- Email: [email protected]
Trabecular architecture in women and men of similar bone mass with and without vertebral fracture: I. two-dimensional histology
Bone Vol. 27, No. 2 August 2000:271–276
Trabecular Architecture in Women and Men of Similar Bone Mass With and Without Vertebral Fracture: I. Two-Dimensional Histology L. D. HORDON,1 M. RAISI,1 J. E. AARON,1 S. K. PAXTON,1 M. BENETON,2 and J. A. KANIS2 1
School of Biomedical Sciences, Worsley Medical and Dental Building, University of Leeds, Leeds, UK Centre for Metabolic Bone Diseases at Sheffield (WHO Collaborating Centre), Medical School, University of Sheffield, Sheffield, UK
2
Key Words: Osteoporosis; Fracture; Bone mineral density; Bone histomorphometry; Microarchitecture; Two-dimensional (2D) image analysis.
While osteoporosis is characterized by a low bone mass there is a well-recognized overlap in bone mineral density (BMD) measurements between groups of subjects with and without vertebral fracture. To investigate whether differences in trabecular architecture may contribute to the presence or absence of fractures independent of the bone mass, fracture and nonfracture groups matched for age, gender, and BMD were assembled. Transiliac biopsies and corresponding lumbar spine BMD measurements from 31 women and 16 men with vertebral fracture were compared with those from 22 women and 11 men without fracture. Lumbar BMD (L1– 4) was measured using a Hologic 2000 densitometer. The lumbar BMD was similar in women with and without fracture (0.63 g/cm3 ⴞ 0.10 SD and 0.71 g/cm3 ⴞ 0.17 SD, n.s.) and in men with and without fracture (0.72 g/cm3 ⴞ 0.12 SD and 0.76 g/cm3 ⴞ 0.17 SD, n.s.). Undecalcified iliac crest biopsy sections, 8 m thick, were analyzed for remodeling variables and trabecular architecture using OsteoMeasure and TAS image analysis systems. No significant difference was found in either gender between fracture and nonfracture groups in percent bone volume (mean 10% in all groups), or in the wide range of remodeling and architectural variables measured, including the trabecular width, number, and separation, mean trabecular plate density and fractal dimension, as well as several indirect indices of connectivity including the node: terminus ratio, marrow star volume, and trabecular pattern factor. On the basis of this evidence it was concluded that there is no difference in the trabecular architecture between patients with crush fracture and controls when account is taken of bone mass. This suggests that microanatomical disruption is a predictable intrinsic feature of bone loss. However, there remains the possibility that the two-dimensional character of the structural deterioration measured indirectly is not sufficiently sensitive for the complex cancellous system. This is considered further in part II. (Bone 27: 271–276; 2000) © 2000 by Elsevier Science Inc. All rights reserved.
Introduction There is extensive evidence in the literature to suggest that the mechanical strength of bone is not solely a function of its mass or bone mineral density (BMD). While the BMD at the spine and the hip is directly related to the risk of fracture at those sites,8 there is considerable overlap in the BMD measured in groups of subjects with and without fracture, despite the higher mean value in the latter.11 Both the initial BMD and the presence of existing vertebral fractures together apparently combine to predict more reliably the occurrence of future vertebral fractures.17 The complex architecture of cancellous bone also determines the mechanical strength,14 and differences in microstructural arrangement between aging men and women may contribute to the lower rates of vertebral and hip fractures in men.2 It follows that differences in cancellous architecture may explain why some women with a low vertebral BMD have vertebral fractures, whereas others with an identical BMD do not. However, examination of this aspect in matched groups, with and without vertebral fractures using either histomorphometric10,15 or microradiographic techniques7 have produced differing results. In endeavoring to resolve this, the aim of the present study was to compare cancellous architecture at the iliac crest in groups of subjects with and without vertebral fracture who were matched for age, gender, and lumbar BMD, to determine whether the difference in cancellous architecture might explain the difference in fracture status. Materials and Methods Patients who had undergone both iliac crest biopsies and lumbar BMD measurements were identified from databases at the Royal Hallamshire Hospital, Sheffield. The records of these patients were reviewed and any who had received drugs for the treatment of osteoporosis (e.g., fluoride, bisphosphonates, and anabolic steroids) prior to biopsy or BMD measurements were excluded, as were those who were receiving or had received oral corticosteroids, or who had primary hyperparathyroidism or renal impairment (defined when serum creatinine exceeded the normal range). The patients were classified into groups with and without vertebral fracture. A total of 89 patients were eligible for study. Seven were excluded because of an inadequate biopsy, and two were omitted as they could not lie flat for densitometry. The final
Address for correspondence and reprints: Dr. J. E. Aaron, School of Biomedical Sciences, Worsley Medical and Dental Building, University of Leeds, Leeds LS2 9JT, UK. E-mail: [email protected] © 2000 by Elsevier Science Inc. All rights reserved.
271
8756-3282/00/$20.00 PII S8756-3282(00)00329-X
272
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
Bone Vol. 27, No. 2 August 2000:271–276
Figure 1. Automated trabecular analysis system (TAS) showing the binary image and its thinned (skeletonized) counterpart together with the area of interest defined by the adjustable rectangle, as well as typical printouts. Examples from the nonfracture (upper) and fracture (lower) groups of women are shown.
female fracture group comprised 31 women, 25 of whom had more than one vertebral fracture (median four), and six of whom had a single fracture (defined as ⬎20% loss of anterior, middle, or posterior vertebral height). The male fracture group comprised 11 men with more than one vertebral fracture (median four), and 5 men with a single fracture. Lumbar spine fractures were present in 12 women and 9 men. The patients without a vertebral fracture consisted of 11 men and 22 women. They were not entirely healthy skeletally as they had come to medical attention for reasons including peripheral fracture, radiological osteopenia, back pain, family history of osteoporosis, or concern about the possibility of having osteoporosis. Transiliac bone biopsies were taken after double tetracy-
cline labeling, using local anesthetic and an 8-mm-diameter Bordier trephine. Bone mineral density was assessed at L1– 4 using a Hologic 2000 densitometer within 3 months of the bone biopsy and generally during the same admission. Severely fractured vertebrae were excluded from the measurement area. The biopsies were preserved in 70% ethanol, dehydrated, and embedded using established procedures.2 Sections, 8 m thick, were cut on a Jung K heavy-duty microtome, stained by the Goldner tetrachrome method, and structural and remodeling (i.e., formation and resorption) variables analyzed using the semiautomated OsteoMeasure system (OsteoMetrics, Inc., Atlanta, GA) and digitizing pad. Unstained sections were examined under an epifluorescence microscope and the mineralization rate determined by mea-
Bone Vol. 27, No. 2 August 2000:271–276
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
Table 1. Age, lumbar bone mineral density, and other characteristics of women with and without vertebral fracture (mean ⫾ SD)
Age (years) BMD (g/cm2) Age at menopause (years) Peripherala osteoporotic fracture Present or past use of hormone replacement therapy
Women with vertebral fracture (n ⫽ 31)
Women without vertebral fracture (n ⫽ 22)
66.0 ⫾ 9.4 0.63 ⫾ 0.10 46.4 ⫾ 6.5 (n ⫽ 28) n ⫽ 14
63.5 ⫾ 8.7 (n.s.) 0.71 ⫾ 0.17 (n.s.) 48.2 ⫾ 5.1 (n ⫽ 17) (n.s.) n ⫽ 4 (n.s.)
n⫽3
n ⫽ 4 (n.s.)
KEY: BMD, bone mineral density; n.s., nonsignificant. a Defined as previous Colles, femoral neck, proximal humerus, or pubic ramus fractures.
suring the distance between the fluorescent bands that were temporally 10 days apart. Additionally, four to six sections stained either by the Goldner method or in 0.1% toluidine blue, pH 3.5, for 30 min were analyzed by means of our automated trabecular analysis system (TAS), which measures a comprehensive range of structural variables1 using the intact binary image and its thinned counterpart (Figure 1). There is also the option to include both the marrow star volume19 (using either the random generation of a selected number of points or a grid of points from which to automatically measure the mean radial path length scanned through the cavity to the surrounding trabecular surface) and the fractal dimension21 (a mathematical descriptor of structural complexity). The area of interest was defined by an adjustable window and was selected to encompass as much intact cancellous bone in the section as possible and to exclude the cortical tissue. An editing facility enabled minor artifacts, such as sectioning cracks, to be removed before analysis (see Aaron et al.1 for details). Once the groups had been assembled, histomorphometry was performed blind with respect to fracture status. Statistical analysis was carried out using Student’s t-test and the chi-square test with Yates correction.
The characteristics of the patients together with their lumbar BMD are shown in Tables 1 and 2. No significant difference was seen between the fracture and nonfracture groups of either gender in terms of age, lumbar BMD, or peripheral fracture, and in the women there was no difference in the age at menopause or use of hormone replacement therapy. When the OsteoMeasure analysis system was applied to the bone biopsies, the similarity Table 2. Age, lumbar bone mineral density (BMD), and peripheral osteoporotic fracture in men with and without vertebral fracture (mean ⫾ SD)
KEY: n.s., nonsignificant.
Table 3. Comparison of histomorphometric data in women with and without vertebral fracture using OsteoMeasure (mean ⫾ SD)
BV/TV (%) OS/BS (%) ObS/BS (%) ES/BS (%) OC.S/BS (%) Tb.Wi (m) O.Th (m) Tb.Sp (m) Tb.N (/mm) MAR (m/day) Calcification front/OS (%)
Women with vertebral fracture (n ⫽ 31)
Women without vertebral fracturea (n ⫽ 22)
11.76 ⫾ 5.79 9.17 ⫾ 6.49 0.28 ⫾ 0.50 4.01 ⫾ 3.62 0.40 ⫾ 0.66 110.70 ⫾ 30.07 7.09 ⫾ 2.70 800.3 ⫾ 314.1 1.26 ⫾ 0.49 0.51 ⫾ 0.27 76 ⫾ 29
11.21 ⫾ 3.83 9.37 ⫾ 8.81 0.50 ⫾ 1.05 3.50 ⫾ 3.39 0.44 ⫾ 0.66 104.10 ⫾ 16.56 8.04 ⫾ 2.87 766.2 ⫾ 274.8 1.30 ⫾ 0.43 0.57 ⫾ 0.23 62 ⫾ 34
See Results for abbreviations. a All differences nonsignificant.
in bone status of the fracture and nonfracture groups was confirmed by their identical mean relative bone volume measurement of 11% (Tables 3 and 4) . No significant difference was found between the remodeling variables. The fracture group tended to have fewer, more separated and thicker trabeculae as indicated by OsteoMeasure, but again the difference did not reach significance. Among the men, remodeling tended to be marginally higher in the fracture group and the trabeculae marginally fewer, more separated, and thicker, but not significantly so. The application of TAS to the histological sections again confirmed the well-matched cancellous bone status of the groups, the relative bone volume now measuring around 10% throughout (Tables 5 and 6). As with OsteoMeasure, the architectural variables measured by TAS failed to demonstrate any structural difference, despite their increased range and diverse character. Before statistical analysis, there seemed to be a consistent trend in the data, whereby the female fracture group (Table 5) tended to have a lower bone surface (BS), wider trabeculae (Tb.Wi) and shorter total trabecular length (Tb.Le), fewer trabeculae (strut Table 4. Comparison of histomorphometric data in men with and without vertebral fracture using OsteoMeasure (mean ⫾ SD)
Results
Age (years) BMD (g/cm2) Peripheral osteoporotic fracture
273
Men with vertebral fracture (n ⫽ 16)
Men without vertebral fracture (n ⫽ 11)
60.9 ⫾ 11.8 0.72 ⫾ 0.12 n⫽4
50.2 ⫾ 16.3 (n.s.) 0.76 ⫾ 0.17 (n.s.) n ⫽ 0 (n.s.)
BV/TV (%) OS/BS (%) Ob.S/BS (%) ES/BS (%) Oc.S/BS (%) Tb.Wi (m) O.Th (m) Tb.Sp (m) Tb.N (/mm) MAR (m/day) Calcification front/OS (%)
Men with vertebral fracture (n ⫽ 16)
Men without vertebral fracturea (n ⫽ 11)
10.86 ⫾ 4.40 9.41 ⫾ 12.70 1.35 ⫾ 2.67 6.91 ⫾ 9.23 0.72 ⫾ 1.56 121.04 ⫾ 33.70 7.53 ⫾ 4.21 904.7 ⫾ 316.7 1.09 ⫾ 0.33 0.66 ⫾ 0.24 63 ⫾ 33
10.94 ⫾ 3.1 8.52 ⫾ 10.09 0.31 ⫾ 0.47 4.13 ⫾ 3.58 0.54 ⫾ 0.51 117.98 ⫾ 14.53 7.33 ⫾ 3.66 864.1 ⫾ 277.4 1.12 ⫾ 0.30 0.69 ⫾ 0.30 63 ⫾ 35
KEY: BS, bone surface; BV, bone volume; ES, eroded surface; MAR, mineral apposition rate; Ob.S, osteoblast surface; Oc.S, osteoclast surface; OS, osteoid surface; O.Th, osteoid thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation; Tb.Wi, trabecular width; TV, tissue volume. a All values nonsignificant.
274
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
Bone Vol. 27, No. 2 August 2000:271–276
Table 5. Comparison of trabecular architecture between women with and without vertebral fracture using TAS (mean ⫾ SD)
BV/TV (%) Tb.Wi (m) Tb.Sp (m) BS (mm/mm2) N.Nd (/field) N.Tm (/field)b N.Nd/N.Tm Tb.Le (m) Tb.PF (/mm) Total strut number (/field) MTPD (Tb.N/mm) Fractal dimension MSV (mm3)
Women with vertebral fracture (n ⫽ 31)
Women without vertebral fracturea (n ⫽ 22)
9.93 ⫾ 4.10 107.02 ⫾ 24.27 932.00 ⫾ 365.40 1.83 ⫾ 0.62 2.86 ⫾ 2.28 20.83 ⫾ 11.98 0.20 ⫾ 0.27 9.24 ⫾ 3.81 3.15 ⫾ 2.22 14.69 ⫾ 7.52
10.24 ⫾ 3.74 104.55 ⫾ 20.64 881.90 ⫾ 383.00 1.99 ⫾ 0.81 3.45 ⫾ 3.05 19.41 ⫾ 13.70 0.23 ⫾ 0.19 10.34 ⫾ 5.17 2.96 ⫾ 3.24 16.92 ⫾ 10.43
1.11 ⫾ 0.37 1.45 ⫾ 0.10 23.6 ⫾ 15.5
1.20 ⫾ 0.49 1.45 ⫾ 0.10 21.3 ⫾ 12.8
KEY: MSV, marrow star volume; MTPD, mean trabecular plate density; N.Nd, node number; N.Nd/N.Tm, node:terminus ratio; N.Tm, terminus number; Tb.Le, trabecular length; Tb.PF, trabecular pattern factor. See Table 4 for other abbreviations. a All differences nonsignificant. b Standardized to arbitrary field size 10 mm2.
number), lower mean trabecular plate density (MTPD; now known as trabecular number/mm, or Tb.N/mm), fewer trabecular junctions (N.Nd), more termini (N.Tm), wider trabecular separation (Tb.Sp), and a large marrow star volume and trabecular pattern factor (Tb.PF), all collectively amounting to a combination that would be consistent with diminished connectivity (Figure 2). However, any difference suspected did not reach statistical significance. At the same time, there was no suggestion of any coherent trend toward trabecular disconnection in the men (Table 6), where the differences were not only nonsignificant but also at variance. For example, in the fracture group, the decline in the node:terminus ratio (used as an index of connectivity) and the higher trabecular pattern factor, if real, would have suggested Table 6. Comparison of trabecular architecture between men with and without vertebral fracture using TAS (mean ⫾ SD)
BV/TV (%) Tb.Wi (m) Tb.Sp (m) BS (mm/mm2) N.Nd (/field) N.Tm (/field)b N.Nd/N.Tm Tb.Le (m) Tb.PF (/mm) MTPD (Tb.N/mm) Total strut number (/field) Fractal dimension MSV (mm3)
Men with vertebral fracture (n ⫽ 16)
Men without vertebral fracturea (n ⫽ 11)
10.74 ⫾ 4.5 120.43 ⫾ 29.17 935.50 ⫾ 382.20 1.80 ⫾ 0.59 3.00 ⫾ 3.47 15.88 ⫾ 8.42 0.24 ⫾ 0.32 9.24 ⫾ 3.63 2.12 ⫾ 1.31 1.10 ⫾ 0.36 13.48 ⫾ 8.09
10.12 ⫾ 3.73 110.64 ⫾ 18.09 925.00 ⫾ 350.00 1.81 ⫾ 0.56 3.48 ⫾ 2.03 13.13 ⫾ 4.82 0.31 ⫾ 0.23 9.61 ⫾ 3.28 1.69 ⫾ 1.59 1.08 ⫾ 0.34 13.10 ⫾ 3.82
1.40 ⫾ 0.18 21.7 ⫾ 1.30
1.39 ⫾ 0.14 23.3 ⫾ 1.63
KEY: MSV, marrow star volume; MTPD, mean trabecular plate density; N.Nd, node number; N.Nd/N.Tm, node:terminus ratio; N.Tm; terminus number; Tb.Le, trabecular length; Tb.PF, trabecular pattern factor. See Table 4 for other abbreviations. a All differences nonsignificant. b Standardized to arbitrary field size 10 mm2.
Figure 2. Photomicrographs of typical undecalcified transilial sections, 8 m thick, from the osteopenic bone biopsy of (top) a vertebral fracture patient showing an apparently disconnected trabecular network and (bottom) a nonfracture patient subjectively suggestive of a better trabecular interconnection despite the similarity in the relative bone volume (BV). Toluidine blue stain; original magnification: ⫻8.
a biomechanically weaker arrangement due to disconnection and the loss of cross ties, whereas the decline in the marrow star volume (another index of connectivity alleged to be of more rigorous mathematical derivation than the above) and unchanged bone surface measurement, would not. As a single vertebral fracture may sometimes be due to trauma, and the diagnostic assessment of vertebral fracture can be difficult,13 the data were reanalyzed excluding subjects (six women and five men) with a solitary vertebral fracture. However, this made no difference in any of the findings, nor the combination of men and women when the departure from significance was increased. Discussion The present examination of cancellous microanatomy based on the two-dimensional imaging of undecalcified histological sections, 8 m thick, showed no significant difference at the iliac crest of groups of women and men of similar vertebral BMD with and without vertebral fracture. On the other hand, in a comparable microradiographic study, Flautre and Hardouin7 found a significant difference in the trabecular pattern factor (Tb.PF) between fracture subjects and controls in a direction that implied, paradoxically, a lower connectivity in the group without crush fracture. In contrast, Kleerekoper et al.10 found that vertebral fracture subjects had fewer but thicker trabecular plates, which were more widely separated
Bone Vol. 27, No. 2 August 2000:271–276
compared with normal controls. Unlike the groups in the present study, their vertebral fracture subjects were atypical in that they were deliberately selected to have a normal bone volume and were compared with normal controls, whereas the present patients were unselected other than by fracture and tended to have a low bone volume. Similarly Recker15 compared highly selected vertebral fracture patients (mean relative bone volume [BV/TV] 19.1% as compared with mean BV/TV 9.9% as measured by TAS herein) with normal controls and found a significant difference in the star volume (greater in fracture subjects) and in the number of trabecular nodes and free ends consistent with a loss of connectivity in fracture subjects, as Kleerekoper et al.10 had also indicated. However, our evidence suggests that it may not be possible to extrapolate results where the cancellous bone remains adequate in quantity to the average osteoporotic crush fracture subject of low bone volume. Croucher et al.5 compared men and women with osteoporosis (the majority of whom had vertebral fracture) with healthy controls. When their subjects were matched for cancellous area, no difference was seen in trabecular width, number, or separation, and of all the variables measured only the terminus count was significantly increased in the fracture group. Direct comparison of this with the evidence presented herein is difficult as men and women were analyzed together, not all osteoporotic subjects had a vertebral fracture, and subjects and controls were matched either for age or cancellous area but not both. There are several possible explanations for the lack of a significant difference between the vertebral fracture subjects and their equally osteopenic controls described here. For example, although measurements of iliac crest bone volume, compressive strength, ash density, and marrow star volume apparently correlate with measurements at the vertebrae,3,12,18 architectural changes (or lack of them) in the iliac crest biopsy may not adequately reflect events in the vertebral body. Local patterns of bone loss in the spine may be influential, as demonstrated by others in postmortem specimens, which showed marked variability in both bone volume and structure between adjacent vertebrae.3,4 Moreover, there is also the likelihood that, when the trabecular bone volume is as low as 10%, the contribution of cortical bone, not considered here, to fracture risk may be relatively more important.9,16 Consistent with this are recent in vitro studies of vertebral compressive strength, which suggest that both cortical thickness and trabecular star volume are determinants.20 The role of biomechanical factors such as anisotropy and other histological factors such as microfractures and their contribution to differences in fragility between fracture and nonfracture groups remains uncertain6 as are possible changes in collagen composition and mineral character. Two weaknesses in the present study are recognized that may have biased the outcome. The first concerns the sample size and the possibility that this was too low for the modest trends to reach significance. Although when all the variables were considered together there seemed to be a trend toward decreased connectivity in the female fracture group, differences between individual variables was small and power/sample size calculation suggested that more than 300 subjects would be needed to reach significance. The second concerns the normality of the control subjects who were not entirely healthy. On the other hand, a strength of the investigation relative to those of others is the matching of subjects for lumbar bone mineral density as well as ilial cancellous bone volume. Without this information the microarchitectural changes described in the literature may be related to a reduction in bone mineral density rather than a separate process contributing to skeletal fragility. Finally, although the present evidence is based on the
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
275
widest range of microanatomical variables that could practicably be accommodated, it is generally recognized that there are aspects of cancellous structure that are not best demonstrated in a two-dimensional section of what is a threedimensional structure. A simple histological method that enables the two-dimensional imaging to be viewed within its three-dimensional context has recently been developed (Shore et al.22). Its application to the present material is described in part II of this study and suggests that the dismissal of an independent architectural component to vertebral fracture risk on the basis of the 2D evidence alone may be premature. Acknowledgments: This project was funded by Action Research. The authors are grateful to Patricia A. Shore, M.Sc., for skilled technical assistance and W. Sidwell for typing the manuscript. Dr. D. R. Johnson, together with S. K. Paxton, developed TAS, and the software is available from S. K. Paxton (e-mail: [email protected]).
References 1. Aaron, J. E., Johnson, D. R., Kanis, J. A., Oakley, B. A., O’Higgins, P., and Paxton, S. K. An automated method for the analysis of trabecular bone structure. Comput Biomed Res 25:1–16; 1992. 2. Aaron, J. E., Makins, N. B., and Sagreiya, K. The microanatomy of trabecular bone loss in normal aging men and women. Clin Orthop 215:260 –172; 1987. 3. Amling, M., Grote, H. J., Posl, M., Hahn, M., and Delling, G. Polyostotic heterogeneity of the spine in osteoporosis. Quantitative analysis and 3-dimensional morphology. Bone Miner 27:193–208; 1994. 4. Amling, M., Posl, M., Ritzel, H., Hahn, M., Vogel, M., Wening, V. J., and Delling, G. Architecture and distribution of cancellous bone yield vertebral fracture clues. Arch Orthop Trauma Surg 115:262–269; 1996. 5. Croucher, P. I., Garrahan, N. J., and Compston, J. E. Structural mechanisms of trabecular bone loss in primary osteoporosis: Specific disease mechanism or early aging? Bone Miner 25:111–121; 1994. 6. Fazzalari, N. L. Trabecular microfracture. Calcif Tissue Int 53(Suppl. 1):S143– S147; 1993. 7. Flautre, B. and Hardouin, P. Microradiographic aspect on iliac bone tissue in postmenopausal women with and without vertebral crush fracture. Bone 15: 477– 481; 1994. 8. Hui, S. L., Slemenda, C. W., and Johnston, C. C., Jr. Baseline measurement of bone mass predicts fracture in white women. Ann Intern Med 111:355–361; 1989. 9. Ito, M., Hayashi, K., Kawahara, Y., Uetani, M., and Imaizumi, Y. The relationship of trabecular and cortical bone mineral density to spinal fractures. Invest Radiol 28:573–580; 1993. 10. Kleerekoper, M., Villaneuva, A. R., Stancui, J., Rao, D. S., and Parfitt, A. M. The role of 3-dimensional trabecular microstructure in the pathogenesis of vertebral compression fractures. Calcif Tissue Int 37:594 –597; 1985. 11. Melton, L. J., Kan, S. H., Frye, M. A., Wahner, H. W., O’Fallon, W. M., and Riggs, B. L. Epidemiology of vertebral fractures in women. Am J Epidemiol 129:1000 –1011; 1989. 12. Mosekilde, Li. and Mosekilde, Le. Iliac crest trabecular bone volume as a predictor for vertebral compressive strength, ash density and trabecular bone volume in normal individuals. Bone 9:195–199; 1988. 13. National Osteoporosis Foundation Working Groups on Vertebral Fractures. Assessing vertebral fractures. J Bone Miner Res 10:518 –523; 1995. 14. Pugh, J. W., Rose, R. M., and Radin, E. L. Elastic and viscoelastic properties of trabecular bone, dependence on structure. Biomechanics 6:475– 485; 1973. 15. Recker, R. R. Architecture and vertebral fracture. Calcif Tissue Int 53(Suppl. 1):S139 –S142; 1993. 16. Ritzel, H., Amling, M., Posl, M., Hahn, M., and Delling, G. The thickness of human vertebral cortical bone and its changes in ageing and osteoporosis. A histomorphometric analysis of the complete spinal column from 37 autopsy specimens. J Bone Miner Res 12:89 –95; 1997. 17. Ross, P. D., Davis, J. W., Epstein, R. S., and Wasnich, R. D. Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 115:919 –923; 1991. 18. Vesterby, A. Star volume of marrow space and trabeculae in iliac crest, sampling procedure and correlation to star volume of first lumbar vertebra. Bone 11:149 –155; 1990.
276
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
19. Vesterby, A., Gundersen, H. J. G., and Melsen, F. Star volume of marrow space and trabeculae of the first lumbar vertebrae. Sampling efficiency and biological variation. Bone 10:7–13; 1989. 20. Vesterby, A., Mosekilde, L., Gunderson, J. G., Melson, F., Mosekilde, L., Holme, K., and Sorensen, S. Biologically meaningful determinants of the in vitro strength of lumbar vertebrae. Bone 12:219 –224; 1991. 21. Weinstein, R. S. and Majumdar, S. Fractal geometry and vertebral compression fractures. J Bone Miner Res 9:1797–1802; 1994. 22. Shore, P. A., Shore, R. C., and Aaron, J. E. A 3-D histological method to
Bone Vol. 27, No. 2 August 2000:271–276 directly determine the number of trabecular termini in cancellous bone. Biotech Histochem. In press.
Date Received: August 13, 1999 Date Revised: March 22, 2000 Date Accepted: March 22, 2000
Trabecular Architecture in Women and Men of Similar Bone Mass With and Without Vertebral Fracture: I. Two-Dimensional Histology L. D. HORDON,1 M. RAISI,1 J. E. AARON,1 S. K. PAXTON,1 M. BENETON,2 and J. A. KANIS2 1
School of Biomedical Sciences, Worsley Medical and Dental Building, University of Leeds, Leeds, UK Centre for Metabolic Bone Diseases at Sheffield (WHO Collaborating Centre), Medical School, University of Sheffield, Sheffield, UK
2
Key Words: Osteoporosis; Fracture; Bone mineral density; Bone histomorphometry; Microarchitecture; Two-dimensional (2D) image analysis.
While osteoporosis is characterized by a low bone mass there is a well-recognized overlap in bone mineral density (BMD) measurements between groups of subjects with and without vertebral fracture. To investigate whether differences in trabecular architecture may contribute to the presence or absence of fractures independent of the bone mass, fracture and nonfracture groups matched for age, gender, and BMD were assembled. Transiliac biopsies and corresponding lumbar spine BMD measurements from 31 women and 16 men with vertebral fracture were compared with those from 22 women and 11 men without fracture. Lumbar BMD (L1– 4) was measured using a Hologic 2000 densitometer. The lumbar BMD was similar in women with and without fracture (0.63 g/cm3 ⴞ 0.10 SD and 0.71 g/cm3 ⴞ 0.17 SD, n.s.) and in men with and without fracture (0.72 g/cm3 ⴞ 0.12 SD and 0.76 g/cm3 ⴞ 0.17 SD, n.s.). Undecalcified iliac crest biopsy sections, 8 m thick, were analyzed for remodeling variables and trabecular architecture using OsteoMeasure and TAS image analysis systems. No significant difference was found in either gender between fracture and nonfracture groups in percent bone volume (mean 10% in all groups), or in the wide range of remodeling and architectural variables measured, including the trabecular width, number, and separation, mean trabecular plate density and fractal dimension, as well as several indirect indices of connectivity including the node: terminus ratio, marrow star volume, and trabecular pattern factor. On the basis of this evidence it was concluded that there is no difference in the trabecular architecture between patients with crush fracture and controls when account is taken of bone mass. This suggests that microanatomical disruption is a predictable intrinsic feature of bone loss. However, there remains the possibility that the two-dimensional character of the structural deterioration measured indirectly is not sufficiently sensitive for the complex cancellous system. This is considered further in part II. (Bone 27: 271–276; 2000) © 2000 by Elsevier Science Inc. All rights reserved.
Introduction There is extensive evidence in the literature to suggest that the mechanical strength of bone is not solely a function of its mass or bone mineral density (BMD). While the BMD at the spine and the hip is directly related to the risk of fracture at those sites,8 there is considerable overlap in the BMD measured in groups of subjects with and without fracture, despite the higher mean value in the latter.11 Both the initial BMD and the presence of existing vertebral fractures together apparently combine to predict more reliably the occurrence of future vertebral fractures.17 The complex architecture of cancellous bone also determines the mechanical strength,14 and differences in microstructural arrangement between aging men and women may contribute to the lower rates of vertebral and hip fractures in men.2 It follows that differences in cancellous architecture may explain why some women with a low vertebral BMD have vertebral fractures, whereas others with an identical BMD do not. However, examination of this aspect in matched groups, with and without vertebral fractures using either histomorphometric10,15 or microradiographic techniques7 have produced differing results. In endeavoring to resolve this, the aim of the present study was to compare cancellous architecture at the iliac crest in groups of subjects with and without vertebral fracture who were matched for age, gender, and lumbar BMD, to determine whether the difference in cancellous architecture might explain the difference in fracture status. Materials and Methods Patients who had undergone both iliac crest biopsies and lumbar BMD measurements were identified from databases at the Royal Hallamshire Hospital, Sheffield. The records of these patients were reviewed and any who had received drugs for the treatment of osteoporosis (e.g., fluoride, bisphosphonates, and anabolic steroids) prior to biopsy or BMD measurements were excluded, as were those who were receiving or had received oral corticosteroids, or who had primary hyperparathyroidism or renal impairment (defined when serum creatinine exceeded the normal range). The patients were classified into groups with and without vertebral fracture. A total of 89 patients were eligible for study. Seven were excluded because of an inadequate biopsy, and two were omitted as they could not lie flat for densitometry. The final
Address for correspondence and reprints: Dr. J. E. Aaron, School of Biomedical Sciences, Worsley Medical and Dental Building, University of Leeds, Leeds LS2 9JT, UK. E-mail: [email protected] © 2000 by Elsevier Science Inc. All rights reserved.
271
8756-3282/00/$20.00 PII S8756-3282(00)00329-X
272
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
Bone Vol. 27, No. 2 August 2000:271–276
Figure 1. Automated trabecular analysis system (TAS) showing the binary image and its thinned (skeletonized) counterpart together with the area of interest defined by the adjustable rectangle, as well as typical printouts. Examples from the nonfracture (upper) and fracture (lower) groups of women are shown.
female fracture group comprised 31 women, 25 of whom had more than one vertebral fracture (median four), and six of whom had a single fracture (defined as ⬎20% loss of anterior, middle, or posterior vertebral height). The male fracture group comprised 11 men with more than one vertebral fracture (median four), and 5 men with a single fracture. Lumbar spine fractures were present in 12 women and 9 men. The patients without a vertebral fracture consisted of 11 men and 22 women. They were not entirely healthy skeletally as they had come to medical attention for reasons including peripheral fracture, radiological osteopenia, back pain, family history of osteoporosis, or concern about the possibility of having osteoporosis. Transiliac bone biopsies were taken after double tetracy-
cline labeling, using local anesthetic and an 8-mm-diameter Bordier trephine. Bone mineral density was assessed at L1– 4 using a Hologic 2000 densitometer within 3 months of the bone biopsy and generally during the same admission. Severely fractured vertebrae were excluded from the measurement area. The biopsies were preserved in 70% ethanol, dehydrated, and embedded using established procedures.2 Sections, 8 m thick, were cut on a Jung K heavy-duty microtome, stained by the Goldner tetrachrome method, and structural and remodeling (i.e., formation and resorption) variables analyzed using the semiautomated OsteoMeasure system (OsteoMetrics, Inc., Atlanta, GA) and digitizing pad. Unstained sections were examined under an epifluorescence microscope and the mineralization rate determined by mea-
Bone Vol. 27, No. 2 August 2000:271–276
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
Table 1. Age, lumbar bone mineral density, and other characteristics of women with and without vertebral fracture (mean ⫾ SD)
Age (years) BMD (g/cm2) Age at menopause (years) Peripherala osteoporotic fracture Present or past use of hormone replacement therapy
Women with vertebral fracture (n ⫽ 31)
Women without vertebral fracture (n ⫽ 22)
66.0 ⫾ 9.4 0.63 ⫾ 0.10 46.4 ⫾ 6.5 (n ⫽ 28) n ⫽ 14
63.5 ⫾ 8.7 (n.s.) 0.71 ⫾ 0.17 (n.s.) 48.2 ⫾ 5.1 (n ⫽ 17) (n.s.) n ⫽ 4 (n.s.)
n⫽3
n ⫽ 4 (n.s.)
KEY: BMD, bone mineral density; n.s., nonsignificant. a Defined as previous Colles, femoral neck, proximal humerus, or pubic ramus fractures.
suring the distance between the fluorescent bands that were temporally 10 days apart. Additionally, four to six sections stained either by the Goldner method or in 0.1% toluidine blue, pH 3.5, for 30 min were analyzed by means of our automated trabecular analysis system (TAS), which measures a comprehensive range of structural variables1 using the intact binary image and its thinned counterpart (Figure 1). There is also the option to include both the marrow star volume19 (using either the random generation of a selected number of points or a grid of points from which to automatically measure the mean radial path length scanned through the cavity to the surrounding trabecular surface) and the fractal dimension21 (a mathematical descriptor of structural complexity). The area of interest was defined by an adjustable window and was selected to encompass as much intact cancellous bone in the section as possible and to exclude the cortical tissue. An editing facility enabled minor artifacts, such as sectioning cracks, to be removed before analysis (see Aaron et al.1 for details). Once the groups had been assembled, histomorphometry was performed blind with respect to fracture status. Statistical analysis was carried out using Student’s t-test and the chi-square test with Yates correction.
The characteristics of the patients together with their lumbar BMD are shown in Tables 1 and 2. No significant difference was seen between the fracture and nonfracture groups of either gender in terms of age, lumbar BMD, or peripheral fracture, and in the women there was no difference in the age at menopause or use of hormone replacement therapy. When the OsteoMeasure analysis system was applied to the bone biopsies, the similarity Table 2. Age, lumbar bone mineral density (BMD), and peripheral osteoporotic fracture in men with and without vertebral fracture (mean ⫾ SD)
KEY: n.s., nonsignificant.
Table 3. Comparison of histomorphometric data in women with and without vertebral fracture using OsteoMeasure (mean ⫾ SD)
BV/TV (%) OS/BS (%) ObS/BS (%) ES/BS (%) OC.S/BS (%) Tb.Wi (m) O.Th (m) Tb.Sp (m) Tb.N (/mm) MAR (m/day) Calcification front/OS (%)
Women with vertebral fracture (n ⫽ 31)
Women without vertebral fracturea (n ⫽ 22)
11.76 ⫾ 5.79 9.17 ⫾ 6.49 0.28 ⫾ 0.50 4.01 ⫾ 3.62 0.40 ⫾ 0.66 110.70 ⫾ 30.07 7.09 ⫾ 2.70 800.3 ⫾ 314.1 1.26 ⫾ 0.49 0.51 ⫾ 0.27 76 ⫾ 29
11.21 ⫾ 3.83 9.37 ⫾ 8.81 0.50 ⫾ 1.05 3.50 ⫾ 3.39 0.44 ⫾ 0.66 104.10 ⫾ 16.56 8.04 ⫾ 2.87 766.2 ⫾ 274.8 1.30 ⫾ 0.43 0.57 ⫾ 0.23 62 ⫾ 34
See Results for abbreviations. a All differences nonsignificant.
in bone status of the fracture and nonfracture groups was confirmed by their identical mean relative bone volume measurement of 11% (Tables 3 and 4) . No significant difference was found between the remodeling variables. The fracture group tended to have fewer, more separated and thicker trabeculae as indicated by OsteoMeasure, but again the difference did not reach significance. Among the men, remodeling tended to be marginally higher in the fracture group and the trabeculae marginally fewer, more separated, and thicker, but not significantly so. The application of TAS to the histological sections again confirmed the well-matched cancellous bone status of the groups, the relative bone volume now measuring around 10% throughout (Tables 5 and 6). As with OsteoMeasure, the architectural variables measured by TAS failed to demonstrate any structural difference, despite their increased range and diverse character. Before statistical analysis, there seemed to be a consistent trend in the data, whereby the female fracture group (Table 5) tended to have a lower bone surface (BS), wider trabeculae (Tb.Wi) and shorter total trabecular length (Tb.Le), fewer trabeculae (strut Table 4. Comparison of histomorphometric data in men with and without vertebral fracture using OsteoMeasure (mean ⫾ SD)
Results
Age (years) BMD (g/cm2) Peripheral osteoporotic fracture
273
Men with vertebral fracture (n ⫽ 16)
Men without vertebral fracture (n ⫽ 11)
60.9 ⫾ 11.8 0.72 ⫾ 0.12 n⫽4
50.2 ⫾ 16.3 (n.s.) 0.76 ⫾ 0.17 (n.s.) n ⫽ 0 (n.s.)
BV/TV (%) OS/BS (%) Ob.S/BS (%) ES/BS (%) Oc.S/BS (%) Tb.Wi (m) O.Th (m) Tb.Sp (m) Tb.N (/mm) MAR (m/day) Calcification front/OS (%)
Men with vertebral fracture (n ⫽ 16)
Men without vertebral fracturea (n ⫽ 11)
10.86 ⫾ 4.40 9.41 ⫾ 12.70 1.35 ⫾ 2.67 6.91 ⫾ 9.23 0.72 ⫾ 1.56 121.04 ⫾ 33.70 7.53 ⫾ 4.21 904.7 ⫾ 316.7 1.09 ⫾ 0.33 0.66 ⫾ 0.24 63 ⫾ 33
10.94 ⫾ 3.1 8.52 ⫾ 10.09 0.31 ⫾ 0.47 4.13 ⫾ 3.58 0.54 ⫾ 0.51 117.98 ⫾ 14.53 7.33 ⫾ 3.66 864.1 ⫾ 277.4 1.12 ⫾ 0.30 0.69 ⫾ 0.30 63 ⫾ 35
KEY: BS, bone surface; BV, bone volume; ES, eroded surface; MAR, mineral apposition rate; Ob.S, osteoblast surface; Oc.S, osteoclast surface; OS, osteoid surface; O.Th, osteoid thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation; Tb.Wi, trabecular width; TV, tissue volume. a All values nonsignificant.
274
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
Bone Vol. 27, No. 2 August 2000:271–276
Table 5. Comparison of trabecular architecture between women with and without vertebral fracture using TAS (mean ⫾ SD)
BV/TV (%) Tb.Wi (m) Tb.Sp (m) BS (mm/mm2) N.Nd (/field) N.Tm (/field)b N.Nd/N.Tm Tb.Le (m) Tb.PF (/mm) Total strut number (/field) MTPD (Tb.N/mm) Fractal dimension MSV (mm3)
Women with vertebral fracture (n ⫽ 31)
Women without vertebral fracturea (n ⫽ 22)
9.93 ⫾ 4.10 107.02 ⫾ 24.27 932.00 ⫾ 365.40 1.83 ⫾ 0.62 2.86 ⫾ 2.28 20.83 ⫾ 11.98 0.20 ⫾ 0.27 9.24 ⫾ 3.81 3.15 ⫾ 2.22 14.69 ⫾ 7.52
10.24 ⫾ 3.74 104.55 ⫾ 20.64 881.90 ⫾ 383.00 1.99 ⫾ 0.81 3.45 ⫾ 3.05 19.41 ⫾ 13.70 0.23 ⫾ 0.19 10.34 ⫾ 5.17 2.96 ⫾ 3.24 16.92 ⫾ 10.43
1.11 ⫾ 0.37 1.45 ⫾ 0.10 23.6 ⫾ 15.5
1.20 ⫾ 0.49 1.45 ⫾ 0.10 21.3 ⫾ 12.8
KEY: MSV, marrow star volume; MTPD, mean trabecular plate density; N.Nd, node number; N.Nd/N.Tm, node:terminus ratio; N.Tm, terminus number; Tb.Le, trabecular length; Tb.PF, trabecular pattern factor. See Table 4 for other abbreviations. a All differences nonsignificant. b Standardized to arbitrary field size 10 mm2.
number), lower mean trabecular plate density (MTPD; now known as trabecular number/mm, or Tb.N/mm), fewer trabecular junctions (N.Nd), more termini (N.Tm), wider trabecular separation (Tb.Sp), and a large marrow star volume and trabecular pattern factor (Tb.PF), all collectively amounting to a combination that would be consistent with diminished connectivity (Figure 2). However, any difference suspected did not reach statistical significance. At the same time, there was no suggestion of any coherent trend toward trabecular disconnection in the men (Table 6), where the differences were not only nonsignificant but also at variance. For example, in the fracture group, the decline in the node:terminus ratio (used as an index of connectivity) and the higher trabecular pattern factor, if real, would have suggested Table 6. Comparison of trabecular architecture between men with and without vertebral fracture using TAS (mean ⫾ SD)
BV/TV (%) Tb.Wi (m) Tb.Sp (m) BS (mm/mm2) N.Nd (/field) N.Tm (/field)b N.Nd/N.Tm Tb.Le (m) Tb.PF (/mm) MTPD (Tb.N/mm) Total strut number (/field) Fractal dimension MSV (mm3)
Men with vertebral fracture (n ⫽ 16)
Men without vertebral fracturea (n ⫽ 11)
10.74 ⫾ 4.5 120.43 ⫾ 29.17 935.50 ⫾ 382.20 1.80 ⫾ 0.59 3.00 ⫾ 3.47 15.88 ⫾ 8.42 0.24 ⫾ 0.32 9.24 ⫾ 3.63 2.12 ⫾ 1.31 1.10 ⫾ 0.36 13.48 ⫾ 8.09
10.12 ⫾ 3.73 110.64 ⫾ 18.09 925.00 ⫾ 350.00 1.81 ⫾ 0.56 3.48 ⫾ 2.03 13.13 ⫾ 4.82 0.31 ⫾ 0.23 9.61 ⫾ 3.28 1.69 ⫾ 1.59 1.08 ⫾ 0.34 13.10 ⫾ 3.82
1.40 ⫾ 0.18 21.7 ⫾ 1.30
1.39 ⫾ 0.14 23.3 ⫾ 1.63
KEY: MSV, marrow star volume; MTPD, mean trabecular plate density; N.Nd, node number; N.Nd/N.Tm, node:terminus ratio; N.Tm; terminus number; Tb.Le, trabecular length; Tb.PF, trabecular pattern factor. See Table 4 for other abbreviations. a All differences nonsignificant. b Standardized to arbitrary field size 10 mm2.
Figure 2. Photomicrographs of typical undecalcified transilial sections, 8 m thick, from the osteopenic bone biopsy of (top) a vertebral fracture patient showing an apparently disconnected trabecular network and (bottom) a nonfracture patient subjectively suggestive of a better trabecular interconnection despite the similarity in the relative bone volume (BV). Toluidine blue stain; original magnification: ⫻8.
a biomechanically weaker arrangement due to disconnection and the loss of cross ties, whereas the decline in the marrow star volume (another index of connectivity alleged to be of more rigorous mathematical derivation than the above) and unchanged bone surface measurement, would not. As a single vertebral fracture may sometimes be due to trauma, and the diagnostic assessment of vertebral fracture can be difficult,13 the data were reanalyzed excluding subjects (six women and five men) with a solitary vertebral fracture. However, this made no difference in any of the findings, nor the combination of men and women when the departure from significance was increased. Discussion The present examination of cancellous microanatomy based on the two-dimensional imaging of undecalcified histological sections, 8 m thick, showed no significant difference at the iliac crest of groups of women and men of similar vertebral BMD with and without vertebral fracture. On the other hand, in a comparable microradiographic study, Flautre and Hardouin7 found a significant difference in the trabecular pattern factor (Tb.PF) between fracture subjects and controls in a direction that implied, paradoxically, a lower connectivity in the group without crush fracture. In contrast, Kleerekoper et al.10 found that vertebral fracture subjects had fewer but thicker trabecular plates, which were more widely separated
Bone Vol. 27, No. 2 August 2000:271–276
compared with normal controls. Unlike the groups in the present study, their vertebral fracture subjects were atypical in that they were deliberately selected to have a normal bone volume and were compared with normal controls, whereas the present patients were unselected other than by fracture and tended to have a low bone volume. Similarly Recker15 compared highly selected vertebral fracture patients (mean relative bone volume [BV/TV] 19.1% as compared with mean BV/TV 9.9% as measured by TAS herein) with normal controls and found a significant difference in the star volume (greater in fracture subjects) and in the number of trabecular nodes and free ends consistent with a loss of connectivity in fracture subjects, as Kleerekoper et al.10 had also indicated. However, our evidence suggests that it may not be possible to extrapolate results where the cancellous bone remains adequate in quantity to the average osteoporotic crush fracture subject of low bone volume. Croucher et al.5 compared men and women with osteoporosis (the majority of whom had vertebral fracture) with healthy controls. When their subjects were matched for cancellous area, no difference was seen in trabecular width, number, or separation, and of all the variables measured only the terminus count was significantly increased in the fracture group. Direct comparison of this with the evidence presented herein is difficult as men and women were analyzed together, not all osteoporotic subjects had a vertebral fracture, and subjects and controls were matched either for age or cancellous area but not both. There are several possible explanations for the lack of a significant difference between the vertebral fracture subjects and their equally osteopenic controls described here. For example, although measurements of iliac crest bone volume, compressive strength, ash density, and marrow star volume apparently correlate with measurements at the vertebrae,3,12,18 architectural changes (or lack of them) in the iliac crest biopsy may not adequately reflect events in the vertebral body. Local patterns of bone loss in the spine may be influential, as demonstrated by others in postmortem specimens, which showed marked variability in both bone volume and structure between adjacent vertebrae.3,4 Moreover, there is also the likelihood that, when the trabecular bone volume is as low as 10%, the contribution of cortical bone, not considered here, to fracture risk may be relatively more important.9,16 Consistent with this are recent in vitro studies of vertebral compressive strength, which suggest that both cortical thickness and trabecular star volume are determinants.20 The role of biomechanical factors such as anisotropy and other histological factors such as microfractures and their contribution to differences in fragility between fracture and nonfracture groups remains uncertain6 as are possible changes in collagen composition and mineral character. Two weaknesses in the present study are recognized that may have biased the outcome. The first concerns the sample size and the possibility that this was too low for the modest trends to reach significance. Although when all the variables were considered together there seemed to be a trend toward decreased connectivity in the female fracture group, differences between individual variables was small and power/sample size calculation suggested that more than 300 subjects would be needed to reach significance. The second concerns the normality of the control subjects who were not entirely healthy. On the other hand, a strength of the investigation relative to those of others is the matching of subjects for lumbar bone mineral density as well as ilial cancellous bone volume. Without this information the microarchitectural changes described in the literature may be related to a reduction in bone mineral density rather than a separate process contributing to skeletal fragility. Finally, although the present evidence is based on the
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
275
widest range of microanatomical variables that could practicably be accommodated, it is generally recognized that there are aspects of cancellous structure that are not best demonstrated in a two-dimensional section of what is a threedimensional structure. A simple histological method that enables the two-dimensional imaging to be viewed within its three-dimensional context has recently been developed (Shore et al.22). Its application to the present material is described in part II of this study and suggests that the dismissal of an independent architectural component to vertebral fracture risk on the basis of the 2D evidence alone may be premature. Acknowledgments: This project was funded by Action Research. The authors are grateful to Patricia A. Shore, M.Sc., for skilled technical assistance and W. Sidwell for typing the manuscript. Dr. D. R. Johnson, together with S. K. Paxton, developed TAS, and the software is available from S. K. Paxton (e-mail: [email protected]).
References 1. Aaron, J. E., Johnson, D. R., Kanis, J. A., Oakley, B. A., O’Higgins, P., and Paxton, S. K. An automated method for the analysis of trabecular bone structure. Comput Biomed Res 25:1–16; 1992. 2. Aaron, J. E., Makins, N. B., and Sagreiya, K. The microanatomy of trabecular bone loss in normal aging men and women. Clin Orthop 215:260 –172; 1987. 3. Amling, M., Grote, H. J., Posl, M., Hahn, M., and Delling, G. Polyostotic heterogeneity of the spine in osteoporosis. Quantitative analysis and 3-dimensional morphology. Bone Miner 27:193–208; 1994. 4. Amling, M., Posl, M., Ritzel, H., Hahn, M., Vogel, M., Wening, V. J., and Delling, G. Architecture and distribution of cancellous bone yield vertebral fracture clues. Arch Orthop Trauma Surg 115:262–269; 1996. 5. Croucher, P. I., Garrahan, N. J., and Compston, J. E. Structural mechanisms of trabecular bone loss in primary osteoporosis: Specific disease mechanism or early aging? Bone Miner 25:111–121; 1994. 6. Fazzalari, N. L. Trabecular microfracture. Calcif Tissue Int 53(Suppl. 1):S143– S147; 1993. 7. Flautre, B. and Hardouin, P. Microradiographic aspect on iliac bone tissue in postmenopausal women with and without vertebral crush fracture. Bone 15: 477– 481; 1994. 8. Hui, S. L., Slemenda, C. W., and Johnston, C. C., Jr. Baseline measurement of bone mass predicts fracture in white women. Ann Intern Med 111:355–361; 1989. 9. Ito, M., Hayashi, K., Kawahara, Y., Uetani, M., and Imaizumi, Y. The relationship of trabecular and cortical bone mineral density to spinal fractures. Invest Radiol 28:573–580; 1993. 10. Kleerekoper, M., Villaneuva, A. R., Stancui, J., Rao, D. S., and Parfitt, A. M. The role of 3-dimensional trabecular microstructure in the pathogenesis of vertebral compression fractures. Calcif Tissue Int 37:594 –597; 1985. 11. Melton, L. J., Kan, S. H., Frye, M. A., Wahner, H. W., O’Fallon, W. M., and Riggs, B. L. Epidemiology of vertebral fractures in women. Am J Epidemiol 129:1000 –1011; 1989. 12. Mosekilde, Li. and Mosekilde, Le. Iliac crest trabecular bone volume as a predictor for vertebral compressive strength, ash density and trabecular bone volume in normal individuals. Bone 9:195–199; 1988. 13. National Osteoporosis Foundation Working Groups on Vertebral Fractures. Assessing vertebral fractures. J Bone Miner Res 10:518 –523; 1995. 14. Pugh, J. W., Rose, R. M., and Radin, E. L. Elastic and viscoelastic properties of trabecular bone, dependence on structure. Biomechanics 6:475– 485; 1973. 15. Recker, R. R. Architecture and vertebral fracture. Calcif Tissue Int 53(Suppl. 1):S139 –S142; 1993. 16. Ritzel, H., Amling, M., Posl, M., Hahn, M., and Delling, G. The thickness of human vertebral cortical bone and its changes in ageing and osteoporosis. A histomorphometric analysis of the complete spinal column from 37 autopsy specimens. J Bone Miner Res 12:89 –95; 1997. 17. Ross, P. D., Davis, J. W., Epstein, R. S., and Wasnich, R. D. Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 115:919 –923; 1991. 18. Vesterby, A. Star volume of marrow space and trabeculae in iliac crest, sampling procedure and correlation to star volume of first lumbar vertebra. Bone 11:149 –155; 1990.
276
L. D. Hordon et al. Two-dimensional histology and vertebral fracture
19. Vesterby, A., Gundersen, H. J. G., and Melsen, F. Star volume of marrow space and trabeculae of the first lumbar vertebrae. Sampling efficiency and biological variation. Bone 10:7–13; 1989. 20. Vesterby, A., Mosekilde, L., Gunderson, J. G., Melson, F., Mosekilde, L., Holme, K., and Sorensen, S. Biologically meaningful determinants of the in vitro strength of lumbar vertebrae. Bone 12:219 –224; 1991. 21. Weinstein, R. S. and Majumdar, S. Fractal geometry and vertebral compression fractures. J Bone Miner Res 9:1797–1802; 1994. 22. Shore, P. A., Shore, R. C., and Aaron, J. E. A 3-D histological method to
Bone Vol. 27, No. 2 August 2000:271–276 directly determine the number of trabecular termini in cancellous bone. Biotech Histochem. In press.
Date Received: August 13, 1999 Date Revised: March 22, 2000 Date Accepted: March 22, 2000