085 healthy individuals were of the Caucasian ethnic group and received digital radiographs of the nondominant hand or wrist for the DXR estimates. All patients underwent measurements of BMD (g/cm2), cortical thickness (cm), bone width (BW) (cm) and metacarpal index (MCI), a dimensionless parameter (based on the ratio of cortical thickness to bone diameter) using DXR technology. Mean age was 36.1 yr (men: 33.5 years 6 17.4; women: 39.1 6 21.7) with a standard deviation of 19.4 yr, and an age range of 6.6 to 105.6 yr (men: 7.3e99.6 yr; women: 6.6e105.6 yr). The following exclusion criteria were considered based on the extensive questionnaire: visible metallic material (i.e., splints and material after osteosynthesis; n 5 899); signs of fracture (n 5 4867) and amputation (n 5 38); endocrinological diseases known to affect bone metabolism (e.g., hyper/hypoparathyroidism, Cushing disease; n 5 245) and genetic diseases (n 5 23); rheumatic diseases (e.g., rheumatoid arthritis; n 5 1237) renal disorders (n 5 1084), and oncological diseases (n 5 1255); medication with bone-influencing drugs (e.g., steroids, vitamin D, or calcium intake; n 5 182); and only Caucasian German patients were enrolled (e.g., verification that birthplaces of the subjects and also their parents and grandparents were located in Germany). A further exclusion criteria was the existence of fractures in the dominant upper extremities, which frequently induce osteopenia caused by immobility on the unmeasured side of fracture, but otherwise cause accentuated mineralization on the nondominant healthy side based on raised physical activity. The population was divided into 16 age-related groups differentiated for gender.
DXR The Pronosco X-Posure System (version 2.0, Sectra Pronosco A/S, Herlev, Denmark), which involves a radiogrammetric and textural analysis of the 3 middle metacarpal bones, was used to determine DXR BMD and MCI based on DXR, requiring radiographs of the nondominant hand. All plain radiographs were acquired by a Polydoros SX 80 (Siemens, Munich, Germany) with the following standardized conditions: filter with 1.0-mm thickness related to aluminium 80, tube voltage 42 kV, exposure level 4 mAs, film focus distance 100 cm (Scopix Laser 2 B 400; Agfa GmbH and Cie. KG, Cologne, Germany). The digital radiographs were printed and subsequently scanned into the system at a resolution of 300 dpi, corresponding to 118 pixels per cm (16). The system itself checked the quality of the scanned images and interrupted Volume 9, 2006
DXR Normative Data the examination in case of inadequate quality. The computer algorithms automatically defined regions of interest (ROIs) around the narrowest bone parts of the metacarpal II, III, and IV and subsequently specified the outer and inner cortical edges of the studied cortical bone parts. To locate the diaphysis of the 3 middle metacarpals in the radiograph (16,24), the DXR system applies a model-based algorithm known as the active shape model. After each diaphysis has been identified, RO’s are placed automatically for the 3 metacarpals and the endosteal (inner) and periosteal (outer) edges are automatically found. Apart from placing the radiography on the chargedcoupled device-based desktop flatbed scanner, otherwise there is no operator interaction connected to the DXR measurement. The analyzed images and their ROIs are displayed on the computer monitor. After calculation of the mean of the cortical thickness, the cortical volume per area (VPA) is computed for each metacarpal assuming a cylindrically-shaped bone based on cortical thickness and mean outer bone diameter as given as follows by Lazenby (25) and Rosholm et al. (16): The final DXR BMD, based on the mean VPA, is then calculated with the following correction for the estimated porosity index (P): DXR BMD 5 c VPA ð1 PÞ: The scaling constant c is determined so that DXR BMD on average is equal to that of the mid-distal forearm region of the Hologic QDR 2000 densitometer (Hologic, Waltham, MA). The porosity index is a technical parameter given as a value between 1 and 19, which is derived from the area percentage of local intensity minima found in the cortical part of the bone relative to the entire cortical area (16). Consequently, the porosity index represents an estimated 3-dimensional cortical porosity, aimed to be the fraction of the cortical bone volume that is not occupied by bone. The MCI obtains the mean cortical thickness normalized with the mean outer bone diameter (BW) for each bone part (16,23,26).
Ethics All examinations were performed in accordance with the rules and regulations of the local human research and ethics committee. As a special note, the authors emphasize that all radiographs used for DXR calculations were performed as part of routine clinical care (e.g., exclusion of fractures due to relevant trauma). No additional radiographs were obtained only for study purposes.
Data Analysis The objective of the statistical analysis was to establish normative values for both DXR BMD and MCI used in Tscore and Z-score calculations. To pool data from the sites into the normative reference database, homogeneity of the data across the sites was investigated. This intention was Journal of Clinical Densitometry
343 accomplished by using a regression model of BMD versus study site indicator variables, age, weight, and height on data for young women. A F-test was calculated if the differences between the sites were significantly different from zero. The modeling of the reference curves for BMD and MCI were achieved by applying the power of the mean model. Prior to regression analysis, a standard test for heteroscedasticity was performed. The reference curves were based on a regression of DXR BMD and MCI versus age and age2 (second order polynomials). The significance of sex-dependent changes was calculated with the ManneWhitney U test. Regarding the significance of differences between X-rayhand and X-raywrist, changes were calculated with the t test for unpaired groups to verify the parity of both groups and in an opposite direction with the ManneWhitney U test to identify significant differences between both X-ray settings. The statistical analysis was performed using SPSS, version 10.13 (SPSS, Chicago, IL) and using SAS System, version 8 (SAS Institute Inc, Cary, NC).
Results Reproducibility of DXR The inter-radiograph reproducibility (evaluated by one DXR analysis of 10 repeated acquired radiographs of the same hand with repositioning under standard X-ray settings): DXR BMD : 0:31%
MCI : 0:24%
The intra-radiograph reproducibility (evaluated by a single digital image that underwent 10 repeated DXR analyses): DXR BMD : 0:09%
MCI : 0:14%
DXR Reference Data See Tables 1 and 2, and Figs. 1 and 2.
DXR Parameters and Age-Related Alterations The DXR system reliably provides recognition of the metacarpals of subjects with a bone age of 6 years and older. This study shows a close relationship between age and all measured DXR parameters. The DXR BMD significantly increased in men (þ41.1%) from 0.460 6 0.072 g/cm2 to 0.649 6 0.067 g/cm2 up to an age of 45 years in a continuous manner. In women, DXR BMD showed a significant and continuous increase (þ30.0%) up to an age of 35 yr from 0.450 6 0.057 g/cm2 to 0.585 6 0.066 g/cm2. Equal results could be verified for MCI with a significant increase from 0.374 6 0.053 to 0.477 6 0.062 in men (þ27.5%), and from 0.399 6 0.065 to 0.511 6 0.063 in women (þ28.1%). A significant increase (þ47.9%) for DXR cortical thickness in men from 0.144 cm 6 0.029 cm (age: 10 yr) to 0.213 cm 6 0.029 (age: 45 yr) was observed. The relative increase of DXR cortical thickness for the female cohort was þ38.3%. Volume 9, 2006
344
Bo¨ttcher et al. Table 1 Mean and Standard Deviation Regarding the Parameters of Digital X-Ray Radiogrammetry Men, X-rayhand
Age (yr) !15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
n
BMD (g/cm2) Mean (SD)
93 200 183 121 93 104 75 65 50 46 23 24 19 17 7 11 1131
0.460 0.596 0.619 0.625 0.625 0.634 0.641 0.649 0.627 0.626 0.615 0.585 0.566 0.559 0.501 0.478 0.603
(0.072) (0.074) (0.078) (0.077) (0.091) (0.079) (0.069) (0.067) (0.080) (0.079) (0.053) (0.076) (0.079) (0.084) (0.040) (0.089) (0.091)
and wrist,
n 5 1131
MCI Mean (SD) 0.374 0.459 0.476 0.482 0.481 0.481 0.465 0.477 0.463 0.441 0.430 0.415 0.382 0.367 0.337 0.335 0.456
(0.053) (0.056) (0.054) (0.064) (0.066) (0.056) (0.060) (0.062) (0.054) (0.061) (0.058) (0.060) (0.056) (0.057) (0.063) (0.070) (0.069)
CT (cm) Mean (SD) 0.144 0.195 0.204 0.206 0.205 0.209 0.209 0.213 0.205 0.203 0.192 0.185 0.175 0.170 0.147 0.147 0.196
(0.029) (0.028) (0.030) (0.030) (0.036) (0.030) (0.029) (0.029) (0.029) (0.030) (0.022) (0.030) (0.027) (0.029) (0.024) (0.024) (0.035)
BW (cm) Mean (SD) 0.762 0.849 0.859 0.863 0.855 0.870 0.901 0.895 0.876 0.922 0.916 0.893 0.923 0.938 0.911 0.924 0.864
(0.082) (0.091) (0.090) (0.100) (0.097) (0.098) (0.078) (0.080) (0.081) (0.084) (0.068) (0.085) (0.106) (0.107) (0.103) (0.049) (0.097)
Abbr: BMD, bone mineral density; BW, bone width; CT, cortical thickness; MCI, metacarpal index; SD, standard deviation.
In both sexes, DXR BW showed a gender-specific relative increase with þ17.5 % in men versus þ7.0% in women. Both women and men show a continuous decline of DXR BMD in subjects older than 50 yr, which was accentuated in the group of women.
Comparison of DXR Parameters in Men and Women See Table 3 and Fig. 3. The results show that the German women cohort has a significantly lower age-related DXR BMD (mean: 12.8%) compared with the male cohort. The female peak bone mass for DXR BMD is reached in the age group of 30e34 yr (0.585 6 0.066 g/cm2). The male peak bone mass reveals their highest values in the age interval of 45e49 yr (0.649 6 0.067 g/cm2).
Comparison of Hand and Wrist Radiographs The normative values of DXR BMD revealed no significant difference (3.2%; t test: p ! 0.01 and ManneWhitney U test: p O 0.05) between radiographs of the male wrist (0.618 g/ cm2 6 0.078) and X-rays of the hand (0.599 g/ cm2 6 0.094). Comparable results are documented for the MCI in men with 2.7% (0.459 6 0.068 vs. 0.447 6 0.070). No significant difference between X-rayshand (DXR BMD: 0.521 g/cm2 6 0.089; MCI: 0.448 6 0.085) and X-rayswrist (DXR BMD: 0.537 g/cm2 6 0.082; MCI: 0.439 6 0.090) was observed for DXR BMD (3.1%) and MCI (2.0%) in women. Similar data were revealed for DXR cortical thickness and also for DXR BW in males (2.5% and 5.3%) as well as in the female cohort (3.0% and 5.0%). Journal of Clinical Densitometry
Annual Bone Loss See Tables 4 and 5. Both men and women show an accentuated increase of cortical bone mass in the age group from 15 to 19 yr for DXR BMD (men: 29.6%; women: 20.4%). Both the data for men and women revealed a continuous increase of the DXR BMD up to the peak bone mass (annual bone gain for men: 0.29%; annual bone gain for women: 0.51%). After surpassing the peak bone mass, a continuous decline of the DXR BMD could be observed that was accentuated in women (annual bone loss for men: 0.91%; annual bone loss for women: 0.86%). A most relevant cortical demineralization was found in the age range from 60 to 69 yr with an annual bone loss of 1.62% in postmenopausal women; in men, two delayed peaks of demineralization could be described in the age group from 65 to 69 yr (5.1%) and between the ages of 80 and 84 yr (11.6%). Comparable results were revealed for the MCI.
Discussion The intention of this study was to estimate normative German reference data for DXR in men and women. Currently, reference data for DXR parameters have been established for Caucasian North-American (27), African American (28), Scandinavian (29), German (30,31), Chinese (32), Hispanic (33), Korean (34), and Indian populations. In contrast to this study, most of the published reference data are only available for women (27,28,30). Recently it has been recognized that osteoporosis represents an important public health issue in men as well (35). Since the epidemiological data have Volume 9, 2006
DXR Normative Data
345
Table 2 Mean and Standard Deviation Regarding the Parameters of Digital X-Ray Radiogrammetry Women, X-Rayhand Age (yr)
BMD (g/cm2) mean (SD)
n 96 106 127 82 75 72 56 59 57 41 37 42 22 25 26 31 954
!15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
0.450 0.542 0.552 0.574 0.585 0.573 0.565 0.566 0.552 0.527 0.490 0.451 0.456 0.429 0.415 0.385 0.526
and wrist;
n 5 954
MCI mean (SD)
(0.057) (0.066) (0.064) (0.069) (0.066) (0.068) (0.056) (0.066) (0.081) (0.070) (0.073) (0.061) (0.069) (0.064) (0.080) (0.057) (0.087)
0.399 0.473 0.477 0.496 0.511 0.494 0.486 0.466 0.465 0.438 0.384 0.357 0.343 0.325 0.317 0.278 0.445
CT (cm) mean (SD)
(0.065) (0.061) (0.070) (0.068) (0.063) (0.071) (0.056) (0.060) (0.058) (0.057) (0.052) (0.046) (0.054) (0.043) (0.054) (0.044) (0.087)
0.141 0.179 0.181 0.192 0.195 0.190 0.186 0.185 0.181 0.170 0.152 0.139 0.139 0.128 0.123 0.116 0.171
BW (cm) mean (SD)
(0.022) (0.025) (0.026) (0.029) (0.028) (0.028) (0.024) (0.026) (0.031) (0.024) (0.026) (0.024) (0.026) (0.021) (0.028) (0.029) (0.035)
0.714 0.762 0.761 0.776 0.764 0.771 0.775 0.792 0.780 0.773 0.786 0.775 0.807 0.792 0.772 0.803 0.768
(0.076) (0.073) (0.071) (0.077) (0.073) (0.077) (0.077) (0.084) (0.098) (0.080) (0.096) (0.101) (0.080) (0.098) (0.096) (0.092) (0.083)
Abbr: BMD, bone mineral density; BW, bone width; CT, cortical thickness; MCI, metacarpal index; SD, standard deviation.
shown that male osteoporosis occurs half as frequently as female osteoporosis, and that men are often afflicted with secondary osteopenia, awareness of this has increased. Consequently, male reference values are strongly recommended as necessary for a reliable identification of male osteoporosis. Therefore our study closes the gap of normative Caucasian data, as estimated by the new software version of DXR, and includes both German males and females.
Comparison of DXR BMD Between German Caucasian Women and Men Our results demonstrate that the mean DXR BMD for German Caucasian women (0.526 6 0.087 g/cm2) is considerably lower than mean DXR BMD for German Caucasian men
(0.603 6 0.091 g/cm2). Wuster et al. (31) demonstrated comparable values of DXR BMD for German Caucasian women (0.554 6 0.067 g/cm2) and men (0.654 6 0.072 g/cm2) with the obsolete DXR version 1.0 in a small cohort. Another German reference cohort, which included only women, and is evaluated by Toledo and Jergas (30), reported higher reference values in all age groups. The reliability of these data is debatable, because Toledo and Jergas (30) preferred a retrospective study design, and consequently only patients with visible fractures were excluded. Otherwise, also individuals with disease-related and drug-induced alterations of the peripheral bone status were included (in particular, subjects with intake of bone affecting drugs [e.g., vitamin D or calcium] were considered); such patients were excluded in our study design.
0.9 0.8 0.7 0.5
BMD
BMD
0.6 0.4 0.3 0.2 0.1 0 0
10
20
30
40
50
60
70
80
90
100
Age
Fig. 1. Bone mineral density (BMD) reference curve for men (X-rayhand and wrist; n 5 1131). Journal of Clinical Densitometry
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
10
20
30
40
50
60
70
80
90
100
110
Age
Fig. 2. Bone mineral density (BMD) reference curve for women (X-rayhand and wrist; n 5 954). Volume 9, 2006
346
Bo¨ttcher et al. Table 3 Comparison of Mean Values for DXR Parameters Between Men and Women (n 5 2085) Men mean (SD)
BMD (g/cm2) MCI CT (cm) BW (cm)
0.603 0.456 0.196 0.864
Women mean (SD)
(0.091) (0.069) (0.035) (0.097)
0.526 0.445 0.171 0.768
(0.087) (0.087) (0.035) (0.083)
Relative difference in %
Significance (ManneWhitney U test)
12.8% 2.4% 12.8% 11.1%
p ! 0.01 p ! 0.05 p ! 0.01 p ! 0.01
Abbr: BMD, bone mineral density; BW, bone width; CT, cortical thickness; DXR, digital X-ray radiogrammetry; MCI, metacarpal index.
A possible limitation of our own study design should not remain unmentioned. The reason for our lower reference values could be induced by a selection bias, because our study population only consisted of diseased subjects who requested help in an emergency unit. Although our questionnaire considered all relevant bone affecting results, this population may represent a cohort with a higher risk for accidents, such as the use of alcohol or inclusive of patients with recurrent falls due to cardiovascular disorders or also osteoporosis. The influence regarding a different standard of living and nutrition in the Eastern part of Germany before 1989 is unlikely (regular government aid for sports and calcium supply via milk), but should be not completely negated. A further limitation of our study design is the lack of comparative technology and of anthropometric details. Our study showed a peak bone mass for women between the ages of 30 and 34 yr and for men it was delayed and between the ages of 45 and 49 yr. This gender-specific observation may be understood as caused by the occurrence of puberty at an earlier stage in women. Similarly Hyldstrup and Nielsen (36) presented normative data of MCI estimated by DXR (version 2.0) for 384 healthy Danish women, which also revealed the maximum gain of MCI between the ages of 30 and 39 yr, whereas the mean value of MCI with 0.486 was somewhat higher compared with our results. In contrast 0.7
BMD in g/cm2
0.6 0.5 0.4 BMD-women
0.3
BMD-men
84 80
–
74 – 70
–
64
54 60
50
–
44 40
–
34 – 30
– 20
<
51
24
0.2
Age groups
Fig. 3. Comparison of age-related bone mineral density (BMD) between men and women (X-rayhand and wrist; n 5 2085). Journal of Clinical Densitometry
to these results, Toledo and Jergas (30) reported the peak bone mass for women between the ages of 40 and 44 yr. The data of Wuster et al. (31) verified the peak bone mass for DXR BMD between the ages of 40 and 49 yr in women and 30 and 39 yr in men. This difference may probably be due to a selection bias, and it may also be significantly influenced by variations regarding the statistical procedures (e.g., grouping of the study population). Furthermore, an important limitation of the studies from Toledo and Jergas (30) and Wuster et al. (31) was the small number of subjects. In the present study, the quantity of subjects in the female cohort (n 5 954) was duplicated compared with the cohort presented by Toledo and Jergas (30). When the results regarding MCI from the present study are compared with published data (26), the same tendency could be shown. Pre-menopausal women reveal higher MCI values than men within the comparable age groups, whereas postmenopausal women have lower age-related MCI data than men. The observation that pre-menopausal women have a higher MCI than men is an apparent contradiction to the fact that the same women (in general) have a smaller cortical thickness as well as lower BMD. The apparent contradiction is explained by the fact that women also have a smaller bone diameter, and therefore, the MCI is higher for pre-menopausal women. When women reach menopause, the subsequent relatively fast bone loss results in a thinning of the cortices that lowers the MCI much faster than in men. A few years after menopause, the postmenopausal women are then observed to have a lower MCI than the age-matched men (31). It is common knowledge that women have a reduced cortical thickness and a smaller metacarpal bone diameter in comparison with male bone proportions. The MCI offers only a minimal gender-specific difference due to the specific calculation of this parameter as ratio of the mean cortical thickness normalized with the mean outer bone diameter. A further, but still neglected parameter of DXR is the total DXR BW of the metacarpals; our data revealed a continuos progressive increase of DXR BW with aging accentuated for men. For women, the maximum of bone width was reached between the ages of 70 and 74 yr, whereas men demonstrated a delayed peak level in the age range of 75e79 yr. Furthermore young men (age range: 15e19 yr) already showed a greater metacarpal bone width as compared with women Volume 9, 2006
DXR Normative Data
347
Table 4 Age-Related Relative Changes of DXR BMD and MCI (Men, X-rayhand Age (yr) !15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
BMD (g/cm2) Mean (SD)
n 93 200 183 121 93 104 75 65 50 46 23 24 19 17 7 11 1131
0.460 0.596 0.619 0.625 0.625 0.634 0.641 0.649 0.627 0.626 0.615 0.585 0.566 0.559 0.501 0.478 0.603
BMD relative changes
(0.072) (0.074) (0.078) (0.077) (0.091) (0.079) (0.071) (0.067) (0.080) (0.079) (0.053) (0.076) (0.079) (0.084) (0.040) (0.089) (0.091)
n 5 1131)
MCI mean (SD) 0.374 0.459 0.476 0.482 0.481 0.481 0.465 0.477 0.463 0.441 0.430 0.415 0.382 0.367 0.337 0.335 0.456
þ29.6% þ3.9% þ1.0% 0.0% þ1.4% þ1.1% þ1.2% 3.5% 0.2% 1.8% 5.1% 3.4% 1.3% 11.6% 4.8%
and wrist,
MCI relative changes
(0.053) (0.056) (0.054) (0.064) (0.066) (0.056) (0.060) (0.062) (0.054) (0.061) (0.058) (0.060) (0.056) (0.057) (0.063) (0.070) (0.069)
þ22.7% þ3.7% þ1.3% 0.2% 0.0% 3.4% þ2.6% 3.0% 5.0% 2.6% 3.6% 8.6% 4.1% 8.2% 0.6%
Abbr: BMD, bone mineral density; CT, cortical thickness; DXR, digital X-ray radiogrammetry; MCI, metacarpal index.
of the same age and magnitude of the age-associated increment (i.e., BW), which was higher in men than in women. Both gender also offered a remarkable increase of DXR BW in the adolescent years (from ages ! 15 yr to an age range of 15e19 yr). Our findings could be confirmed by Russo et al. (37) using measurements of peripheral quantitative
computed tomography. The authors suggested that the increase of bone width in aging men may contribute to the maintenance of adequate bone mechanical competence in the face of declining cortical bone mass. In women, this compensatory mechanism appears to be less efficient, and consequently, the bone mechanical competence is more impaired with age.
Table 5 Age-Related Relative Changes of DXR BMD and MCI (Women, X-Rayhand Age (yr) !15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
n 96 106 127 82 75 72 56 59 57 41 37 42 22 25 26 31 954
BMD (g/cm2) Mean (SD) 0.450 0.542 0.552 0.574 0.585 0.573 0.565 0.566 0.552 0.527 0.490 0.451 0.456 0.429 0.415 0.385 0.526
(0.057) (0.066) (0.064) (0.069) (0.066) (0.068) (0.056) (0.066) (0.081) (0.070) (0.073) (0.061) (0.069) (0.064) (0.080) (0.057) (0.087)
BMD þ20.4% þ1.8% þ4.0% þ1.9% 2.1% 1.4% þ0.2% 2.5% 4.7% 7.6% 8.6% þ1.1% 6.3% 3.4% 7.8%
and wrist,
n 5 954)
MCI Mean (SD) 0.399 0.473 0.477 0.496 0.511 0.494 0.486 0.466 0.465 0.438 0.384 0.357 0.343 0.325 0.317 0.278 0.445
(0.065) (0.061) (0.070) (0.068) (0.063) (0.071) (0.056) (0.060) (0.058) (0.057) (0.052) (0.046) (0.054) (0.043) (0.054) (0.044) (0.087)
MCI þ18.5% þ0.8% þ4.0% þ3.0% 3.4% 1.6% 4.3% 0.2% 6.2% 14.1% 7.6% 4.1% 5.5% 2.5% 14.0%
Abbr: BMD, bone mineral density; CT, cortical thickness; DXR, digital X-ray radiogrammetry; MCI, metacarpal index. Journal of Clinical Densitometry
Volume 9, 2006
348 This geometric adaptation of increasing bone width, which occurs parallel to the cortical bone loss detected by MCI and DXR BMD, could be an important component in the assessment of bone mechanical properties and for the diagnosis of osteoporosis. These coherences, which are methodologically overlooked by planar densitometry, were also described by other studies focusing on micro-radioscopy and micro-densitometry (38,39). In detail, recent studies revealed a cortical bone loss in postmenopausal as well as in age-associated demineralization caused by increased intra-cortical porosity and reduced cortical thickness, which is related to endosteal and juxta-endosteal bone resorption (40). These changes tend to be compensated by periosteal apposition (i.e., increase of metacarpal DXR BW) in both genders, but more obviously in men than in women with the previously mentioned consequences (37,40). As expected, women showed an accentuated total DXR BMD decline than men after passing through the peak bone mass period (men: 31.7%; women: 43.1%). In addition, men revealed two age-related peaks of cortical bone loss (age group: 65e69 and 80e84 yr) indicated by a decrease of cortical thickness, MCI, and DXR BMD in our study. For very elderly men, the second accentuated maximum of bone loss could be explained by the combined age-induced effect of decreasing testosterone, growth hormone, and insulin-like growth factor I (38), whereas in women, our results documented more of a continuous decline of DXR BMD, MCI, and cortical thickness. The interdependency of DXR parameters, in particular focusing on cortical thickness and DXR BW, in patients with primary and secondary osteoporosis, is most relevant for the understanding of various physiological processes regarding the cortical bone partition deserving further intensive investigations.
Influence of Bone Age on DXR BMD Measurement Several other studies have reported the same finding that a quantification of BMD could be achieved by the DXR system, with a bone age of 6 yr and older (30,41,42). In children with a skeletal age younger than 6 six yr, the DXR system can not reliably recognize the bone structure and detect the edges of the metacarpals, and thus automatically discontinues the DXR analysis.
Impact of Different X-Ray Settings (X-rayhand Versus X-raywrist) on DXR Measurement It has been reported that DXR provides calculations of DXR BMD and MCI with high intra-radiograph and inter-radiograph reproducibility of conventional, as well as digitally performed radiographs (14). This technique has been shown to be insensitive toward alterations of film sensitivity, film brand, film-focus distance, and exposure level (14,15). Only variation in the tube voltage and the use of different Xray equipment demonstrated a significant influence on the calculation process of DXR BMD and MCI for conventional and Journal of Clinical Densitometry
Bo¨ttcher et al. particularly for digital printouts (14,15). Toledo and Jergas (30) included data calculated from hand and wrist radiographs in their study, which were performed with different tube voltage (hand X-rays: 46 kV; wrist X-rays: 50 kV), and therefore a limited reliability of these data should be discussed. In our study, the image-capturing conditions of hand and wrist radiographs were similar, and the same X-ray device was used. The present study evaluated the impact of different X-ray settings (X-rayhand vs. X-raywrist) on DXR-measurement; this circumstance should be evaluated, because the process of contour finding and edge detection during DXR measurements could be impaired by different X-ray settings in spite of constant image-capturing parameters. The performance of different radiographs focused at the wrist or hand results from a various fracture localization; the equivalent use of these different X-ray settings (without alteration of the DXR results) facilitates the identification of osteoporosis in combination with the onset of fractures at the distal upper extremities. The results show no significant difference between the two types of images, which were performed in the identical anterior-posterior projection. Hence, it is not mandatory to use only one X-ray setting in clinical practice, but it is advisable to use consistent technical conditions to maintain the best reliability, in particular when normative data were compared with DXR measurements of patients suffering from various bone affecting diseases. Toledo and Jergas (30) revealed an important fact: awareness that DXR BMD and DXR cortical thickness of the dominant hand (right handedness in 93% of a population) show significantly higher values than for the nondominant hand. Therefore, only DXR estimates of the nondominant hand provide reliable data as considered in this study, because measurements of the dominant hand could show increased cortical bone mineralization caused by increased physical activity.
Comparison of DXR-BMD Between Different Ethnic Groups Recent studies have shown higher results for DXR BMD in Scandinavian (0.555 6 0.070 g/cm2; 29), Indian (0.544 6 0.066 g/cm2; 35), and North American Caucasian women (0.558 6 0.065 g/cm2; 27) compared with our own DXR data. This observation may be caused by the use of the version 1.0 of the DXR system. The obsolete version (1.0) had calculated the cortical width and overall BW at 5 locations: the second, third, and forth metacarpal bone, as well as the distal radius (radial side, only cortical width), and the distal ulna (ulnar side, only cortical width). In contrast with the older version, the newer version 2.0 of the DXR system estimates DXR BMD only on the second, third, and fourth metacarpals of the nondominant hand. In various studies, the extensive ethnic differences of BMD have been evaluated (e.g., BMD measured by DXA on the proximal femur) and showed an accentuated increase in non-Hispanic blacks compared with Mexican Americans Volume 9, 2006
DXR Normative Data and non-Hispanic whites (43). Significantly higher results of DXR (þ14.0%) have also been observed for the AfricanAmerican females (0.597 6 0.055 g/cm2) compared with German females (28). Further ethnic differences in DXR BMD values of Caucasian, Korean, and Indian men, as well as North-American, African-American, Scandinavian, and Indian women (32e34,36), emphasize the essential requirement to evaluate and establish a gender-specific, ethnically differentiated database for each osteodensitometric technique, including DXR. Based on the availability of gender-specific normative data for various ethnic populations and promoted by the excellent reproducibility, without a possible influence of soft tissue alterations during the measurement, DXR is at the forefront of competing osteodensitometric techniques in detecting patients with high fracture risk due to primary or secondary osteoporosis. The cutting edge of DXR could be confirmed by Boonen et al. (1), who demonstrated a better sensitivity for identifying women both with and without osteoporosis (e.g., positive and negative predictive values) by DXR compared with quantitative ultrasound, and suggest that DXR BMD and phalangeal radiographic absorptiometry may be the most effective for DXA testing in high-risk postmenopausal women.
Conclusion The development of digital imaging techniques has promoted the precise measurement of several radiogeometrical features, and DXR has provided a reliable diagnostic approach for the quantification of cortical bone loss in the clinical routine. The present study has estimated normative reference values for DXR in German Caucasian women and men. The verified influence of ethnic groups and gender on normative values has revealed the need for regional reference data to achieve a reliable interpretation of DXR measurements. Furthermore, our normative values insure the applicability of the T-score by comparing estimated DXR BMD with the normative values in young healthy women or men, and also the Z-score that focused on age-matched controls. The now available T- and Z-scores enable application of DXR as a screening tool to identify patients who may benefit the most from axial DXA measurements. Further prospective studies will be required to determine the predictive value of DXR to distinguish patients with high fracture risk, and to introduce the DXR technique as a cost-effective, widely available, and precise method in the clinical routine.
Acknowledgments We would like to thank Monika Arens, (managing director, Arewus GmbH) for the use of the X-posure equipment and R€ udiger Vollandt, PhD for the statistical advice. Finally, the authors would also like to thank Dieter Felsenberg, MD (Berlin, Germany), Claus C. Gluer, PhD (Kiel, Germany), Stephan Grampp, MD, and Herwig Imhof, MD (Wien, Austria) for their comments regarding this study.
Journal of Clinical Densitometry
349
References 1. Boonen S, Nijs J, Borghs H, Peeters H, Vanderschueren D, Luyten FP. 2005 Identifying postmenopausal women with osteoporosis by calcaneal ultrasound, metacarpal digital X-ray radiogrammetry and phalangeal radiographic absorptiometry: a comparative study. Osteoporos Int 16:93e100. 2. Bouxsein ML, Palermo L, Yeung C, Black DM. 2002 Digital X-ray radiogrammetry predicts hip, wrist and vertebral fracture in elderly women: a prospective analysis from the study of osteoporotic fractures. Osteoporos Int 13:358e365. 3. Hyldstrup L, Jorgensen JT, Sorensen TK, Baeksgaard L. 2001 Response of cortical bone to antiresorptive treatment. Calcif Tissue Int 68:135e139. 4. Ozcakar L, Guven GS, Unal S, Akinci A. 2005 Osteoporosis in Turkish HIV/AIDS patients: comparative analysis by dual energy X-ray absorptiometry and digital X-ray radiogrammetry. Osteoporos Int 16:1363e1367. 5. Ward KA, Cotton J, Adams JE. 2003 A technical and clinical evaluation of digital X-ray radiogrammetry. Osteoporos Int 14: 389e395. 6. Barnett E, Nordin B. 1960 The radiological diagnosis of osteoporosis: a new approach. Clin Radiol 11:166e174. 7. Meema HE. 1962 The occurrence of cortical bone atrophy in old age and osteoporosis. J Can Assoc Radiol 13:27e32. 8. Meema HE, Meema S. 1963 Measurable roentgenologic changes in some peripheral bones in senile osteoporosis. J Am Geriatr Soc 11:1170e1182. 9. Garn S, Poznanski A, Nagy J. 1971 Bone measurement in the differential diagnosis of osteopenia and osteoporosis. Radiology 100:509e518. 10. Exton-Smith A, Millard P, Payne P, Wheeler E. 1969 Method of measuring quantity of bone. Lancet 2:1153e1157. 11. Adams P, Davies G, Sweetnam P. 1969 Observer error and measurements of the metacarpal. Br J. Radiol 42:192e197. 12. Chan E, Pandith V, Towheed TE, Brouillard D, Zee B, Anastassiades TP. 1998 Comparison of the combined cortical thickness of the second metacarpal with the Sharp’s method for scoring hand microradiographs in rheumatoid arthritis. J Rheumatol 25:1290e1294. 13. Aguado F, Revilla M, Villa LF, Rico H. 1997 Cortical bone resorption in osteoporosis. Calcif Tissue Int 60:323e326. 14. Bo¨ttcher J, Pfeil A, Rosholm A, et al. 2005 Influence of imagecapturing parameters on Digital X-Ray Radiogrammetry. J Clin Densitom 8:87e94. 15. Malich A, Bo¨ttcher J, Pfeil A, et al. 2004 The influence of technical parameters of Xeray imaging on reproducibility and precision of digital computer-assisted radiogrammetry (DXR). Skeletal Radiol 33:473e482. 16. Rosholm A, Hyldstrup L, Baeksgaard L, Grunkin M, Thodberg HH. 2001 Estimation of bone mineral density by digital X-ray radiogrammetry: theoretical background and clinical testing. Osteoporos Int 12:961e969. 17. Bo¨ttcher J, Pfeil A, Rosholm A, et al. 2006 Computerized digital imaging techniques provided by digital X-ray radiogrammetry as new diagnostic tool in rheumatoid arthritis. J Digit Imaging April 26 (DOI: 10.1007/s10278-006-0263-y). 18. Bo¨ttcher J, Malich A, Pfeil A, et al. 2004 Potential clinical relevance of digital radiogrammetry for quantification of periarticular bone demineralization in patients suffering from rheumatoid arthritis dependent on severity and compared with DXA. Eur Radiol 14:631e637. 19. Bo¨ttcher J, Pfeil A, Heinrich B, et al. 2005 Digital radiogrammetry as a new osteodensitometric technique and diagnostic tool in rheumatoid arthritis. Rheumatol Int 25:457e464.
Volume 9, 2006
350 20. Bo¨ttcher J, Pfeil A, Rosholm A, et al. 2005 Digital X-ray radiogrammetry combined with semi-automated analysis of joint space distances as a new diagnostic approach in rheumatoid arthritisea cross-sectional and longitudinal study. Arthritis Rheum 52:3850e3859. 21. Stewart A, Mackenzie LM, Black AJ, Reid DM. 2004 Predicting erosive disease in rheumatoid arthritis. A longitudinal study of changes in bone density using digital X-ray radiogrammetry: a pilot study. Rheumatalogy 43:1561e1564. 22. Mentzel HJ, John U, Bo¨ttcher J, et al. 2005 Evaluation of bonemineral density by digital X-ray radiogrammetry (DXR) in pediatric renal transplant recipients. Ped Radiol 35:489e494. 23. Van Rijn RR, Boot A, Wittenberg R, et al. 2006 Direct X-ray radiogrammetry versus dual-energy X-ray absorptiometry: assessment of bone density in children treated for acute lymphoblastic leukemia and growth hormone deficiency. Ped Radiol 24:1e6. 24. Cootes TF, Taylor CJ, Cooper D, Graham J. 1995 Active shape models: their training and application. Comp Vision Image Understand 61:38e59. 25. Lazenby R. 1995 Brief communication: non-circular geometry and radiogrammetry of the second metacarpal. Am J Phys Anthropol 97:323e327. 26. Jorgensen JT, Andersen PB, Rosholm A, Bjarnason NH. 2000 Digital X-ray radiogrammetry: a new appendicular bone densitometric method with high precision. Clin Physiol 20:330e335. 27. Black DM, Palermo L, Sorensen T, et al. 2001 A normative reference database study for Pronosco X-posure system. J Clin Densitom 4:5e12. 28. Shepherd JA, Meta M, Landau J, et al. 2005 Metacarpal index and bone mineral density in healthy African-American women. Osteoporos Int 16:1621e1626. 29. Cummings S, Hyldstrup L, Liu Z, Wuster C, Maffei L. 2000 Normative reference curves for the Pronosco X-posure system established in five ethnic groups. J Bone Miner Res 15(Suppl 1):321. 30. Toledo VAM, Jergas M. 2005 Age-related changes in cortical bone mass: data from a German female cohort. Eur Radiol 8:1e7. 31. Wuster C, Wenzler M, Kappes J, Rehm C, G€ uhring T, Arnbjerg D. 2000 Digital X-ray radiogrammetry as a clinical method for estimating bone mineral densityea German reference database. J Bone Miner Res 15(Suppl 1):298.
Journal of Clinical Densitometry
Bo¨ttcher et al. 32. Jiang Y, Miao Li, Zhonghou L. 2000 Normal values of forearm bone BMD and incidence of primary osteoporosis in chinese women. Chinese Journal of Osteoporosis 6:30e31. 33. Maffei L, Venarotti HO, Treviso JC, Sorensen TK, Nissen DS. 2000 Digital X-ray radiogrammetry: a Hispanic normative reference database for the Pronosco X-posure System. J Bone Miner Res 15(Suppl 1):304. 34. Yang SO, Ham SY, Kim YI, et al. 2001 Comparison of Korean female and male reference data using DXR. J Bone Miner Res 16(Suppl 1):304. 35. Wuster C, Hadji P. 2001 Use of quantitative ultrasound densitometry (QUS) in male osteoporosis. Calcif Tissue Int 69: 225e228. 36. Hyldstrup L, Nielsen SP. 2001 Metacarpal index by digital X-ray radiogrammetry: normative reference values and comparison with dual X-ray absorptiometry. J Clin Densitom 4:229e306. 37. Russo CR, Lauretani F, Bandinelli S, et al. 2003 Aging bone in men and women: beyond changes in bone mineral density. Osteoporos Int 14:531e538. 38. Iwamoto J, Takeda T, Ichimura S, Tsukimura Y, Toyama Y. 2000 Age-related changes in cortical bone in men: metacarpal bone mass measurement study. J Orthop Sci 5:4e9. 39. Meema HE, Meema S. 1987 Longitudinal microradioscopic comparison on endosteal and juxtaendosteal bone loss in premenopausal and postmenopausal women, and in those with end-stage renal disease. Bone 8:343e350. 40. Russo CR, Lauretani F, Seeman E, et al. 2006 Structural adaptations to bone in aging men and women. Bone 38:112e118. 41. Malich A, Freesmeyer MG, Mentzel HJ, et al. 2003 Normative values of bone parameters of children and adolescents using digital computer-assisted radiogrammetry (DXR). J Clin Densitom 6:103e112. 42. van Rijn RR, Grootfaam DS, Lequin MH, et al. 2004 Digital radiogrammetry of the hand in a pediatric and adolescent Dutch Caucasian population: normative data and measurements in children with inflammatory bowel disease and juvenile chronic arthritis. Calcif Tissue Int 74:342e350. 43. Looker AC, Wahner HW, Dunn WL, et al. 1995 Proximal femur bone mineral levels of US adults. Osteoporos Int 5: 389e409.
Volume 9, 2006
Original Article
Normative Data for Digital X-Ray Radiogrammetry From a Female and Male German Cohort Joachim Bo¨ttcher,*,1 Alexander Pfeil,1 Max L. Scha¨fer,1 Alexander Petrovitch,1 Bettina E. Seidl,1 Hans-Joachim Mentzel,1 Gabriele Lehmann,2 Ansgar Malich,3 Jens-Peter Heyne,1 Gert Hein,2 Gunter Wolf,2 and Werner A. Kaiser1 1
Institute of Diagnostic and Interventional Radiology, Friedrich-Schiller-University Jena, Jena, Germany; 2Department of Rheumatology and Osteology, Clinic of Internal Medicine III, Friedrich-Schiller-University Jena, Jena, Germany; and 3 Sued-Harz Klinikum, Department of Radiology, Nordhausen, Germany
Abstract This study presents German reference data for digital X-ray radiogrammetry (DXR) differentiated by males as well as females, and quantifies for gender-specific and age-related differences including all DXR parameters. This study also documents the effects of different X-ray settings (e.g., radiographs of the wrist or the hand) on DXR measurements. There were 2085 patients who were prospectively enrolled (954 females and 1131 males) from a data pool of 11,915 patients with radiographs of the nondominant hand or wrist. All patients underwent measurements of bone mineral density (BMD), cortical thickness, bone width, and the metacarpal index (MCI) using DXR technology. These data showed a continuous age-related increase of the DXR parameters to the point of peak bone mass, then a continuous decline beyond the peak bone mass with accentuated age-related cortical bone loss in women. Peak bone mass is reached at approximately 30e34 yr for women and 45e49 yr for men. In addition, men had a significantly higher DXR BMD (mean: þ12.8%) compared with woman in all age groups. Regarding the impact of various X-ray settings (e.g., X-raywrist vs. X-rayhand), no significant difference was observed between both groups, men as well as women. The development of digital imaging technology has enabled more precise measurements of several radio-geometric features. The present study estimated normative reference values for DXR in German Caucasian women and men. Based on this reference data, a valid and reliable quantification of disease-related demineralization based on measurements of DXR BMD and MCI is now available for the Caucasian ethnic group. Key Words: Bone mineral density; bone width; digital X-ray radiogrammetry; metacarpal index; normative data.
is a radio-geometrical technique used to estimate the thickness of the cortical bone partition. For more than 40 yr, conventional radiogrammetry has been the primary method for assessment of bone status from hand radiographs (6e8). This method measures total width of a bone and the medullary width at the midpoint of the second metacarpal of the nondominant hand. The measured data are typically used to calculate various indices, such as the metacarpal index (i.e., the ratio of total bone width and cortical thickness), the Barnett and Nordin index (i.e., the percentage of cortical thickness) (6), the Garn index (i.e., the cortical area) (9), and the Exton-Smith index (i.e., related to the cortical area and the surface area) (10), and thus to quantify changes of the cortical bone.
Introduction Digital X-ray radiogrammetry (DXR) is a new diagnostic tool for the measurement of cortical bone mineral density (BMD) on metacarpals using digitized radiographs (1e5). The DXR is based on conventional radiogrammetry, which
Received 01/18/06; Revised 05/22/06; Accepted 06/03/06. *Address correspondence to: Joachim Bo¨ttcher, MD, Institute of Diagnostic and Interventional Radiology, Friedrich-Schiller-University Jena, Erlanger Allee 101, Jena, 07747 Germany. E-mail: [email protected]
341
342 Conventional radiogrammetry and the previously mentioned indices are most sensitive to cortical bone changes (i.e., periosteal apposition and especially endosteal resorption). Despite conventional radiogrammetry use being relatively inexpensive and widely available, it is limited because of its imprecision (from 2.4% up to 11%) due to the difficulty in identification of the endosteal margin and precise marking of the midshaft location by the operator-dependent radiogrammetry measurements (11e13). Digital X-ray radiogrammetry is a new operator-independent diagnostic tool, providing automated measurement of cortical BMD on the metacarpals using digitized radiographs with an excellent intra-radiograph (0.05e0.33%) and interradiograph reproducibility (0.26e1.54%). The high precision is based on refinement, computerization, and the implementation of algorithms for automatic and operator-independent image analysis by the DXR technology (14,15), which is based on a combined computerized radiogrammetry and textural analysis of the 3 middle metacarpal bones. Although DXR BMD estimates were not taken at a point known for high incidence of fracture, in many studies DXR has confirmed its predictive value regarding fracture risk (1,2), has provided evidence for therapeutic effects in the treatment of postmenopausal osteoporosis (3), and has revealed closed associations to dual-energy X-ray absorptiometry (DXA) measurements of the spine, femur, and forearm (5,16). The DXR-technique seems to bridge the gap between radiogrammetry and densitometry by estimating BMD from radiogrammetry measurements in case of primary and secondary osteoporosis. Another promising application of DXR is the quantification of periarticular bone loss in rheumatoid arthritis. In comparison with other osteodensitometric techniques (DXA, peripheral quanitative computer tomography, quantitative ultrasound), DXR could detect and quantify the diseaserelated cortical bone loss caused by the course and severity of rheumatoid arthritis (17e21). A third potential application of DXR technology is the determination of peripheral bone status in childhood (22,23). Children with cystic fibrosis, Turner syndrome, Marfan syndrome, osteogenesis imperfecta, inflammatory bowel diseases, and renal disorders often have disease-related or therapy-induced cortical osteopenia develop, frequently requiring radiographs of the hand for disease monitoring (based on bone age determination because of identification of diseaserelated retardation), which could be used for DXR estimates. The aim of this study was to present German reference data of DXR for Caucasian men and women to quantify the gender-specific as well as age-related differences of DXR BMD, and to evaluate the effects of different X-ray settings on DXR measurements.
Patients and Methods Patients This prospective study enrolled 2085 patients (954 females and 1131 males) between 2001 and 2005 from a data pool of Journal of Clinical Densitometry
Bo¨ttcher et al. 11,915 patients with hand or wrist radiographs, which were performed during admission in the emergency unit of a university in the Eastern part of Germany to exclude fractures due to trauma. As an important part of our study design, a detailed questionnaire was completed by all subjects to record possible exclusion criteria mentioned as follows. The remaining 2085 healthy individuals were of the Caucasian ethnic group and received digital radiographs of the nondominant hand or wrist for the DXR estimates. All patients underwent measurements of BMD (g/cm2), cortical thickness (cm), bone width (BW) (cm) and metacarpal index (MCI), a dimensionless parameter (based on the ratio of cortical thickness to bone diameter) using DXR technology. Mean age was 36.1 yr (men: 33.5 years 6 17.4; women: 39.1 6 21.7) with a standard deviation of 19.4 yr, and an age range of 6.6 to 105.6 yr (men: 7.3e99.6 yr; women: 6.6e105.6 yr). The following exclusion criteria were considered based on the extensive questionnaire: visible metallic material (i.e., splints and material after osteosynthesis; n 5 899); signs of fracture (n 5 4867) and amputation (n 5 38); endocrinological diseases known to affect bone metabolism (e.g., hyper/hypoparathyroidism, Cushing disease; n 5 245) and genetic diseases (n 5 23); rheumatic diseases (e.g., rheumatoid arthritis; n 5 1237) renal disorders (n 5 1084), and oncological diseases (n 5 1255); medication with bone-influencing drugs (e.g., steroids, vitamin D, or calcium intake; n 5 182); and only Caucasian German patients were enrolled (e.g., verification that birthplaces of the subjects and also their parents and grandparents were located in Germany). A further exclusion criteria was the existence of fractures in the dominant upper extremities, which frequently induce osteopenia caused by immobility on the unmeasured side of fracture, but otherwise cause accentuated mineralization on the nondominant healthy side based on raised physical activity. The population was divided into 16 age-related groups differentiated for gender.
DXR The Pronosco X-Posure System (version 2.0, Sectra Pronosco A/S, Herlev, Denmark), which involves a radiogrammetric and textural analysis of the 3 middle metacarpal bones, was used to determine DXR BMD and MCI based on DXR, requiring radiographs of the nondominant hand. All plain radiographs were acquired by a Polydoros SX 80 (Siemens, Munich, Germany) with the following standardized conditions: filter with 1.0-mm thickness related to aluminium 80, tube voltage 42 kV, exposure level 4 mAs, film focus distance 100 cm (Scopix Laser 2 B 400; Agfa GmbH and Cie. KG, Cologne, Germany). The digital radiographs were printed and subsequently scanned into the system at a resolution of 300 dpi, corresponding to 118 pixels per cm (16). The system itself checked the quality of the scanned images and interrupted Volume 9, 2006
DXR Normative Data the examination in case of inadequate quality. The computer algorithms automatically defined regions of interest (ROIs) around the narrowest bone parts of the metacarpal II, III, and IV and subsequently specified the outer and inner cortical edges of the studied cortical bone parts. To locate the diaphysis of the 3 middle metacarpals in the radiograph (16,24), the DXR system applies a model-based algorithm known as the active shape model. After each diaphysis has been identified, RO’s are placed automatically for the 3 metacarpals and the endosteal (inner) and periosteal (outer) edges are automatically found. Apart from placing the radiography on the chargedcoupled device-based desktop flatbed scanner, otherwise there is no operator interaction connected to the DXR measurement. The analyzed images and their ROIs are displayed on the computer monitor. After calculation of the mean of the cortical thickness, the cortical volume per area (VPA) is computed for each metacarpal assuming a cylindrically-shaped bone based on cortical thickness and mean outer bone diameter as given as follows by Lazenby (25) and Rosholm et al. (16): The final DXR BMD, based on the mean VPA, is then calculated with the following correction for the estimated porosity index (P): DXR BMD 5 c VPA ð1 PÞ: The scaling constant c is determined so that DXR BMD on average is equal to that of the mid-distal forearm region of the Hologic QDR 2000 densitometer (Hologic, Waltham, MA). The porosity index is a technical parameter given as a value between 1 and 19, which is derived from the area percentage of local intensity minima found in the cortical part of the bone relative to the entire cortical area (16). Consequently, the porosity index represents an estimated 3-dimensional cortical porosity, aimed to be the fraction of the cortical bone volume that is not occupied by bone. The MCI obtains the mean cortical thickness normalized with the mean outer bone diameter (BW) for each bone part (16,23,26).
Ethics All examinations were performed in accordance with the rules and regulations of the local human research and ethics committee. As a special note, the authors emphasize that all radiographs used for DXR calculations were performed as part of routine clinical care (e.g., exclusion of fractures due to relevant trauma). No additional radiographs were obtained only for study purposes.
Data Analysis The objective of the statistical analysis was to establish normative values for both DXR BMD and MCI used in Tscore and Z-score calculations. To pool data from the sites into the normative reference database, homogeneity of the data across the sites was investigated. This intention was Journal of Clinical Densitometry
343 accomplished by using a regression model of BMD versus study site indicator variables, age, weight, and height on data for young women. A F-test was calculated if the differences between the sites were significantly different from zero. The modeling of the reference curves for BMD and MCI were achieved by applying the power of the mean model. Prior to regression analysis, a standard test for heteroscedasticity was performed. The reference curves were based on a regression of DXR BMD and MCI versus age and age2 (second order polynomials). The significance of sex-dependent changes was calculated with the ManneWhitney U test. Regarding the significance of differences between X-rayhand and X-raywrist, changes were calculated with the t test for unpaired groups to verify the parity of both groups and in an opposite direction with the ManneWhitney U test to identify significant differences between both X-ray settings. The statistical analysis was performed using SPSS, version 10.13 (SPSS, Chicago, IL) and using SAS System, version 8 (SAS Institute Inc, Cary, NC).
Results Reproducibility of DXR The inter-radiograph reproducibility (evaluated by one DXR analysis of 10 repeated acquired radiographs of the same hand with repositioning under standard X-ray settings): DXR BMD : 0:31%
MCI : 0:24%
The intra-radiograph reproducibility (evaluated by a single digital image that underwent 10 repeated DXR analyses): DXR BMD : 0:09%
MCI : 0:14%
DXR Reference Data See Tables 1 and 2, and Figs. 1 and 2.
DXR Parameters and Age-Related Alterations The DXR system reliably provides recognition of the metacarpals of subjects with a bone age of 6 years and older. This study shows a close relationship between age and all measured DXR parameters. The DXR BMD significantly increased in men (þ41.1%) from 0.460 6 0.072 g/cm2 to 0.649 6 0.067 g/cm2 up to an age of 45 years in a continuous manner. In women, DXR BMD showed a significant and continuous increase (þ30.0%) up to an age of 35 yr from 0.450 6 0.057 g/cm2 to 0.585 6 0.066 g/cm2. Equal results could be verified for MCI with a significant increase from 0.374 6 0.053 to 0.477 6 0.062 in men (þ27.5%), and from 0.399 6 0.065 to 0.511 6 0.063 in women (þ28.1%). A significant increase (þ47.9%) for DXR cortical thickness in men from 0.144 cm 6 0.029 cm (age: 10 yr) to 0.213 cm 6 0.029 (age: 45 yr) was observed. The relative increase of DXR cortical thickness for the female cohort was þ38.3%. Volume 9, 2006
344
Bo¨ttcher et al. Table 1 Mean and Standard Deviation Regarding the Parameters of Digital X-Ray Radiogrammetry Men, X-rayhand
Age (yr) !15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
n
BMD (g/cm2) Mean (SD)
93 200 183 121 93 104 75 65 50 46 23 24 19 17 7 11 1131
0.460 0.596 0.619 0.625 0.625 0.634 0.641 0.649 0.627 0.626 0.615 0.585 0.566 0.559 0.501 0.478 0.603
(0.072) (0.074) (0.078) (0.077) (0.091) (0.079) (0.069) (0.067) (0.080) (0.079) (0.053) (0.076) (0.079) (0.084) (0.040) (0.089) (0.091)
and wrist,
n 5 1131
MCI Mean (SD) 0.374 0.459 0.476 0.482 0.481 0.481 0.465 0.477 0.463 0.441 0.430 0.415 0.382 0.367 0.337 0.335 0.456
(0.053) (0.056) (0.054) (0.064) (0.066) (0.056) (0.060) (0.062) (0.054) (0.061) (0.058) (0.060) (0.056) (0.057) (0.063) (0.070) (0.069)
CT (cm) Mean (SD) 0.144 0.195 0.204 0.206 0.205 0.209 0.209 0.213 0.205 0.203 0.192 0.185 0.175 0.170 0.147 0.147 0.196
(0.029) (0.028) (0.030) (0.030) (0.036) (0.030) (0.029) (0.029) (0.029) (0.030) (0.022) (0.030) (0.027) (0.029) (0.024) (0.024) (0.035)
BW (cm) Mean (SD) 0.762 0.849 0.859 0.863 0.855 0.870 0.901 0.895 0.876 0.922 0.916 0.893 0.923 0.938 0.911 0.924 0.864
(0.082) (0.091) (0.090) (0.100) (0.097) (0.098) (0.078) (0.080) (0.081) (0.084) (0.068) (0.085) (0.106) (0.107) (0.103) (0.049) (0.097)
Abbr: BMD, bone mineral density; BW, bone width; CT, cortical thickness; MCI, metacarpal index; SD, standard deviation.
In both sexes, DXR BW showed a gender-specific relative increase with þ17.5 % in men versus þ7.0% in women. Both women and men show a continuous decline of DXR BMD in subjects older than 50 yr, which was accentuated in the group of women.
Comparison of DXR Parameters in Men and Women See Table 3 and Fig. 3. The results show that the German women cohort has a significantly lower age-related DXR BMD (mean: 12.8%) compared with the male cohort. The female peak bone mass for DXR BMD is reached in the age group of 30e34 yr (0.585 6 0.066 g/cm2). The male peak bone mass reveals their highest values in the age interval of 45e49 yr (0.649 6 0.067 g/cm2).
Comparison of Hand and Wrist Radiographs The normative values of DXR BMD revealed no significant difference (3.2%; t test: p ! 0.01 and ManneWhitney U test: p O 0.05) between radiographs of the male wrist (0.618 g/ cm2 6 0.078) and X-rays of the hand (0.599 g/ cm2 6 0.094). Comparable results are documented for the MCI in men with 2.7% (0.459 6 0.068 vs. 0.447 6 0.070). No significant difference between X-rayshand (DXR BMD: 0.521 g/cm2 6 0.089; MCI: 0.448 6 0.085) and X-rayswrist (DXR BMD: 0.537 g/cm2 6 0.082; MCI: 0.439 6 0.090) was observed for DXR BMD (3.1%) and MCI (2.0%) in women. Similar data were revealed for DXR cortical thickness and also for DXR BW in males (2.5% and 5.3%) as well as in the female cohort (3.0% and 5.0%). Journal of Clinical Densitometry
Annual Bone Loss See Tables 4 and 5. Both men and women show an accentuated increase of cortical bone mass in the age group from 15 to 19 yr for DXR BMD (men: 29.6%; women: 20.4%). Both the data for men and women revealed a continuous increase of the DXR BMD up to the peak bone mass (annual bone gain for men: 0.29%; annual bone gain for women: 0.51%). After surpassing the peak bone mass, a continuous decline of the DXR BMD could be observed that was accentuated in women (annual bone loss for men: 0.91%; annual bone loss for women: 0.86%). A most relevant cortical demineralization was found in the age range from 60 to 69 yr with an annual bone loss of 1.62% in postmenopausal women; in men, two delayed peaks of demineralization could be described in the age group from 65 to 69 yr (5.1%) and between the ages of 80 and 84 yr (11.6%). Comparable results were revealed for the MCI.
Discussion The intention of this study was to estimate normative German reference data for DXR in men and women. Currently, reference data for DXR parameters have been established for Caucasian North-American (27), African American (28), Scandinavian (29), German (30,31), Chinese (32), Hispanic (33), Korean (34), and Indian populations. In contrast to this study, most of the published reference data are only available for women (27,28,30). Recently it has been recognized that osteoporosis represents an important public health issue in men as well (35). Since the epidemiological data have Volume 9, 2006
DXR Normative Data
345
Table 2 Mean and Standard Deviation Regarding the Parameters of Digital X-Ray Radiogrammetry Women, X-Rayhand Age (yr)
BMD (g/cm2) mean (SD)
n 96 106 127 82 75 72 56 59 57 41 37 42 22 25 26 31 954
!15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
0.450 0.542 0.552 0.574 0.585 0.573 0.565 0.566 0.552 0.527 0.490 0.451 0.456 0.429 0.415 0.385 0.526
and wrist;
n 5 954
MCI mean (SD)
(0.057) (0.066) (0.064) (0.069) (0.066) (0.068) (0.056) (0.066) (0.081) (0.070) (0.073) (0.061) (0.069) (0.064) (0.080) (0.057) (0.087)
0.399 0.473 0.477 0.496 0.511 0.494 0.486 0.466 0.465 0.438 0.384 0.357 0.343 0.325 0.317 0.278 0.445
CT (cm) mean (SD)
(0.065) (0.061) (0.070) (0.068) (0.063) (0.071) (0.056) (0.060) (0.058) (0.057) (0.052) (0.046) (0.054) (0.043) (0.054) (0.044) (0.087)
0.141 0.179 0.181 0.192 0.195 0.190 0.186 0.185 0.181 0.170 0.152 0.139 0.139 0.128 0.123 0.116 0.171
BW (cm) mean (SD)
(0.022) (0.025) (0.026) (0.029) (0.028) (0.028) (0.024) (0.026) (0.031) (0.024) (0.026) (0.024) (0.026) (0.021) (0.028) (0.029) (0.035)
0.714 0.762 0.761 0.776 0.764 0.771 0.775 0.792 0.780 0.773 0.786 0.775 0.807 0.792 0.772 0.803 0.768
(0.076) (0.073) (0.071) (0.077) (0.073) (0.077) (0.077) (0.084) (0.098) (0.080) (0.096) (0.101) (0.080) (0.098) (0.096) (0.092) (0.083)
Abbr: BMD, bone mineral density; BW, bone width; CT, cortical thickness; MCI, metacarpal index; SD, standard deviation.
shown that male osteoporosis occurs half as frequently as female osteoporosis, and that men are often afflicted with secondary osteopenia, awareness of this has increased. Consequently, male reference values are strongly recommended as necessary for a reliable identification of male osteoporosis. Therefore our study closes the gap of normative Caucasian data, as estimated by the new software version of DXR, and includes both German males and females.
Comparison of DXR BMD Between German Caucasian Women and Men Our results demonstrate that the mean DXR BMD for German Caucasian women (0.526 6 0.087 g/cm2) is considerably lower than mean DXR BMD for German Caucasian men
(0.603 6 0.091 g/cm2). Wuster et al. (31) demonstrated comparable values of DXR BMD for German Caucasian women (0.554 6 0.067 g/cm2) and men (0.654 6 0.072 g/cm2) with the obsolete DXR version 1.0 in a small cohort. Another German reference cohort, which included only women, and is evaluated by Toledo and Jergas (30), reported higher reference values in all age groups. The reliability of these data is debatable, because Toledo and Jergas (30) preferred a retrospective study design, and consequently only patients with visible fractures were excluded. Otherwise, also individuals with disease-related and drug-induced alterations of the peripheral bone status were included (in particular, subjects with intake of bone affecting drugs [e.g., vitamin D or calcium] were considered); such patients were excluded in our study design.
0.9 0.8 0.7 0.5
BMD
BMD
0.6 0.4 0.3 0.2 0.1 0 0
10
20
30
40
50
60
70
80
90
100
Age
Fig. 1. Bone mineral density (BMD) reference curve for men (X-rayhand and wrist; n 5 1131). Journal of Clinical Densitometry
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
10
20
30
40
50
60
70
80
90
100
110
Age
Fig. 2. Bone mineral density (BMD) reference curve for women (X-rayhand and wrist; n 5 954). Volume 9, 2006
346
Bo¨ttcher et al. Table 3 Comparison of Mean Values for DXR Parameters Between Men and Women (n 5 2085) Men mean (SD)
BMD (g/cm2) MCI CT (cm) BW (cm)
0.603 0.456 0.196 0.864
Women mean (SD)
(0.091) (0.069) (0.035) (0.097)
0.526 0.445 0.171 0.768
(0.087) (0.087) (0.035) (0.083)
Relative difference in %
Significance (ManneWhitney U test)
12.8% 2.4% 12.8% 11.1%
p ! 0.01 p ! 0.05 p ! 0.01 p ! 0.01
Abbr: BMD, bone mineral density; BW, bone width; CT, cortical thickness; DXR, digital X-ray radiogrammetry; MCI, metacarpal index.
A possible limitation of our own study design should not remain unmentioned. The reason for our lower reference values could be induced by a selection bias, because our study population only consisted of diseased subjects who requested help in an emergency unit. Although our questionnaire considered all relevant bone affecting results, this population may represent a cohort with a higher risk for accidents, such as the use of alcohol or inclusive of patients with recurrent falls due to cardiovascular disorders or also osteoporosis. The influence regarding a different standard of living and nutrition in the Eastern part of Germany before 1989 is unlikely (regular government aid for sports and calcium supply via milk), but should be not completely negated. A further limitation of our study design is the lack of comparative technology and of anthropometric details. Our study showed a peak bone mass for women between the ages of 30 and 34 yr and for men it was delayed and between the ages of 45 and 49 yr. This gender-specific observation may be understood as caused by the occurrence of puberty at an earlier stage in women. Similarly Hyldstrup and Nielsen (36) presented normative data of MCI estimated by DXR (version 2.0) for 384 healthy Danish women, which also revealed the maximum gain of MCI between the ages of 30 and 39 yr, whereas the mean value of MCI with 0.486 was somewhat higher compared with our results. In contrast 0.7
BMD in g/cm2
0.6 0.5 0.4 BMD-women
0.3
BMD-men
84 80
–
74 – 70
–
64
54 60
50
–
44 40
–
34 – 30
– 20
<
51
24
0.2
Age groups
Fig. 3. Comparison of age-related bone mineral density (BMD) between men and women (X-rayhand and wrist; n 5 2085). Journal of Clinical Densitometry
to these results, Toledo and Jergas (30) reported the peak bone mass for women between the ages of 40 and 44 yr. The data of Wuster et al. (31) verified the peak bone mass for DXR BMD between the ages of 40 and 49 yr in women and 30 and 39 yr in men. This difference may probably be due to a selection bias, and it may also be significantly influenced by variations regarding the statistical procedures (e.g., grouping of the study population). Furthermore, an important limitation of the studies from Toledo and Jergas (30) and Wuster et al. (31) was the small number of subjects. In the present study, the quantity of subjects in the female cohort (n 5 954) was duplicated compared with the cohort presented by Toledo and Jergas (30). When the results regarding MCI from the present study are compared with published data (26), the same tendency could be shown. Pre-menopausal women reveal higher MCI values than men within the comparable age groups, whereas postmenopausal women have lower age-related MCI data than men. The observation that pre-menopausal women have a higher MCI than men is an apparent contradiction to the fact that the same women (in general) have a smaller cortical thickness as well as lower BMD. The apparent contradiction is explained by the fact that women also have a smaller bone diameter, and therefore, the MCI is higher for pre-menopausal women. When women reach menopause, the subsequent relatively fast bone loss results in a thinning of the cortices that lowers the MCI much faster than in men. A few years after menopause, the postmenopausal women are then observed to have a lower MCI than the age-matched men (31). It is common knowledge that women have a reduced cortical thickness and a smaller metacarpal bone diameter in comparison with male bone proportions. The MCI offers only a minimal gender-specific difference due to the specific calculation of this parameter as ratio of the mean cortical thickness normalized with the mean outer bone diameter. A further, but still neglected parameter of DXR is the total DXR BW of the metacarpals; our data revealed a continuos progressive increase of DXR BW with aging accentuated for men. For women, the maximum of bone width was reached between the ages of 70 and 74 yr, whereas men demonstrated a delayed peak level in the age range of 75e79 yr. Furthermore young men (age range: 15e19 yr) already showed a greater metacarpal bone width as compared with women Volume 9, 2006
DXR Normative Data
347
Table 4 Age-Related Relative Changes of DXR BMD and MCI (Men, X-rayhand Age (yr) !15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
BMD (g/cm2) Mean (SD)
n 93 200 183 121 93 104 75 65 50 46 23 24 19 17 7 11 1131
0.460 0.596 0.619 0.625 0.625 0.634 0.641 0.649 0.627 0.626 0.615 0.585 0.566 0.559 0.501 0.478 0.603
BMD relative changes
(0.072) (0.074) (0.078) (0.077) (0.091) (0.079) (0.071) (0.067) (0.080) (0.079) (0.053) (0.076) (0.079) (0.084) (0.040) (0.089) (0.091)
n 5 1131)
MCI mean (SD) 0.374 0.459 0.476 0.482 0.481 0.481 0.465 0.477 0.463 0.441 0.430 0.415 0.382 0.367 0.337 0.335 0.456
þ29.6% þ3.9% þ1.0% 0.0% þ1.4% þ1.1% þ1.2% 3.5% 0.2% 1.8% 5.1% 3.4% 1.3% 11.6% 4.8%
and wrist,
MCI relative changes
(0.053) (0.056) (0.054) (0.064) (0.066) (0.056) (0.060) (0.062) (0.054) (0.061) (0.058) (0.060) (0.056) (0.057) (0.063) (0.070) (0.069)
þ22.7% þ3.7% þ1.3% 0.2% 0.0% 3.4% þ2.6% 3.0% 5.0% 2.6% 3.6% 8.6% 4.1% 8.2% 0.6%
Abbr: BMD, bone mineral density; CT, cortical thickness; DXR, digital X-ray radiogrammetry; MCI, metacarpal index.
of the same age and magnitude of the age-associated increment (i.e., BW), which was higher in men than in women. Both gender also offered a remarkable increase of DXR BW in the adolescent years (from ages ! 15 yr to an age range of 15e19 yr). Our findings could be confirmed by Russo et al. (37) using measurements of peripheral quantitative
computed tomography. The authors suggested that the increase of bone width in aging men may contribute to the maintenance of adequate bone mechanical competence in the face of declining cortical bone mass. In women, this compensatory mechanism appears to be less efficient, and consequently, the bone mechanical competence is more impaired with age.
Table 5 Age-Related Relative Changes of DXR BMD and MCI (Women, X-Rayhand Age (yr) !15 15e19 20e24 25e29 30e34 35e39 40e44 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80e84 O85 Total
n 96 106 127 82 75 72 56 59 57 41 37 42 22 25 26 31 954
BMD (g/cm2) Mean (SD) 0.450 0.542 0.552 0.574 0.585 0.573 0.565 0.566 0.552 0.527 0.490 0.451 0.456 0.429 0.415 0.385 0.526
(0.057) (0.066) (0.064) (0.069) (0.066) (0.068) (0.056) (0.066) (0.081) (0.070) (0.073) (0.061) (0.069) (0.064) (0.080) (0.057) (0.087)
BMD þ20.4% þ1.8% þ4.0% þ1.9% 2.1% 1.4% þ0.2% 2.5% 4.7% 7.6% 8.6% þ1.1% 6.3% 3.4% 7.8%
and wrist,
n 5 954)
MCI Mean (SD) 0.399 0.473 0.477 0.496 0.511 0.494 0.486 0.466 0.465 0.438 0.384 0.357 0.343 0.325 0.317 0.278 0.445
(0.065) (0.061) (0.070) (0.068) (0.063) (0.071) (0.056) (0.060) (0.058) (0.057) (0.052) (0.046) (0.054) (0.043) (0.054) (0.044) (0.087)
MCI þ18.5% þ0.8% þ4.0% þ3.0% 3.4% 1.6% 4.3% 0.2% 6.2% 14.1% 7.6% 4.1% 5.5% 2.5% 14.0%
Abbr: BMD, bone mineral density; CT, cortical thickness; DXR, digital X-ray radiogrammetry; MCI, metacarpal index. Journal of Clinical Densitometry
Volume 9, 2006
348 This geometric adaptation of increasing bone width, which occurs parallel to the cortical bone loss detected by MCI and DXR BMD, could be an important component in the assessment of bone mechanical properties and for the diagnosis of osteoporosis. These coherences, which are methodologically overlooked by planar densitometry, were also described by other studies focusing on micro-radioscopy and micro-densitometry (38,39). In detail, recent studies revealed a cortical bone loss in postmenopausal as well as in age-associated demineralization caused by increased intra-cortical porosity and reduced cortical thickness, which is related to endosteal and juxta-endosteal bone resorption (40). These changes tend to be compensated by periosteal apposition (i.e., increase of metacarpal DXR BW) in both genders, but more obviously in men than in women with the previously mentioned consequences (37,40). As expected, women showed an accentuated total DXR BMD decline than men after passing through the peak bone mass period (men: 31.7%; women: 43.1%). In addition, men revealed two age-related peaks of cortical bone loss (age group: 65e69 and 80e84 yr) indicated by a decrease of cortical thickness, MCI, and DXR BMD in our study. For very elderly men, the second accentuated maximum of bone loss could be explained by the combined age-induced effect of decreasing testosterone, growth hormone, and insulin-like growth factor I (38), whereas in women, our results documented more of a continuous decline of DXR BMD, MCI, and cortical thickness. The interdependency of DXR parameters, in particular focusing on cortical thickness and DXR BW, in patients with primary and secondary osteoporosis, is most relevant for the understanding of various physiological processes regarding the cortical bone partition deserving further intensive investigations.
Influence of Bone Age on DXR BMD Measurement Several other studies have reported the same finding that a quantification of BMD could be achieved by the DXR system, with a bone age of 6 yr and older (30,41,42). In children with a skeletal age younger than 6 six yr, the DXR system can not reliably recognize the bone structure and detect the edges of the metacarpals, and thus automatically discontinues the DXR analysis.
Impact of Different X-Ray Settings (X-rayhand Versus X-raywrist) on DXR Measurement It has been reported that DXR provides calculations of DXR BMD and MCI with high intra-radiograph and inter-radiograph reproducibility of conventional, as well as digitally performed radiographs (14). This technique has been shown to be insensitive toward alterations of film sensitivity, film brand, film-focus distance, and exposure level (14,15). Only variation in the tube voltage and the use of different Xray equipment demonstrated a significant influence on the calculation process of DXR BMD and MCI for conventional and Journal of Clinical Densitometry
Bo¨ttcher et al. particularly for digital printouts (14,15). Toledo and Jergas (30) included data calculated from hand and wrist radiographs in their study, which were performed with different tube voltage (hand X-rays: 46 kV; wrist X-rays: 50 kV), and therefore a limited reliability of these data should be discussed. In our study, the image-capturing conditions of hand and wrist radiographs were similar, and the same X-ray device was used. The present study evaluated the impact of different X-ray settings (X-rayhand vs. X-raywrist) on DXR-measurement; this circumstance should be evaluated, because the process of contour finding and edge detection during DXR measurements could be impaired by different X-ray settings in spite of constant image-capturing parameters. The performance of different radiographs focused at the wrist or hand results from a various fracture localization; the equivalent use of these different X-ray settings (without alteration of the DXR results) facilitates the identification of osteoporosis in combination with the onset of fractures at the distal upper extremities. The results show no significant difference between the two types of images, which were performed in the identical anterior-posterior projection. Hence, it is not mandatory to use only one X-ray setting in clinical practice, but it is advisable to use consistent technical conditions to maintain the best reliability, in particular when normative data were compared with DXR measurements of patients suffering from various bone affecting diseases. Toledo and Jergas (30) revealed an important fact: awareness that DXR BMD and DXR cortical thickness of the dominant hand (right handedness in 93% of a population) show significantly higher values than for the nondominant hand. Therefore, only DXR estimates of the nondominant hand provide reliable data as considered in this study, because measurements of the dominant hand could show increased cortical bone mineralization caused by increased physical activity.
Comparison of DXR-BMD Between Different Ethnic Groups Recent studies have shown higher results for DXR BMD in Scandinavian (0.555 6 0.070 g/cm2; 29), Indian (0.544 6 0.066 g/cm2; 35), and North American Caucasian women (0.558 6 0.065 g/cm2; 27) compared with our own DXR data. This observation may be caused by the use of the version 1.0 of the DXR system. The obsolete version (1.0) had calculated the cortical width and overall BW at 5 locations: the second, third, and forth metacarpal bone, as well as the distal radius (radial side, only cortical width), and the distal ulna (ulnar side, only cortical width). In contrast with the older version, the newer version 2.0 of the DXR system estimates DXR BMD only on the second, third, and fourth metacarpals of the nondominant hand. In various studies, the extensive ethnic differences of BMD have been evaluated (e.g., BMD measured by DXA on the proximal femur) and showed an accentuated increase in non-Hispanic blacks compared with Mexican Americans Volume 9, 2006
DXR Normative Data and non-Hispanic whites (43). Significantly higher results of DXR (þ14.0%) have also been observed for the AfricanAmerican females (0.597 6 0.055 g/cm2) compared with German females (28). Further ethnic differences in DXR BMD values of Caucasian, Korean, and Indian men, as well as North-American, African-American, Scandinavian, and Indian women (32e34,36), emphasize the essential requirement to evaluate and establish a gender-specific, ethnically differentiated database for each osteodensitometric technique, including DXR. Based on the availability of gender-specific normative data for various ethnic populations and promoted by the excellent reproducibility, without a possible influence of soft tissue alterations during the measurement, DXR is at the forefront of competing osteodensitometric techniques in detecting patients with high fracture risk due to primary or secondary osteoporosis. The cutting edge of DXR could be confirmed by Boonen et al. (1), who demonstrated a better sensitivity for identifying women both with and without osteoporosis (e.g., positive and negative predictive values) by DXR compared with quantitative ultrasound, and suggest that DXR BMD and phalangeal radiographic absorptiometry may be the most effective for DXA testing in high-risk postmenopausal women.
Conclusion The development of digital imaging techniques has promoted the precise measurement of several radiogeometrical features, and DXR has provided a reliable diagnostic approach for the quantification of cortical bone loss in the clinical routine. The present study has estimated normative reference values for DXR in German Caucasian women and men. The verified influence of ethnic groups and gender on normative values has revealed the need for regional reference data to achieve a reliable interpretation of DXR measurements. Furthermore, our normative values insure the applicability of the T-score by comparing estimated DXR BMD with the normative values in young healthy women or men, and also the Z-score that focused on age-matched controls. The now available T- and Z-scores enable application of DXR as a screening tool to identify patients who may benefit the most from axial DXA measurements. Further prospective studies will be required to determine the predictive value of DXR to distinguish patients with high fracture risk, and to introduce the DXR technique as a cost-effective, widely available, and precise method in the clinical routine.
Acknowledgments We would like to thank Monika Arens, (managing director, Arewus GmbH) for the use of the X-posure equipment and R€ udiger Vollandt, PhD for the statistical advice. Finally, the authors would also like to thank Dieter Felsenberg, MD (Berlin, Germany), Claus C. Gluer, PhD (Kiel, Germany), Stephan Grampp, MD, and Herwig Imhof, MD (Wien, Austria) for their comments regarding this study.
Journal of Clinical Densitometry
349
References 1. Boonen S, Nijs J, Borghs H, Peeters H, Vanderschueren D, Luyten FP. 2005 Identifying postmenopausal women with osteoporosis by calcaneal ultrasound, metacarpal digital X-ray radiogrammetry and phalangeal radiographic absorptiometry: a comparative study. Osteoporos Int 16:93e100. 2. Bouxsein ML, Palermo L, Yeung C, Black DM. 2002 Digital X-ray radiogrammetry predicts hip, wrist and vertebral fracture in elderly women: a prospective analysis from the study of osteoporotic fractures. Osteoporos Int 13:358e365. 3. Hyldstrup L, Jorgensen JT, Sorensen TK, Baeksgaard L. 2001 Response of cortical bone to antiresorptive treatment. Calcif Tissue Int 68:135e139. 4. Ozcakar L, Guven GS, Unal S, Akinci A. 2005 Osteoporosis in Turkish HIV/AIDS patients: comparative analysis by dual energy X-ray absorptiometry and digital X-ray radiogrammetry. Osteoporos Int 16:1363e1367. 5. Ward KA, Cotton J, Adams JE. 2003 A technical and clinical evaluation of digital X-ray radiogrammetry. Osteoporos Int 14: 389e395. 6. Barnett E, Nordin B. 1960 The radiological diagnosis of osteoporosis: a new approach. Clin Radiol 11:166e174. 7. Meema HE. 1962 The occurrence of cortical bone atrophy in old age and osteoporosis. J Can Assoc Radiol 13:27e32. 8. Meema HE, Meema S. 1963 Measurable roentgenologic changes in some peripheral bones in senile osteoporosis. J Am Geriatr Soc 11:1170e1182. 9. Garn S, Poznanski A, Nagy J. 1971 Bone measurement in the differential diagnosis of osteopenia and osteoporosis. Radiology 100:509e518. 10. Exton-Smith A, Millard P, Payne P, Wheeler E. 1969 Method of measuring quantity of bone. Lancet 2:1153e1157. 11. Adams P, Davies G, Sweetnam P. 1969 Observer error and measurements of the metacarpal. Br J. Radiol 42:192e197. 12. Chan E, Pandith V, Towheed TE, Brouillard D, Zee B, Anastassiades TP. 1998 Comparison of the combined cortical thickness of the second metacarpal with the Sharp’s method for scoring hand microradiographs in rheumatoid arthritis. J Rheumatol 25:1290e1294. 13. Aguado F, Revilla M, Villa LF, Rico H. 1997 Cortical bone resorption in osteoporosis. Calcif Tissue Int 60:323e326. 14. Bo¨ttcher J, Pfeil A, Rosholm A, et al. 2005 Influence of imagecapturing parameters on Digital X-Ray Radiogrammetry. J Clin Densitom 8:87e94. 15. Malich A, Bo¨ttcher J, Pfeil A, et al. 2004 The influence of technical parameters of Xeray imaging on reproducibility and precision of digital computer-assisted radiogrammetry (DXR). Skeletal Radiol 33:473e482. 16. Rosholm A, Hyldstrup L, Baeksgaard L, Grunkin M, Thodberg HH. 2001 Estimation of bone mineral density by digital X-ray radiogrammetry: theoretical background and clinical testing. Osteoporos Int 12:961e969. 17. Bo¨ttcher J, Pfeil A, Rosholm A, et al. 2006 Computerized digital imaging techniques provided by digital X-ray radiogrammetry as new diagnostic tool in rheumatoid arthritis. J Digit Imaging April 26 (DOI: 10.1007/s10278-006-0263-y). 18. Bo¨ttcher J, Malich A, Pfeil A, et al. 2004 Potential clinical relevance of digital radiogrammetry for quantification of periarticular bone demineralization in patients suffering from rheumatoid arthritis dependent on severity and compared with DXA. Eur Radiol 14:631e637. 19. Bo¨ttcher J, Pfeil A, Heinrich B, et al. 2005 Digital radiogrammetry as a new osteodensitometric technique and diagnostic tool in rheumatoid arthritis. Rheumatol Int 25:457e464.
Volume 9, 2006
350 20. Bo¨ttcher J, Pfeil A, Rosholm A, et al. 2005 Digital X-ray radiogrammetry combined with semi-automated analysis of joint space distances as a new diagnostic approach in rheumatoid arthritisea cross-sectional and longitudinal study. Arthritis Rheum 52:3850e3859. 21. Stewart A, Mackenzie LM, Black AJ, Reid DM. 2004 Predicting erosive disease in rheumatoid arthritis. A longitudinal study of changes in bone density using digital X-ray radiogrammetry: a pilot study. Rheumatalogy 43:1561e1564. 22. Mentzel HJ, John U, Bo¨ttcher J, et al. 2005 Evaluation of bonemineral density by digital X-ray radiogrammetry (DXR) in pediatric renal transplant recipients. Ped Radiol 35:489e494. 23. Van Rijn RR, Boot A, Wittenberg R, et al. 2006 Direct X-ray radiogrammetry versus dual-energy X-ray absorptiometry: assessment of bone density in children treated for acute lymphoblastic leukemia and growth hormone deficiency. Ped Radiol 24:1e6. 24. Cootes TF, Taylor CJ, Cooper D, Graham J. 1995 Active shape models: their training and application. Comp Vision Image Understand 61:38e59. 25. Lazenby R. 1995 Brief communication: non-circular geometry and radiogrammetry of the second metacarpal. Am J Phys Anthropol 97:323e327. 26. Jorgensen JT, Andersen PB, Rosholm A, Bjarnason NH. 2000 Digital X-ray radiogrammetry: a new appendicular bone densitometric method with high precision. Clin Physiol 20:330e335. 27. Black DM, Palermo L, Sorensen T, et al. 2001 A normative reference database study for Pronosco X-posure system. J Clin Densitom 4:5e12. 28. Shepherd JA, Meta M, Landau J, et al. 2005 Metacarpal index and bone mineral density in healthy African-American women. Osteoporos Int 16:1621e1626. 29. Cummings S, Hyldstrup L, Liu Z, Wuster C, Maffei L. 2000 Normative reference curves for the Pronosco X-posure system established in five ethnic groups. J Bone Miner Res 15(Suppl 1):321. 30. Toledo VAM, Jergas M. 2005 Age-related changes in cortical bone mass: data from a German female cohort. Eur Radiol 8:1e7. 31. Wuster C, Wenzler M, Kappes J, Rehm C, G€ uhring T, Arnbjerg D. 2000 Digital X-ray radiogrammetry as a clinical method for estimating bone mineral densityea German reference database. J Bone Miner Res 15(Suppl 1):298.
Journal of Clinical Densitometry
Bo¨ttcher et al. 32. Jiang Y, Miao Li, Zhonghou L. 2000 Normal values of forearm bone BMD and incidence of primary osteoporosis in chinese women. Chinese Journal of Osteoporosis 6:30e31. 33. Maffei L, Venarotti HO, Treviso JC, Sorensen TK, Nissen DS. 2000 Digital X-ray radiogrammetry: a Hispanic normative reference database for the Pronosco X-posure System. J Bone Miner Res 15(Suppl 1):304. 34. Yang SO, Ham SY, Kim YI, et al. 2001 Comparison of Korean female and male reference data using DXR. J Bone Miner Res 16(Suppl 1):304. 35. Wuster C, Hadji P. 2001 Use of quantitative ultrasound densitometry (QUS) in male osteoporosis. Calcif Tissue Int 69: 225e228. 36. Hyldstrup L, Nielsen SP. 2001 Metacarpal index by digital X-ray radiogrammetry: normative reference values and comparison with dual X-ray absorptiometry. J Clin Densitom 4:229e306. 37. Russo CR, Lauretani F, Bandinelli S, et al. 2003 Aging bone in men and women: beyond changes in bone mineral density. Osteoporos Int 14:531e538. 38. Iwamoto J, Takeda T, Ichimura S, Tsukimura Y, Toyama Y. 2000 Age-related changes in cortical bone in men: metacarpal bone mass measurement study. J Orthop Sci 5:4e9. 39. Meema HE, Meema S. 1987 Longitudinal microradioscopic comparison on endosteal and juxtaendosteal bone loss in premenopausal and postmenopausal women, and in those with end-stage renal disease. Bone 8:343e350. 40. Russo CR, Lauretani F, Seeman E, et al. 2006 Structural adaptations to bone in aging men and women. Bone 38:112e118. 41. Malich A, Freesmeyer MG, Mentzel HJ, et al. 2003 Normative values of bone parameters of children and adolescents using digital computer-assisted radiogrammetry (DXR). J Clin Densitom 6:103e112. 42. van Rijn RR, Grootfaam DS, Lequin MH, et al. 2004 Digital radiogrammetry of the hand in a pediatric and adolescent Dutch Caucasian population: normative data and measurements in children with inflammatory bowel disease and juvenile chronic arthritis. Calcif Tissue Int 74:342e350. 43. Looker AC, Wahner HW, Dunn WL, et al. 1995 Proximal femur bone mineral levels of US adults. Osteoporos Int 5: 389e409.
Volume 9, 2006