Abstract
Vertebral fractures are the most common osteoporotic fracture, and patients with prevalent vertebral fractures have a greater risk of future fractures. However, radiographically determined vertebral fractures are not identified as a distinct risk factor in the World Health Organization (WHO) fracture risk assessment tool. The objective of this study was to evaluate and compare potential risk factors including morphometric spine fracture status and the WHO risk factors for predicting 5-yr fracture risk. We hypothesized that spine fracture status provides prognostic information in addition to consideration of the WHO risk factors alone. A randomly selected, population-based community cohort of 2761 noninstitutionalized men and women ≥50 yr of age living within 50 km of one of nine regional centers was enrolled in the Canadian Multicentre Osteoporosis Study (CaMOS), a prospective and longitudinal cohort study following subjects for 5 yr. Prevalent and incident spine fractures were identified from lateral spine radiographs. Incident nonvertebral fragility fractures were determined by an annual, mailed fracture questionnaire with validation, and nonvertebral fragility fracture was defined by investigators as a fracture with minimal trauma. A model considering the WHO risk factors plus spine fracture status provided greater prognostic information regarding future fracture risk than a model considering the WHO risk factors alone. In univariate analyses, age, BMD, and spine fracture status had the highest gradient of risk. A model considering these three risk factors captured almost all of the predictive information provided by a model considering spine fracture status plus the WHO risk factors and provided greater predictive information than a model considering the WHO risk factors alone. The use of spine fracture status along with age and BMD predicted future fracture risk with greater simplicity and higher prognostic accuracy than consideration of the risk factors included in the WHO tool.
Keywords: vertebral fracture, World Health Organization tool, fracture prediction, osteoporosis
INTRODUCTION
The prediction of future fracture risk provides clinicians and patients with information on which to make lifestyle and treatment choices. Recently, a fracture risk assessment tool (FRAX) was developed by the World Health Organization (WHO).(1) The WHO fracture risk assessment tool considers clinical risk factors for future fracture, including age, prior clinical fracture, current smoking, alcohol use, parental history of hip fracture, glucocorticoid use, rheumatoid arthritis, and BMD, to assign a 10-yr absolute fracture risk.(2)
Vertebral fractures are the most common type of fragility fracture occurring in postmenopausal women with osteoporosis.(3–6) Many studies have shown that prevalent vertebral fractures increase the risk of new vertebral and nonvertebral fractures in postmenopausal women.(7–12) Future vertebral fracture risk is positively associated with the number of prevalent vertebral fractures(10,13–15) and the severity of the vertebral deformity.(10,14,15) The number and severity of vertebral fractures is captured in the spinal deformity index (SDI), a summary measure of spine fracture burden that is predictive of future fracture risk.(16,17) Recently, Cauley et al.(18) found that women with a prevalent vertebral fracture at baseline were more than four times more likely to experience an incident vertebral fracture over 15 yr of follow-up compared with women without a prevalent vertebral fracture. Furthermore, Siris et al.(17) showed that, at any particular value for BMD, spine fracture status increased future vertebral or nonvertebral fragility fracture risk by up to 7-fold.
It is acknowledged by the authors of the WHO tool(19) that a prior clinical vertebral fracture is an especially strong risk factor. It is also acknowledged that a fracture detected as a radiographic observation alone (a morphometric vertebral fracture) should count as a previous fracture.(19) However, most of the epidemiology studies from which this tool was developed did not include spine imaging, and therefore spine fracture status information was not available for study or for inclusion in the tool. The objectives of this analysis were to determine whether information regarding morphometric vertebral fracture status at baseline would improve prognosis of future fracture risk when added to the WHO risk factors, and, in addition, whether a more parsimonious model including some of the WHO risk factors plus morphometric vertebral fracture could provide a prognosis of future fracture risk as well as or better than the WHO risk factors.
MATERIALS AND METHODS
Study participants and population
The Canadian Multicentre Osteoporosis Study (CaMOS) is a prospective cohort study following a randomly selected, population-based community cohort of 9423 non-institutionalized men and women ≥25 yr of age living within 50 km of one of nine regional centers. Details of the objectives, purpose, and methodology of the CaMOS study have been reported elsewhere.(20) Briefly, recruitment for the cohort began in February 1996 and ended in September 1997. The study was approved by all regional institutional ethics review boards. All participants provided written informed consent in accordance with the Helsinki Declaration. This analysis was limited to those subjects who were ≥50 yr of age who had completed analyses of spine radiographs at baseline and year 5.
BMD
Lumbar spine (L1–L4) and hip BMD was measured in all subjects by DXA. BMD was with Hologic QDR 1000, 2000, or 4500 densitometers at seven centers and Lunar DPX densitometers at two centers. BMD results were converted to a Hologic standard, using the method described by Genant et al.(21) Each year, a semianthropomorphic spine phantom (Siemens, Munich, Germany) was measured at each site for cross-calibration purposes.(22) BMD results used in this analysis are femoral neck measurements from the baseline assessment. The T-scores for the femoral neck were derived from CaMOS reference data.(23)
Clinical risk factor measurement
Participants completed an extensive interviewer-administered questionnaire to assess for osteoporosis and fracture-related risk factors at baseline. A second intensive interview was conducted 5 yr after enrollment to reassess these risk factors. All clinical risk factors were derived from the baseline interview except for parental history of hip fracture, which was obtained from the year 5 questionnaires. Subject responses were coded to indicate if they were current cigarette smokers, if they had used systemic glucocorticoid therapy for >3 mo without regard to dose, if they had sustained a minimal trauma fracture after 50 yr of age,(24) if their parents had hip fracture, or if they consumed >2 units of alcohol/d. A subject was considered to have rheumatoid arthritis if he or she self-reported a physician diagnosis of rheumatoid arthritis.
Fracture diagnosis
Lateral thoracic and lumbar spine radiographs obtained on all subjects ≥50 yr of age at baseline and after 5 yr were digitized and assessed at a central site (Department of Radiology, University of Alberta, Edmonton, Alberta, Canada). Fracture severity was determined by quantitative morphometric analysis of T4–L4 vertebrae, with each vertebrae being assigned a severity score.(6) Deformities were graded as 0 if the height ratio was ≤3 SD below the mean of respective uninvolved vertebrae by sex; grade 1 for a mild deformity (>3 but ≤4 SD); and grade 2 for a moderate deformity (>4 SD). A new vertebral fracture was recorded if a normal vertebra (grade 0) became deformed (grade ≥ 1). The SDI for a subject was defined as the sum of the individual vertebral deformity scores for each vertebra from T4 to L4.
In this study, a nonvertebral fragility fracture (hip, forearm/wrist, ribs, pelvis, and other) was defined as a fracture with minimal trauma that was confirmed by either radiographic or medical report and a vertebral fracture was a fracture detected by spine radiography. The WHO fracture risk assessment tool defines a prevalent fracture as a previous fracture in adult life occurring spontaneously, or a fracture arising from trauma which, in a healthy individual, would not have resulted in a fracture. The WHO risk fracture assessment tool predicts the risk for hip fractures and of a major osteoporotic fracture (clinical spine, forearm, hip, or shoulder fracture). In our study, the risk of any fragility fracture refers to the risk of a participant experiencing either an incident vertebral fracture detected by spine radiography and/or a nonvertebral fragility fracture by annual questionnaire and confirmed by either radiographic or medical report.
Statistical analysis
A series of logistic regression analyses were performed to determine the importance of spine fracture status and the WHO risk factors for predicting the 5-yr risk of any future vertebral or nonvertebral fragility fracture. Although the WHO fracture risk assessment tool provides a 10-yr fracture risk, it is stated that, in individuals with a low mortality, the 1-yr probability is ~10% of the 10-yr probability. To test the hypothesis that inclusion of spine fracture status along with the WHO risk factors would improve the prediction of future fracture risk, a logistic regression model that included the WHO risk factors only was compared with models including the WHO risk factors plus the following: vertebral fracture (yes/no), vertebral fracture severity (grade 0, 1, or 2), prevalent vertebral fracture number (0, 1, or ≥2), and SDI score (0, 1, 2, or ≥3). To pool the data from the sexes, the models included the interaction effects of risk factors with sex at the 10% level of significance. Because there were no statistically significant interactions, the relationship between all the risk factors and incident fracture risk was statistically consistent between the two sexes. The performance of each model was assessed as the gradient of risk (GR; i.e., the increase in fracture risk per SD [GR/SD]); this assessment was used in the development of the WHO fracture risk tool.(25) GR was also transformed as area under the receiver operating characteristic (ROC) curve (AUC) as detailed elsewhere.(26)
After these models showed that consideration of spine fracture status plus the WHO risk factors improved the prediction of future fracture risk beyond consideration of the WHO risk factors alone, further analyses were conducted to determine the predictive ability of sequential addition of the most important WHO risk factors and spine fracture status. To do this, univariate analysis was used to investigate the association between individual risk factors (age, BMD, prior fragility fractures, spine fracture status, current smoking, alcohol use, parental history of hip fracture, glucocorticoid use, rheumatoid arthritis) and future fracture risk. Spine fracture status was considered in the four different ways described above. The gradient of fracture risk was examined in different models by sequential addition of the most important risk factors determined from the univariate analyses. Five-year absolute fracture risk was estimated using the logistic regression model including the risk factors age, femoral neck T-score, and spine fracture (yes/no). All analyses are reported for pooled data using SAS version 8.2 (SAS Institute, Cary, NC, USA).
RESULTS
Subject characteristics
Of 7753 subjects ≥50 yr of age in the original CaMOS cohort, 4744 had spine radiographs both at baseline and 5 yr, and analyses of vertebral deformities were complete in 2761 (58%). The mean age of the sample population was 64.3 yr for men (n = 776) and 64.4 yr for women (n = 1985). Compared with men, women had significantly lower BMD T-scores, a higher proportion had prior clinical fracture, and fewer used alcohol (Table 1). Whereas a greater proportion of women than men experienced incident non-vertebral fractures, similar proportions experienced new vertebral fractures and total fragility fractures at 5 yr (Table 1).
Table 1.
Demographics of CaMOS Participants ≥50 yr of Age Who Had Assessments at Baseline and 5 yr*
| Women (N = 1985) | Men (N = 776) | Total (N = 2761) | |
|---|---|---|---|
| Age (yr) at baseline | 64.4 ± 0.19 | 64.3 ± 0.30 | 64.3 ± 0.16 |
| Prevalent morphometric vertebral fracture(s) at baseline (% yes) | 20.9 | 19.8 | 20.6 |
| Number of prevalent morphometric vertebral fractures at baseline (%) | |||
| 0 | 79.1 | 80.2 | 79.4 |
| 1 | 14.5 | 14.3 | 14.5 |
| ≥2 | 6.4 | 5.5 | 6.2 |
| Severity of prevalent morphometric vertebral fracture(s) at baseline (%) | |||
| 0 | 79.1 | 80.2 | 79.4 |
| Grade 1 | 12.8 | 13.1 | 12.9 |
| Grade 2 | 8.2 | 6.7 | 7.8 |
| Spinal deformity index at baseline (%)† | |||
| 0 | 79.1 | 80.2 | 79.4 |
| 1 | 10.5 | 10.8 | 10.6 |
| 2 | 5.9 | 5.3 | 5.8 |
| ≥3 | 4.5 | 3.7 | 4.3 |
| Femoral neck BMD T-score at baseline‡ | −1.24 ± 0.02 | −0.91 ± 0.03 | −1.15 ± 0.02 |
| Prior clinical fracture at baseline (% yes)‡ | 15.5 | 6.7 | 13.0 |
| Prior glucocorticoid use at baseline (% yes) | 1.2 | 1.6 | 1.3 |
| Parental history of hip fracture at baseline (% yes) | 11.7 | 9.7 | 11.2 |
| Current smoking at baseline (% yes) | 12.5 | 13.9 | 12.9 |
| Consume >2 units of alcohol/d‡ | 1.4 | 7.1 | 3.0 |
| Rheumatoid arthritis at baseline (% yes) | 6.3 | 4.7 | 5.9 |
| Incident fragility fracture(s) at 5 yr | |||
| Morphometric vertebral fracture(s) [n (%)] | 238 (12.0) | 105 (13.5) | 343 (12.4) |
| Nonvertebral fragility fracture(s) [n (%)]‡ | 170 (8.6) | 30 (3.9) | 200 (7.24) |
| Total fractures [n (%)]§ | 381 (19.2) | 127 (16.4) | 508 (18.4) |
Values are mean ± SE unless otherwise stated.
Sum of maximal morphometric vertebral deformity severity scores in T4–L4 vertebrae.
p < 0.01 between women and men.
Subjects who had either vertebral or nonvertebral fragility fractures.
Comparison of models considering WHO risk factors alone versus WHO risk factors plus spine fracture status
The performance characteristics of the models were expressed as GR/SD change in the risk indicator. The GR for the original WHO risk factors was 1.88 with an AUC of 0.67. Including spine fracture status as determined by four different methods improved the GR and AUC (Table 2). Inclusion of the vertebral fracture (yes/no), vertebral fracture severity (grade 0, 1, or 2), prevalent vertebral fracture number (0, 1, or ≥2), and SDI score (0, 1, 2, or ≥3) in the WHO model increased the GR to 2.08, 2.11, 2.10, and 2.11, respectively.
Table 2.
Comparison of Predictive Ability of the WHO Risk Factors vs. the WHO Risk Factors Plus Spine Fracture Status
| Model | GR/SD (95% CI)* |
|---|---|
| WHO clinical risk factors alone | 1.88 (1.69–2.10) [0.67] |
| WHO clinical risk factors + spine fracture (yes/no) | 2.08 (1.87–2.33) [0.70] |
| WHO clinical risk factors + number of vertebral fractures | 2.11 (1.89–2.36) [0.70] |
| WHO clinical risk factors + severity of vertebral fractures | 2.10 (1.89–2.35) [0.70] |
| WHO clinical risk factors + SDI† | 2.11 (1.90–2.37) [0.70] |
To determine which model offers the greatest ability for predicting the fracture risk, we also conducted a receiver operator characteristic (ROC) analysis, with the overall discriminatory ability assessed by the areas under the ROC curves (AUC) shown in square brackets.
SDI, spinal deformity index; sum of morphometric vertebral deformity severity scores for vertebrae T4–L4.
Univariate analyses for 5-yr risk of new fractures
In univariate analyses, age provided the highest GR, followed by femoral neck BMD T-score and spine fracture status. Prior clinical fracture had the next highest GR, and other risk factors provided relatively lower GR (data not shown). Because the results were similar in men and women, multivariable analyses were performed on the combined set of men and women.
Multivariable analyses for 5-yr risk of new fractures
The performance characteristics of models with sequential addition of the most important risk factors are shown in Table 3, expressed as GR/SD change in the risk indicator. For fracture prediction, a model that included age, BMD T-score, and presence or absence of spine fracture had a GR of 1.99. After these three risk factors were included in the model, the increment in the GR/SD by adding the six additional risk factors described in the WHO risk assessment tool was 0.09.
Table 3.
GR/SD Change in Risk Score (95% CIs) and AUC for Different Models
| Model | Variable | GR/SD* |
|---|---|---|
| 1 | Age | 1.62 (1.46–1.81) [0.63] |
| 2 | Age + FN BMD T-score | 1.75 (1.58–1.96) [0.65] |
| 3 | Age + FN BMD T-score + spine fracture (yes/no) | 1.99 (1.78–2.22) [0.69] |
| 4 | Age + FN BMD T-score + spine fracture (yes/no) + prior clinical fracture | 2.01 (1.81–2.25) [0.69] |
| 5 | Age + FN BMD T-score + spine fracture (yes/no) + prior clinical fracture + parental history of hip fracture | 2.06 (1.85–2.31) [0.70] |
| 6 | Age + FN BMD T-score + spine fracture (yes/no) + prior clinical fracture + parental history of hip fracture + current smoking | 2.06 (1.85–2.31) [0.70] |
| 7 | Age + FN BMD T-score + spine fracture (yes/no) + prior clinical fracture + parental history of hip fracture + current smoking + prior glucocorticoid use | 2.07 (1.85–2.31) [070] |
| 8 | Age + FN BMD T-score + spine fracture (yes/no) + prior clinical fracture + parental history of hip fracture + current smoking + prior glucocorticoid use + rheumatoid arthritis | 2.08 (1.87–2.33) [0.70] |
| 9 | Age + FN BMD T-score + spine fracture (yes/no) + prior clinical fracture + parental history of hip fracture + current smoking + prior glucocorticoid use + rheumatoid arthritis + consume >2 units of alcohol per day | 2.08 (1.87–2.33) [0.70] |
To determine which model offers the greatest ability for predicting the fracture risk, we also conducted a receiver operator characteristic (ROC) analysis, with the overall discriminatory ability assessed by the areas under the ROC curves (AUC) shown in square brackets.
FN BMD, femoral neck BMD.
Absolute risk of fracture based on age, femoral neck T-score, and spine fracture status
The 5-yr absolute risk of incident fragility fracture in the CaMOS population based on age, femoral neck T-score, and presence or absence of spine fracture is shown for women (Table 4) and men (Table 5). Results for total hip and lumbar spine BMD were similar to results for femoral neck BMD (data not shown). The fracture risk increased in both men and women with increasing age, more negative T-score, and presence of spine fracture.
Table 4.
Five-Year Risk of Incident Fragility Fracture in the CaMOS Population of Women Based on Age, Femoral Neck T-Score, and Spine Fracture (No/Yes)
| Femoral neck T-score | Spine fracture | Age (yr)
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| 50 | 55 | 60 | 65 | 70 | 75 | 80 | 85 | ||
| −1 | No | 7.6 | 9.4 | 11.4 | 13.9 | 16.8 | 20.1 | 23.9 | 28.2 |
| Yes | 17.9 | 21.4 | 25.4 | 29.8 | 34.7 | 39.9 | 45.3 | 50.9 | |
| −1.5 | No | 8.6 | 10.6 | 12.9 | 15.6 | 18.7 | 22.4 | 26.4 | 31.0 |
| Yes | 20.0 | 23.8 | 28.0 | 32.7 | 37.8 | 43.2 | 48.7 | 54.3 | |
| −2 | No | 9.8 | 11.9 | 14.5 | 17.4 | 20.9 | 24.8 | 29.2 | 34.0 |
| Yes | 22.2 | 26.3 | 30.9 | 35.8 | 41.1 | 46.5 | 52.1 | 57.6 | |
| −2.5 | No | 11.0 | 13.4 | 16.2 | 19.5 | 23.2 | 27.4 | 32.0 | 37.1 |
| Yes | 24.7 | 29.0 | 33.8 | 39.0 | 44.4 | 49.9 | 55.5 | 60.9 | |
| −3 | No | 12.4 | 15.1 | 18.1 | 21.7 | 25.7 | 30.2 | 35.1 | 40.3 |
| Yes | 27.3 | 31.9 | 36.9 | 42.2 | 47.7 | 53.3 | 58.8 | 64.0 | |
| −3.5 | No | 14.0 | 16.9 | 20.2 | 24.1 | 28.4 | 33.1 | 38.2 | 43.6 |
| Yes | 30.0 | 34.9 | 40.1 | 45.6 | 51.1 | 56.6 | 62.0 | 67.1 | |
| −4 | No | 15.7 | 18.9 | 22.5 | 26.6 | 31.2 | 36.2 | 41.4 | 46.9 |
| Yes | 33.0 | 38.0 | 43.4 | 48.9 | 54.5 | 59.9 | 65.1 | 70.0 | |
Table 5.
Five-Year Risk of Incident Fragility Fracture in the CaMOS Population of Men Based on Age, Femoral Neck T-Score, and Spine Fracture (No/Yes)
| Femoral neck T-score | Spine fracture | Age (yr)
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| 50 | 55 | 60 | 65 | 70 | 75 | 80 | 85 | ||
| −1 | No | 10.5 | 11.5 | 12.6 | 13.8 | 15.0 | 16.4 | 17.9 | 19.5 |
| Yes | 21.2 | 23.0 | 24.9 | 26.8 | 28.9 | 31.1 | 33.4 | 35.7 | |
| −1.5 | No | 11.6 | 12.7 | 13.9 | 15.2 | 16.6 | 18.1 | 19.7 | 21.3 |
| Yes | 23.2 | 25.1 | 27.1 | 29.2 | 31.4 | 33.6 | 36.0 | 38.4 | |
| −2 | No | 12.8 | 14.0 | 15.3 | 16.7 | 18.2 | 19.8 | 21.5 | 23.3 |
| Yes | 25.3 | 27.3 | 29.4 | 31.6 | 33.9 | 36.3 | 38.7 | 41.2 | |
| −2.5 | No | 14.2 | 15.5 | 16.9 | 18.4 | 20.0 | 21.7 | 23.6 | 25.5 |
| Yes | 27.5 | 29.7 | 31.9 | 34.2 | 36.5 | 39.0 | 41.5 | 44.0 | |
| −3 | No | 15.6 | 17.1 | 18.6 | 20.2 | 21.9 | 23.8 | 25.7 | 27.7 |
| Yes | 29.9 | 32.1 | 34.4 | 36.8 | 39.3 | 41.8 | 44.3 | 46.9 | |
| −3.5 | No | 17.2 | 18.8 | 20.4 | 22.1 | 24.0 | 25.9 | 28.0 | 30.1 |
| Yes | 32.4 | 34.7 | 37.1 | 39.5 | 42.0 | 44.6 | 47.2 | 49.8 | |
| −4 | No | 18.9 | 20.6 | 22.3 | 24.2 | 26.2 | 28.2 | 30.4 | 32.6 |
| Yes | 35.0 | 37.4 | 39.8 | 42.3 | 44.9 | 47.5 | 50.1 | 52.7 | |
DISCUSSION
We found that consideration of spine fracture status along with the WHO risk factors provided additional information compared with considering the WHO risk factors alone. In univariate analysis, we found that spine fracture status was one of the most significant predictors of 5-yr fracture risk. In addition, we assessed models for predicting future fracture risk by sequentially adding the most important risk factors and found that a model including age, BMD T-score, and presence or absence of spine fracture provided almost as much information as consideration of all WHO fracture risk tool risk factors plus presence or absence of spine fracture. Moreover, we found that this model provided more prognostic information than consideration of the WHO risk factors alone.
In the absence of knowledge about prevalent spine fracture status, assessments based on the WHO risk factors may under- or overestimate the true risk of an individual experiencing an incident fracture. This is similar to the experience of Siris et al.,(17) who observed that, in the absence of knowledge about spine fracture status, assessments based on BMD alone may under- or overestimate the true fracture risk.
This is the first analysis to compare the importance of the WHO fracture risk factors and spine fracture status for predicting future fracture risk. We assessed spine fracture burden as a predictor of future fractures in four different ways, comparing the impact of dichotomization of the presence or absence of vertebral fractures; the number of vertebral fractures; the severity of vertebral fractures; and the SDI. Although SDI provided the best discrimination of future fracture risk in the univariate analyses, the four different methods of assessing spine fracture status provided roughly similar information in multivariate analyses, and for simplicity, we used only the presence or absence of spine fracture in the final model. Like several of the WHO risk factors that are dichotomized, clearly patients with greater spine fracture burdens have greater risk than those patients with lesser spine fracture burdens.
We previously found that a simplified risk factor system comprising age, BMD, a history of prior clinical fracture, and systemic glucocorticoid use performed approximately as well as the eight clinical risk factors comprising the WHO tool in predicting absolute fracture risk.(27) This analysis showed that age, BMD, and presence or absence of spine fracture were the most important risk factors for predicting future fracture in this population-based cohort. Consideration of these three risk factors alone provided greater predictive capacity than the risk factors included in the WHO tool. Furthermore, consideration of age, BMD, and presence or absence of spine fracture was sufficient, and little more useful risk prediction was obtained by consideration of the other risk factors in the WHO model.
An advantage of including only three variables in the assessment of future fracture risk is that predicted absolute fracture risk can be reported in simple tables such as Table 4 for women and Table 5 for men. These tables highlight the prognostic significance of spine fracture status. For example, in a 55-yr-old woman having a femoral neck T-score of −1, fracture risk was 9.4% for subjects with no spine fractures and was 21.4% for subjects with spine fractures. For those patients with age or BMD between the intervals provided in the tables, the risk is intermediate. Practitioners assessing patients similar to those in our study for osteoporosis can therefore use age, BMD, and presence or absence of spine fracture to predict 5-yr fracture risk using these tables.
Several differences between these analyses and those performed to develop the WHO fracture risk assessment tool bear mentioning. CaMOS included serial assessments of spine radiographs, allowing us to consider spine fracture status as a determinant of future fracture risk. Additionally, inclusion of serial spine imaging in CaMOS provided the opportunity to define an endpoint of vertebral fracture and/or nonvertebral fragility fracture. It is important to note that the WHO tool provides information regarding the probability of different endpoints, including hip fracture or of a major osteoporotic fracture, defined as any hip, clinical spine, humerus, or forearm fracture.(26) Our analyses included only one cohort of patients, whereas nine cohorts were used to develop the WHO fracture risk assessment tool. As such, our results may not be as generalizable.
Our study included 5 yr of follow-up, whereas the WHO fracture risk assessment tool provides 10-yr fracture risk. In this regard, it is notable that some of the most elderly individuals in this cohort did not survive to complete spine radiographs at 5 yr, highlighting a possible limitation of using a 10-yr interval in providing osteoporotic fracture risk assessment in elderly patients.
Prevalent vertebral fracture status was assessed in our study by conventional lateral spine radiography, a technique considered the standard for identifying vertebral fractures. Nevertheless, patient inconvenience and radiation exposure could limit the performance of this test in all patients with low bone mass or osteoporosis. Furthermore, access to spine imaging may be hard to find in some jurisdictions because of limitations of equipment or expertise, and the lack of inclusion of radiographic data may have been an attempt by the WHO to keep some degree of standardization across borders, a laudable goal. Nevertheless, these issues are less likely to be impediments in regions where BMD measurements are routinely available, and BMD has been incorporated as an important component of the WHO fracture risk prediction tool in many countries in North America, Europe, and Asia. To maximize the diagnostic yield of radiographic assessment, some screening of those most in need of spine imaging might include an assessment of height loss.(28,29) Serial measurements of height showing 2 cm of height loss or history showing 4 cm of height loss should prompt spine imaging. Also, whereas morphometric vertebral fractures are sometimes described as “silent,” there is considerable innervation of the periosteum, and rather than being asymptomatic, the pain from spine fractures may resolve spontaneously and not reach medical attention. Consequently, questioning of patients about back pain, especially positional back pain, may occasionally show a history consistent with a vertebral fracture. Nevertheless, whereas symptoms of vertebral fracture may resolve or indeed be silent, and it has been estimated that only about one quarter of incident radiographic vertebral deformities are diagnosed as new clinical vertebral fractures,(30) their occurrence remains permanently detectable by lateral spine imaging, which therefore represents an important mechanism for maximizing the detection of an important risk factor. A possible alternative to lateral spine radiography, vertebral fracture assessment (VFA)—lateral spine imaging performed by DXA—at the time of BMD testing, involves less radiation exposure and cost and is less subject to parallax distortion; however, it historically has generated images of lower resolution.(31,32)
Our study therefore showed that the use of prevalent vertebral fracture status along with age and BMD has the capacity to predict future fracture risk at least as well as or better than the risk factors included in the WHO tool but with greater simplicity. For patients for whom spine imaging is practical, our findings emphasize the degree to which spine fracture burden offers future fracture risk prediction, shows the importance of having such information as part of the routine evaluation for osteoporosis, and provides a practical approach for using this information.
Acknowledgments
The authors thank all those participants in CaMOS whose careful responses and attendance made this analysis possible. This study was funded by the Canadian Institutes of Health Research (CIHR); Merck Frosst Canada; Eli Lilly Canada; Novartis Pharmaceuticals; The Alliance for Better Bone Health (sanofiaventis and Procter & Gamble Pharmaceuticals Canada); The Dairy Farmers of Canada; and The Arthritis Society.
APPENDIX: CaMOS RESEARCH GROUP
CaMOS Coordinating Center, McGill University, Montreal, Quebec: Suzette Poliquin (national coordinator), Suzanne Godmaire (research assistant), Claudie Berger (study statistician), Lawrence Joseph (consultant statistician).
Memorial University, St John’s Newfoundland: Carol Joyce (director), Christopher Kovacs (co-director), Emma Sheppard (coordinator).
Dalhousie University, Halifax, Nova Scotia: Susan Kirkland, Stephanie Kaiser (co-directors), Barbara Stanfield (coordinator).
Laval University, Quebec City, Quebec: Louis Bessette (co-director), Marc Gendreau (coordinator).
Queen’s University, Kingston, Ontario: Tassos Anastassiades (director), Tanveer Towheed (co-director), Barbara Matthews (coordinator).
University of Toronto, Toronto, Ontario: Sophie Jamal(co-director), Tim Murray (past director), Barbara Gardner-Bray (coordinator)
McMaster University, Hamilton, Ontario: Alexandra Papaioannou (co-director), Laura Pickard (coordinator).
University of Saskatchewan, Saskatoon, Saskatchewan: K. Shawn Davison (co-director), Jola Thingvold (coordinator).
University of Calgary, Calgary, Alberta: David A. Hanley (director), Jane Allan (coordinator).
University British Columbia, Vancouver, British Columbia: Yvette Vigna (coordinator); Brian C. Lentle (radiologist).
Footnotes
Parts of the manuscript were presented at the 29th Annual Meeting of the American Society for Bone and Mineral Research in Honolulu, HI, September 16–20 2007.
References
- 1.Kanis JA, Johnell O, Oden A, Johansson H, McCloskey E. FRAX™ and the assessment of fracture probability in men and women from the UK. Osteoporos Int. 2008;19:385–397. doi: 10.1007/s00198-007-0543-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kanis JA, Borgstrom F, De Laet C, Johansson H, Johnell O, Jonsson B, Oden A, Zethraeus N, Pfleger B, Khaltaev N. Assessment of fracture risk. Osteoporos Int. 2005;16:581–589. doi: 10.1007/s00198-004-1780-5. [DOI] [PubMed] [Google Scholar]
- 3.Riggs BL, Melton LJ., III The worldwide problem of osteoporosis: Insights afforded by epidemiology. Bone. 1995;17(Suppl 5):S505–S511. doi: 10.1016/8756-3282(95)00258-4. [DOI] [PubMed] [Google Scholar]
- 4.Melton LJ, III, Kan SH, Frye MA, Wahner HW, O’Fallon WM, Riggs BL. Epidemiology of vertebral fractures in women. Am J Epidemiol. 1989;129:1000–1011. doi: 10.1093/oxfordjournals.aje.a115204. [DOI] [PubMed] [Google Scholar]
- 5.Kanis JA, Johnell O, Oden A, Borgstrom F, Zethraeus N, DeLaet C, Jonsson B. The risk and burden of vertebral fractures in Sweden. Osteoporos Int. 2004;15:20–26. doi: 10.1007/s00198-003-1463-7. [DOI] [PubMed] [Google Scholar]
- 6.Jackson SA, Tenenhouse A, Robertson L the CaMOS Study Group. Vertebral fracture definition from population-based data: Preliminary results from the Canadian Multicenter Osteoporosis Study (CaMOS) Osteoporos Int. 2000;11:680–687. doi: 10.1007/s001980070066. [DOI] [PubMed] [Google Scholar]
- 7.Ettinger B, Black DM, Mitlak BH, Knickerbocker RK, Nickelsen T, Genant HK, Christiansen C, Delmas PD, Zanchetta JR, Stakkestad J, Gluer CC, Krueger K, Cohen FJ, Eckert S, Ensrud KE, Avioli LV, Lips P, Cummings SR. Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene: Results from a 3-year randomized clinical trial. Multiple Outcomes of Raloxifene Evaluation (MORE) Investigators. JAMA. 1999;282:637–645. doi: 10.1001/jama.282.7.637. [DOI] [PubMed] [Google Scholar]
- 8.Black DM, Cummings SR, Karpf DB, Cauley JA, Thompson DE, Nevitt MC, Bauer DC, Genant HK, Haskell WL, Marcus R, Ott SM, Torner JC, Quandt SA, Reiss TF, Ensrud KE. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Fracture Intervention Trial Research Group. Lancet. 1996;348:1535–1541. doi: 10.1016/s0140-6736(96)07088-2. [DOI] [PubMed] [Google Scholar]
- 9.Cummings SR, Black DM, Thompson DE, Applegate WB, Barrett-Connor E, Musliner TA, Palermo L, Prineas R, Rubin SM, Scott JC, Vogt T, Wallace R, Yates AJ, LaCroix AZ. Effect of alendronate on risk of fracture in women with low bone density but without vertebral fractures: Results from the Fracture Intervention Trial. JAMA. 1998;280:2077–2082. doi: 10.1001/jama.280.24.2077. [DOI] [PubMed] [Google Scholar]
- 10.Black DM, Arden NK, Palermo L, Pearson J, Cummings SR. Prevalent vertebral deformities predict hip fractures and new vertebral deformities but not wrist fractures. Study of Osteoporotic Fractures Research Group. J Bone Miner Res. 1999;14:821–828. doi: 10.1359/jbmr.1999.14.5.821. [DOI] [PubMed] [Google Scholar]
- 11.Lindsay R, Silverman SL, Cooper C, Hanley DA, Barton I, Broy SB, Licata A, Benhamou L, Geusens P, Flowers K, Stracke H, Seeman E. Risk of new vertebral fracture in the year following a fracture. JAMA. 2001;285:320–323. doi: 10.1001/jama.285.3.320. [DOI] [PubMed] [Google Scholar]
- 12.Ross PD, Davis JW, Epstein RS, Wasnich RD. Pre-existing fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med. 1999;114:919–923. doi: 10.7326/0003-4819-114-11-919. [DOI] [PubMed] [Google Scholar]
- 13.Gallagher JC, Genant HK, Crans GG, Vargas SJ, Krege JH. Teriparatide reduces the fracture risk associated with increasing number and severity of osteoporotic fractures. J Clin Endocrinol Metab. 2005;90:1583–1587. doi: 10.1210/jc.2004-0826. [DOI] [PubMed] [Google Scholar]
- 14.Delmas PD, Genant HK, Crans GG, Stock JL, Wong M, Siris E, Adachi JD. Severity of prevalent vertebral fractures and the risk of subsequent vertebral and nonvertebral fractures: Results from the MORE Trial. Bone. 2003;33:522–532. doi: 10.1016/s8756-3282(03)00241-2. [DOI] [PubMed] [Google Scholar]
- 15.Pongchaiyakul C, Nguyen ND, Jones G, Center JR, Eisman JA, Nguyen TV. Asymptomatic vertebral deformity as a major risk factor for subsequent fractures and mortality: A long-term prospective study. J Bone Miner Res. 2005;8:1349–1355. doi: 10.1359/JBMR.050317. [DOI] [PubMed] [Google Scholar]
- 16.Crans GG, Genant HK, Krege JH. Prognostic utility of a semiquantitative spinal deformity index. Bone. 2005;37:175–179. doi: 10.1016/j.bone.2005.04.003. [DOI] [PubMed] [Google Scholar]
- 17.Siris ES, Genant HK, Laster AJ, Chen P, Misurski DA, Krege JH. Enhanced prediction of fracture risk combining vertebral fracture status and BMD. Osteoporos Int. 2007;18:761–770. doi: 10.1007/s00198-006-0306-8. [DOI] [PubMed] [Google Scholar]
- 18.Cauley JA, Hochberg MC, Lui L, Palermo L, Ensrud KE, Hillier TA, Nevitt MC, Cummings SR. Long-term risk of incident vertebral fractures. JAMA. 2007;298:2761–2767. doi: 10.1001/jama.298.23.2761. [DOI] [PubMed] [Google Scholar]
- 19.World Health Organization. [Accessed May 5, 2008];FRAX WHO fracture risk assessment tool. 2008 Available at www.shef.ac.uk/FRAX/index.htm.
- 20.Kreiger N, Tenenhouse A, Joseph L, Machenzie T, Poliquin S, Brown JP, Prior JC, Rittmaster RS. Research notes: The Canadian multicentre osteoporosis study (CaMOS): Background, rationale, methods. Can J Aging. 1999;18:376–387. [Google Scholar]
- 21.Genant HK, Grampp S, Gluer CC, Faulkner KG, Jergas M, Engelke K, Hagiwara S, Van Kuijk C. Universal standardization for dual X-ray absorptiometry—patient and phantom cross-calibration results. J Bone Miner Res. 1994;9:1503–1514. doi: 10.1002/jbmr.5650091002. [DOI] [PubMed] [Google Scholar]
- 22.Pearson J, Dequeker J, Henley M, Bright J, Reeve J, Kalender W, Laval-Jeantet AM, Ruegsegger P, Felsenberg D, Adams J. European semi-anthropomorphic spine phantom for the calibration of bone densitometers—assessment of precision, stability and accuracy—the European Quantitation of Osteoporosis Study Group. Osteoporos Int. 1995;5:174–184. doi: 10.1007/BF02106097. [DOI] [PubMed] [Google Scholar]
- 23.Tenenhouse A, Joseph L, Kreiger N, Poliquin S, Murray TM, Blondeau L, Berger C, Hanley DA, Prior JC CaMOS Research Group; Canadian Multicentre Osteoporosis Study. Estimation of the prevalence of low bone density in Canadian women and men using a population-specific DXA reference standard: The Canadian Multicentre Osteoporosis Study (CaMOS) Osteoporos Int. 2000;11:897–904. doi: 10.1007/s001980070050. [DOI] [PubMed] [Google Scholar]
- 24.Kanis JA, Johnell O, De Laet C, Johansson H, Oden A, Delmas P, Eisman J, Fujiwara S, Garnero P, Kroger H, McCloskey EV, Mellstrom D, Melton LJ, Pols H, Reeve J, Sillman A, Tenenhouse A. A meta-analysis of previous fracture and subsequent fracture risk. Bone. 2004;35:375–382. doi: 10.1016/j.bone.2004.03.024. [DOI] [PubMed] [Google Scholar]
- 25.De Laet C, Oden A, Johansson H, Johnell O, Jonsson B, Kanis JA. The impact of the use of multiple risk indicators for fracture on case-finding strategies: A mathematical approach. Osteoporos Int. 2005;16:313–318. doi: 10.1007/s00198-004-1689-z. [DOI] [PubMed] [Google Scholar]
- 26.Kanis JA, Oden A, Johnell O, Johansson H, De Laet C, Brown J, Burckhardt P, Cooper C, Christiansen C, Cummings S, Eisman JA, Fujiwara S, Gluer C, Goltzman D, Hans D, Krieg MA, La Croix A, McCloskey E, Melstrom D, Melton LJ, III, Pols H, Reeve J, Sanders K, Schott AM, Silman A, Torgerson D, van Staa T, Watts NB, Yoshimura N. The use of clinical risk factors enhances the performance of BMD in the prediction of hip and osteoporotic fractures in men and women. Osteoporos Int. 2007;18:1033–1046. doi: 10.1007/s00198-007-0343-y. [DOI] [PubMed] [Google Scholar]
- 27.Richards JB, Leslie WD, Joseph L, Siminoski K, Hanley DA, Adachi JD, Brown JP, Morin S, Papaioannou A, Josse RG, Prior JC, Davison KS, Tenenhouse A, Goltzman D CaMOS Study Group. Changes to osteoporosis prevalence according to method of risk assessment. J Bone Miner Res. 2007;22:228–234. doi: 10.1359/jbmr.061109. [DOI] [PubMed] [Google Scholar]
- 28.Krege JH, Siminoski K, Adachi JD, Misurski DA, Chen P. A simple method for determining the probability a new vertebral fracture is present in postmenopausal women with osteoporosis. Osteoporos Int. 2006;17:379–386. doi: 10.1007/s00198-005-2005-2. [DOI] [PubMed] [Google Scholar]
- 29.Siminoski K, Jiang G, Adachi JD, Hanley DA, Cline G, Ioannidis G, Hodsman A, Josse RG, Kendler D, Olszymski WP, Ste Marie LG, Eastell R. Accuracy of height loss during prospective monitoring for detection of incident vertebral fractures. Osteoporos Int. 2005;16:403–410. doi: 10.1007/s00198-004-1709-z. [DOI] [PubMed] [Google Scholar]
- 30.Fink HA, Milavetz DL, Palermo L, Nevitt MC, Cauley JA, Genant HK, Black DM, Ensrud KE Fracture Intervention Trial Research Group. What proportion of incident radiographic vertebral deformities is clinically diagnosed and vice versa? J Bone Miner Res. 2005;20:1216–1222. doi: 10.1359/JBMR.050314. [DOI] [PubMed] [Google Scholar]
- 31.Guermazi A, Mohr A, Grigorian M, Taouli B, Genant HK. Identification of vertebral fractures in osteoporosis. Semin Musculoskelet Radiol. 2002;6:241–252. doi: 10.1055/s-2002-36722. [DOI] [PubMed] [Google Scholar]
- 32.Faulkner KG. Clinical use of bone densitometry. In: Marcus R, Feldman D, Kelsey J, editors. Osteoporosis. Academic Press; San Diego, CA, USA: 2001. pp. 433–458. [Google Scholar]