OSTEOPOROSIS, A MULTIFACTORIAL clinical state

0021-972X/04/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 89(6):2776 2782 Printed in U.S.A. Copyright 2004 by The Endocrine Society doi: 10.1210/jc.2003-030551 Volumetric Bone Density at the Femoral Neck as a Common Measure of Hip Fracture Risk for Men and Women JACQUELINE R. CENTER, TUAN V. NGUYEN, NICK A. POCOCK, AND JOHN A. EISMAN Bone and Mineral Research Program (J.R.C., T.V.N., J.A.E.), Garvan Institute of Medical Research, St. Vincent s Hospital and University of New South Wales; and Department of Nuclear Medicine (N.A.P.), St Vincent s Hospital, Sydney, New South Wales 2010, Australia Measurements of bone density using dual-energy x-ray absorptiometry are generally based on the areal projection, which incompletely accounts for size. The larger areal bone density in older men compared with older women is primarily due to their larger bone size, conferring a biomechanical advantage that may be a major factor contributing to lower hip fracture rates. The aim of this study was to evaluate estimated volumetric bone density at the hip in men and women with and without fractures to better determine the role of estimated volumetric density vs. size in hip fracture risk. This prospective population-based study compared 852 women and 635 men without fractures with 73 women and 23 men with hip fractures. As expected, areal bone mineral density (BMD) and cross-sectional area were lower in women than men, and areal bone density was lower in those with hip fractures compared with nonfracture subjects. However, estimated volumetric BMD was the only parameter, apart from OSTEOPOROSIS, A MULTIFACTORIAL clinical state predisposing to fracture, is dependent not only on bone density, as assessed by dual-energy x-ray absorptiometry (DXA), but also on measures such as bone size, geometry, and quality (1). Fracture, the ultimate consequence of osteoporosis, is more recently being evaluated not only in relation to bone density and structure but also in conjunction with a number of nonskeletal risks such as falls, age, sex, and genetic predisposition. Most studies have examined fracture risk in women rather than men due to their 2- to 3-fold higher fracture incidence. This has been thought to be due to the lower areal bone density in women than men (2 4). However, there are a number of factors interrelated with areal bone density that are known to influence fracture risk. Firstly, areal bone density is a two-dimensional representation of a three-dimensional quantity, with the third dimension (depth) being ignored within the measure. Hence, for the same volumetric density, a larger bone will have a greater areal bone density. The generally larger areal bone density in older men compared with older women is primarily due to their larger bone Abbreviations: BMC, Bone mineral content; BMD, bone mineral density; CSA, cross-sectional area; DXA, dual-energy x-ray absorptiometry; FNAL, femoral neck axis length; HAL, hip axis length. JCEM is published monthly by The Endocrine Society (http://www. endo-society.org), the foremost professional society serving the endocrine community. 2776 age, that was the same in women and men both without hip fractures (0.31 0.06 and 0.31 0.06 g/cm 3, respectively) and with hip fractures (0.25 0.04 and 0.26 0.04 g/cm 3, respectively). Using the World Health Organization 2.5 SD cut-off for osteoporosis for hip fracture prediction, estimated volumetric BMD was more sensitive than areal BMD in men (70 vs. 43%; P 0.04) and similar to that in women, in whom sensitivity was similar for both areal (73%) and estimated volumetric (78%) BMD cut-offs. Thus, men and women have hip fractures at the same estimated femoral neck volumetric BMD, which is largely independent of the size artifact inherent in areal BMD. This aspect of estimated femoral neck volumetric BMD suggests that it can provide a single measure that could be used in men and women. It needs further exploration for a role in assessment of hip fracture risk across the sexes and particularly in men. (J Clin Endocrinol Metab 89: 2776 2782, 2004) size (1, 5 7). In contrast, estimates of volumetric density are similar in men and women (8 11). The larger bone size in men, translating into a larger cross-sectional area (CSA), confers a biomechanical advantage and results in a greater breaking strength (12 15). Thus, it could be hypothesized that larger size rather than any differences in bone density per se would be a major protective factor in men compared with women. Focusing specifically on hip fractures, geometry of the hip is another risk factor. Each sd increase in hip axis length (HAL) is associated with an almost 2-fold increase in hip fracture risk in women (16, 17), although the data are less clear for measurements excluding the pelvic brim, i.e. femoral neck axis length (FNAL), and for men (18 20). HAL is correlated with height, another possible risk factor for hip fracture (21, 22). As might be expected, men are taller than women and have substantially longer HAL. It could be argued that the increased risk of a longer HAL in men is counterbalanced by the protective nature of their greater bone size with respect to the risk of hip fracture. There is substantial epidemiological and clinical data supporting the use of a bone density of more than 2.5 sd below the young normal range (T-score) in women to identify a group at relatively high risk of fracture and in whom preventive treatment is warranted (23). With relatively little evidence, a similar T-score-based stratification has been suggested for men (24). The aim of this study was to evaluate estimated volumetric

Center et al. Volumetric Bone Density and Hip Fracture Risk J Clin Endocrinol Metab, June 2004, 89(6):2776 2782 2777 as well as areal bone density at the femoral neck in men and women with and without hip fractures to help examine the importance of volumetric density, as estimated by DXA, vs. size and geometry as factors in hip fracture risk. Subjects Subjects and Methods Subjects were all participants of the Dubbo Osteoporosis Epidemiology Study, a community-based longitudinal study of osteoporosis and fractures in Dubbo, a semiurban city 400 km northwest of Sydney, Australia. The study commenced in 1989, as previously described (25 27), and is ongoing. Briefly, the population is approximately 32,000 and, although the study is open to the whole population, all the participants have been of Caucasian background (98.6% of the population). The community has centralized health services and is relatively stable, making it suitable for long-term, prospective epidemiological studies. All subjects aged 60 yr and over as of January 1, 1989, obtained from the electoral roll and local approaches (25), were approached by letter to participate in the study. Informed consent, using a signed consent form, was obtained from every subject before participation in the study. The present analysis compared men and women with hip fractures with all subjects who had not sustained a fracture by February 2002. Only those hip fracture subjects who had had a bone mineral density (BMD) scan before or within 3 months of sustaining a fracture were included in this study. The 73 women and 23 men with hip fractures who met the above criteria were compared with 853 women and 635 men without fractures. Clinical data collection Subjects were interviewed by a nurse coordinator at initial and subsequent visits, at approximately two-yearly intervals. A structured questionnaire was used to collect data including age and lifestyle factors. Anthropometric variables including weight and current height were measured at each visit. Hip fractures were identified by review of all radiology reports from the two radiology services serving the Dubbo area. The circumstances surrounding the fracture were determined by personal interview after the fracture. Only low-trauma hip fractures caused by a fall from a standing height or less were included in the present analysis. Subjects with underlying conditions that could cause pathological fractures, such as malignancy or metabolic bone disease, were excluded. BMD was measured at the femoral neck by DXA with a LUNAR DPX-L densitometer (GE-LUNAR, Madison, WI). In those subjects in whom BMD had been measured more than once before a hip fracture, the most recent DXA scan before the event was used for analysis. All clinical data were taken from the appropriate BMD visit. The coefficient of variation for the femoral neck BMD measurement in normal subjects was 3.5% (28). Twelve women and seven men with hip fractures who had had a bone density assessment more than 3 months after the hip fracture were excluded from analysis. However, inclusion of these subjects did not alter the results. FNAL, the length between the lateral aspect of the greater trochanter and the medial border of the femoral head, had been measured in 105 TABLE 1. Characteristics of nonfracture and fracture subjects women and 122 men (29). The length of the femoral neck was 2 sd longer in men than women (102.1 6.0 vs. 90.2 5.1 mm; P 0.0001). Estimation of volumetric BMD Volumetric BMD (in grams per centimeter cubed) was estimated assuming the femoral neck to be a cylinder. The average diameter (d)of the femoral neck was obtained from the LUNAR software that uses a fixed length (k 1.5 cm) along the femoral neck for the measurement of areal BMD (in grams per centimeter squared) and bone mineral content (BMC) (in grams). The area measured, calculated as BMC/BMD, is equivalent to k d. This average neck diameter was used to estimate the CSA [ (d/2) 2 ] of the femoral neck and thus its volume [ (d/2) 2 k]. Volumetric BMD (vbmd) was estimated as BMC per unit volume BMC/( (d/2) 2 k), i.e. vbmd ((BMD) 2 /BMC) (4k/ ). Statistical analysis Comparisons between subjects with and without hip fractures and between women and men were made using the unpaired t test. A result was considered significant for P 0.05. Correlations between measures of bone content and anthropometric values were obtained using the Pearson correlation coefficient (r). All statistical analyses were performed using the Statistical Analysis System (SAS, SAS Institute, Cary, NC). All measures of bone content (areal BMD and estimated volumetric BMD, CSA, and BMC) were analyzed as independent variables in univariate proportional hazards models in relation to the risk of hip fracture. Women and men were analyzed separately, and these models were subsequently adjusted for age. Multivariate stepwise analysis, with the above measures of BMC, as well as anthropometric variables, was used to determine the best independent predictor of hip fracture. Prevalence of osteoporosis was calculated from the whole Dubbo Osteoporosis Epidemiology Study sample, based on a bone density 2.5 sd or more below young normal values. The young normal values (and sd) were obtained from a previous study of healthy twins (30) aged between 20 and 35 yr and were comparable with published LUNAR data. Thus, for men, the osteoporotic cut-off for areal BMD was higher than for women ( 0.74 vs. 0.70 g/cm 2 ). However, estimated volumetric BMD was similar in young women and men (0.40 0.05 g/cm 3 ), and thus the osteoporotic cut-off value ( 2.5 sd) was 0.275 g/cm 3. Sensitivity and specificity for osteoporosis as a predictor of hip fracture were calculated using 2 2 contingency tables. Significance differences between sensitivities were evaluated using Poisson distribution. Results Subject characteristics Seventy-three women and 23 men with low-trauma hip fractures were compared with 852 women and 635 men who did not sustain any fractures (Table 1). Nine of the women and three of the men had a hip fracture within 3 months before the bone density scan. Exclusion of this subset did not Nonfracture Hip fracture Women Men Women Men n 853 635 73 23 Age (yr) 70 7 d 69 6 c 77 7 76 8 Height (cm) 160 6 a,d 174 7 c 156 7 a 170 7 Weight (kg) 66 12 a,d 79 12 b 55 10 a 73 12 Areal FN BMD (g/cm 2 ) 0.81 0.12 a,d 0.93 0.15 d 0.63 0.11 a 0.76 0.14 Vol FN BMD (g/cm 3 ) 0.31 0.06 d 0.31 0.06 d 0.25 0.04 0.26 0.04 FN BMC (g) 4.0 0.7 a,d 5.3 1.0 d 3.1 0.7 a 4.3 1.1 CSA (cm 2 ) 8.6 1.5 a,c 11.3 1.8 8.1 1.5 a 11.2 2.3 FN, Femoral neck. For differences between women and men: a P 0.0001. For differences between nonfracture and hip fracture: b P 0.01; c P 0.001, d P 0.0001.

2778 J Clin Endocrinol Metab, June 2004, 89(6):2776 2782 Center et al. Volumetric Bone Density and Hip Fracture Risk alter the results. Hip fracture subjects were older, shorter, and weighed less than the nonfracture subjects. Women, as expected, were consistently shorter and lighter than men. However, there was no age difference between the sexes in either the hip fracture or nonfracture groupings (Table 1). Relationships between bone measurements and anthropometric characteristics Height was associated with areal BMD (r 0.31 for women and r 0.23 for men; P 0.0001), BMC (r 0.40 for women and r 0.39 for men; P 0.0001), CSA (r 0.31 for women and r 0.42 for men; P 0.0001), and FNAL (r 0.41 for women and r 0.57 for men; P 0.0001). However, height was only weakly correlated with estimated volumetric BMD in women (r 0.13; P 0.0001) and not related in men (r 0.02; P 0.52). Weight was positively associated with all the bone measures (r 0.23 0.57; P 0.0001), whereas age was negatively correlated with all the bone parameters except for CSA. For CSA, there were weakly positive correlations with age in women (r 0.07; P 0.05) and men (r 0.13; P 0.0009) and with FNAL in men (r 0.37; P 0.0001) but not in women (r 0.14; P 0.16). Sex differences in bone measurements Areal and estimated volumetric femoral neck bone density was lower in subjects with hip fractures compared with nonfracture subjects (P 0.0001; Table 1). Estimates of bone size, such as CSA, were also lower in women (P 0.001) with hip fractures, although not in men. Areal BMD, BMC, and CSA were lower in women than men for both hip fracture and nonfracture subjects. Estimated volumetric bone density Estimated volumetric density at the femoral neck was the same for women and men without hip fractures (0.31 0.06 g/cm 3 ). It was lower and again the same in women and men with hip fractures (0.25 0.04 and 0.26 0.04 g/cm 3 for women and men, respectively). The distribution of femoral neck areal and estimated volumetric bone densities (Fig. 1, A and B) confirms the lower mean values in the hip fracture group for both areal and estimated volumetric BMD. However, whereas areal bone density was consistently lower in women than men within hip fracture and nonfracture groups, estimated volumetric density had the same distribution in women and men within fracture and nonfracture groups. Areal and volumetric BMD and hip fracture risk Both areal and estimated volumetric BMD were predictors of hip fracture with similar magnitudes of risk. For a 1 sd (0.13 g/cm 2 ) decrease in areal BMD, relative risk (95% confidence interval) of hip fracture was 3.8 (3.1 4.8) in women and 3.3 (2.2 5.0) in men. For estimated volumetric BMD, the relative risks were 3.3 (2.6 4.2) in women and 3.5 (2.1 5.8) in men. These relative risks remained significant after adjustment for age (Table 2). There was no significant difference between the relative risks for estimated volumetric compared with areal BMD for women or men. Sensitivity and specificity of areal and volumetric BMD for hip fracture prediction The sensitivity and specificity for hip fracture prediction were compared using areal and estimated volumetric BMD with a 2.5 sd cut-off for osteoporosis. In women, sensitivity was similar for both areal and estimated volumetric BMD (P 0.22), although specificity was slightly but significantly better using the areal BMD cut-off (Table 3). However, in men, estimated volumetric BMD was more sensitive than areal BMD for hip fracture prediction (70 vs. 43%; P 0.04). Although the specificity was lower, it was similar to that in women (Table 3). Lowering the areal cut-off to 2 sd for men improved the sensitivity only marginally (to 57%), with a lowering of specificity to 83%. The 2.5 sd cut-off was the same in women and men using estimated volumetric BMD but higher in men than women using areal BMD due to the overall higher areal bone density in men (Fig. 2). The difference between the 2.5 sd areal and estimated volumetric cut-offs, particularly in men, resulted in different osteoporosis prevalence estimates. For men, the overall prevalence of osteoporosis in this sample was 9% using areal BMD and 25% using estimated volumetric BMD. For women, the difference was similar but less marked: 23 and 30%, respectively. For areal BMD, this resulted in a large proportion of men with hip fractures above the 2.5 sd cut-off (Fig. 2). However, risk of hip fracture at any specific areal BMD was similar in men and women. In contrast, for estimated volumetric BMD, the majority of hip fractures were below the 2.5 sd cut-off in both in women and men (78 and 70%, respectively; P 0.22). However, as expected, the incidence of hip fracture at any specific estimated volumetric BMD was higher in men than women (Fig. 2). This difference in incidence rate at the same estimated femoral neck volumetric BMD can be explained by the higher number of fractures overall in women compared with men (Fig. 1). For example, at an estimated volumetric BMD between 0.20 and 0.25 g/cm 3, the hip fracture incidence in women was 31% and in men was 11% (P 0.009). Below 0.2 g/cm 3, the incidence was higher in women and men (82 and 51%, respectively), but the difference did not reach statistical significance, possibly due to the small number of hip fractures in men (P 0.6). Discussion In the present study, the difference in bone density between men and women using the usual areal BMD measure disappeared when an estimate of volumetric BMD at the femoral neck was used. Not only did women and men in general have the same estimated volumetric BMD, but women and men with hip fractures had similar and lower estimated volumetric femoral neck bone densities than their nonfractured counterparts. In addition, the sensitivity of estimated femoral neck volumetric BMD, when used to define a cut-off for hip fracture prediction, was better in men than areal BMD and comparable with that in women. These re-

Center et al. Volumetric Bone Density and Hip Fracture Risk J Clin Endocrinol Metab, June 2004, 89(6):2776 2782 2779 FIG. 1. Top, Areal femoral neck bone density in hip fracture and nonfracture subjects (women, f; and men, u). Bottom, Volumetric femoral neck bone density in hip fracture and nonfracture subjects. TABLE 2. Risk for hip fracture (relative risk and 95% confidence intervals) SD unit of change Univariate Adjusted for age Women Areal BMD 0.13 g/cm 2 3.8 (3.1, 4.8) 3.0 (2.3, 3.9) Vol BMD 0.05 g/cm 3 3.3 (2.6, 4.2) 2.4 (1.8, 3.2) BMC 0.8 g 3.9 (3.1, 5.0) 4.4 (2.8, 6.9) CSA 1.5 cm 1.5 (1.1, 1.9) 1.6 (1.2, 2.0) Men Areal BMD 0.13 g/cm 2 3.3 (2.2, 5.0) 2.8 (1.7, 4.7) Vol BMD 0.05 g/cm 3 3.5 (2.1, 5.8) 3.2 (2.0, 5.0) BMC 0.8 g 2.4 (1.7, 3.3) 2.4 (1.7, 3.5) CSA 1.5 cm 1.1 (0.8, 1.6) 1.2 (0.9, 1.7) Bold relative risk values are significant. sults suggest that hip fractures occur at the same estimated femoral neck volumetric BMD between the sexes and that bone density per se does not explain the difference in hip TABLE 3. Sensitivity and specificity for hip fracture prediction using areal and volumetric BMD (using 2.5 SD cut-off) Areal BMD (%) Volumetric BMD (%) Women Sensitivity 73 78 Specificity 81 75 a Men Sensitivity 43 70 a Specificity 92 77 a a For comparison between areal and volumetric BMD, P 0.05. fracture rate between the sexes. In addition, when determining cut-off values for decisions regarding hip fracture risk across the sexes or where there are significant size differences, volumetric BMD estimated at the femoral neck may be an informative measure. This study has clearly demonstrated that estimated fem-

2780 J Clin Endocrinol Metab, June 2004, 89(6):2776 2782 Center et al. Volumetric Bone Density and Hip Fracture Risk FIG. 2. Hip fracture incidence using areal and volumetric BMD (women, f; and men, u). The arrows represent the 2.5 SD cut-off for osteoporosis. p-yrs, Person-years. oral neck volumetric bone density is the same between men and women without fractures as has been noted by others (8, 31). However, in the present study, estimated femoral neck volumetric bone density was not only lower in men and women with hip fractures but also virtually identical in the two sexes. Thus, once size was removed from the equation, men and women fractured at the same estimated volumetric bone densities. The increased CSA in men (2 sd greater than in women) may explain, at least in part, the decreased fracture incidence in men. It presumably counterbalances the increased fracture risk one would have expected from the increased FNAL in men (2 sd longer than in women). In this study, estimated hip cross-section (CSA) was an independent predictor of hip fracture in women but not in men, suggesting that the bone size /hip fracture relationship is not linear. When using either areal or estimated volumetric bone density to define cut-off values for osteoporosis, this study showed significantly different results between the sexes. Using a cut-off value of 2.5 sd below the young normal range, as has been widely used for women to define osteoporosis, estimated volumetric and areal BMD gave similar sensitivities (78 and 73%, respectively) although somewhat different specificities (75 vs. 81%, respectively) for hip fracture prediction for the women in this population. In men, estimated volumetric BMD was considerably more sensitive than areal BMD (70 vs. 43%), whereas its specificity (77%) was similar to that in women, albeit lower than with areal BMD (92%). However, the higher specificity for areal BMD was only achieved with an unacceptable low sensitivity. The cut-off value of more than 2.5 sd below the young normal range to define osteoporosis in women was selected by the World Health Organization because it identified a proportion of women comparable to the expected lifetime fracture risk (32). There have been many epidemiological data to support use of this value (23). However, when using cut-off values for clinical decision-making, there is always a trade-off between sensitivity and specificity. Despite the 2.5 sd cut-off in women identifying those at higher risk of fracture, it is clear that a large proportion of fractures occur in those who have T-scores above the 2.5 sd threshold (33). In men, the optimal cut-off value is an even more contentious issue. It is still debated whether it should be based upon the female or male young normal values (24, 34, 35). Part of the problem lies not only in the relative paucity of data in men but also in the size effect in the areal BMD measure. To further complicate matters, although some studies have suggested that the relationship between areal BMD and fracture risk in women and men is similar (36 38), others have reported sex differences between BMD and fracture risk (9), and one recent review concluded that the interpretation of BMD levels in men remains unclear (24). In the present study, for a given estimated volumetric femoral neck BMD measurement, hip fracture incidence rates were higher in women than men, consistent with overall hip fracture incidence data, and the 2.5 sd cut-off captured the majority of the hip fractures in men. This was in contrast with areal BMD where, for the same areal BMD, hip fracture rates were similar between the sexes, consistent with findings by others (36 39); however, the majority of the hip fractures in men occurred above the 2.5 sd cut-off. This low sensitivity is in keeping with a recent review that concluded that use of the 2.5 areal BMD T-score criterion underestimated the prevalence of osteoporosis in men (35). The higher

Center et al. Volumetric Bone Density and Hip Fracture Risk J Clin Endocrinol Metab, June 2004, 89(6):2776 2782 2781 hip fracture rates in women than men at the same estimated femoral neck volumetric BMD can be explained by the overall higher number of fractures in women than men at any level of BMD. There has been increasing interest in using absolute vs. relative risk (T-score) for the institution of therapy (39). The formulation underlying this concept may eventually involve more than just age and BMD. It is not yet apparent whether the same absolute risk should be used for women and men because outcomes after fracture differ between the sexes (40). The present study, which has focused on sensitivity and specificity for clinical cut-off decision-making, has demonstrated a broadly significant difference between areal and estimated volumetric BMD between the sexes. The issues raised by this study need to be considered in an independent population when exploring concepts of absolute risk. This study has a number of advantages but also some limitations. Data were collected prospectively on a large number of men and women. Volumetric bone density at the femoral neck was estimated from the DXA scan and not a true volumetric measurement. Parallel measurements using quantitative computed tomography were not performed in this study, and we are not aware of any similar study. However, in a small study of 28 men using magnetic resonance imaging, measured magnetic resonance imaging volume of the femoral neck and the derived DXA volume were moderately correlated (r 0.83; P 0.001) (41). The number of hip fractures in men was small (n 23); thus calculated incidence rates for the lowest levels of areal or estimated volumetric BMD have a large sd. However, this aspect was not the main issue of this paper, and this number of hip fractures is comparable to other major prospective epidemiological studies of hip fracture in men (36). Moreover, the sd of the BMD measurements in hip fracture subjects was similar to that of the nonfracture subjects, suggesting consistency within the data. The present data relate only to estimated volumetric bone density at the femoral neck, and thus, similar conclusions cannot be drawn regarding estimates of volumetric BMD of the spine. However, femoral BMD is subject to less degenerative artifact than spinal BMD and is generally regarded as a better single measure for hip fracture prediction (42). In summary, this study has demonstrated that volumetric BMD estimated at the femoral neck is the same in men and women without hip fractures and similar but lower in those men and women with hip fractures. By minimizing the size artifact in the areal BMD measure, comparison of hip fracture risk between the sexes is reduced in a mechanistic sense to a density estimate only. Indeed, estimated volumetric BMD was more sensitive than areal BMD when used to define an osteoporotic cut-off level in men and comparable to areal BMD in women. Clearly, this area requires more work before estimates of volumetric BMD can be recommended for cutoff decisions. However, estimated femoral neck volumetric BMD, which is the same between the sexes, highlights some of the problems with areal BMD, which differs between the sexes. Thus, it may provide additional useful information by which to assess hip fracture risk in both men and women. Acknowledgments Received April 1, 2003. Accepted February 23, 2004. Address all correspondence and requests for reprints to: Dr. Jacqueline Center, Bone and Mineral Research Program, Garvan Institute of Medical Research, St Vincent s Hospital, 384 Victoria Street, Sydney, New South Wales 2010, Australia. E-mail: j.center@garvan.org.au. References 1. Seeman E 2002 Pathogenesis of bone fragility in women and men. Lancet 359:1841 1850 2. Nguyen T, Sambrook P, Kelly P, Jones G, Lord S, Freund J, Eisman J 1993 Prediction of osteoporotic fractures by postural instability and bone density. BMJ 307:1111 1115 3. Looker AC, Wahner HW, Dunn WL, Calvo MS, Harris TB, Heyse SP, Johnston Jr CC, Lindsay R 1998 Updated data on proximal femur bone mineral levels of US adults. Osteoporos Int 8:468 489 4. Melton LJ, Khosla S, Achenbach SJ, O Connor MK, O Fallon WM, Riggs BL 2000 Effects of body size and skeletal site on the estimated prevalence of osteoporosis in women and men. Osteoporos Int 11:977 983 5. Gilsanz V, Boechat MI, Gilsanz R, Loro ML, Roe TF, Goodman WG 1994 Gender differences in vertebral sizes in adults: biomechanical implications. Radiology 190:678 682 6. Mosekilde L, Mosekilde L 1990 Sex differences in age-related changes in vertebral body size, density and biomechanical competence in normal individuals. Bone 11:67 73 7. Looker AC, Beck TJ, Orwoll ES 2001 Does body size account for gender differences in femur bone density and geometry? J Bone Miner Res 16:1291 1299 8. Faulkner RA, McCulloch RG, Fyke SL, Coteau WE, McKay HA, Bailey DA, Houston CS, Wilkinson AA 1995 Comparison of areal and estimated volumetric bone mineral density values between older men and women. Osteoporos Int 5:271 275 9. Melton LJ, Atkinson J, O Connor MK, O Fallon WM, Riggs BL 1998 Bone density and fracture risk in men. J Bone Miner Res 13:1915 1923 10. Kröger H, Lunt M, Reeve J, Dequecker J, Adams JE, Birkenhager JC, Curiel MD, Felsenberg D, Hyldstrup L, Kotzki P, Lavai-Jeantet A-M, Lips P, Louis O, Perez Cano R, Reiners C, Ribot C, Ruegsegger P, Schneider P, Braillon P, Pearson J 1999 Bone density reduction in various measurement sites in men and women with osteoporotic fractures of spine and hip: the European quantitation of osteoporosis study. Calcif Tissue Int 64:191 199 11. Seeman E, Duan Y, Fong C, Edmonds J 2001 Fracture site-specific deficits in bone size and volumetric density in men with spine or hip fractures. J Bone Miner Res 16:120 127 12. Cheng XG, Lowett G, Boonen S, Nicholson PHF, Brys P, Nijs J, Dequecker J 1997 Assessment of the strength of proximal femur in vitro: relationship to femoral bone mineral density and femoral geometry. Bone 20:213 218 13. Cordey J, Schneider M, Belendez C, Zeigler WJ, Rahn BA, Perren SM 1992 Effect of bone size, not density, on the stiffness of the proximal part of the normal and osteoporotic human femora. J Bone Miner Res 7(Suppl 2):S437 S444 14. Lotz JC, Hayes WC 1990 The use of quantitative computed tomography to estimate the risk of fracture of the hip from falls. J Bone Joint Surg Am 72:689 700 15. Beck TJ, Ruff CB, Warden KE, Scott WWJ, Rao GU 1990 Predicting femoral neck strength from bone mineral data: a structural approach. Invest Rad 25:6 18 16. Faulkner KG, Cummings SR, Black D, Palermo L, Gluer C-C, Genent HK 1993 Simple measurement of femoral geometry predicts hip fracture: the study of osteoporotic fractures. J Bone Miner Res 8:1211 1217 17. Nakamura T, Turner CH, Yoshikawa T, Slemder CW, Peacock M, Burr DB, Mizuno Y, Orimo H, Ouchi Y, Johnston Jr CC 1994 Do variations in hip geometry explain differences in hip fracture risk between Japanese and White Americans? J Bone Miner Res 9:1071 1076 18. Peacock M, Turner CH, Liu G, Manatunga AK, Timmerman L, Johnston CCJ 1995 Better discrimination of hip fracture using bone density, geometry and architecture. Osteoporos Int 5:167 173 19. Karlsson KM, Serbno I, Obrant KJ, Redlund-Johnell I, Johnell O 1996 Femoral neck geometry and radiographic signs of osteoporosis as predictors of hip fracture. Bone 18:327 330 20. Center JR, Nguyen TV, Pocock NA, Noakes K, Eisman JA, Sambrook PN 1998 Femoral neck axis length, peak height, height loss and risk of hip fracture in males and females. Osteoporos Int 8:75 81 21. Meyer HE, Tverdal A, Falch JA 1993 Risk factors for hip fracture in middleaged Norwegian women and men. Am J Epidemiol 137:1203 1211 22. Hemenway D, Feskanich D, Colditz GA 1995 Body height and hip fracture: a cohort study of 90,000 women. Int J Epidemiol 24:783 786 23. Kanis JA, Melton LJI, Christiansen C, Johnston CC, Khaltaev N 1994 The diagnosis of osteoporosis. J Bone Miner Res 9:1137 1141

2782 J Clin Endocrinol Metab, June 2004, 89(6):2776 2782 Center et al. Volumetric Bone Density and Hip Fracture Risk 24. Orwoll ES 2000 Assessing bone density in men. J Bone Miner Res 15:1867 1870 25. Simons LA, McCallum J, Simons J, Powell I, Ruys J, Heller R, Lerba C 1990 The Dubbo study: an Australian prospective community study of the health of the elderly. Aust NZ J Med 20:783 789 26. Jones G, Nguyen T, Sambrook PN, Kelly PJ, Gilbert C, Eisman JA 1994 Symptomatic fracture incidence in elderly men and women: the Dubbo Osteoporosis Epidemiology Study (DOES). Osteoporos Int 4:277 282 27. Nguyen TV, Kelly PJ, Sambrook PN, Gilbert C, Pocock NA, Eisman JA 1994 Lifestyle factors and bone density in the elderly: implications for osteoporosis prevention. J Bone Miner Res 9:1339 1346 28. Nguyen TV, Sambrook PN, Eisman JA 1997 Sources of variability in bone mineral density measurements: implications for study design and analysis of bone loss. J Bone Miner Res 12:124 135 29. Center JR, Nguyen TV, Sambrook PN, Noakes K, Pocock NA, Eisman JA 1995 Relationship between femoral neck axis length, height loss, and risk of hip fracture in males and females. J Bone Miner Res 10:S366 30. Pocock NA, Eisman JE, Mazess RB, Sambrook PN, Yeates MG, Freund J 1988 Bone mineral density in Australia compared with the United States. J Bone Miner Res 3:601 604 31. Compston JE, Evans WD, Crawley EO, Evans C 1988 Bone mineral content in normal UK subjects. Br J Radiol 61:631 636 32. World Health Organization Study Group 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Technical Report Series, No. 843. Geneva: World Health Organization 33. Wainwright SA, Phipps KR, Stone JV, Cauley JA, Vogt MT, Black DM, Orwoll ES 2001 A large proportion of fractures in postmenopausal women occur with baseline bone mineral density T-score -2.5. J Bone Miner Res 16 (Suppl 1):S155 34. Melton LJ, Orwoll ES, Wasnich RD 2001 Does bone density predict fractures comparably in men and women. Osteoporos Int 12:707 709 35. Faulkner KG, Orwoll ES 2002 Implications in the use of T-scores for the diagnosis of osteoporosis in men. J Clin Densitom 5:87 93 36. De Laet CE, Van Hout BA, Burger H, Weel AE, Hofman A, Pols HA 1998 Hip fracture prediction in elderly men and women: validation of the Rotterdam Study. J Bone Miner Res 13:1587 1593 37. Cheng S, Suominen H, Sakari-Rantala R, Laukkanen P, Avikainen V, Heikkinen E 1997 Calcaneal bone density predicts fracture occurrence: a five year follow-up study in elderly people. J Bone Miner Res 12:1075 1082 38. Selby PL, Davies D, Adams JE 2000 Do men and women fracture at similar bone densities? Osteoporos Int 11:153 157 39. De Laet CEDH, Van der Klift M, Hofman A, Pols HAP 2002 Osteoporosis in men and women: a story about bone mineral density thresholds and hip fracture risk. J Bone Miner Res 17:2231 2236 40. Center JR, Nguyen TV, Schneider P, Sambrook PN, Eisman JA 1999 Mortality after all major types of osteoporotic fracture in men and women: an observational study. Lancet 353:878 882 41. Arokoski MH, Arokoski JP, Vainio P, Niemitukia LH, Kroger H, Jurvelin JS 2002 Comparison of DXA and MRI methods for interpreting femoral neck bone mineral density. J Clin Densitom 5:289 296 42. Melton 3rd LJ, Atkinson EJ, O Fallon WM, Wahner HW, Riggs BL 1993 Long-term fracture prediction by bone mineral assessed at different skeletal sites. J Bone Miner Res 8:1227 1233 JCEM is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.