SEARCH

SEARCH BY CITATION

Keywords:

  • osteoporosis;
  • absolute risk;
  • intervention thresholds;
  • prevalence;
  • fracture probability

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The impact of clinical risk factor–based absolute risk methods on the prevalence of high risk for osteoporotic fracture is unknown. We applied absolute risk methods to 6646 subjects and found that the prevalence of elderly women deemed to be at high risk increased substantially, whereas the overall prevalence was highly dependent on the threshold used to designate high risk.

Introduction: Many groups have advocated using absolute risk methods that incorporate clinical risk factors to target patients for osteoporosis therapy. We examined how the application of such absolute risk classification systems influences the prevalence of those considered to be at high risk for osteoporotic fracture and compared these systems to one based solely on BMD.

Materials and Methods: Using 6646 subjects from the Canadian Multicentre Osteoporosis Study (CaMos), a prospective, randomly selected, population-based cohort, we assessed three different systems for determining prevalence of high risk for osteoporotic fracture: a BMD-based system; a simplified risk factor system incorporating age, sex, BMD, and two clinical risk factors; and a comprehensive system, incorporating age, sex, BMD, and seven clinical risk factors. The 10-year absolute risks of incident fragility fracture were compared across systems using three different high-risk thresholds.

Results: The prevalence of a T score ≤ −2.5 was 18.8% (95% CI: 17.7–19.9%) in women and 3.9% (95% CI: 3.0–4.7%) in men. Using a 15% 10-year risk of fracture threshold, the prevalence of women at high risk increased to 46.9% (95% CI: 45.4–48.4) and 42.5% (95% CI: 41.1–43.9) when the comprehensive and simplified risk factor classification systems were used, respectively. Using a 25% 10-year absolute risk threshold, the prevalence of high risk was similar to that of the BMD-based system, whereas the 20% threshold gave intermediate rates. All thresholds analyzed resulted in an increased prevalence of older women at high risk for fracture, whereas only the 15% 10-year risk of fracture threshold resulted in an increase in the prevalence of men at high risk.

Conclusions: The application of risk factor–based systems results in an increased prevalence of older women at high risk. The prevalence of individuals at high risk may increase with changes to the methods used to determine those who are eligible for therapy. These data have important implications for the pattern of care and costs of treating osteoporotic fractures.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Osteoporosis is a disease characterized by an increased propensity to fracture caused by a loss of bone strength.(1) In 1994, the World Health Organization (WHO) set the criterion for diagnosing osteoporosis in white women as a BMD 2.5 or more SD below the normal BMD for young healthy white women.(2) However, the application of the BMD criterion in isolation serves to identify only a minority of individuals who eventually fracture, because of its low sensitivity, whereas at the same time identifying many false positives.(3) Thus, the International Osteoporosis Foundation (IOF), the WHO, and Osteoporosis Canada (OC) have recommended the incorporation of clinical risk factors with BMD to yield a metric that better predicts the absolute risk of osteoporotic fracture.(4–6)

The WHO Collaborating Center has recently identified a comprehensive set of seven clinical risk factors (prior fragility fracture, a parental history of hip fracture, smoking, use of systemic corticosteroids, excess alcohol intake, body mass index, and rheumatoid arthritis), which in addition to age and sex, contribute to fracture risk independently of BMD.(4) A simplified approach suggested by Osteoporosis Canada incorporates age, sex, prior fragility fracture, and systemic steroid use together with BMD to define an absolute fracture risk.(5) Both risk factor assessment strategies aim to derive an absolute risk of fracture to improve the ability of clinicians to correctly identify those at high risk. It is currently not known how these absolute risk factors systems will impact the prevalence of osteoporosis.

Assessment of absolute risk for coronary heart disease has long been advocated as part of standard care for presymptomatic individuals and has been proposed as a guide to instituting therapy to prevent myocardial infarction.(7,8) Although randomized controlled trials for therapies used for coronary heart disease prevention have generally not been tested using these absolute risk methods, the efficacy of these therapies has been regarded as sufficient evidence to recommend them in treatment strategies for patients designated to be at high risk for myocardial infarction. By analogy, it is likely that absolute fracture risk assessment in osteoporosis may well guide therapeutic intervention using available therapeutics based on existing randomized controlled trials.

To determine the prevalence of women and men at high risk for fracture based on these absolute risk approaches, we used data from a large, randomly selected, population-based community cohort. We contrasted a comprehensive absolute risk system (using age, sex, BMD, and all seven clinical risk factors suggested by the WHO Collaborating Center) and a simplified absolute risk system (using only age, sex, BMD, prior fragility fracture, and systemic steroid use, as suggested by the OC) to the 1994 WHO diagnostic classification based solely on BMD.(2)

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Study design and population

The Canadian Multicentre Osteoporosis Study (CaMos) is a prospective cohort study following a randomly selected, population-based community cohort of 9423 noninstitutionalized men and women ≤25 years 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.(9) Briefly, recruitment for the cohort began in February 1996 and ended in September 1997. At the time of recruitment, BMD was measured in all available subjects, and participants were interviewed by a trained interviewer to assess for osteoporosis and fracture-related risk factors. After the baseline interview, incident fragility fractures were determined by annually mailing each subject a fracture questionnaire, and a clinical interview was conducted at the fifth anniversary of the commencement of the study. This analysis was limited to those subjects who were ≤50 years of age at the initial assessment. The study was approved by regional institutional ethics review boards, and participants provided written informed consent in accordance with the Helsinki Declaration.

BMD

Seven of nine centers measured BMD of the lumbar spine (L1–L4) and hip using DXA with Hologic QDR 1000, 2000, or 4500, while two centers used Lunar DPX densitometers. All BMD results were converted to a Hologic standard, using the method described by Genant et al.(10) Each year, a semianthropomorphic European Spine Phantom (Siemens, Munich, Germany) was measured at each site, for cross-calibration purposes.(11) BMD results used in this analysis are those measured at the baseline assessment and include total hip, femoral neck, trochanter, and lumbar spine (L1–L4). The T scores used are derived from CaMos female reference data.(12) This method is consistent with the simplified absolute risk system, as proposed by Osteoporosis Canada.(5)

Clinical risk factor measurement

All clinical risk factors were derived from the baseline interview. To apply the comprehensive risk factor system, subject responses were coded to indicate if they were current cigarette smokers, if they had ever used systemic corticosteroids, if they had sustained a minimal trauma fracture after 50 years of age,(13) if their parents had sustained any fracture at any age, or if they consumed greater than two units of alcohol per day. A subject was considered to have rheumatoid arthritis if they self-reported a physician-made diagnosis of rheumatoid arthritis. Weight and height of each participant were measured, and body mass index (BMI) was calculated by dividing the weight of the subject in kilograms, by the square of his/her height in meters. A low BMI was defined as <20 kg/m2.

To apply the simplified risk factor system, subject responses were coded to indicate whether they had sustained a minimal trauma fracture after age 40(13) and whether they had used systemic glucocorticoid therapy for >3 months without regard to dose.

Absolute fracture risk estimation and statistical methods

Each subject was assigned an estimated 10-year risk of osteoporotic fracture based on the comprehensive absolute risk system (using age, sex, BMD, and all seven clinical risk factors suggested by the WHO Collaborating Center) and the simplified absolute risk system (using only age, sex, BMD, prior fragility fracture, and systemic steroid use).

To apply the simplified risk factor system an estimated 10-year probability of fragility fracture based on age, sex, and the lowest T score was assigned to each subject. This estimate was derived using fracture data from the Malmö, Sweden, population.(14) If the subject reported a prevalent fragility fracture after age 40 or if they had used systemic glucocorticoid therapy for >3 months, the 10-year probability of fracture was increased by 10%, which is equivalent to an increase in risk category.(5) This simplified risk factor system was based on the risk factor system advocated by Osteoporosis Canada.(5)

To apply the comprehensive risk factor system, each subject was again assigned a 10-year probability of fragility fracture based on age, sex, and femoral neck T score, again using the same previously published fracture data.(14) To determine the independent effect of each of the seven clinical risk factors, a logistic regression model was constructed. The outcome of interest was incident clinically reported fragility fractures at any site. The period of follow-up for fragility fractures was 5 years. Incident fragility fractures were defined as those fractures that occurred because of minimal trauma (spontaneously or caused by a fall from standing height or lower). All incident fractures were clinically reported fragility fractures. The logistic regression model included the age- and sex-specific 10-year risk of fracture based on femoral neck T score and the seven clinical risk factors listed above.

The intercept and β coefficients for each risk factor from this logistic regression were used to calculate the logit for incident fragility fracture for each subject. The 5-year probability of fracture for each subject was determined from the inverse logit: exp[logit]/(1 + exp[logit]). The 10-year probability of fracture was estimated from two consecutive 5-year probabilities; the second 5-year period adjusted for the 5-year increase in age and the probability of a fracture in the first 5 years. This method was preferred to simply doubling the 5-year risk for two reasons: First, risk increases with age, so performing a second calculation for each subject accounting for their having aged 5 years results in a more accurate prediction. Second, we were interested in the probability of one fracture and not the total number of fractures for each subject. Simple doubling of risk ignores the fact that, if a subject experienced the event in the first 5 years, the risk of a first fracture in the second 5 years becomes zero.

To apply the densitometric system of identifying candidates for osteoporosis therapy, the lowest T score at spine, total hip, trochanter, or femoral neck was identified. If this BMD was 2.5 or more SD below the young adult mean, the subject was considered a candidate for osteoporosis therapy. For consistency with the absolute risk approaches, T scores used in the densitometric classification system were also derived from the CaMos female reference population.(12) Overall prevalences in men and women were directly standardized to the age structure of the Canadian general population for persons ≥50 years of age.(15) 95% CIs for the population proportions were calculated.(16) The logistic regression modeling was performed using the Proc Logistic procedure from SAS (version 9.1).

Thresholds to define high risk

Because absolute risk-based systems are highly dependent on the threshold used to define high risk, we examined three 10-year risk thresholds, above which subjects were considered to be at high risk for fracture: >15%, >20%, and >25% 10-year risk of osteoporotic fracture. A cut-off of >20% was the threshold suggested by Osteoporosis Canada to designate high risk.(5) To determine the sensitivity of our analyses to this risk threshold, we also evaluated higher and lower cut-offs. A threshold of >25% 10-year risk of fragility fracture was chosen because it most closely approximated the overall prevalence of high risk for fragility fracture derived from the densitometric-based intervention threshold. Finally a >15% 10-year risk of fracture was assessed to better understand how a lower threshold would affect the prevalence of those at high risk for fracture.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

There were 7753 subjects ≥50 years of age in the CaMos cohort. We limited our analysis to the 6646 (85.7% of the original cohort) subjects who underwent baseline BMD testing. The majority of persons included in the analysis were women (71.2%) and of self-identified white ethnicity (95.6%). The mean age of the sample population was 65.3 years for men and 65.9 years for women. The clinical risk factor with the highest prevalence was a parental history of fracture (32.9% for women and 27.3% for men). The majority of those subjects reporting previous use of systemic corticosteroids had used these medications for <3 months. The prevalence of all risk factors, except current cigarette smoking and alcohol intake, were higher in women (Table 1). Over the 5 years of follow-up, there were 495 fragility fractures reported among subjects ≥50 years old at baseline.

Table Table 1.. Prevalence of Risk Factors at Baseline Assessment in Subjects ≤50 Years of Age
Thumbnail image of

Based on BMD T score alone, the Canadian population-adjusted prevalence of women at high risk for osteoporotic fracture was 18.8% (95% CI: 17.7–19.9%). Using a 15% 10-year risk of fracture, this prevalence increased to 42.5% (95% CI: 41.1–43.9%) when the simplified risk factor system was applied, whereas application of the comprehensive risk factor system increased the prevalence of women at high osteoporotic fracture risk to 46.9% (95% CI: 45.4–48.4%). Using a 25% 10-year risk of fracture as an intervention threshold, the prevalence of high risk women was 18.2% (95% CI: 17.2–19.3%) when applying the simplified risk factor system and 18.3% (95% CI: 17.1–19.5%) when applying the comprehensive risk factor system. The 20% 10-year risk of fracture threshold provided prevalence rates intermediate to that of the 15% and 25% thresholds. In contrast to the results in women, only when the lowest threshold for high risk designation was applied was there an appreciable increase in the prevalence of men at high osteoporotic fracture risk. All of these overall prevalence rates were age-adjusted to the Canadian general population. (Fig. 1).

thumbnail image

Figure Figure 1. Effect of different risk classification systems on the prevalence of high risk for osteoporotic fractures in (A) women and (B) men ≤50 years of age, according to absolute risk threshold (age-adjusted to the general Canadian population; 95% CIs).

Download figure to PowerPoint

Figure 2 shows the effect of each risk classification system on the prevalence of high risk for fracture in different age groups of women using the three different risk thresholds. In women, the use of both simplified and comprehensive absolute risk classifications systems resulted in a substantially increased prevalence of high risk in elderly women relative to the densitometric classification. On the other hand, this prevalence was decreased in women <60 years old, when using 20% and 25% 10-year absolute risk thresholds. Using the 15% 10-year absolute risk threshold, the prevalence of those at high risk increased for all age groups.

thumbnail image

Figure Figure 2. Prevalence of high risk for osteoporotic fractures by age group in women using (A) 15%, (B) 20%, and (C) 25% 10-year absolute risk thresholds (*95% CI for the difference between the T score ≤ −2.5 SD system and the clinical risk factor systems excludes zero).

Download figure to PowerPoint

The effect of risk classification systems on the prevalence of high risk for fragility fracture in men is shown in Fig. 3. Applying the 15% 10-year absolute risk threshold increased the prevalence of those at high risk in all age groups exceeding 65 years. However, both the 20% and 25% 10-year absolute risk thresholds produced little appreciable change in the prevalence of high risk for osteoporotic fractures.

thumbnail image

Figure Figure 3. Prevalence of high risk for osteoporotic fractures by age group in men using (A) 15%, (B) 20%, and (C) 25% 10-year absolute risk thresholds (*95% CI for the difference between the T score ≤ −2.5 SD system and the clinical risk factor systems excludes zero).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

In the large randomly selected, population-based cohort we examined, the application of absolute risk classification systems resulted in substantially increased rates of older women designated to be at high fracture risk compared with a classification system based on a T score ≤ −2.5. In general, clinical risk factor based systems tended to decrease the prevalence of younger women at high risk for fracture. In men, these same risk classification systems resulted in an increase in prevalence only when the lowest threshold for high risk classification was applied, and overall, the number of men at high risk for fracture was much lower than in women across all methods. The prevalence of high risk for fracture was strongly dependent on the 10-year absolute risk threshold used to designate high risk. These data suggest that application of absolute risk factor systems may substantially affect the costs associated with the treatment of osteoporosis.

BMD testing alone is insufficiently sensitive to predict incident fragility fractures,(3) and including clinical risk factors may improve test accuracy.(4–6,17) possibly by adding an index of bone quality to the bone mass metric provided by BMD testing. Absolute risk classification systems can also overcome many of the limitations posed by relative risk classification systems(18) and are intuitive to both patients and clinicians. Consequently, because of their improved test characteristics, it is therefore likely that absolute risk classification systems will become central to the treatment of osteoporosis.

When using risk factor classification systems, the overall prevalence of those at high risk for fractures was highly dependent on which threshold was used. To assess the impact of different thresholds of risk on prevalence of those at high risk for osteoporotic fracture, we examined 10-year probability intervention thresholds of 15%, 20%, and 25%. Using a 15% 10-year risk of fracture as a threshold to designate high risk, −45% of Canadian women over the age of 50 would be deemed to be at high risk of fracture, a substantial increase over those that would be diagnosed as having osteoporosis based on a T score ≤ −2.5. This is consistent with a report using a cohort of 7806 men and women from Rotterdam that found that the majority of fractures in a population-based cohort occur at a T score that is > −2.5.(19) Our results are also concordant with previously reported work describing an increase in the number of elderly women with a high risk of osteoporotic fracture when absolute fracture risk methods were compared with a densitometric classification system without inclusion of clinical risk factors.(20) Although dependent on the type of intervention planned, our results suggest that the pattern of patient care and the costs associated with the treatment of osteoporosis may change substantially if clinical risk factor based absolute risk strategies are implemented'and that these costs will depend on which threshold of risk is used to prompt treatment. These costs must of course be balanced against the costs of not treating those at high risk of fracture.

There are several limitations to this study. First, when applying the comprehensive risk factor system, 10-year absolute risks were approximated from 5-year incident fracture data. However, we have accounted for the temporal effects of age and the fact that a fracture in the first 5 years eliminates the risk of fracture in the second 5 years. Our results are applicable at the population level and may not accurately predict incident fragility fractures in individuals. Second, different groups define the risk associated with systemic steroid use differently; Osteoporosis Canada stipulates that the subject must have taken ≤7.5 mg of prednisone daily or equivalent, whereas the WHO Collaborating Center makes no reference to dose. Furthermore, recent analyses have indicated that prednisone doses as low as 2.5 mg/day may increase fracture rates.(21) In our analysis, the dose of steroid was not available, and thus the prevalence of high fracture risk in the simplified risk factor system might have been falsely elevated. However, when we reanalyzed our data excluding systemic steroid therapy entirely, the prevalence estimates changed only minimally. Third, the WHO Collaborating Center includes parental history of hip fracture as a clinical risk factor, whereas in our analysis, the site of parental fracture was not known. However, Kanis et al.(22) recently reported the risk ratio for parental history of hip and parental history of any fracture to be similar. A self-reported diagnosis of rheumatoid arthritis may not be entirely reliable. Finally, our work has not attempted to provide evidence to refute or support the contention that these new risk classification systems should be administered in clinical practice. Prior randomized controlled trial inclusion criteria for the treatment of osteoporosis were not based on these absolute risk classification systems and further work will be needed to clarify whether it would be appropriate to use these absolute risk systems in the absence of clinical trials using these methods.

Data from the NHANES III study indicate that when using a hip BMD T score of ≤ −2.5 SD as a diagnostic threshold,(23) the number of American women >50 years of age with osteoporosis ranges from 4 to 6 million, depending on which hip site was measured. Our data indicate that the prevalence of high risk women >50 years of age in the United States may increase to >10 million persons, depending on which absolute risk factor system and intervention threshold are used. Decreasing the intervention threshold for all osteoporotic fractures to 15% or lower would result in a remarkable increase in osteoporosis intervention rates for both men and women, which could have significant consequences for health care costs. Consequently, when using absolute risk methods, decisions will have to be made about which threshold of absolute risk to use to recommend initiation of therapy. This decision will be at least partly driven by considerations of cost-effectiveness. Nevertheless, studies have indicated that cost-effectiveness may be achieved when the 10-year probability of hip fracture is >10% in women, although a lower threshold may be equally cost-effective if other osteoporotic fractures are considered in addition to hip fracture.(24,25) Thresholds for intervention will therefore be dependent on local and national considerations, such as the public health burden of osteoporotic fractures in a specific jurisdiction, ability and willingness to pay for therapeutics to prevent fractures, and other competing health priorities. The cost-savings of preventing osteoporotic fractures with anti-osteoporotic medications must also be considered in any cost-effectiveness evaluation. National osteoporosis societies that develop absolute risk-based osteoporosis guidelines should carefully consider these issues.

In summary, the administration of absolute risk systems to classify fracture probability resulted in a substantial increase in the prevalence of older women designated at high risk for osteoporotic fracture. It is likely that the introduction of these classification systems will have an important impact on the pattern of care of patients with osteoporosis and on the future costs of osteoporosis therapy.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The Canadian Multicentre Osteoporosis Study was funded by the Canadian Institutes of Health Research (CIHR); Merck Frosst Canada Ltd.; Eli Lilly Canada Inc.; Novartis Pharmaceuticals Inc.; The Alliance: Sanofi-Aventis and Procter & Gamble Pharmaceuticals Canada Inc.; The Dairy Farmers of Canada; and The Arthritis Society. These funding sources had no role in the conception of this analysis, statistical methods, interpretation of the data, or drafting of this paper.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Raisz LG 2005 Pathogenesis of osteoporosis: Concepts, conflicts, and prospects. J Clin Invest 115: 33183325.
  • 2
    World Health Organization 1994 Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Report of a WHO Study Group. World Health Organ Tech Rep Ser 843: 1129.
  • 3
    Marshall D, Johnell O, Wedel H 1996 Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 312: 12541259.
  • 4
    Kanis JA, Borgstrom F, De Laet C, Johansson H, Johnell O, Jonsson B, Oden A, Zethraeus N, Pfleger B, Khaltaev N 2005 Assessment of fracture risk. Osteoporos Int 16: 581589.
  • 5
    Siminoski K, Leslie WD, Frame H, Hodsman A, Josse RG, Khan A, Lentle BC, Levesque J, Lyons DJ, Tarulli G, Brown JP 2005 Recommendations for bone mineral density reporting in Canada. Can Assoc Radiol J 56: 178188.
  • 6
    Kanis JA, Gluer CC 2000 An update on the diagnosis and assessment of osteoporosis with densitometry. Osteoporos Int 11: 192202.
  • 7
    Brindle PM, McConnachie A, Upton MN, Hart CL, Smith GD, Watt GCM 2005 The accuracy of the Framingham risk-score in different socioeconomic groups: A prospective study. Br J Gen Pract 55: 838845.
  • 8
    Grundy SM, Pasternak R, Greenland P, Smith S, Fuster V 1999 Assessment of cardiovascular risk by use of multiple-risk-factor assessment equations—a statement for healthcare professionals from the American Heart Association and the American College of Cardiology. J Am Coll Cardiol 34: 13481359.
  • 9
    Kreiger N, Tenenhouse A, Joseph L, Mackenzie T, Poliquin S, Brown JP, Prior JC, Rittmaster RS 1999 Research notes: The Canadian multicentre osteoporosis study (CaMos): Background, rationale, methods. Can J Aging 18: 376387.
  • 10
    Genant HK, Grampp S, Gluer CC, Faulkner KG, Jergas M, Engelke K, Hagiwara S, Vankuijk C 1994 Universal standardization for deual X-ray absorptiometry—patient and phantom cross-calibration results. J Bone Miner Res 9: 15031514.
  • 11
    Pearson J, Dequeker J, Henley M, Bright J, Reeve J, Kalender W, Lavaljeantet AM, Ruegsegger P, Felsenberg D, Adams J, Birkenhager JC, Braillon P, Curiel MD, Fischer M, Galan F, Geusens P, Hyldstrup L, Jaeger P, Jonson R, Kalefezras J, Kotzki P, Kroger H, Vanlingen A, Nilsson S, Osteaux M, Cano RP, Reid DM, Reiners C, Ribot C, Schneider P, Slosman DO, Wittenberg G 1995 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 5: 174184.
  • 12
    Tenenhouse A, Joseph L, Kreiger N, Poliquin S, Murray TM, Blondeau L, Berger C, Hanley DA, Prior JC 2000 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 11: 897904.
  • 13
    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, Silman A, Tenenhouse A 2004 A meta-analysis of previous fracture and subsequent fracture risk. Bone 35: 375382.
  • 14
    Kanis JA, Johnell O, Oden A, Dawson A, De Laet C, Jonsson B 2001 Ten year probabilities of osteoporotic fractures according to BMD and diagnostic thresholds. Osteoporos Int 12: 989995.
  • 15
    Stastistics Canada 2005 Canadian Population by Age and Sex. Available at http://www40.statcan.ca/l01/cst01/demo10a.htm?sdi=age. Accessed December 20, 2005.
  • 16
    Moore SD, McCabe PG 2003 Introduction to the Practice of Statistics, 4th ed. W.H. Freeman, New York, NY, USA.
  • 17
    Kanis JA, Black D, Cooper C, Dargent P, Dawson-Hughes B, De Laet C, Delmas P, Eisman J, Johnell O, Jonsson B, Melton L, Oden A, Papapoulos S, Pols H, Rizzoli R, Silman A, Tenenhouse A 2002 A new approach to the development of assessment guidelines for osteoporosis. Osteoporos Int 13: 527536.
  • 18
    Kanis JA, Johnell O, Oden A, Jonsson B, Dawson A, Dere W 2000 Risk of hip fracture derived from relative risks: An analysis applied to the population of Sweden. Osteoporos Int 11: 120127.
  • 19
    Schuit SCE, van der Klift M, Weel AEAM, de Laet CEDH, Burger H, Seeman E, Hofman A, Uitterlinden AG, van Leeuwen JPTM, Pols HAP 2004 Fracture incidence and association with bone mineral density in elderly men and women: The Rotterdam Study. Bone 34: 195202.
  • 20
    Leslie WD, Siminoski K, Brown JP 2005 Comparison of ten-year absolute fracture risk and WHO risk categorization in menopausal females. J Bone Miner Res 20: S11.
  • 21
    van Staa TP, Leufkens HGM, Abenhaim L, Zhang B, Cooper C 2000 Oral corticosteroids and fracture risk: Relationship to daily and cumulative doses. Rheumatol 39: 13831389.
  • 22
    Kanis JA, Johansson H, Oden A, Johnell O, De Laet C, Eisman JA, McCloskey EV, Mellstrom D, Melton LJ III, Pols HA, Reeve J, Silman AJ, Tenenhouse A 2004 A family history of fracture and fracture risk: A meta-analysis. Bone 35: 10291037.
  • 23
    Looker AC, Orwoll ES, Johnston CC Jr, Lindsay RL, Wahner HW, Dunn WL, Calvo MS, Harris TB, Heyse SP 1997 Prevalence of low femoral bone density in older U.S. adults from NHANES III. J Bone Miner Res 12: 17611768.
  • 24
    Kanis JA, Dawson A, Oden A, Johnell O, De Laet C, Jonsson B 2001 Cost-effectiveness of preventing hip fracture in the general female population. Osteoporos Int 12: 356361.
  • 25
    Kanis JA, Johnell O, Oden A, De Laet C, Oglesby A, Jonsson B 2002 Intervention thresholds for osteoporosis. Bone 31: 2631.