Vertebral body density role in determining vertebral osteoporotic fracture type and its progression.

INTRODUCTION

Vertebral osteoporotic fractures (VOF) are among the major health concerns nowadays, especially with the increase in life expectancy (Ballane et al., 2017 ; Holroyd et al., 2008 ). These fractures commonly lead to an impaired quality of life with an increased mortality risk and a high economic burden on the health system and the individual.

Osteoporosis is characterized by a reduction in bone mass and alteration in bone architecture, resulting in increased bone fragility and fracture risk (Ensrud, 2017 ). The World Health Organization defines osteoporosis as a bone mineral density (BMD) value of more than 2.5 standard deviations below the mean for a young white female reference group. Dual‐energy x‐ray absorptiometry (DEXA) is the gold standard for measuring BMD, and low BMD is considered a strong risk factor for sustaining a fracture (Blake & Fogelman, 2007 ). Consequently, drug development is directed at improving BMD to reduce the risk of new osteoporosis‐related fractures (Nogués & Martinez‐Laguna, 2018 ).

Among the most common osteoporotic fractures are those of the spine. These fractures include several types, which were classified by the German Society for Orthopedics and Trauma (DGOU) in the AO Spine‐DGOU classification for osteoporotic Thoracolumbar Fractures (Schnake et al., 2018 ). Nevertheless, it is unclear whether this classification method has a prognostic value. Moreover, BMD contributes to predicting the risk of osteoporotic vertebral fracture (Cann et al., 1985 ; Wang et al., 2012 ), whereby improved BMD reduces the risk of vertebral fractures in osteoporotic patients (Bouxsein et al., 2019 ; Cummings et al., 2002 ). However, the effect of BMD on vertebral fracture type and progression (i.e., deterioration) over time is unclear.

Previous studies have shown that bone density in Hounsfield units (HU) obtained from computed tomography (CT) can predict BMD (Aydin Ozturk et al., 2021 ; Lee et al., 2013 ; Pinto et al., 2022 ; Schreiber et al., 2011 ; Schreiber et al., 2014 ). Furthermore, HU measurements may be superior to DEXA in evaluating spinal bone quality since they are less influenced by degenerative changes, vascular calcifications, and spinal deformity (Pinto et al., 2022 ). Recently, several studies have evaluated HU measurements to predict proximal junctional failure (Meredith et al., 2013 ), pedicle screw loosening (Bredow et al., 2016 ; Schwaiger et al., 2014 ), osteoporotic vertebral fractures (Schwaiger et al., 2014 ), cage subsidence (Mi et al., 2017 ; Xi et al., 2020 ), and vertebral collapse fractures in multiple myeloma patients (Zijlstra et al., 2020 ). Studying the correlation between BMD, as reflected in HU measurements in patients with osteoporotic compression spine fractures, may elucidate whether BMD influences the fracture type and corresponds to the definition in the AO Spine‐DGOU Classification for Osteoporotic Thoracolumbar Fractures (Schnake et al., 2018 ).

The major aims of this study were: (1) To determine whether significant differences exist in the bone density (measured in HU) of fractured and non‐fractured vertebral bodies between the various types of osteoporotic vertebral fractures according to the definition of the AO Spine‐DGOU Classification for Osteoporotic Thoracolumbar Fractures (Schnake et al., 2018 ). (2) To examine whether vertebral body bone density (in HU) or fracture classification is associated with initial and progressive collapse of the vertebral body following acute osteoporotic compression spine fractures. (3) To provide predictive measures for fracture progression.

We hypothesized that the BMD of non‐fractured vertebral bodies would play a significant role in determining the fracture type; namely, a lower BMD would be associated with a more severe fracture. However, we hypothesized the opposite for the fractured vertebra. That is, a higher BMD will be associated with a more severe fracture, and that BMD will be negatively associated with the magnitude of the collapse. Therefore, we hypothesized that BMD and the magnitude of vertebral collapse would serve as predictive variables for fracture progression.

MATERIALS AND METHODS

Materials

The study included 124 patients (40 males and 84 females between 50 and 93 years) with acute osteoporotic compression spine fractures who underwent a computerized tomography (CT) scan between 2008 and 2013. The study was approved by the Institutional Review Board.

Patients were included in the study if they were 50 years or older and had an acute osteoporotic compression spine fracture. The diagnosis of acute fracture was based on the clinical presentation of acute pain, usually after minimal trauma, typical appearance on the CT scan with new onset fracture lines, and comparison to old radiographs when available (Smorgick et al., 2020 ). All patients included in the study underwent a CT scan at the time of diagnosis and plain radiography at least 3 months later. The mean length of follow‐up was 17.5 months, with a range of 3 to 70 months. All patients were treated with a spinal orthosis for 3 months. Patients were excluded if they presented with pathological fractures, transverse or spinous process fractures, and fractures due to high‐energy trauma. Patients with old fractures and those with a radiological follow‐up of less than 3 months were also excluded from the study. Consequently, of 305 patients, 34 were excluded due to having a transverse process or a spinous process fracture, 34 due to high‐energy trauma, 44 did not meet the age criterion (i.e., were <50 years old), 21 had old fractures, and five were excluded due to having pathological fractures. Of the 167 remaining patients, 43 did not have the required three‐month follow‐up radiograph.

Fracture classification

Vertebral body fractures were classified by a spine surgeon according to the AO spine‐DGOU classification for osteoporotic thoracolumbar fractures (Schnake et al., 2018 ). Based on the AO spine‐DGOU classification for osteoporotic thoracolumbar fractures, we divided the fractures into three major types, which were also the most frequent (95% of the fractures) (Schnake et al., 2018 ): Impression fracture (OF2): deformation with no or only minor involvement of the posterior wall. The fracture involved a single endplate. Incomplete burst fracture (OF3): deformation with distinct involvement of the posterior wall. It affected only one endplate but showed a distinct involvement of the anterior and posterior walls. Complete burst fracture (OF4): loss of integrity of the vertebral frame structure, vertebral body collapse, or pincer‐type fracture. In cases of loss of integrity of the vertebral frame structure, both endplates and the posterior wall were involved.

OF1 was excluded from our fracture criteria because it can be diagnosed only via MRI images and OF5 was excluded because it is surgically treated.

Measurements of the vertebral body

The vertebral body height (mm) was measured using the Sectra PACS (v. 23.2.6, 2022, Sweden). Vertebral body density (HU) and homogeneity of the density (HU) were measured using the IntelliSpace Portal, Philips (v. 11.2) (Table 1 ). All measurements were obtained for the fractured vertebra and adjacent (above and below) non‐fractured vertebrae. Vertebral body height was measured at diagnosis and at the follow‐up examination. Additional variables were calculated from the height measurements, including the collapse magnitude at diagnosis and follow‐up and the progress (deterioration) of collapse over time (i.e., the delta between the values at diagnosis and follow‐up examination). The density and homogeneity measurements were obtained only from CT scans performed at diagnosis (Table 1 ).

STATISTICAL ANALYSIS

Interobserver variation was performed for all variables measured. Two independent researchers blindly performed the measurements. Intraclass correlation coefficient (ICC) analysis revealed excellent results (0.942 ≤ ICC ≤ 0.968).

Differences in age and vertebral characteristics between males and females were examined using independent sample t‐tests. A chi‐square test was performed to determine the association between sex and fracture type. Spearman or Pearson correlations (according to variable distribution, SI Table 1 ) were carried out between age and other linear measurements. Kruskal‐Wallis tests were performed to examine the differences in age, follow‐up duration, and vertebral variables between fracture types. Paired‐sample t‐tests were performed to examine differences in variables between the fractured vertebra and the average calculated from the adjacent vertebrae. Multinomial logistic regression and logistic regression (with the Forward LR method using dummy variables) were performed to determine the predictive variables for each fracture type. Only measurements measured at the time of fracture diagnosis were included in these analyses. A linear regression (Forward method) was performed to examine the predictive variables for the deterioration in collapse between diagnosis and follow‐up, for the combined fracture types as well as for each condition separately. In these analyses, variables measured at the time of diagnosis and the duration of the follow‐up were included.

RESULTS

The type of fracture was found to be sex‐independent (χ 2 = 2.960, df = 2, p = 0.228). Similar frequencies of OF2 were found for males and females (35.0% and 33.3%, respectively). OF3 was slightly more frequent among females (46.4% vs. 32.5%), whereas OF4 was more frequent among males (32.5% vs. 20.2%). Nevertheless, fracture type was age‐dependent ( p = 0.005). Accordingly, the mean age of patients with fracture type OF3 and OF4 was higher (71.3 ± 9.41 and 70.1 ± 10.83 years, respectively) than those with a fracture type OF2 (65.0 ± 8.73 years). Therefore, further analyses were performed while controlling for age. No significant differences were found in the duration of the follow‐up between different fracture types ( p = 0.230).

A significant difference in the average height of the non‐fractured vertebrae was found between males and females, yet not in any other characteristic of the vertebrae nor in the deterioration of collapse over time (Table 2 ). However, significant associations were found between the different vertebral variables and age. At diagnosis, associations with age were found only for vertebral density, especially with the average density of non‐fractured vertebrae (r = −0.514, p < 0.001). At follow‐up, all variables measured were significantly associated with age, with the strongest correlation between age and the height of the fractured vertebra (r = −0.427, p < 0.001), followed by the average height of the adjacent vertebrae and the magnitude of the collapse (r = −0.323 and r = 0.349, p < 0.001, respectively). The deterioration in the magnitude of the vertebral collapse was also significantly correlated with age (r = 0.414, p < 0.001) (SI Table 2 ).

The mean height of the fractured vertebrae of the group with an OF4 fracture type was significantly lower than that of the other two groups, and the collapse was significantly larger (Table 3 ). In the follow‐up examination, significant differences in the fractured vertebral body height were found between all fracture types. The magnitude of collapse increased over time and differed significantly between all fracture types (Table 3 and Figure 1 ). Furthermore, significant differences were found in vertebral body collapse between the diagnosis and the follow‐up, whereby the deterioration over time was the largest in the OF4 type, followed by the OF3 type. The smallest progression was evident in the OF2 type, reaching significance when compared with the other two groups (Table 3 and Figure 1 ).

The fractured vertebra was significantly denser than the average density calculated from the adjacent non‐fractured vertebrae, independent of fracture type ( p < 0.01). However, a tendency of increment in the density of the fractured vertebra was found between OF2 and OF4 types, with a significant reduction in the density homogeneity (SD) in the OF4 type (Table 3 ). Nevertheless, in the adjacent non‐fractured vertebrae, no significant differences in bone density or homogeneity were found between fracture types (Table 3 ).

Following the multinomial logistic regression, age and vertebrae height (fractured and the average of the adjacent ones) differentiated between OF 2 and OF3. Yet, only age differentiated between OF2 and OF4. Nevertheless, sex, density of the fractured and average of the adjacent vertebrae, density homogeneity of the fractured vertebra, and rate of collapse at diagnosis differentiated between OF3 and OF4 (Table 4 ). In the binary logistic regression, different variables were included to predict each type of fracture (Table 4 ). The probability of manifesting the OF2 type decreased with increasing age (Exp(B) = 0.922), as well as with the increase in the magnitude of collapse at diagnosis (Exp(B) = 0.902). The probability of manifesting the OF3 type increased slightly with increasing age (Exp(B) = 1.050), whereas the OF4 type manifestation increased with increasing magnitude of collapse at diagnosis (Exp(B) = 1.112), regardless of the age of the individual (excluded from the discriminant function).

The main factor that predicted the progression of compression fracture over time (i.e., deterioration) was the fracture type; the more severe the fracture, the more it tended to progress. However, the predictive factors for each fracture type vary. The progression of collapse among patients with an OF2 fracture was larger with increased age; however, the collapse magnitude at diagnosis served as a protective factor (the larger the collapse, the smaller the progression). Among patients with OF3, the density of the fractured vertebra served as a protective factor against the progression of collapse over time. In patients with OF4, increased age and reduced density homogeneity advanced the progression of the collapse, whereas increased vertebral body density was a protective factor (Table 5 ).

DISCUSSION

Osteoporotic vertebral fractures are multifactorial and vary in definition and diagnosis (Ballane et al., 2017 ). These characteristics hinder their prediction and prevention. This study is the first to examine the predictive value of the AO Spine‐DGOU Classification for Osteoporotic Thoracolumbar Fractures (Schnake et al., 2018 ) for progressive collapse. It is also the first study to investigate the role of vertebral body density (in HU) in determining the type of osteoporotic fracture defined by the AO spine‐DGOU classification for osteoporotic thoracolumbar fractures (Schnake et al., 2018 ), as well as in predicting the progression of vertebral body collapse. Interestingly, our findings refuted our hypothesis that the density of non‐fractured vertebrae will predict the severity of the fracture and will help predict its progression. However, our hypotheses that the density of the fractured vertebra and its magnitude of collapse predict the type of fracture, as well as its progression, were validated.

Our study supports previous findings indicating that the mean BMD of fractured vertebrae, including the old and recent fracture subgroups, was significantly higher than that of the non‐fractured vertebrae (Hsu et al., 2021 ). In addition, we found that when the fracture was more severe, the density of the fractured vertebra was higher, although it did not reach significance. A possible explanation for the high density of fractured vertebrae in general, and with a more severe manifestation, might be the degree of compression. A more compressed vertebra involves less air space between the trabeculae. Moreover, vertebral body density homogeneity was significantly lower in the OF4 fracture type. The reduced homogeneity in OF4 could be due to the various forms of fractures that are included in this category, as well as their comminution nature (Schnake et al., 2018 ).

Similar to previous studies (Pinto et al., 2022 ), and probably due to the higher incidence of osteoporosis in females compared to males (Holroyd et al., 2008 ), the number of females with an osteoporotic vertebral fracture in our study cohort was more than twice of males. However, there was no significant difference between males and females concerning the fracture type, especially in stable impression fractures (OF2). As opposed to OF2, both OF3 and OF4 are considered unstable fractures (Schnake et al., 2018 ). These differences in the fracture type stability were evident in our results. Significant differences in the deterioration in collapse magnitude (between the time of diagnosis and follow‐up) were found between OF2 and the other two types, with no significant differences between the OF3 and OF4 types. However, in each fracture type, a larger compression had a protective effect on the progression of the fracture. It is noteworthy that males and females exhibited a similar progression of vertebral collapse following acute osteoporotic compression spine fracture.

Advanced age significantly contributed to a more severe manifestation of fracture type and its progression over time. This is probably due to the reduction in the number, thickness, and interconnectivity of trabeculae in the bones of older patients with osteoporosis (Rao & Singrakhia, 2003 ). The deterioration in bone quality with age is supported by our study. We found a strong and significant negative correlation between age and bone density of the non‐fractured vertebrae. However, the density of the non‐fractured vertebra did not correlate significantly with fracture progression over time.

Although osteoporosis manifestation and its severity increase the risk of vertebral fracture (Holroyd et al., 2008 ), according to our results, it did not increase the risk of a specific fracture type since we did not find any significant difference in the density of the non‐fractured vertebrae between the fracture types. Hence, other explanatory factors that affect the magnitude of the collapse of fractured vertebrae should be considered. As opposed to hip fractures, which usually occur due to a fall, vertebral fractures can result from various activities, including lifting, bending, and reaching (Holroyd et al., 2008 ; Myers & Wilson, 1997 ). Therefore, the immediate cause of the fracture (e.g., trauma or specific activity) and vertebral characteristics, such as the cross‐sectional area, might have a larger effect on determining the type of vertebral fracture (Gilsanz et al., 1995 ; Holroyd et al., 2008 ; Myers & Wilson, 1997 ; Silman et al., 1997 ).

This study has several limitations, including the age distribution of the patients; therefore, for some of the analyses, we had to control for age by studying a subsample. However, for most analyses, we examined the effect of age on the fracture type or severity of deterioration over time. Another limitation was the inability to measure bone density at the follow‐up examination, as only X‐rays were available.

In conclusion, identifying the type of fracture using the AO spine‐DGOU classification for osteoporotic thoracolumbar fractures classification is beneficial in determining the prognosis of patients with fractures. Furthermore, the density of the fractured vertebra, its magnitude of collapse, and the age of the individual are beneficial for predicting fracture progression.

FUNDING INFORMATION

The authors declare no financial support.

CONFLICT OF INTEREST STATEMENT

Yossi Smorgick, Ruth Pelleg‐Kallevag, Yoram Anekstein, Sergey Goldstein, and Hila May, declare that they have no conflict of interest.

Supporting information

Table S1: Kolmogorov–Smirnov test for examining normality of distribution.

Table S2: Spearman correlations between age and vertebral variables at diagnosis and follow‐up.

ACKNOWLEDGEMENTS

We thank Mrs. Ariana Dan for editing this article.

DATA AVAILABILITY STATEMENT

The datasets generated during and analyzed during the current study are available from the corresponding authors.

REFERENCES

Aydin Ozturk , P. , Arac , E. , Ozturk , U. , & Arac , S. ( 2021 ). Estimation of bone mineral density with Hounsfield unit measurement . British Journal of Neurosurgery , 38 , 1 – 4 . 10.1080/02688697.2021.1888877 33629886

Ballane , G. , Cauley , J. A. , Luckey , M. M. , & El‐Hajj Fuleihan , G. ( 2017 ). Worldwide prevalence and incidence of osteoporotic vertebral fractures . Osteoporosis International , 28 ( 5 ), 1531 – 1542 . 28168409 10.1007/s00198-017-3909-3

Blake , G. M. , & Fogelman , I. ( 2007 ). The role of DXA bone density scans in the diagnosis and treatment of osteoporosis . Postgraduate Medical Journal , 83 ( 982 ), 509 – 517 . 17675543 10.1136/pgmj.2007.057505 PMC2600106

Bouxsein , M. L. , Eastell , R. , Lui , L. , et al. ( 2019 ). Change in bone density and reduction in fracture risk: A meta‐regression of published trials . Journal of Bone and Mineral Research , 34 ( 4 ), 632 – 642 . 30674078 10.1002/jbmr.3641

Bredow , J. , Boese , C. K. , Werner , C. M. L. , Siewe , J. , Löhrer , L. , Zarghooni , K. , Eysel , P. , & Scheyerer , M. J. ( 2016 ). Predictive validity of preoperative CT scans and the risk of pedicle screw loosening in spinal surgery . Archives of Orthopaedic and Trauma Surgery , 136 ( 8 ), 1063 – 1067 . 27312862 10.1007/s00402-016-2487-8

Cann , C. E. , Genant , H. K. , Kolb , F. O. , & Ettinger , B. ( 1985 ). Quantitative computed tomography for prediction of vertebral fracture risk . Bone , 6 ( 1 ), 1 – 7 . 3994856 10.1016/8756-3282(85)90399-0

Cummings , S. R. , Karpf , D. B. , Harris , F. , Genant , H. K. , Ensrud , K. , LaCroix , A. Z. , & Black , D. M. ( 2002 ). Improvement in spine bone density and reduction in risk of vertebral fractures during treatment with antiresorptive drugs . The American Journal of Medicine , 112 ( 4 ), 281 – 289 . 11893367 10.1016/s0002-9343(01)01124-x

Ensrud , K. E. ( 2017 ). Crandall CJ: Osteoporosis . Annals of Internal Medicine , 167 ( 3 ), ITC17 – ITC31 . 28761958 10.7326/AITC201708010

Gilsanz , V. , Loro , M. L. , Roe , T. F. , Sayre , J. , Gilsanz , R. , & Schulz , E. E. ( 1995 ). Vertebral size in elderly women with osteoporosis. Mechanical implications and relationship to fractures . The Journal of Clinical Investigation , 95 ( 5 ), 2332 – 2337 . 7738196 10.1172/JCI117925 PMC295847

Goldstein , S. , Smorgick , Y. , Mirovsky , Y. , Anekstein , Y. , Blecher , R. , & Tal , S. ( 2016 ). Clinical and radiological factors affecting progressive collapse of acute osteoporotic compression spinal fractures . Journal of Clinical Neuroscience , 31 , 122 – 126 . 27387197 10.1016/j.jocn.2016.02.020

Holroyd , C. , Cooper , C. , & Dennison , E. ( 2008 ). Epidemiology of osteoporosis. Best Practice & Research . Clinical Endocrinology & Metabolism , 22 ( 5 ), 671 – 685 . 19028351 10.1016/j.beem.2008.06.001

Hsu , Y. , Hsieh , T. J. , Ho , C. H. , Lin , C. H. , & Chen , C. H. ( 2021 ). Effect of compression fracture on trabecular bone score at lumbar spine . Osteoporosis International , 32 ( 5 ), 961 – 970 . 33104822 10.1007/s00198-020-05707-3

Lee , S. , Chung , C. K. , Oh , S. H. , & Park , S. B. ( 2013 ). Correlation between bone mineral density measured by dual‐energy x‐ray absorptiometry and Hounsfield units measured by diagnostic CT in lumbar spine . Journal of Korean Neurosurgical Association , 54 ( 5 ), 384 – 389 . 10.3340/jkns.2013.54.5.384 PMC3873350 24379944

Meredith , D. S. , Schreiber , J. J. , Taher , F. , Cammisa , F. P. , Jr. , & Girardi , F. P. ( 2013 ). Lower preoperative Hounsfield unit measurements are associated with adjacent segment fracture after spinal fusion . Spine (Phila Pa 1976) , 38 ( 5 ), 415 – 418 . 22926280 10.1097/BRS.0b013e31826ff084

Mi , J. , Li , K. , Zhao , X. , Zhao , C.‐Q. , Li , H. , & Zhao , J. ( 2017 ). Vertebral body Hounsfield units are associated with cage subsidence after transforaminal lumbar interbody fusion with unilateral pedicle screw fixation . Clinical Spine Surgery , 30 ( 8 ), E1130 – E1136 . 27906743 10.1097/BSD.0000000000000490

Myers , E. R. , & Wilson , S. E. ( 1997 ). Biomechanics of osteoporosis and vertebral fracture . Spine (Phila Pa 1976) , 22 ( 24 ), 25S – 31S . 9431641 10.1097/00007632-199712151-00005

Nogués , X. , & Martinez‐Laguna , D. ( 2018 ). Update on osteoporosis treatment . Medicina Clínica (English Ed) , 150 ( 12 ), 479 – 486 . 10.1016/j.medcli.2017.10.019 29179892

Pinto , E. M. , Neves , J. R. , Teixeira , A. , Frada , R. , Atilano , P. , Oliveira , F. , Veigas , T. , & Miranda , A. ( 2022 ). Efficacy of Hounsfield units measured by lumbar computer tomography on bone density assessment: A systematic review . Spine (Phila Pa 1976) , 47 ( 9 ), 702 – 710 . 34468433 10.1097/BRS.0000000000004211

Rao , R. D. , & Singrakhia , M. D. ( 2003 ). Painful osteoporotic vertebral fracture: Pathogenesis, evaluation, and roles of vertebroplasty and kyphoplasty in its management . JBJS , 85 ( 10 ), 2010 – 2022 . 14563813

Schnake , K. J. , Blattert , T. R. , Hahn , P. , Franck , A. , Hartmann , F. , Ullrich , B. , Verheyden , A. , Mörk , S. , Zimmermann , V. , Gonschorek , O. , Müller , M. , Katscher , S. , Saman , A. E. , Pajenda , G. , Morrison , R. , Schinkel , C. , Piltz , S. , Partenheimer , A. , Müller , C. W. , … the Spine Section of the German Society for Orthopaedics and Trauma . ( 2018 ). Classification of osteoporotic thoracolumbar spine fractures: Recommendations of the spine section of the German Society for Orthopaedics and Trauma (DGOU) . Global the Spine Journal , 8 ( suppl 2 ), 46S – 49S . 10.1177/2192568217717972 PMC6130101 30210960

Schreiber , J. J. , Anderson , P. A. , & Hsu , W. K. ( 2014 ). Use of computed tomography for assessing bone mineral density . Neurosurgical Focus , 37 ( 1 ), E4 . 10.3171/2014.5.FOCUS1483 24981903

Schreiber , J. J. , Anderson , P. A. , Rosas , H. G. , Buchholz , A. L. , & Au , A. G. ( 2011 ). Hounsfield units for assessing bone mineral density and strength: A tool for osteoporosis management . JBJS , 93 ( 11 ), 1057 – 1063 . 10.2106/JBJS.J.00160 21655899

Schwaiger , B. J. , Gersing , A. S. , Baum , T. , Noël , P. B. , Zimmer , C. , & Bauer , J. S. ( 2014 ). Bone mineral density values derived from routine lumbar spine multidetector row CT predict osteoporotic vertebral fractures and screw loosening . American Journal of Neuroradiology , 35 ( 8 ), 1628 – 1633 . 24627455 10.3174/ajnr.A3893 PMC7964446

Silman , A. J. , O'Neill , T. W. , Cooper , C. , Kanis , J. , & Felsenberg , D. ( 1997 ). Influence of physical activity on vertebral deformity in men and women: Results from the European vertebral osteoporosis study . Journal of Bone and Mineral Research , 12 ( 5 ), 813 – 819 . 9144348 10.1359/jbmr.1997.12.5.813

Smorgick , Y. , Geftler , A. , Goldstein , S. , Mirovsky , Y. , Blecher , R. , & Anekstein , Y. ( 2020 ). Response to: Determination of any correlation between sagittal spinopelvic configuration and progressive collapse of acute osteoporotic compression spine fractures: A retrospective radiological analysis . Asian Spine Journal , 14 ( 5 ), 770 – 771 . 33108844 10.31616/asj.2020.0453.r2 PMC7595807

Wang , X. , Sanyal , A. , Cawthon , P. M. , Palermo , L. , Jekir , M. , Christensen , J. , Ensrud , K. E. , Cummings , S. R. , Orwoll , E. , Black , D. M. , for the Osteoporotic Fractures in Men (MrOS) Research Group , & Keaveny , T. M. ( 2012 ). Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans . Journal of Bone and Mineral Research , 27 ( 4 ), 808 – 816 . 22190331 10.1002/jbmr.1539 PMC3510751

Xi , Z. , Mummaneni , P. V. , Wang , M. , Ruan , H. , Burch , S. , Deviren , V. , Clark , A. J. , Berven , S. H. , & Chou , D. ( 2020 ). The association between lower Hounsfield units on computed tomography and cage subsidence after lateral lumbar interbody fusion . Neurosurgical Focus , 49 ( 2 ), E8 . 10.3171/2020.5.FOCUS20169 32738801

Zijlstra , H. , Wolterbeek , N. , Drost , R. W. , Koene , H. R. , van der Woude , H. J. , Terpstra , W. E. , Delawi , D. , & Kempen , D. H. R. ( 2020 ). Identifying predictive factors for vertebral collapse fractures in multiple myeloma patients . The Spine Journal , 20 ( 11 ), 1832 – 1839 . 32673729 10.1016/j.spinee.2020.07.004