BJR
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
First published online September 17, 2007
British Journal of Radiology (2007) 80, 893-897
© 2007 British Institute of Radiology
doi: 10.1259/bjr/37401526
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gnudi, S
Right arrow Articles by Fiumi, N
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gnudi, S
Right arrow Articles by Fiumi, N

Full paper

Bone density and geometry in assessing hip fracture risk in post-menopausal women

S Gnudi, MD E Sitta, MD and N Fiumi, MD

Modulo Dipartimentale di Medicina Interna, Istituti Ortopedici Rizzoli, Via Pupilli 1, 40136 Bologna, Italy

Correspondence: Dr Saverio Gnudi, Istituti Ortopedici Rizzoli, Via Pupilli 1, 40136 Bologna, Italy. E-mail: saverio.gnudi{at}ior.it


   Abstract
Top
 Abstract
Introduction
 Methods and materials
 Results
 Discussion
 References
 
We used femoral neck structural parameters (FNSPs), calculated from bone mineral density (BMD) measurements of the femoral neck by dual X-ray absorptiometry, to discriminate osteoporotic fractures of the proximal femur in post-menopausal women. We compared 1646 women without fracture and 429 women with hip fractures, including 273 with femoral neck (FN) and 156 with trochanter (TR) fractures. The association between the studied parameters and the fractures was modelled using multiple logistic regression, and included age, height and weight. Fracture-predicted probability (FPP) was also calculated for each predictor tested. Receiver operating characteristic (ROC) curve areas with their standard errors (SEs) were calculated for the fracture status, having the calculated FPP as a test variable. The areas were compared by the Hanley–McNeil test. Hip fracture had lower BMD, cross-sectional area (CSA), section modulus (SM) and cortical thickness (CT), and higher buckling ratio (BR), than controls. To the same extent as FN BMD, BR best predicted the risk for each fracture, showing ROC curve areas of 0.809 (SE 0.011) for hip fracture, 0.789 (SE 0.014) for FN fracture, and 0.848 (SE 0.016) for TR fracture. The association of BR with fracture risk did not differ from that of FN BMD, which has a ROC curve area of 0.801 (SE 0.011) for hip fracture, 0.778 (SE 0.014) for FN fracture, and 0.852 (SE 0.016) for TR fracture. Both FN BMD and BR predicted TR fracture significantly better than they did FN fracture. FNSPs, although interesting in understanding the biomechanics of bone fragility, do not appear to add diagnostic value to the simple measurement of BMD.


   Introduction
Top
 Abstract
 Introduction
Methods and materials
 Results
 Discussion
 References
 
Osteoporotic fractures are an increasing problem for healthcare services because their incidence is high, and therefore they are expensive to treat [1, 2]. Therefore, effective methods are needed to diagnose those at risk in order to target fracture prevention treatment [3–5], particularly with regard to hip fracture, which is the osteoporotic fracture with the worst outcome [6, 7]. Bone mineral density (BMD) has been found to be the best predictor of osteoporotic fracture [8, 9], although other risk factors [10, 11] have also been shown to be important, such as those leading to an increase in the propensity to fall [12–15]. For hip fracture in women, proximal femur geometry (PFG) has also been highlighted as a risk factor [16–19], with particular reference to the hip axis length (HAL) [16], the neck shaft angle (NSA) [18] and the femoral neck diameter (FND) [19]. Recently, femoral neck structural parameters (FNSPs) have also been tested for association with hip fracture risk [20–24]. However, whether their ability to predict hip fracture is better than that of BMD, or whether they can improve the hip fracture risk assessment of the latter parameter, is still debated. As FNSPs are estimable from dual X-ray absorptiometry (DXA), we evaluate in this study the ability of FNSPs to assess hip fracture risk in comparison with, and/or in addition to, BMD, in order to add to the estimation of their potential in predicting future fracture.


   Methods and materials
Top
 Abstract
 Introduction
 Methods and materials
Results
 Discussion
 References
 
After local ethics committee approval, we examined the data of 2075 Italian post-menopausal women living in the district of Bologna who were referred to our centre for DXA examination. They consisted of 1646 consecutive outpatients without medical history of previous fracture, and 429 consecutive women with hip fracture — either femoral neck (FN) (n = 273) or trochanter (TR) (n = 156) fracture — which occurred within 30 days of DXA examination. Subjects having a reported history of malignancies, rheumatoid arthritis, hyperthyroidism, hyperparathyroidism, heart, lung, liver or kidney failure, Paget's disease of bone, and those treated with corticosteroids were excluded from the present study. Age, age at menopause, height, weight, BMD, bone mineral content (BMC) and the area of the measured region of interest were recorded for each subject and included in the study data. A pencil-beam densitometer (Norland XR 36; Coopersurgical Co, Fort Atkinson, WI) was used to perform DXA scans of the FN region of the hip in all of the women and to measure their BMD. The women were positioned on the scanning table according to the manufacturer's recommendations with legs internally rotated, which was standardized by using the Norland intra-rotating legs fixture. Densitometry measurements were performed on the left hip or on the contralateral one in patients who had had a fracture of the left hip. At our centre, the repeated measurement coefficient of variation of the FN BMD was 1.9% (standard deviation (SD) = 0.7) [25]. Calculation of the FNSP was based on the assumption that density of fully mineralized bone is 1050 mg cm–3 [26] and that the cortical mass proportion (CMP) in the FN is 60% [27–29]. FND was estimated by dividing the area of the FN region by the constant height of 1.5 cm of the region of interest. Subsequently, the structural parameters of the femoral neck geometry (i.e. the cross-sectional area (CSA), the cross-sectional moment of inertia (CSMI), the section modulus (SM), the cortical thickness (CT) and the buckling ratio (BR)) were calculated according to the equations described in a paper by Filardi et al [30].

Statistics
The unpaired t-test was used to compare variables between women with hip fractures and those without. One-way analysis of variance (ANOVA) and the Scheffé post hoc test for pairwise comparisons were used to compare variables among women with TR fractures, FN fractures and those without fractures.

Pearson's correlation was used to assess the relationship between the studied variables. The association between the BMD and FNSPs (standardized for 1 SD change from the healthy subjects' mean value) when tested as predictors, and the dependent variable hip fracture status, were modelled using multiple logistic regression adjusting for age, height and weight. The relative probabilities of fracture were expressed by an odds ratio (OR). The group of predictors giving the best fracture prediction was modelled by logistic regression using the stepwise forward method. From the logistic equations, the subjects' fracture-predicted probability (FPP) was calculated for each tested predictor after adjusting for age, weight and height. Receiver operating characteristic (ROC) curves were then constructed by using the presence/absence of fracture as the status variable and the calculated FPP as the test variable. The calculated areas of different ROC curves were compared using the Hanley–McNeil [31, 32] test, in order to assess the ability of the fracture predictors to discriminate between subjects with and without fracture.


   Results
Top
 Abstract
 Introduction
 Methods and materials
 Results
Discussion
 References
 
The women who had FN or TR fracture were older and lighter than those without fracture. They also had lower BMD, CSA, SM and CT, and higher BR, than controls (Table 1Go). Women with TR fracture were also smaller and had lower CSMI than controls and had lower BMD, CSA, CSMI, SM and CT, and higher BR, than those with cervical fracture. FN BMD was positively correlated with CT (r = 0.99), CSA (r = 0.91), SM (r = 0.63) and CSMI (r = 0.43), and negatively with BR (r = –0.90) (p<0.001 for each structural parameter's correlation with FN BMD).


View this table:
[in this window]
[in a new window]

  		Table 1. Estimated parameters of the studied cases

 
Logistic regression showed that, after adjusting for age, weight and height and standardizing for 1 SD decrease in BMD, CSA, CSMI and SM and 1 SD increase in BR, each of these parameters predicted hip fracture and both types of proximal femur fractures, except for CSMI, which was associated with TR but not with FN fracture (Table 2Go). Table 3Go shows the areas of the ROC curves, built by using the FPP (obtained from each of the five logistic models reported in Table 2Go) as the test variable and the fracture status as the status variable. The ROC curve areas derived from the logistic models having FN BMD and BR as predictors did not differ significantly from each other according to the Hanley–McNeil test, but were significantly higher than those derived from models having CSA, CSMI and SM as predictors in hip and FN fractures, and those having CSMI in TR fracture. For each studied parameter, the ROC curve areas of the trochanteric fracture were significantly greater (Hanley–McNeil test) than those of the FN fracture (Table 3Go).


View this table:
[in this window]
[in a new window]

  		Table 2. Odds ratios(OR) and 95% confidence interval (95% CI) obtained from five different logistic regressions, each one including age, height, weight and one of the test variables

 

View this table:
[in this window]
[in a new window]

  		Table 3. ROC curve areas estimating the fracture probability by the test variable FPP calculated by logistic models having FN BMD and single FNSPs as predictors(adjusted for age, height and weight).

 
When BMD and FNSP were forced together into the logistic forward regression with age, age at menopause, weight and height, only BR among the FNSPs entered the model to predict hip fracture — BR: OR 1.73 (95% confidence interval (CI) 1.42–2.10); FN BMD: OR 1.82 (95% CI 1.29–2.11); age: OR 1.031 (95% CI 1.017–1.046); height: OR 1.03 (95% CI 1.01–1.06).

The area of the ROC curve, built by using the FPP derived from this forward logistic regression model as the test variable, was 0.812 (SE 0.011). It was not significantly different (Hanley–McNeil test) from the ROC curve areas calculated by using the FPP from the logistic regression testing BR and FN BMD separately.


   Discussion
Top
 Abstract
 Introduction
 Methods and materials
 Results
 Discussion
References
 
In this study, we evaluated the ability of FNSPs to discriminate between subjects with and without proximal femur fracture owing to minimal trauma. We found a statistically significant association between hip fracture risk and all of the tested FNSPs, after adjusting for age, age at menopause, weight and height, except for CSMI and FN fracture. Among the tested structural parameters, BR was the best predictor of the fracture risk. It predicted fracture significantly better than all of the other FNSPs and equally as well as FN BMD. These data agree with recent reports by Melton et al [23] and Ahlborg et al [24] who showed a similar ability of standard BMD measurements of the proximal femur and FNSPs to predict the fracture risk in osteoporotic women. The finding that BR is the best FNSP to predict fracture is particularly attractive because it derives from a measurement related to bone mass and femoral neck geometry (i.e. FND), both of which have been shown to influence bone strength [33, 34]. It provides a mechanism to explain the limit of geometric compensation (i.e. FND expansion) by bone cortex thinning, which leads to bone wall instability. Although structural parameters are interesting for understanding the biomechanics of bone fragility, they do not seem to add diagnostic value to the simple measurement of BMD. In fact, by conditional logistic regression, the best model for separating women with and without hip fractures, which contains both BMD and BR as predictors, does not predict fracture better than either considered separately. Conversely, the combined use of BMD and simple geometrical measurements of the FN has been shown to give better results, with modest but statistically significant improvements on the hip fracture prediction by BMD [19, 35–37].

Analysing hip fracture according to anatomical type, differences were found between FN and TR fracture. Lower FN BMD was found to characterize TR fracture better than FN [38, 39]; statistically significant differences in structural parameters between cervical and trochanteric fractures were also found. It follows, according to others [37], that FN BMD is better at separating women with and without TR fractures than those with and without FN fracture. Perhaps owing to their correlation with BMD, structural parameters are also better predictors of TR fracture than of FN fracture, in agreement with Szulc et al [22]. Among FNSPs, BR was the best predictor of both FN and TR fractures, but was most accurate at separating TR fracture, thus showing the same fracture prediction pattern as FN BMD. The fact that BR was best at predicting TR fracture highlights the importance of cortical bone instability in this type of fracture that occurs in women with very low bone mass. These data indicate differences in the structural basis of the two different types of proximal femur fracture, which should lead to FN and TR fracture being considered separately when evaluating individuals at risk of fracture.

This study has some limitations. The controls were densitometry-referred individuals, and therefore they may not reflect the general population. BMD measurement of women with fractures was performed retrospectively and current BMD may not reflect that at the time of fracture, although the short time that had elapsed between fracture and BMD measurement should have minimized the bias [40]. Measurement of both BMD and FNSPs lacks validation across equipment. In addition, the estimation of structural parameters from a two-dimensional image makes some assumptions that are likely to lead to inaccuracy in the evaluation of the structure indices, which could be reduced by using a three-dimensional technique. Despite these limitations, our study shows that some FNSPs and BMD discriminate proximal femur fractures to the same extent, and that their combined use does not significantly improve the predictive ability of each separately. It also shows that both FN BMD and FNSPs are stronger predictors of TR fracture than of FN fracture, and that BR is, among FNSPs, the best predictor of fracture and is equally as good as BMD. Although FNSPs do not predict hip fracture better than BMD, their evaluation has been shown to be useful in understanding the biomechanics of hip fractures. Further studies are needed to evaluate whether more advanced and precise methods for FNSP extraction, i.e. using a three-dimensional technique, can improve the assessment of fracture risk in osteoporotic women.


   Acknowledgments
 
We wish to thank Mrs Elettra Pignotti for statistical analysis and Mr Keith Smith for linguistic help.

Received for publication October 17, 2006. Revision received January 17, 2007. Accepted for publication January 30, 2007.


   References
Top
 Abstract
 Introduction
 Methods and materials
 Results
 Discussion
 References
 

  1. Schutte HE. Social and economic impact of osteoporosis. A review of the literature. Eur J Radiol 1995;20:165–9.[CrossRef][Medline]
  2. Schwartz AV, Kelsey JL, Maggi S, Tuttleman M, Ho SC, Jonsson PV, et al. International variation in the incidence of hip fractures: cross-national project on osteoporosis for the World Health Organization Program for Research on Aging. Osteoporos Int 1999;9:242–53.[CrossRef][Medline]
  3. Kanis JA, Melton LJ 3rd, Christiansen C, Johnston CC, Khaltaev N. The diagnosis of osteoporosis. J Bone Miner Res 1994;9:1137–41.[Medline]
  4. Lu Y, Genant HK, Shepherd J, Zhao S, Mathur A, Fuerst TP, et al. Classification of osteoporosis based on bone mineral densities. J Bone Miner Res 2001;16:901–10.[CrossRef][Medline]
  5. Kanis JA, Glüer CC. An update on the diagnosis and assessment of osteoporosis with densitometry. Committee of Scientific Advisors, International Osteoporosis Foundation. Osteoporos Int 2000;11:192–202.[CrossRef][Medline]
  6. Keene GS, Parker MJ, Pryor GA. Mortality and morbidity after hip fractures. BMJ 1993;307:1248–50.[Abstract/Free Full Text]
  7. Baudoin C, Fardellone P, Bean K, Ostertag-Ezembe A, Hervy F. Clinical outcomes and mortality after hip fracture: a 2-year follow-up study. Bone 1996;18(3 Suppl):149S–57S.
  8. Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 1996;312:1254–9.[Abstract/Free Full Text]
  9. Kanis JA, Johnell O, Oden A, Dawson A, De Laet C, Jonsson B. Ten-year probabilities of osteoporotic fractures according to BMD and diagnostic thresholds. Osteoporos Int 2001;12:989–95.[CrossRef][Medline]
  10. Cooper C, Barker DJ. Risk factors for hip fracture. N Engl J Med 1995;332:814–5.[Free Full Text]
  11. Cummings SR, Nevitt MC, Browner WS, Stone K, Fox KM, Ensrud KE, et al. Risk factors for hip fracture in white women. Study of Osteoporotic Fractures Research Group. N Engl J Med 1995;332:767–73.[Abstract/Free Full Text]
  12. Albrand G, Munoz F, Sornay-Rendu E, DuBoeuf F, Delmas PD. Independent predictors of all osteoporosis-related fractures in healthy postmenopausal women: the OFELY study. Bone 2003;32:78–85.[Medline]
  13. Grisso JA, Kelsey JL, Strom BL, Chiu GY, Maislin G, O'Brien LA, et al. Risk factors for falls as a cause of hip fracture in women. The Northeast Hip Fracture Study Group. N Engl J Med 1991. 324:1326–31.
  14. Pinilla TP, Boardman KC, Bouxsein ML, Myers ER, Hayes WC. Impact direction from a fall influences the failure load of the proximal femur as much as age-related bone loss. Calcif Tissue Int 1996;58:231–35.[Medline]
  15. Greenspan SL, Myers ER, Maitland LA, Resnick NM, Hayes WC. Fall severity and bone mineral density as risk factors for hip fracture in ambulatory elderly. JAMA 1994;271:128–33.[Abstract/Free Full Text]
  16. Faulkner KG, Cummings SR, Black D, Palermo L, Glüer CC, Genant HK. Simple measurement of femoral geometry predicts hip fracture: the study of osteoporotic fractures. J Bone Miner Res 1993;8:1211–17.[Medline]
  17. Peacock M, Turner CH, Liu G, Manatunga AK, Timmerman L, Johnston CC Jr. Better discrimination of hip fracture using bone density, geometry and architecture. Osteoporos Int 1995;5:167–73.[CrossRef][Medline]
  18. Gnudi S, Ripamonti C, Gualtieri G, Malavolta N. Geometry of the proximal femur in the prediction of hip fracture in osteoporotic women. Br J Radiol 1999;72:729–33.[Abstract]
  19. Alonso CG, Curiel MD, Carranza FH, Cano RP, Peréz AD. Femoral bone mineral density, neck-shaft angle and mean femoral neck width as predictors of hip fracture in men and women. Multicenter Project for Research in Osteoporosis. Osteoporos Int 2000;11:714–20.[CrossRef][Medline]
  20. Crabtree NJ, Kroger H, Martin A, Pols HA, Lorenc R, Nijs J, et al. Improving risk assessment: hip geometry, bone mineral distribution and bone strength in hip fracture cases and controls. The EPOS study. European Prospective Osteoporosis Study. Osteoporos Int 2002;13:48–54.[CrossRef][Medline]
  21. Gregory JS, Stewart A, Undrill PE, Reid DM, Aspden RM. Bone shape, structure, and density as determinants of osteoporotic hip fracture: a pilot study investigating the combination of risk factors. Invest Radiol 2005;40:591–7.[CrossRef][Medline]
  22. Szulc P, Duboeuf F, Schott AM, Dargent-Molina P, Meunier PJ, Delmas PD. Structural determinants of hip fracture in elderly women: re-analysis of the data from the EPIDOS study. Osteoporos Int 2006;17:231–6.[CrossRef][Medline]
  23. Melton LJ 3rd, Beck TJ, Amin S, Khosla S, Achenbach SJ, Oberg AL, et al. Contributions of bone density and structure to fracture risk assessment in men and women. Osteoporos Int 2005;16:460–7.[CrossRef][Medline]
  24. Ahlborg HG, Nguyen ND, Nguyen TV, Center JR, Eisman JA. Contribution of hip strength indices to hip fracture risk in elderly men and women. J Bone Miner Res 2005;20:1820–7.[CrossRef][Medline]
  25. Gnudi S, Sitta E. Clinical risk factor evaluation to defer postmenopausal women from bone mineral density measurement: an Italian study. J Clin Densitom 2005;8:199–205.[CrossRef][Medline]
  26. Martin RB, Burr DB. Non-invasive measurement of long bone cross-sectional moment of inertia by photon absorptiometry. J Biomech 1984;17:195–201.[CrossRef][Medline]
  27. Bohr H, Schaadt O. Bone mineral content of the femoral neck and shaft: relation between cortical and trabecular bone. Calcif Tissue Int 1985;37:340–4.[CrossRef][Medline]
  28. Bell KL, Loveridge N, Power J, Garrahan N, Stanton M, Lunt M, et al. Structure of the femoral neck in hip fracture: cortical bone loss in inferoanterior to superoposterior axis. J Bone Miner Res 1999;14:111–9.[CrossRef][Medline]
  29. Kuiper JW, Van Kuijk C, Grashuis JL. Distribution of trabecular and cortical bone related to geometry. A quantitative computed tomography study of the femoral neck. Invest Radiol 1997;32:83–9.[CrossRef][Medline]
  30. Filardi S, Zebaze RM, Duan Y, Edmonds J, Beck T, Seeman E. Femoral neck fragility in women has its structural and biomechanical basis established by periosteal modeling during growth and endocortical remodeling during aging. Osteoporos Int 2004;15:103–7.[CrossRef][Medline]
  31. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29–36.[Abstract/Free Full Text]
  32. Hanley JA, McNeil BJ. A method of comparing the areas under the receiver operating characteristic curves derived from the same cases. Radiology 1983;148:839–43.[Abstract/Free Full Text]
  33. McCreadie BR, Goldstein SA. Biomechanics of fracture: is bone mineral density sufficient to assess risk? J Bone Miner Res 2000;15:2305–8.[CrossRef][Medline]
  34. Seeman E. From density to structure: growing up and growing old on the surfaces of bone. J Bone Miner Res 1997;12:509–21.[CrossRef][Medline]
  35. Duboeuf F, Hans D, Schott AM, Kotzki PO, Favier F, Marcelli C, et al. Different morphometric and densitometric parameters predict cervical and trochanteric hip fracture: the EPIDOS Study. J Bone Miner Res 1997;12:1895–1902.[CrossRef][Medline]
  36. Gnudi S, Ripamonti C, Lisi L, Fini M, Giardino R, Giavaresi G. Proximal femur geometry to detect and distinguish femoral neck fractures from trochanteric fractures in postmenopausal women. Osteoporos Int 2002;13:69–73.[CrossRef][Medline]
  37. Pulkkinen P, Partanen J, Jalovaara P, Jämsä T. Combination of bone mineral density and upper femur geometry improves the prediction of hip fracture. Osteoporos Int 2004;15:274–80.[CrossRef][Medline]
  38. Greenspan SL, Myers ER, Maitland LA, Kido TH, Krasnov MB, Hayes WC. Trochanteric bone mineral density is associated with type of hip fracture in the elderly. J Bone Miner Res 1994;9:1889–94.[Medline]
  39. Vega E, Mautalen C, Gomez H, Garrido A, Melo L, Sahores AO. Bone mineral density in patients with cervical and trochanteric fractures of the proximal femur. Osteoporos Int 1991;1:81–6.[CrossRef][Medline]
  40. Jørgensen L, Crabtree NJ, Reeve J, Jacobsen BK. Ambulatory level and asymmetrical weight bearing after stroke affects bone loss in the upper and lower part of the femoral neck differently: bone adaptation after decreased mechanical loading. Bone 2000;27:701–07.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gnudi, S
Right arrow Articles by Fiumi, N
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gnudi, S
Right arrow Articles by Fiumi, N

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
BJR DMFR IMAGING  ALL BIR JOURNALS