- Email: [email protected]
Vertebral Deformities Identified by Vertebral Fracture Assessment
Journal of Clinical Densitometry, vol. 8, no. 3, 267–272, 2005 © Copyright 2005 by Humana Press Inc. All rights of any nature whatsoever reserved. 1094-6950/05/8:267–272/$30.00 DOI: 10.1220/1094-6950
Original Article
Vertebral Deformities Identified by Vertebral Fracture Assessment Associations With Clinical Characteristics and Bone Mineral Density
Dana Jacobs-Kosmin,1 Nora Sandorfi,*,1 Heather Murray,2 and John L. Abruzzo1 1Division
of Rheumatology and 2Division of Pulmonary Medicine, Department of Medicine, Thomas Jefferson University, Philadelphia, PA 19107
Abstract Whether vertebral fractures identified on radiographs are painful or not, they are associated with increased morbidity and mortality. Vertebral fractures on X-rays correlate with low bone mineral density (BMD) at the spine and hip in addition to several clinical characteristics. Evidence suggests that vertebral deformities detected by X-ray and by vertebral fracture assessment (VFA) show good agreement. We examined the relationship between VFA-detected vertebral deformities and patient characteristics as well as BMD by analyzing the records of 432 patients who had undergone dual-energy X-ray absorptiometry (DXA) scans with VFA. Patients’ demographic data and T-scores were obtained from patient questionnaires and DXA scans. We categorized vertebral deformities by type and severity. Patients with vertebral deformities were significantly older and more likely to report a history of fracture after childhood. Significantly more estrogen use was reported in patients without deformity. Those with deformities had significantly lower T-scores at the femoral neck and total hip but not at the spine. Increased severity and number of deformities correlated with lower T-scores at the total hip and femoral neck but not the spine. In conclusion, vertebral deformities detected by VFA, like those on X-ray, correlate with both clinical characteristics and reduced bone mass at the hip. These relationships, in addition to rapid performance, convenience, and minimal radiation exposure, indicate VFA-detected vertebral deformities are a valuable adjunct in identifying patients in need of additional evaluation and treatment. Key Words: Vertebral fracture; vertebral fracture assessment; osteoporosis.
Introduction
Additionally, as the number and severity of vertebral fractures increase, so does the risk for new fracture (6,8). Vertebral fractures can be identified on conventional lateral radiographs or by vertebral fracture assessment (VFA) of X-ray absorptiometry scans. VFA has shown good agreement with conventional radiographs in identifying vertebral deformities (10,11). Studies have shown that vertebral fractures on conventional X-rays correlate with low BMD at the spine and hip (12–14). The relative risk of a new fracture is greater within three adjacent vertebrae of an existing deformity than the risk in more distant vertebrae (15,16). Finally, female gender, postmenopausal state, advanced age, Caucasian or Asian race, low body mass index (BMI), smoking, and the use of certain medications,
Many vertebral fractures remain unrecognized by clinicians and, therefore, undiagnosed (1). Painful or not, vertebral fractures are associated with diminished pulmonary function, difficulty with activities of daily living, and increased mortality (2–5). The presence of a vertebral fracture is a predictor for the occurrence of new osteoporotic spine and hip fractures, independent of age and bone mineral density (BMD) (6–9).
Received 11/16/04; Revised 02/22/05; Accepted 03/10/05. *Address correspondence to: Nora Sandorfi, 1015 Walnut St., Suite 613, Philadelphia, PA 19107. E-mail: [email protected]
267
268 such as corticosteroids and anticonvulsants, are risk factors for low BMD and osteoporotic fractures on X-rays (13,17–21). The association of these clinical factors with vertebral deformities identified by VFA has not been investigated. The aim of this study was to examine whether there is a relationship between VFA-detected vertebral deformities and patients’ clinical characteristics. We also studied the correlation between vertebral deformities and BMD.
Methods Patient Population and Study Design After obtaining Institutional Review Board (IRB) approval, we searched records at our tertiary osteoporosis center for all dual-energy X-ray absorptiometry (DXA) scans done from February 2002 (the date VFA first became available at our center) to March 2004. Four hundred thirty-six patient charts were identified in which VFA was included. Patients were included if at least 10 out of 11 vertebrae from T6 to L4 could be sufficiently visualized for vertebral deformity assessment. Four patients were excluded from the analysis. All four patients had multiple obscured vertebrae on the scan for reasons including spinal deformity (in two patients), diaphragmatic shadow (in one patient), and a Greenfield filter (in one patient). Patients were divided into those with and without vertebral deformity for the comparisons of clinical characteristics and T-scores. In addition, we conducted planned subgroup analyses to explore the relationships between severity and number of deformities with clinical factors and T-scores.
Performing DXA and VFA Dual-energy X-ray absorptiometry scans had been carried out by a single, experienced X-ray technologist dedicated to DXA testing. A Hologic Delphi C densitometer was used to measure the BMD of the total hip, femoral neck, and lumbar spine (L1–L4). T-scores were calculated for the hip using the NHANES III reference population database and for the spine using the Hologic reference population database. T-scores for African-American patients were calculated using a race-adjusted database. The World Health Organization criteria for osteopenia (T-score between –1 and –2.5) and osteoporosis (T-score less than or equal to –2.5) were used for categorization. The VFA had been performed using a semiquantitative technique combining visual and morphometric assessment. First, visual assessment was used to identify possible deformities. Next, morphometric assessment was used to confirm the presence and grade of deformities. Using DXA system software, scans were visually reviewed by the technologist for vertebral deformity. If vertebral deformity was suspected at any level from T6 to L4, six points were placed on the vertebral body at the anterior, mid, and posterior margins of the superior and inferior end plates. The vertebral anterior (A), mid (M), and posterior (P) heights were calculated and compared with one another. An experienced clinician (the senior author) then visually inspected the scans, noted vertebral deformities, and reviewed the technologist’s point placement. If point placement was Journal of Clinical Densitometry
Jacobs-Kosmin et al. questioned, the technologist was asked to reanalyze the image and place new markers. Deformities of vertebrae were classified as being of the anterior wedge or mid-wedge type based on reduction in A/P or M/P height ratios, respectively. Furthermore, a crush deformity was noted when visual assessment revealed an overall decrease in A, M, and P vertebral heights. Using categories described by Genant et al. for vertebral height loss on X-ray (22), wedge deformities were graded as mild, moderate, or severe based on the degree of reduction of A/P or M/P ratios. Reduction by 20 to 25% was categorized as mild, 26 to 40% as moderate, and greater than 40% as severe. All crush deformities were interpreted as severe.
Clinical Variable Assessment Every patient who receives a DXA scan at the center completes a standard questionnaire. From these, patients’ demographic data were compiled and standardized. Data included patient age, gender, race, postmenopausal state, and history of any type of fracture after childhood. Height and weight, measured at the time of DXA, were used to calculate BMI. The use of corticosteroids, anticonvulsants, estrogen replacement therapy (ERT), calcium (±vitamin D), calcitonin, raloxifene, a bisphosphonate, and teriparatide were noted.
Statistical Methods Statistical analysis was performed using SPSS, version 11. In comparisons involving clinical characteristics, a t-test was used for continuous variables and a chi-square (χ2) test was used for dichotomous variables. One-way analysis of variance and post hoc analysis with the Bonferroni correction was performed to determine the relationship between T-score and deformities. Differences were considered significant when the p-value was less than 0.05.
Results Patient Characteristics Vertebral deformities were identified in 91 (21%) of 432 total patients. Demographic and clinical characteristics of the 91 patients with and 341 patients without vertebral deformities were compared (Table 1). The mean age of patients with deformities was greater by 6 yr, and these patients were more likely to report a history of fracture after childhood. Although 37 patients reported past vertebral fractures, six had no vertebral deformity detected by VFA. Patients with deformity were less likely to report alcohol use. No significant differences were detected in gender, race, height loss, BMI, postmenopausal state, steroid use, anticonvulsant use or smoking between the groups. We examined the osteoporosis treatment use among all patients (Table 2). ERT use was significantly more common among patients without vertebral deformity, while bisphosphonate and calcitonin use were more common among patients with vertebral deformities. There was no significant difference in calcium, raloxifene, or teriparatide use between groups. Of the 91 patients with vertebral deformities, we identified 63 (69%) as having at least one moderate or severe deformity. We compared the clinical characteristics and osteoporosis Volume 8, 2005
Vertebral Deformities Identified by VFA
269
Table 1 Characteristics of Patients With and Without Vertebral Deformities Patients with vertebral deformity (n = 91) Gender Female: n (%) Male: n (%) Race (self-reported) Caucasian: n (%) Black: n (%) Asian: n (%) Mean age (yr) ± SD Mean height loss (in.) ± SD Mean BMI (kg/m2) ± SD Postmenopausal women: n (%) Any fracture after childhood: n (%) Steroid use, ever: n (%) Anticonvulsant use, ever: n (%) Smoker, ever: n (%) Alcohol use, ever: n (%)
Patients without vertebral deformity (n = 341)
p-Value
83 (91.2) 8 (8.8)
312 (91.5) 29 (8.5)
0.986
85 (93.4) 6 (6.6) 0 (0) 75.4 ± 8.6 2.1 ± 1.5 24.2 ± 5.3 82 (98.8) 64 (70.3) 24 (26.4) 5 (5.5) 27 (29.7) 27 (29.7)
309 (90.6) 28 (8.21) 4 (1.2) 69 ± 9.8 1.5 ± 3.8 24.4 ± 4.3 303 (97.1) 151 (44.3) 100 (29.3) 16 (4.7) 96 (28.2) 140 (41.1)
0.502
<0.001 0.127 0.597 0.387 <0.001 0.549 0.733 0.906 0.016
Table 2 Osteoporosis Treatment Use Among Patients With and Without Vertebral Deformities Patients with vertebral deformity (n = 91) Estrogen replacement therapy: n (% women) Calcium (±Vitamin D): n (%) Bisphosphonate: n (%) Calcitonin: n (%) Raloxifene: n (%) Teriparatide: n (%)
35 (42.2) 84 (92.3) 72 (79.1) 28 (30.8) 17 (18.7) 3 (3.3)
treatment use of this group to patients without deformity. Similar to the comparison using all 91 patients with vertebral deformities, the 63 patients were older (mean age 76.7 vs 69.4 yr, p < 0.001), more likely to report a history of fracture (79.4 vs 44.2%, p < 0.001) and less likely to report alcohol use (23.8 vs 41%, p = 0.008) than patients without deformity. No other significant differences in clinical characteristics were detected. When comparing osteoporosis treatment use, ERT was less common among patients with deformity (41.1 vs 55.9%, p = 0.035). Conversely, use of a bisphosphonate (84.1 vs 57.8%, p < 0.001), calcitonin (38.1 vs 19.4%, p = 0.001), as well as teriparatide (4.8 vs 0.9%, p = 0.019) were reported more often by the patients with vertebral deformities. Calcium and raloxifene use were similar in both groups.
Vertebral Deformities A total of 158 vertebral deformities were identified. Deformities occurred most commonly at the L1 level (Table 3). Forty-eight deformities were classified as mild, 72 were moderate, and 38 were severe. The rate of vertebral deformity in women Journal of Clinical Densitometry
Patients without vertebral deformity (n = 341)
p-value
174 (55.8) 294 (86.2) 197 (57.8) 66 (19.4) 66 (19.4) 3 (0.9)
0.04 0.117 <0.001 0.019 0.885 0.377
and men was similar. Fifty-three patients (58.2%) had only 1 vertebral deformity, and of these, 28 (52.8%) were classified as moderate or severe. Among the 38 patients (41.8%) with two or more vertebral deformities, 31 patients (81.6%) had at least two deformities present within three adjacent vertebral levels.
Subgroup Analyses of Patient Characteristics and Vertebral Deformities We examined the relationship between severity and number of vertebral deformities with clinical variables. Among patients with vertebral deformities (including mild, moderate, and severe deformities), a history of fracture was significantly associated with increased severity of deformity (p = 0.044). Significant associations between severity and number of vertebral deformities and all other clinical characteristics were not detected.
T-Scores T-scores of patients with and without vertebral deformities were compared (Table 4). Patients with vertebral deformities Volume 8, 2005
270
Jacobs-Kosmin et al.
Table 3 Distribution of Vertebral Deformities Along the Spine Spinal level
Total number of deformities
T6
6
T7
20
T8
23
T9
13
T10
12
T11
14
T12
18
L1
27
L2
12
L3
10
L4
3
Classified by severity Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe
1 3 2 4 10 6 11 8 4 7 5 1 4 4 4 5 5 4 3 10 5 6 13 8 4 8 0 2 5 3 1 1 1
had lower T-scores at the femoral neck and total hip than patients without deformity; however, no difference was found for the lumbar spine. We performed several additional analyses of patient Tscores. First, T-scores of the 63 patients with at least 1 moderate or severe deformity were compared to patients without deformity. The patients with vertebral deformities had significantly lower T-scores at the femoral neck (mean T-score: –2.5 vs –1.7, p < 0.001) and total hip (–1.9 vs –1.1, p < 0.001). Once again, no significant difference was detected at the lumbar spine (–1.7 vs –1.3, p = 0.115). Next, we compared T-scores of the 63-patient subgroup to patients without deformity using a Caucasian database for all patients. This cohort included 4 African-American patients with and 28 African-American patients without deformity. All of the African-American patients’ T-scores demonstrated Journal of Clinical Densitometry
Table 4 Bone Mass (Mean T-Scores) of Patients With and Without Vertebral Deformities
Lumbar spine Femoral neck Total hip
Patients with vertebral deformity (n = 91)
Patients without vertebral deformity (n = 341)
p-value
–1.5 (±1.7) –2.3 (±1.1) –1.6 (±1.1)
–1.3 (±1.5) –1.7 (±1.0) –1.1 (±1.1)
0.322 <0.001 <0.001
improvement using a Caucasian database and, overall, patients with vertebral deformities had significantly lower T-scores at the femoral neck (–2.5 vs –1.7, p < 0.001) and total hip (–1.8 vs –1.1, p < 0.001), but not at the lumbar spine (–1.6 vs –1.2, p = 0.099). Finally, we performed an analysis using the 28 patients with only mild vertebral deformities. There were no significant differences between T-scores of the patients with mild deformities to those without deformity at the femoral neck (–1.9 vs –1.7, p = 0.344), total hip (–1.1 vs –1.1, p = 0.845) or lumbar spine (–1.2 vs –1.3, p = 0.746).
Subgroup Analyses of T-Scores and Vertebral Deformities Among patients with mild, moderate, and severe vertebral deformities, greater severity was associated with significantly lower T-scores at the femoral neck (p = 0.046) and total hip (p = 0.005). In contrast, the relationship between severity of vertebral deformities and lower T-scores was not significant for the spine (p = 0.177). Patients with multiple vertebral deformities had significantly lower T-scores than patients with a single deformity at the femoral neck (p = 0.003) and total hip (p < 0.001). As expected, at the femoral neck, T-scores of patients with three deformities were significantly lower than in patients with one deformity (p = 0.007). At the total hip, T-scores of patients with three deformities were lower than patients with one (p < 0.001) or two deformities (p = 0.024). The relationship between Tscores and number of vertebral deformities was not significant for the spine (p = 0.322).
Discussion Vertebral fractures increase patient morbidity and mortality but are often unrecognized. BMD is a strong predictor of fracture risk. As there is great overlap in the BMD of patients who will develop fracture and those who will not, BMD is not entirely predictive of individuals who will develop a fracture (22). Plain radiographs are a reliable method of determining the presence of vertebral fractures, yet, they are not commonly included as part of the osteoporosis evaluation in individual patients. VFA inclusion with DXA scan, a rapid method associated with minimal radiation exposure, would be a convenient Volume 8, 2005
Vertebral Deformities Identified by VFA adjunct in the identification of patients in need of additional evaluation and treatment. Prior studies of vertebral deformities detected by VFA have used qualitative, quantitative, and semiquantitative methods to identify vertebral deformities. These varied approaches are used, also to diagnose vertebral fractures on radiographs, and little standardization exists in the field (23). Therefore, studies of VFA-detected vertebral deformities are limited by the lack of a consistent standard by which to compare. A study by Ferrar et al. found that the visual qualitative assessment of X-ray absorptiometry scans showed good agreement with visual assessment of spinal radiographs in identifying vertebral deformities (11). Using a semiquantitative technique, Rea et al. found good agreement of VFA with X-ray in classifying vertebrae as normal or deformed and good sensitivity of VFA in identifying moderate or severe deformities (10). In both studies, most disagreement between VFA and X-ray resulted from the different classification of mild deformities (10,11) and poor visualization of upper thoracic vertebrae by X-ray absorptiometry (10). Using a visual method to assess vertebral deformities on X-ray absorptiometry scans in Caucasian female volunteers, Greenspan et al. found that 11 to 18.7% of subjects with normal BMD at the spine, hip, and/or femoral neck had asymptomatic vertebral deformities (24). Our study, to our knowledge, is the first to examine the association of vertebral deformities identified by VFA with both clinical characteristics and BMD. As had been found in prior studies using X-rays, we showed that vertebral deformities identified by VFA correlated with older age, history of fracture, and lower bone mass at the hip. In addition, lower T-scores were demonstrated in patients whose vertebral deformities were greater in number and severity. Our finding that patients without deformity reported a higher use of ERT was consistent with the utilization of ERT as a preventative treatment. Similarly, the higher reported use of bisphosphonates (and less so, calcitonin) in patients with deformity was consistent with their use as established osteoporosis treatment. In total, we identified 158 vertebral deformities. Our study population, with mean ages of 75 yr in patients with deformity and 69 yr in patients without deformity, was a clinically relevant cohort with high risk for fracture. Of note, the majority (69.6%) of the vertebral deformities identified in our study were moderate or severe. There have been concerns that minor deformities might be confounded by errors of long-term reproducibility and might not be as clinically significant as deformities of greater severity (25). Our data would seem to support this, as we were unable to demonstrate a difference in T-scores of patients with only mild deformities compared to those without deformity. Moreover, when we excluded patients with only mild deformity from the analysis and compared patients with at least one moderate or severe deformity to patients without deformity, there was no change from our original findings, with the exception of higher teriparatide use in patients with deformity. An earlier study by Jones et al. found that the femoral neck BMD was more strongly associated than the spinal BMD with vertebral deformities on spinal radiographs (26). Similarly, our study found that reduction of bone mass at the total hip and Journal of Clinical Densitometry
271 femoral neck correlated with the presence (as well as number and severity) of vertebral deformities as detected by VFA, whereas bone mass at the spine did not. It is known that BMD of the lumbar spine in patients over 70 yr old might be falsely elevated as a result of discogenic sclerosis, aortic calcification, or apophyseal arthropathy (27). We postulated that reasons such as these could account for our results. Without radiographs including oblique views, we could not completely evaluate the presence of degenerative joint disease in our patients. Additionally, without radiographs, we could not evaluate the entire thoracic spine for the presence of fracture. The relatively small number of patients in the study might have limited its power to detect other anticipated differences between groups. The retrospective study design was also vulnerable to selection bias. Patients included in this study were those for whom insurance had covered the cost of VFA. This selection bias could not be overcome in the interpretation of our study results. However, as consideration was not given to clinical status, we believe that this factor likely did not play an important role in determining the diversity of patients studied. The study population was comprosed of mostly Caucasian, postmenopausal women; therefore, the findings might not apply to other patient populations. The data on patient questionnaires were collected by self-report, and we did not confirm this information with medical records or other sources. The questionnaires assessed for exposure to medicines such as corticosteroids but could not quantify the total exposure. Therefore, the presence of clinically significant steroid use or total tobacco and alcohol exposure could not be obtained. Questions requiring remote recall might not have been answered accurately. Of note, though, several studies have observed the self-report of fractures to be generally reliable (28–30). It has been suggested by groups such as the National Osteoporosis Foundation that patients with vertebral fractures be treated for osteoporosis regardless of whether their BMD meets osteoporosis criteria (31). Following the identification of vertebral fracture by VFA, then, regardless of BMD, therapies for osteoporosis treatment, rather than prevention, could be started. It should be noted, however, that this view remains controversial, and data to confirm efficacy of treatment in these patients have not been established. As suggested by Lenchik et al., the presence of vertebral deformities might prompt further evaluation with standard radiographs to distinguish pathologic fracture from osteoporotic fracture (23). As the need for X-ray confirmation in every case of VFA-detected vertebral deformity would limit the cost–benefit ratio of VFA, it would be prudent to develop guidelines for identifying patients at highrisk for pathologic fracture who require X-ray follow-up. In summary, vertebral deformities detected by VFA, like those on X-ray, correlate with clinically relevant patient characteristics and T-scores. The patients tend to be older by an average of 6 yr, more likely to have sustained a prior bone fracture as an adult by history, and less likely to report the use of alcohol. In contrast, patients with vertebral deformities in our study did not differ from patients without deformity by gender, race, height loss, body mass index, steroid use, anticonvulsant use, or smoking. VFA is a convenient method that minimizes Volume 8, 2005
272 patient exposure to radiation while providing valuable information that could change patient management. Prospective studies using VFA are needed to determine if the information it provides will ultimately affect clinical outcomes.
Acknowledgments We gratefully acknowledge Jane Colvin at the Jefferson Osteoporosis Center for her support with BMD studies.
References 1. Cooper C, Atkinson EJ, O’Fallon WM, Melton LJ III. 1992 The incidence of clinically diagnosed vertebral fractures: a population-based study in Rochester, Minnesota, 1985–89. J Bone Miner Res 7:221–227. 2. Schlaich C, Minne HW, Bruckner T, et al. 1998 Reduced pulmonary function in patients with spinal osteoporotic fractures. Osteoporos Int 8:261–267. 3. Nevitt MC, Ettinger B, Black DM, et al. 1998 The association of radiographically detected vertebral fractures with back pain and function: a prospective study. Ann Intern Med 128:793–800. 4. Kado D, Browner W, Palermo L, Nevitt MC, Genant HK, Cummings SR. 1999 Vertebral fractures and mortality in older women: a prospective study. Arch Intern Med 159:1215–1220. 5. Ensrud KE, Thompson DE, Cauley JA, et al. 2000 Prevalent vertebral deformities predict mortality and hospitalization in older women with low bone mass. J Am Geriatrics Soc 48:241–249. 6. Black DM, Arden NK, Palermo L, Pearson J, Cummings SR. 1999 Prevalent vertebral deformities predict hip fractures and new vertebral deformities but not wrist fractures. J Bone Min Res 14:821–828. 7. Lindsay R, Silverman S, Cooper C, et al. 2001 Risk of new vertebral fracture in the year following a fracture. JAMA 285:320–323. 8. Klotzbuecher CM, Ross PD, Landsman PB, Abott III TA, Berger M. 2000 Patients with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. J Bone Min Res 15:721–727. 9. Ross PD, Davis JW, Epstein RS, Wasnich RD. 1991 Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 114:919–923. 10. Rea JA, Li J, Blake GM, Steiger P, Genant HK, Fogelman I. 2000 Visual assessment of vertebral deformity by X-ray absorptiometry: a highly predictive method to exclude vertebral deformity. Osteoporos Int 11(8):660–668 11. Ferrar L, Jiang G, Eastell R, Peel NFA. 2003 Visual identification of vertebral fractures in osteoporosis using morphometric X-ray absorptiometry. J Bone Miner Res 18:933–938. 12. Marshall D, Johnell O, Wedel H. 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. Br Med J 312:1254–1259. 13. Ross PD, Genant HK, Davis JW, Miller PD, Wasnich RD.1993 Predicting vertebral fracture incidence from prevalent fractures and bone density among non-black, osteoporotic women. Osteoporos Int 3:120–126.
Journal of Clinical Densitometry
Jacobs-Kosmin et al. 14. van der Klift M, de Laet CE, McCloskey EV, Hofman A, Pols HA. 2002 The incidence of vertebral fractures in men and women: the Rotterdam Study. J Bone Miner Res 17:1051–1056. 15. Lunt M, O’Neill TW, Felsenberg D, et al. 2003 Characteristics of a prevalent vertebral deformity predict subsequent vertebral fracture: results form the European prospective osteoporosis study (EPOS). Bone 33:505–513. 16. Davis JW, Grove JS, Washich RD, Ross PD. 1999 Spatial relationships between prevalent and incident spine fractures. Bone 24:262–264. 17. Burger H, de Laet CE, van Daele PL, et al. 1998 Risk factors for increased bone loss in an elderly population: the Rotterdam Study. Am J Epidemiol 147:871–879. 18. Kroger H, Tuppurainen M, Honkanen R, Alhava E, Saarikoski S. 1994 Bone mineral density and risk factors for osteoporosis—a population-based study of 1600 perimenopausal women. Calcif Tissue Int 55:1–7. 19. Steinbuch M, Youket TE, Cohen S. 2004 Oral glucocorticoid use is associated with an increased risk of fracture. Osteoporos Int 15:323–328. 20. van Staa TP, Leufkens HG, Cooper C. 2002 Utility of medical and drug history in fracture risk prediction among men and women. Bone 31:508–514. 21. Melton LJ III. 1997 Epidemiology of spinal osteoporosis. Spine 22:2S–11S. 22. Genant HK, Wu CY, Van Kuijk C, Nevitt MC. 1993 Vertebral fracture assessment using a semiquantitative technique. J Bone Min Res 8:1137–1148. 23. Lenchik L, Rogers L, Delmas PD, Genant HK. 2004 Diagnosis of osteoporotic vertebral fractures: importance of recognition and description by radiologists. Am J Roentgenol 183: 949–958. 24. Greenspan SL, vonStetten E, Emond SK, Jones L, Parker RA. 2001 Instant vertebral assessment: a noninvasive dual X-ray absorptiometry technique to avoid misclassification and clinical mismanagement of osteoporosis. J Clin Densitom 4:373–380. 25. McCloskey EV, Spector TD, Eyres KS, et al. 1993 The assessment of vertebral deformity: a method for use in population studies and clinical trials. Osteoporos Int 3:138–147. 26. Jones G, White C, Nguyen T, Sambrook PN, Kelly PJ, Eisman JA. 1996 Prevalent vertebral deformities: relationship to bone mineral density and spinal osteophytosis in elderly men and women. Osteoporos Int 6:233–239. 27. Faulkner KG. 1998 Bone densitometry: choosing the proper skeletal site to measure. J Clin Densitom 1:279–285. 28. Nevitt MC, Cummings SR, Browner WS, et al. 1992 The accuracy of self-report of fractures in elderly women: evidence from a prospective study. Am J Epidemiol 135:490–499. 29. Honkanen K, Honkanen R, Heikkinen L, Kroger H, Saarikoski S. 1999 Validity of self-reports of fractures in perimenopausal women. Am J Epidemiol 150:511–516. 30. Ismail AA, O’Neill TW, Cockerill JD, et al. 2000 Validity of self-report of fractures: results from a prospective study in men and women across Europe. Osteoporos Int 11:248–254. 31. National Osteoporosis Foundation. 2003 Physicians Guide to Prevention and Treatment of Osteoporosis. Available at: http:// www.nof.org/physguide/index.htm
Volume 8, 2005
Original Article
Vertebral Deformities Identified by Vertebral Fracture Assessment Associations With Clinical Characteristics and Bone Mineral Density
Dana Jacobs-Kosmin,1 Nora Sandorfi,*,1 Heather Murray,2 and John L. Abruzzo1 1Division
of Rheumatology and 2Division of Pulmonary Medicine, Department of Medicine, Thomas Jefferson University, Philadelphia, PA 19107
Abstract Whether vertebral fractures identified on radiographs are painful or not, they are associated with increased morbidity and mortality. Vertebral fractures on X-rays correlate with low bone mineral density (BMD) at the spine and hip in addition to several clinical characteristics. Evidence suggests that vertebral deformities detected by X-ray and by vertebral fracture assessment (VFA) show good agreement. We examined the relationship between VFA-detected vertebral deformities and patient characteristics as well as BMD by analyzing the records of 432 patients who had undergone dual-energy X-ray absorptiometry (DXA) scans with VFA. Patients’ demographic data and T-scores were obtained from patient questionnaires and DXA scans. We categorized vertebral deformities by type and severity. Patients with vertebral deformities were significantly older and more likely to report a history of fracture after childhood. Significantly more estrogen use was reported in patients without deformity. Those with deformities had significantly lower T-scores at the femoral neck and total hip but not at the spine. Increased severity and number of deformities correlated with lower T-scores at the total hip and femoral neck but not the spine. In conclusion, vertebral deformities detected by VFA, like those on X-ray, correlate with both clinical characteristics and reduced bone mass at the hip. These relationships, in addition to rapid performance, convenience, and minimal radiation exposure, indicate VFA-detected vertebral deformities are a valuable adjunct in identifying patients in need of additional evaluation and treatment. Key Words: Vertebral fracture; vertebral fracture assessment; osteoporosis.
Introduction
Additionally, as the number and severity of vertebral fractures increase, so does the risk for new fracture (6,8). Vertebral fractures can be identified on conventional lateral radiographs or by vertebral fracture assessment (VFA) of X-ray absorptiometry scans. VFA has shown good agreement with conventional radiographs in identifying vertebral deformities (10,11). Studies have shown that vertebral fractures on conventional X-rays correlate with low BMD at the spine and hip (12–14). The relative risk of a new fracture is greater within three adjacent vertebrae of an existing deformity than the risk in more distant vertebrae (15,16). Finally, female gender, postmenopausal state, advanced age, Caucasian or Asian race, low body mass index (BMI), smoking, and the use of certain medications,
Many vertebral fractures remain unrecognized by clinicians and, therefore, undiagnosed (1). Painful or not, vertebral fractures are associated with diminished pulmonary function, difficulty with activities of daily living, and increased mortality (2–5). The presence of a vertebral fracture is a predictor for the occurrence of new osteoporotic spine and hip fractures, independent of age and bone mineral density (BMD) (6–9).
Received 11/16/04; Revised 02/22/05; Accepted 03/10/05. *Address correspondence to: Nora Sandorfi, 1015 Walnut St., Suite 613, Philadelphia, PA 19107. E-mail: [email protected]
267
268 such as corticosteroids and anticonvulsants, are risk factors for low BMD and osteoporotic fractures on X-rays (13,17–21). The association of these clinical factors with vertebral deformities identified by VFA has not been investigated. The aim of this study was to examine whether there is a relationship between VFA-detected vertebral deformities and patients’ clinical characteristics. We also studied the correlation between vertebral deformities and BMD.
Methods Patient Population and Study Design After obtaining Institutional Review Board (IRB) approval, we searched records at our tertiary osteoporosis center for all dual-energy X-ray absorptiometry (DXA) scans done from February 2002 (the date VFA first became available at our center) to March 2004. Four hundred thirty-six patient charts were identified in which VFA was included. Patients were included if at least 10 out of 11 vertebrae from T6 to L4 could be sufficiently visualized for vertebral deformity assessment. Four patients were excluded from the analysis. All four patients had multiple obscured vertebrae on the scan for reasons including spinal deformity (in two patients), diaphragmatic shadow (in one patient), and a Greenfield filter (in one patient). Patients were divided into those with and without vertebral deformity for the comparisons of clinical characteristics and T-scores. In addition, we conducted planned subgroup analyses to explore the relationships between severity and number of deformities with clinical factors and T-scores.
Performing DXA and VFA Dual-energy X-ray absorptiometry scans had been carried out by a single, experienced X-ray technologist dedicated to DXA testing. A Hologic Delphi C densitometer was used to measure the BMD of the total hip, femoral neck, and lumbar spine (L1–L4). T-scores were calculated for the hip using the NHANES III reference population database and for the spine using the Hologic reference population database. T-scores for African-American patients were calculated using a race-adjusted database. The World Health Organization criteria for osteopenia (T-score between –1 and –2.5) and osteoporosis (T-score less than or equal to –2.5) were used for categorization. The VFA had been performed using a semiquantitative technique combining visual and morphometric assessment. First, visual assessment was used to identify possible deformities. Next, morphometric assessment was used to confirm the presence and grade of deformities. Using DXA system software, scans were visually reviewed by the technologist for vertebral deformity. If vertebral deformity was suspected at any level from T6 to L4, six points were placed on the vertebral body at the anterior, mid, and posterior margins of the superior and inferior end plates. The vertebral anterior (A), mid (M), and posterior (P) heights were calculated and compared with one another. An experienced clinician (the senior author) then visually inspected the scans, noted vertebral deformities, and reviewed the technologist’s point placement. If point placement was Journal of Clinical Densitometry
Jacobs-Kosmin et al. questioned, the technologist was asked to reanalyze the image and place new markers. Deformities of vertebrae were classified as being of the anterior wedge or mid-wedge type based on reduction in A/P or M/P height ratios, respectively. Furthermore, a crush deformity was noted when visual assessment revealed an overall decrease in A, M, and P vertebral heights. Using categories described by Genant et al. for vertebral height loss on X-ray (22), wedge deformities were graded as mild, moderate, or severe based on the degree of reduction of A/P or M/P ratios. Reduction by 20 to 25% was categorized as mild, 26 to 40% as moderate, and greater than 40% as severe. All crush deformities were interpreted as severe.
Clinical Variable Assessment Every patient who receives a DXA scan at the center completes a standard questionnaire. From these, patients’ demographic data were compiled and standardized. Data included patient age, gender, race, postmenopausal state, and history of any type of fracture after childhood. Height and weight, measured at the time of DXA, were used to calculate BMI. The use of corticosteroids, anticonvulsants, estrogen replacement therapy (ERT), calcium (±vitamin D), calcitonin, raloxifene, a bisphosphonate, and teriparatide were noted.
Statistical Methods Statistical analysis was performed using SPSS, version 11. In comparisons involving clinical characteristics, a t-test was used for continuous variables and a chi-square (χ2) test was used for dichotomous variables. One-way analysis of variance and post hoc analysis with the Bonferroni correction was performed to determine the relationship between T-score and deformities. Differences were considered significant when the p-value was less than 0.05.
Results Patient Characteristics Vertebral deformities were identified in 91 (21%) of 432 total patients. Demographic and clinical characteristics of the 91 patients with and 341 patients without vertebral deformities were compared (Table 1). The mean age of patients with deformities was greater by 6 yr, and these patients were more likely to report a history of fracture after childhood. Although 37 patients reported past vertebral fractures, six had no vertebral deformity detected by VFA. Patients with deformity were less likely to report alcohol use. No significant differences were detected in gender, race, height loss, BMI, postmenopausal state, steroid use, anticonvulsant use or smoking between the groups. We examined the osteoporosis treatment use among all patients (Table 2). ERT use was significantly more common among patients without vertebral deformity, while bisphosphonate and calcitonin use were more common among patients with vertebral deformities. There was no significant difference in calcium, raloxifene, or teriparatide use between groups. Of the 91 patients with vertebral deformities, we identified 63 (69%) as having at least one moderate or severe deformity. We compared the clinical characteristics and osteoporosis Volume 8, 2005
Vertebral Deformities Identified by VFA
269
Table 1 Characteristics of Patients With and Without Vertebral Deformities Patients with vertebral deformity (n = 91) Gender Female: n (%) Male: n (%) Race (self-reported) Caucasian: n (%) Black: n (%) Asian: n (%) Mean age (yr) ± SD Mean height loss (in.) ± SD Mean BMI (kg/m2) ± SD Postmenopausal women: n (%) Any fracture after childhood: n (%) Steroid use, ever: n (%) Anticonvulsant use, ever: n (%) Smoker, ever: n (%) Alcohol use, ever: n (%)
Patients without vertebral deformity (n = 341)
p-Value
83 (91.2) 8 (8.8)
312 (91.5) 29 (8.5)
0.986
85 (93.4) 6 (6.6) 0 (0) 75.4 ± 8.6 2.1 ± 1.5 24.2 ± 5.3 82 (98.8) 64 (70.3) 24 (26.4) 5 (5.5) 27 (29.7) 27 (29.7)
309 (90.6) 28 (8.21) 4 (1.2) 69 ± 9.8 1.5 ± 3.8 24.4 ± 4.3 303 (97.1) 151 (44.3) 100 (29.3) 16 (4.7) 96 (28.2) 140 (41.1)
0.502
<0.001 0.127 0.597 0.387 <0.001 0.549 0.733 0.906 0.016
Table 2 Osteoporosis Treatment Use Among Patients With and Without Vertebral Deformities Patients with vertebral deformity (n = 91) Estrogen replacement therapy: n (% women) Calcium (±Vitamin D): n (%) Bisphosphonate: n (%) Calcitonin: n (%) Raloxifene: n (%) Teriparatide: n (%)
35 (42.2) 84 (92.3) 72 (79.1) 28 (30.8) 17 (18.7) 3 (3.3)
treatment use of this group to patients without deformity. Similar to the comparison using all 91 patients with vertebral deformities, the 63 patients were older (mean age 76.7 vs 69.4 yr, p < 0.001), more likely to report a history of fracture (79.4 vs 44.2%, p < 0.001) and less likely to report alcohol use (23.8 vs 41%, p = 0.008) than patients without deformity. No other significant differences in clinical characteristics were detected. When comparing osteoporosis treatment use, ERT was less common among patients with deformity (41.1 vs 55.9%, p = 0.035). Conversely, use of a bisphosphonate (84.1 vs 57.8%, p < 0.001), calcitonin (38.1 vs 19.4%, p = 0.001), as well as teriparatide (4.8 vs 0.9%, p = 0.019) were reported more often by the patients with vertebral deformities. Calcium and raloxifene use were similar in both groups.
Vertebral Deformities A total of 158 vertebral deformities were identified. Deformities occurred most commonly at the L1 level (Table 3). Forty-eight deformities were classified as mild, 72 were moderate, and 38 were severe. The rate of vertebral deformity in women Journal of Clinical Densitometry
Patients without vertebral deformity (n = 341)
p-value
174 (55.8) 294 (86.2) 197 (57.8) 66 (19.4) 66 (19.4) 3 (0.9)
0.04 0.117 <0.001 0.019 0.885 0.377
and men was similar. Fifty-three patients (58.2%) had only 1 vertebral deformity, and of these, 28 (52.8%) were classified as moderate or severe. Among the 38 patients (41.8%) with two or more vertebral deformities, 31 patients (81.6%) had at least two deformities present within three adjacent vertebral levels.
Subgroup Analyses of Patient Characteristics and Vertebral Deformities We examined the relationship between severity and number of vertebral deformities with clinical variables. Among patients with vertebral deformities (including mild, moderate, and severe deformities), a history of fracture was significantly associated with increased severity of deformity (p = 0.044). Significant associations between severity and number of vertebral deformities and all other clinical characteristics were not detected.
T-Scores T-scores of patients with and without vertebral deformities were compared (Table 4). Patients with vertebral deformities Volume 8, 2005
270
Jacobs-Kosmin et al.
Table 3 Distribution of Vertebral Deformities Along the Spine Spinal level
Total number of deformities
T6
6
T7
20
T8
23
T9
13
T10
12
T11
14
T12
18
L1
27
L2
12
L3
10
L4
3
Classified by severity Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe Mild Moderate Severe
1 3 2 4 10 6 11 8 4 7 5 1 4 4 4 5 5 4 3 10 5 6 13 8 4 8 0 2 5 3 1 1 1
had lower T-scores at the femoral neck and total hip than patients without deformity; however, no difference was found for the lumbar spine. We performed several additional analyses of patient Tscores. First, T-scores of the 63 patients with at least 1 moderate or severe deformity were compared to patients without deformity. The patients with vertebral deformities had significantly lower T-scores at the femoral neck (mean T-score: –2.5 vs –1.7, p < 0.001) and total hip (–1.9 vs –1.1, p < 0.001). Once again, no significant difference was detected at the lumbar spine (–1.7 vs –1.3, p = 0.115). Next, we compared T-scores of the 63-patient subgroup to patients without deformity using a Caucasian database for all patients. This cohort included 4 African-American patients with and 28 African-American patients without deformity. All of the African-American patients’ T-scores demonstrated Journal of Clinical Densitometry
Table 4 Bone Mass (Mean T-Scores) of Patients With and Without Vertebral Deformities
Lumbar spine Femoral neck Total hip
Patients with vertebral deformity (n = 91)
Patients without vertebral deformity (n = 341)
p-value
–1.5 (±1.7) –2.3 (±1.1) –1.6 (±1.1)
–1.3 (±1.5) –1.7 (±1.0) –1.1 (±1.1)
0.322 <0.001 <0.001
improvement using a Caucasian database and, overall, patients with vertebral deformities had significantly lower T-scores at the femoral neck (–2.5 vs –1.7, p < 0.001) and total hip (–1.8 vs –1.1, p < 0.001), but not at the lumbar spine (–1.6 vs –1.2, p = 0.099). Finally, we performed an analysis using the 28 patients with only mild vertebral deformities. There were no significant differences between T-scores of the patients with mild deformities to those without deformity at the femoral neck (–1.9 vs –1.7, p = 0.344), total hip (–1.1 vs –1.1, p = 0.845) or lumbar spine (–1.2 vs –1.3, p = 0.746).
Subgroup Analyses of T-Scores and Vertebral Deformities Among patients with mild, moderate, and severe vertebral deformities, greater severity was associated with significantly lower T-scores at the femoral neck (p = 0.046) and total hip (p = 0.005). In contrast, the relationship between severity of vertebral deformities and lower T-scores was not significant for the spine (p = 0.177). Patients with multiple vertebral deformities had significantly lower T-scores than patients with a single deformity at the femoral neck (p = 0.003) and total hip (p < 0.001). As expected, at the femoral neck, T-scores of patients with three deformities were significantly lower than in patients with one deformity (p = 0.007). At the total hip, T-scores of patients with three deformities were lower than patients with one (p < 0.001) or two deformities (p = 0.024). The relationship between Tscores and number of vertebral deformities was not significant for the spine (p = 0.322).
Discussion Vertebral fractures increase patient morbidity and mortality but are often unrecognized. BMD is a strong predictor of fracture risk. As there is great overlap in the BMD of patients who will develop fracture and those who will not, BMD is not entirely predictive of individuals who will develop a fracture (22). Plain radiographs are a reliable method of determining the presence of vertebral fractures, yet, they are not commonly included as part of the osteoporosis evaluation in individual patients. VFA inclusion with DXA scan, a rapid method associated with minimal radiation exposure, would be a convenient Volume 8, 2005
Vertebral Deformities Identified by VFA adjunct in the identification of patients in need of additional evaluation and treatment. Prior studies of vertebral deformities detected by VFA have used qualitative, quantitative, and semiquantitative methods to identify vertebral deformities. These varied approaches are used, also to diagnose vertebral fractures on radiographs, and little standardization exists in the field (23). Therefore, studies of VFA-detected vertebral deformities are limited by the lack of a consistent standard by which to compare. A study by Ferrar et al. found that the visual qualitative assessment of X-ray absorptiometry scans showed good agreement with visual assessment of spinal radiographs in identifying vertebral deformities (11). Using a semiquantitative technique, Rea et al. found good agreement of VFA with X-ray in classifying vertebrae as normal or deformed and good sensitivity of VFA in identifying moderate or severe deformities (10). In both studies, most disagreement between VFA and X-ray resulted from the different classification of mild deformities (10,11) and poor visualization of upper thoracic vertebrae by X-ray absorptiometry (10). Using a visual method to assess vertebral deformities on X-ray absorptiometry scans in Caucasian female volunteers, Greenspan et al. found that 11 to 18.7% of subjects with normal BMD at the spine, hip, and/or femoral neck had asymptomatic vertebral deformities (24). Our study, to our knowledge, is the first to examine the association of vertebral deformities identified by VFA with both clinical characteristics and BMD. As had been found in prior studies using X-rays, we showed that vertebral deformities identified by VFA correlated with older age, history of fracture, and lower bone mass at the hip. In addition, lower T-scores were demonstrated in patients whose vertebral deformities were greater in number and severity. Our finding that patients without deformity reported a higher use of ERT was consistent with the utilization of ERT as a preventative treatment. Similarly, the higher reported use of bisphosphonates (and less so, calcitonin) in patients with deformity was consistent with their use as established osteoporosis treatment. In total, we identified 158 vertebral deformities. Our study population, with mean ages of 75 yr in patients with deformity and 69 yr in patients without deformity, was a clinically relevant cohort with high risk for fracture. Of note, the majority (69.6%) of the vertebral deformities identified in our study were moderate or severe. There have been concerns that minor deformities might be confounded by errors of long-term reproducibility and might not be as clinically significant as deformities of greater severity (25). Our data would seem to support this, as we were unable to demonstrate a difference in T-scores of patients with only mild deformities compared to those without deformity. Moreover, when we excluded patients with only mild deformity from the analysis and compared patients with at least one moderate or severe deformity to patients without deformity, there was no change from our original findings, with the exception of higher teriparatide use in patients with deformity. An earlier study by Jones et al. found that the femoral neck BMD was more strongly associated than the spinal BMD with vertebral deformities on spinal radiographs (26). Similarly, our study found that reduction of bone mass at the total hip and Journal of Clinical Densitometry
271 femoral neck correlated with the presence (as well as number and severity) of vertebral deformities as detected by VFA, whereas bone mass at the spine did not. It is known that BMD of the lumbar spine in patients over 70 yr old might be falsely elevated as a result of discogenic sclerosis, aortic calcification, or apophyseal arthropathy (27). We postulated that reasons such as these could account for our results. Without radiographs including oblique views, we could not completely evaluate the presence of degenerative joint disease in our patients. Additionally, without radiographs, we could not evaluate the entire thoracic spine for the presence of fracture. The relatively small number of patients in the study might have limited its power to detect other anticipated differences between groups. The retrospective study design was also vulnerable to selection bias. Patients included in this study were those for whom insurance had covered the cost of VFA. This selection bias could not be overcome in the interpretation of our study results. However, as consideration was not given to clinical status, we believe that this factor likely did not play an important role in determining the diversity of patients studied. The study population was comprosed of mostly Caucasian, postmenopausal women; therefore, the findings might not apply to other patient populations. The data on patient questionnaires were collected by self-report, and we did not confirm this information with medical records or other sources. The questionnaires assessed for exposure to medicines such as corticosteroids but could not quantify the total exposure. Therefore, the presence of clinically significant steroid use or total tobacco and alcohol exposure could not be obtained. Questions requiring remote recall might not have been answered accurately. Of note, though, several studies have observed the self-report of fractures to be generally reliable (28–30). It has been suggested by groups such as the National Osteoporosis Foundation that patients with vertebral fractures be treated for osteoporosis regardless of whether their BMD meets osteoporosis criteria (31). Following the identification of vertebral fracture by VFA, then, regardless of BMD, therapies for osteoporosis treatment, rather than prevention, could be started. It should be noted, however, that this view remains controversial, and data to confirm efficacy of treatment in these patients have not been established. As suggested by Lenchik et al., the presence of vertebral deformities might prompt further evaluation with standard radiographs to distinguish pathologic fracture from osteoporotic fracture (23). As the need for X-ray confirmation in every case of VFA-detected vertebral deformity would limit the cost–benefit ratio of VFA, it would be prudent to develop guidelines for identifying patients at highrisk for pathologic fracture who require X-ray follow-up. In summary, vertebral deformities detected by VFA, like those on X-ray, correlate with clinically relevant patient characteristics and T-scores. The patients tend to be older by an average of 6 yr, more likely to have sustained a prior bone fracture as an adult by history, and less likely to report the use of alcohol. In contrast, patients with vertebral deformities in our study did not differ from patients without deformity by gender, race, height loss, body mass index, steroid use, anticonvulsant use, or smoking. VFA is a convenient method that minimizes Volume 8, 2005
272 patient exposure to radiation while providing valuable information that could change patient management. Prospective studies using VFA are needed to determine if the information it provides will ultimately affect clinical outcomes.
Acknowledgments We gratefully acknowledge Jane Colvin at the Jefferson Osteoporosis Center for her support with BMD studies.
References 1. Cooper C, Atkinson EJ, O’Fallon WM, Melton LJ III. 1992 The incidence of clinically diagnosed vertebral fractures: a population-based study in Rochester, Minnesota, 1985–89. J Bone Miner Res 7:221–227. 2. Schlaich C, Minne HW, Bruckner T, et al. 1998 Reduced pulmonary function in patients with spinal osteoporotic fractures. Osteoporos Int 8:261–267. 3. Nevitt MC, Ettinger B, Black DM, et al. 1998 The association of radiographically detected vertebral fractures with back pain and function: a prospective study. Ann Intern Med 128:793–800. 4. Kado D, Browner W, Palermo L, Nevitt MC, Genant HK, Cummings SR. 1999 Vertebral fractures and mortality in older women: a prospective study. Arch Intern Med 159:1215–1220. 5. Ensrud KE, Thompson DE, Cauley JA, et al. 2000 Prevalent vertebral deformities predict mortality and hospitalization in older women with low bone mass. J Am Geriatrics Soc 48:241–249. 6. Black DM, Arden NK, Palermo L, Pearson J, Cummings SR. 1999 Prevalent vertebral deformities predict hip fractures and new vertebral deformities but not wrist fractures. J Bone Min Res 14:821–828. 7. Lindsay R, Silverman S, Cooper C, et al. 2001 Risk of new vertebral fracture in the year following a fracture. JAMA 285:320–323. 8. Klotzbuecher CM, Ross PD, Landsman PB, Abott III TA, Berger M. 2000 Patients with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. J Bone Min Res 15:721–727. 9. Ross PD, Davis JW, Epstein RS, Wasnich RD. 1991 Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 114:919–923. 10. Rea JA, Li J, Blake GM, Steiger P, Genant HK, Fogelman I. 2000 Visual assessment of vertebral deformity by X-ray absorptiometry: a highly predictive method to exclude vertebral deformity. Osteoporos Int 11(8):660–668 11. Ferrar L, Jiang G, Eastell R, Peel NFA. 2003 Visual identification of vertebral fractures in osteoporosis using morphometric X-ray absorptiometry. J Bone Miner Res 18:933–938. 12. Marshall D, Johnell O, Wedel H. 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. Br Med J 312:1254–1259. 13. Ross PD, Genant HK, Davis JW, Miller PD, Wasnich RD.1993 Predicting vertebral fracture incidence from prevalent fractures and bone density among non-black, osteoporotic women. Osteoporos Int 3:120–126.
Journal of Clinical Densitometry
Jacobs-Kosmin et al. 14. van der Klift M, de Laet CE, McCloskey EV, Hofman A, Pols HA. 2002 The incidence of vertebral fractures in men and women: the Rotterdam Study. J Bone Miner Res 17:1051–1056. 15. Lunt M, O’Neill TW, Felsenberg D, et al. 2003 Characteristics of a prevalent vertebral deformity predict subsequent vertebral fracture: results form the European prospective osteoporosis study (EPOS). Bone 33:505–513. 16. Davis JW, Grove JS, Washich RD, Ross PD. 1999 Spatial relationships between prevalent and incident spine fractures. Bone 24:262–264. 17. Burger H, de Laet CE, van Daele PL, et al. 1998 Risk factors for increased bone loss in an elderly population: the Rotterdam Study. Am J Epidemiol 147:871–879. 18. Kroger H, Tuppurainen M, Honkanen R, Alhava E, Saarikoski S. 1994 Bone mineral density and risk factors for osteoporosis—a population-based study of 1600 perimenopausal women. Calcif Tissue Int 55:1–7. 19. Steinbuch M, Youket TE, Cohen S. 2004 Oral glucocorticoid use is associated with an increased risk of fracture. Osteoporos Int 15:323–328. 20. van Staa TP, Leufkens HG, Cooper C. 2002 Utility of medical and drug history in fracture risk prediction among men and women. Bone 31:508–514. 21. Melton LJ III. 1997 Epidemiology of spinal osteoporosis. Spine 22:2S–11S. 22. Genant HK, Wu CY, Van Kuijk C, Nevitt MC. 1993 Vertebral fracture assessment using a semiquantitative technique. J Bone Min Res 8:1137–1148. 23. Lenchik L, Rogers L, Delmas PD, Genant HK. 2004 Diagnosis of osteoporotic vertebral fractures: importance of recognition and description by radiologists. Am J Roentgenol 183: 949–958. 24. Greenspan SL, vonStetten E, Emond SK, Jones L, Parker RA. 2001 Instant vertebral assessment: a noninvasive dual X-ray absorptiometry technique to avoid misclassification and clinical mismanagement of osteoporosis. J Clin Densitom 4:373–380. 25. McCloskey EV, Spector TD, Eyres KS, et al. 1993 The assessment of vertebral deformity: a method for use in population studies and clinical trials. Osteoporos Int 3:138–147. 26. Jones G, White C, Nguyen T, Sambrook PN, Kelly PJ, Eisman JA. 1996 Prevalent vertebral deformities: relationship to bone mineral density and spinal osteophytosis in elderly men and women. Osteoporos Int 6:233–239. 27. Faulkner KG. 1998 Bone densitometry: choosing the proper skeletal site to measure. J Clin Densitom 1:279–285. 28. Nevitt MC, Cummings SR, Browner WS, et al. 1992 The accuracy of self-report of fractures in elderly women: evidence from a prospective study. Am J Epidemiol 135:490–499. 29. Honkanen K, Honkanen R, Heikkinen L, Kroger H, Saarikoski S. 1999 Validity of self-reports of fractures in perimenopausal women. Am J Epidemiol 150:511–516. 30. Ismail AA, O’Neill TW, Cockerill JD, et al. 2000 Validity of self-report of fractures: results from a prospective study in men and women across Europe. Osteoporos Int 11:248–254. 31. National Osteoporosis Foundation. 2003 Physicians Guide to Prevention and Treatment of Osteoporosis. Available at: http:// www.nof.org/physguide/index.htm
Volume 8, 2005